EPA Document# EPA-740-R-24-013
November 2024
United States Office of Chemical Safety and
v/crM Environmental Protection Agency Pollution Prevention
Supplement to the Risk Evaluation for
1,4-Dioxane
CASRN 123-91-1
o
November 2024
-------
TABLE OF CONTENTS
ACKNOWLEDGEMENTS 19
EXECUTIVE SUMMARY 20
1 INTRODUCTION 27
1.1 Regulatory Context 27
1.2 Scope 30
1.3 Use Characterization 31
1.3.1 Conceptual Models 31
1.3.1.1 1,4-Dioxane as a Byproduct 33
1.3.1.2 Occupational Exposures 35
1.3.1.3 General Population Exposures 37
1.3.1.3.1 Drinking Water 39
1.3.1.3.2 Air 40
1.3.1.3.3 Aggregate Exposure 40
1.3.2 Potentially Exposed or Susceptible Subpopulations 40
1.4 Systematic Review 41
1.5 Document Outline 41
2 RELEASES AND CONCENTRATIONS 43
2.1 Approach and Methodology 43
2.1,1 Industrial and Commercial Releases 43
2.1.1.1 General Approach and Methodology for Environmental Releases 46
2.1.1.2 Water Release Estimates 46
2.1.1.3 Land Release Estimates 47
2.1.1.4 Air Release Estimates 48
2.1.1.4.1 Pre-screening Analysis 48
2.1.1.4.2 Single-Year Fenceline Analysis 48
2.1.1.4.3 Multi-year Analysis 49
2.2 Environmental Releases 49
2.2,1 Industrial and Commercial Releases 49
2.2.1.1 Release Estimates Summary 49
2.2.1.2 Weight of Scientific Evidence Conclusions for Environmental Releases 54
2.2.1.3 Strengths, Limitations, Assumptions, and Key Sources of Uncertainty for the
Environmental Release Assessment 58
2.3 1,4-Dioxane Environmental Concentrations 60
2.3,1 Surface Water Pathway 60
2.3.1.1 Monitoring Data 60
2.3.1.2 Surface Water and Drinking Water Modeling 66
2.3.1.2.1 Modeling Methodology 66
2.3.1.2.2 EstimatingDown-the-DrainReleases 67
2.3.1.2.3 Hydraulic Fracturing 68
2.3.1.2.4 Proximity to Drinking Water Sources 68
2.3.1.3 Modeling Results 68
2.3.1.3.1 Facility-Specific Results 68
2.3.1.3.2 Concentrations from Down-the-Drain Loading 74
2.3.1.3.3 Concentrations from Hydraulic Fracturing 75
2.3.1.3.4 Aggregate Probabilistic Results 76
Page 2 of 570
-------
2.3.1.4 Comparison of Modeled and Monitored Surface Water Concentrations 79
2.3.1.5 Strengths, Limitations, and Sources of Uncertainty in Assessment Results for
Monitored and Modeled Drinking Water and Surface Water Concentrations 80
2.3.2 Land Pathway (Groundwater) 81
2.3.2.1 Groundwater Monitoring Data 81
2.3.2.2 Disposal via Underground Injection 83
2.3.2.2.1 Summary of Assessment for Disposal to Underground Injection 84
2.3.2.2.2 Strengths, Limitations, and Sources of Uncertainty in Assessment of Disposal to
Underground Injection Wells 84
2.3.2.3 Disposal to Landfills 84
2.3.2.3.1 Summary of Assessment for Disposal to Landfills 88
2.3.2.3.2 Strengths, Limitations, and Sources of Uncertainty in Assessment Results for
Disposal to Landfills 88
2.3.2.4 Disposal of Hydraulic Fracturing Produced Water to Surface Impoundments 89
2.3.2.4.1 Summary of Assessment for Disposal of Hydraulic Fracturing Produced Water.... 89
2.3.2.4.2 Strengths, Limitations, and Sources of Uncertainty in Assessment Results for
Disposal from Hydraulic Fracturing Operations 90
2.3.3 Ambient Air Pathway 90
2.3.3.1 Measured Concentrations in Air 91
2.3.3.2 Modeled Concentrations in Air 91
2.3.3.2.1 Ambient Air: Screening Methodology 92
2.3.3.2.2 Ambient Air: Single Year Methodology (AERMOD) 92
2.3.3.2.3 Ambient Air: Multi-Year Analysis (IIOAC) 96
2.3.3.2.4 Ambient Air: IIOAC Methodology for COUs Without Site-Specific Data
(Hydraulic Fracturing, Industrial, and Institutional Laundry Facilities) 96
2.3.3.3 Strengths, Limitations, and Sources of Uncertainty for Modeled Air Concentrations.... 97
3 HUMAN EXPOSURES 99
3.1 Occupational Exposures 99
3.1.1 Approach and Methodol ogy 100
3.1.1.1 Process Description, Number of Sites, Number of Workers, and ONUs 100
3.1.1.2 Inhalation Exposures Approach and Methodology 101
3.1.1.3 Dermal Exposures Approach and Methodology 102
3.1.1.4 Engineering Controls and Personal Protective Equipment 102
3.1.2 Occupational Exposure Estimates 103
3.1.2.1 Summary of Inhalation Exposure Assessment 103
3.1.2.2 Summary of Dermal Exposures Assessment 104
3.1.2.3 Weight of Scientific Evidence Conclusions for Occupational Exposure Information.. 105
3.1.2.4 Strengths, Limitations, Assumptions, and Key Sources of Uncertainty for the
Occupational Exposure Assessment 107
3.1.2.4.1 Number of Workers 107
3.1.2.4.2 Analysis of Inhalation Exposure Monitoring Data 107
3.1.2.4.3 Modeled Inhalation Exposures 108
3.1.2.4.4 Modeled Dermal Exposures 108
3.2 General Population Exposures 108
3.2.1 Approach and Methodol ogy 110
3.2.2 Drinking Water Exposure Assessment 110
3.2.2.1 Surface Water Exposure Assessment Ill
3.2.2.1.1 Exposures from Individual Facility Releases Ill
Page 3 of 570
-------
3.2.2.1.2 Exposures from Down-the-Drain Releases 114
3.2.2.1.3 Disposal of Hydraulic Fracturing Produced Waters 114
3.2.2.1.4 Aggregate Exposure 115
3.2.2.2 Groundwater Exposure Assessment 116
3.2.2.2.1 Disposal to Landfills 116
3.2.2.2.2 Disposal of Hydraulic Fracturing Produced Waters 117
3,2,3 Air Exposure Assessment 118
3.2.3.1 Industrial COUs Reported to TRI 118
3.2.3.2 Hydrauli c Fracturing 120
3.2.3.3 Industrial and Institutional Laundry Facilities 123
3.3 Weight of Scientific Evidence Conclusions 124
3.3.1 Occupational Exposures 125
3.3.1.1 Inhalation Exposure 125
3.3.1.2 Dermal Exposure 126
3.3.2 Drinking Water 127
3.3.2.1 Drinking Water Exposure Estimates Based on Surface Water Concentrations 127
3.3.2.2 Drinking Water Exposure Estimates Based on Groundwater Concentrations 131
3.3.2.2.1 Groundwater Concentrations Resulting from Disposal to Landfill 131
3.3.2.2.2 Groundwater Concentrations Resulting from Disposal of Hydraulic Fracturing
Waste 132
3.3.3 Air 133
3.3.3.1 Modeled Air Concentrations for Industrial COUs Reported to TRI 133
3.3.3.2 Air Concentrations Modeled near Hydraulic Fracturing Operations and
Industrial/Institutional Laundries 135
4 HUMAN HEALTH HAZARD 137
4.1 Summary of Hazard Endpoints Previously Identified in the 2020 Risk Evaluation 137
4.2 Summary of Adjustments to Previously Established Hazard Values 137
4.2.1 Derivation of Acute/Short-Term Hazard Values 140
4.2.1.1 Inhalation HEC 140
4.2.1.2 Oral and Dermal HEDs 140
4.2.2 Derivation of Chronic Hazard Values 140
4.2.2.1 Inhalation HEC 140
4.2.2.2 Oral HEDs 141
4.2.2.3 Dermal HEDs 141
4.2.3 Derivation of Cancer Hazard Values 141
4.3 Strengths, Limitations, Assumptions, and Key Sources of Uncertainty in the Hazard and
Dose-Response Analysis 142
5 HUMAN HEALTH RISK CHARACTERIZATION 143
5.1 Risk Characterization Approach 144
5.1.1 Estimation of Non-cancer Risks 145
5.1.2 Estimation of Cancer Risks 145
5.2 Human Health Risk Characterization 145
5.2.1 Summary of Risk Estimates for Occupational Exposures 145
5.2.2 Summary of Risk Estimates for the General Population 149
5.2.2.1 Drinking Water - Surface Water Pathway 149
5.2.2.1.1 Risks from Exposure to Drinking Water Concentrations Indicated in Finished
Drinking Water Monitoring Data 150
Page 4 of 570
-------
5.2.2.1.2 Risks from Exposures to Water Concentrations Modeled from Industrial
Releases 151
5.2.2.1.3 Risks from Exposures to Water Concentrations Modeled from DTD Releases
(from POTWs), Assuming No Downstream Dilution 156
5.2.2.1.4 Risks from Exposure to Drinking Water Concentrations Modeled from Disposal
of Hydraulic Fracturing Produced Waters to Surface Water, Assuming No
Downstream Dilution 158
5.2.2.1.5 Aggregate Risks from Drinking Water Exposures Modeled from Multiple
Sources Releasing to Surface Water, Assuming No Downstream Dilution 159
5.2.2.1.6 Integrated Summary of Drinking Water Risk Estimates across Multiple Lines of
Evidence for Surface Water 162
5.2.2.2 Drinking Water - Groundwater and Disposal Pathways 163
5.2.2.3 Air Pathway 165
5.2.2.3.1 Industrial COUs Reported to TRI 165
5.2.2.3.2 Hydraulic Fracturing 170
5.2.2.3.3 Industrial and Institutional Laundry Facilities 173
5.2.3 Potentially Exposed or Susceptible Subpopulations 174
5.2.4 Aggregate and Sentinel Exposures 176
5.2.5 Summary of Overall Confidence and Remaining Uncertainties in Human Health Risk
Characterization 177
5.2.5.1 Risks from Occupational Exposures 178
5.2.5.2 Risks from General Population Exposures through Drinking Water 178
5.2.5.3 Risks from General Population Exposures through Groundwater and Land Disposal
Pathways 180
5.2.5.4 Risks from General Population Exposures through Air 180
REFERENCES 182
APPENDICES 191
Appendix A KEY ABBREVIATIONS AND ACRONYMS 191
Appendix B LIST OF SUPPLEMENTAL DOCUMENTS 194
Appendix C SYSTEMATIC REVIEW PROTOCOL FOR THE DRAFT SUPPLEMENT TO
THE RISK EVALUATION FOR 1,4-DIOXANE 198
C.l Clarifications and Updates to the 2021 Draft Systematic Review Protocol 199
C.l.l Clarifications and Updates 199
C.l Data Search 202
C.2.1 Multi-disciplinary Updates to the Data Search 203
C.2.2 Additional Data Sources Identified 203
C.2.2.1 Additional Data Sources Identified for Environmental Release and Occupational
Exposure 204
C.2.2.2 Additional Data Sources Identified for General Population, Consumer, and
Environmental Exposure 204
C.2.3 Search Strings 205
C.2.3.1 Environmental Release and Occupational Exposure Search Strings 205
C.2.3.2 General Population, Consumer, and Environmental Exposure Search Strings 206
C.3 Data Screening 206
C.3.1 Environmental Release and Occupational Exposure 208
C.3.1.1 Environmental Release and Occupational Exposure Literature Inventory Tree 209
Page 5 of 570
-------
C.3.2 General Population, Consumer, and Environmental Exposure 209
C.3.2.1 General Population, Consumer, and Environmental Exposure Literature Inventory
Tree210
C.4 Data Evaluation and Data Extraction 210
C.4.1 Environmental Release and Occupational Exposure 211
C.4.2 General Population, Consumer, and Environmental Exposure 211
C.4.2.1 Data Quality Evaluation Metric Updates 212
C.4.2.2 Data Evaluation Criteria for Monitoring Data, as Revised 213
C.4.2.3 Data Evaluation Criteria for Experimental Data, as Revised 221
C.4.2.4 Data Evaluation Criteria for Databases, as Revised 228
C.5 Evidence Integration 232
C.5.1 Environmental Release and Occupational Exposure 237
C.5.2 General Population 237
C.5.2.1 General Population: Surface Water 237
C.5.2.2 General Population: Groundwater 237
C.5.2.3 General Population Exposure: Ambient Air 238
Appendix D COU-OES MAPPING AND CROSSWALK 239
D. 1 COU-OES Mapping 239
D.2 COU-OES Crosswalk 241
Appendix E INDUSTRIAL AND COMMERCIAL ENVIRONMENTAL RELEASES 244
E.l Estimates of the Number of Industrial and Commercial Facilities with Environmental
Releases 244
E.2 Estimates of Number of Release Days for Industrial and Commercial Releases 246
E.3 Water Release Assessment 249
E.3.1 Assessment Using TRI and DMR 249
E.3.2 Assessment for OES Without TRI and DMR 252
E.3.3 Water Release Estimates Summary 255
E.3.4 Summary of Weight of Scientific Evidence Conclusions in Water Release Estimates 262
E.4 Land Release Assessment 269
E.4.1 Assessment Using TRI 269
E.4.2 Assessment for OES Without TRI 271
E.4.3 Land Release Estimates Summary 277
E.4.4 Summary of Weight of Scientific Evidence Conclusions in Land Release Estimates 284
E,5 Air Release Assessment 291
E.5.1 Assessment Using TRI 291
E.5.2 Assessment for OESs Without TRI 293
E.5.3 Air Release Estimates Summary 299
E.5.4 Summary of Weight of Scientific Evidence Conclusions in Air Release Estimates 308
E.6 Comparison to PET Life Cycle Analysis 314
E,7 Detailed Strengths, Limitations, Assumptions and Key Sources of Uncertainties for the
Environmental Release Assessment 315
E.8 Weight of Scientific Evidence Conclusions for Environmental Releases 320
E.9 TRI to CDR Crosswalk 324
E. 10 Developing Models that Use Monte Carlo Methods 342
E.10.1 Background on Monte Carlo Methods 342
E.10.2 Implementation of Monte Carlo Methods 342
E.10.3 Building the Model 343
Page 6 of 570
-------
E.10.3.1 Build the Deterministic Model 343
E.10.3.2 Define Probability Distributions for Input Parameters 343
E.10.3.3 Select Model Outputs for Aggregation of Simulation Results 347
E.10.3.4 Select Simulation Settings and Run Model 347
E.10.3.5 Aggregate the Simulation Results and Produce Output Statistics 347
E. 11 Textile Dye Modeling Approach and Parameters for Estimating Environmental Releases 347
E. 11.1 Model Equations 348
E. 11.2 Model Input Parameters 350
E.11.3 Number of Sites 353
E.11.4 Mass Fraction of Dye Containing 1,4-Dioxane 353
E.11.5 Operating Days 353
E.l 1.6 Mass Fraction of 1,4-Dioxane in Dye Formulation 353
E. 11.7 Textile Production Rate 353
E.l 1.8 Mass Fraction of Textiles Treated with Dye 354
E.l 1.9 Mass Fraction of Dye Used per Mass of Textile Dyed 354
E. 11.10 Mass Fraction of the Dye Formulation in the Dyebath 354
E. 11.11 Container Size for Dye Formulation 354
E. 11.12 Container Residual Fraction for Totes 354
E. 11.13 Container Residual Fraction for Drums 355
E.l 1.14 Container Residual Fraction for Pails 355
E.l 1.15 Fraction of Dye Product Affixed to Textile During Dyeing Process Substrate 355
E.l 1.16 Key Strengths, Limitations, Uncertainties, and Sensitivity Analysis 356
E, 12 Laundry Detergent Modeling Approach and Parameters for Estimating Environmental
Releases 358
E.12.1 Model Equations 359
E.12.2 Model Input Parameters 366
E.12.3 Operating Days 370
E.12.4 Mass Fraction of 1,4-Dioxane in Laundry Detergent 370
E.12.5 Daily Use Rate of Detergent 370
E.12.6 Container Size 371
E.12.7 Indoor Air Speed 372
E.12.8 Container Residual Fraction for Totes 372
E.12.9 Container Residual Fraction for Drums 372
E.12.10 Container Residual Fraction for Pails 373
E. 12.11 Container Residual Fraction for Powders 373
E.12.12 Fraction of Laundry Detergents Containing 1,4-Dioxane 373
E. 12.13 Duration of Release for Container Unloading 374
E. 12.14 Fraction of Chemical Lost During Transfer of Solid Powders 374
E. 12.15 Control Efficiency for Dust Control Methods 374
E. 12.16 Capture Efficiency for Dust Capture Methods 375
E.12.17 Number of Sites 375
E.12.18 Diameter of Container Opening 375
E.12.19 Diameter of Wash Opening 375
E.12.20 Dilution Factor 375
E. 12.21 Container Fill Rate 375
E. 12.22 Key Strengths, Limitations, Uncertainties, and Sensitivity Analysis 375
E, 13 Hydraulic Fracturing Modeling Approach and Parameters for Estimating Environmental
Releases 377
E.13.1 Model Equations 379
Page 7 of 570
-------
E.13.2 Model Input Parameters 383
E.13.3 Number of Sites 386
E.13.4 Operating Days 386
E.13.5 Container Size 386
E.13.6 Diameter of Container Opening 387
E.13.7 Diameter of Equipment Opening 387
E.13.8 Air Speed During Equipment Cleaning 387
E.13.9 Equipment Cleaning Loss Fraction 387
E. 13.10 Container Fill Rate 387
E. 13.11 Equipment Cleaning Operating Hours 387
E.13.12 Spill Loss Fraction 387
E.13.13 Annual Use Rate of Fracturing Fluids Containing 1,4-Dioxane 388
E.13.14 Mass Fraction of 1,4-Dioxane in Hydraulic Fracturing Additive/Fluid 388
E. 13.15 Saturation Factor 389
E.13.16 Container Residual Fraction for Totes 389
E.13.17 Container Residual Fraction for Drums 389
E.13.18 Fraction of Injected Fracturing Fluid that Returns to the Surface 390
E.13.19 Key Strengths, Limitations, Uncertainties, and Sensitivity Analysis 390
E. 14 Dish Soap and Dishwasher Detergent Modeling Approach and Parameters for Estimating
Environmental Releases 392
E.14.1 Model Equations 394
E.14.2 Model Input Parameters 397
E.14.3 Facility Daily Throughput - Dish Soap 400
E.14.4 Facility Daily Throughput - Dishwasher Detergent 400
E.14.5 Concentration of 1,4-Dioxane in Dish Soap 400
E.14.6 Concentration of 1,4-Dioxane in Dishwasher Detergent 400
E.14.7 Saturation Factor 401
E.14.8 Container Size 401
E.14.9 Container Residual Loss Fraction 402
E.14.10 Diameter of Sink Opening 402
E. 14.11 Release Duration for Dishwashers 402
E.14.12 Number of Sites 402
E.14.13 Operating Days 403
E. 14.14 Container Unloading Rate 403
E.14.15 Dish Soap Wash Water Temperature 403
E. 14.16 Dishwasher Water Temperature 403
E.14.17 Indoor Air Speed 403
E.14.18 Key Strengths, Limitations, Uncertainties, and Sensitivity Analysis 404
Appendix F OCCUPATIONAL EXPOSURES 406
F.l Calculating Acute and Chronic Inhalation Exposures and Dermal Doses 406
F.2 Approach for Estimating Number of Workers and Occupational Non-users 406
F.3 Occupational Dermal Exposure Assessment Method 409
F.4 Occupational Exposure Scenarios 412
F.4.1 Textile Dye 412
F.4.2 Antifreeze 421
F.4.3 Surface Cleaner 424
F.4.4 Dish Soap 428
F.4.5 Dishwasher Detergent 431
Page 8 of 570
-------
F.4.6 Laundry Detergent (Industrial and Institutional) 433
F.4.7 Paint and Floor Lacquer 440
F.4.8 Spray Foam Application 446
F.4.9 Polyethylene Terephthalate Byproduct 447
F.4.10 Ethoxylation Process Byproduct 463
F.4.11 Hydraulic Fracturing 468
F.5 Summary of Occupational Inhalation Exposures 473
F.6 Summary of Weight of Scientific Evidence Conclusions in Inhalation Exposure Estimates ... 479
F.7 Antifreeze Modeling Approach and Parameters for Estimating Occupational Inhalation
Exposures 482
F.7.1 Model Equations 482
F.7.2 Modeling Input Parameters 484
F.7.3 Container Size 487
F.7.4 Jobs per Day 487
F.7.5 Concentration of 1,4-Dioxane in Antifreeze 487
F.7.6 Ventilation Rate 487
F.7.7 Mixing Factor 487
F.7.8 Saturation Factor 487
F.7.9 Use Rate of Antifreeze per Job 488
F.7.10 Container Fill Rate 488
F.7.11 Key Strengths, Limitations, Uncertainties, and Sensitivity Analysis 488
F.8 Laundry Detergent Modeling Approach and Parameters for Estimating Occupational
Inhalation Exposures 489
F.8.1 Model Equations 491
F.8.2 Model Input Parameters 494
F.8.3 Ventilation Rate 496
F.8.4 Mixing Factor 496
F.8.5 Total Particulate Concentration 496
F.8.6 Respirable Particulate Concentration 496
F.8.7 Key Strengths, Limitations, Uncertainties, and Sensitivity Analysis 497
F.9 Hydraulic Fracturing Modeling Approach and Parameters for Estimating Occupational
Inhalation Exposures 498
F.9.1 Model Equations 500
F.9.2 Model Input Parameters 504
F.9.3 Ventilation Rate 506
F.9.4 Mixing Factor 506
F.9.5 Key Strengths, Limitations, Uncertainties, and Sensitivity Analysis 506
F. 10 Dish Soap and Dishwasher Detergent Modeling Approach and Parameters for Estimating
Occupational Inhalation Exposures 507
F.10.1 Model Equations 509
F. 10.2 Model Input Parameters 511
F.10.3 Ventilation Rate 513
F.10.4 MixingFactor 513
F.10.5 Key Strengths, Limitations, Uncertainties, and Sensitivity Analysis 513
Appendix G SURFACE WATER CONCENTRATIONS 515
G. 1 Surface Water Monitoring Data 515
G. 1.1 Monitoring Data Retrieval and Processing 515
G. 1.2 Raw and Finished Drinking Water 516
Page 9 of 570
-------
G.2 Surface Water Modeling 517
G.2.1 Hydrologic Flow Data 517
G.2.2 Facility-Specific Release Modeling 518
G.2.3 Aggregate and Probabilistic Modeling 519
G. 2.3.1 The Fit-For-Purpose Aggregate Surface Water Model 519
G.2.3.2 Case Studies to Validate Aggregate Model 521
G.2.3.3 The Probabilistic Model 526
G.2.3.4 Modeling Ranges of DTD Contributions 528
G.2.3.5 Modeling Concentrations in Surface Water from Hydraulic Fracturing 530
G.2.4 Assessing Downstream Drinking Water Intakes 531
Appendix H GROUNDWATER CONCENTRATIONS AND DISPOSAL PATHWAYS
FROM LAND RELEASES 536
H. 1 Groundwater Monitoring Data Retrieval and Processing 536
H,2 Review of Land Release Permits 536
H.3 Landfill Analysis Using DRAS 538
H.4 Landfill Analysis Using EPACMTP 541
H.5 Surface Impoundment Analysis for the Disposal of Hydraulic Fracturing Produced Water
Using DRAS 544
Appendix I DRINKING WATER EXPOSURE ESTIMATES 547
I.1 Surface Water Sources of Drinking Water 548
1,2 Groundwater Sources of Drinking Water 548
Appendix J AIR EXPOSURE PATHWAY 549
J. 1 Ambient Air Concentrations and Exposures 549
J. 1.1 Ambient Air: Screening Methodologies and Results Summary - Fenceline 549
J. 1.2 Ambient Air: IIOAC Methodology and Results for COUs Without Site-Specific Data
(Hydraulic Fracturing, Industrial, and Institutional Laundry Facilities) 552
J. 1.3 Ambient Air: Single Year Methodology (AERMOD) 554
J. 1.4 Ambient Air: Multi-Year Analysis Methodology (IIOAC) 557
J,2 Inhalation Exposure Estimates for Fenceline Communities 558
J.3 Land Use Analysis 559
J.4 Aggregate Analysis across Facilities 560
Appendix K SUMMARY OF REVISED ANALYSES COMPLETED IN RESPONSE TO
SACC AND PUBLIC COMMENT 567
Appendix L OCCUPATIONAL EXPOSURE VALUE 570
LIST OF TABLES
Table 2-1. Additional Categories and Subcategories of COUs and Associated OESs Included in the
Scope of the Supplement Due to the Presence of 1,4-Dioxane Produced as a
Byproduct 44
Table 2-2. Summary of the Weight of Scientific Evidence for Environmental Release Estimates by
OES 56
Table 2-3. Summary of PWS Monitoring Datasets of 1,4-Dioxane Monitoring in PWSs Using
Surface Water as a Source 63
Page 10 of 570
-------
Table 2-4. Summary of Surface Water Concentration Results by OES from Facility-Specific
Modeling of Annual Maximum Releases between 2013 and 2019 for 1 Operating Day
per Year 70
Table 2-5. Summary of Surface Water Concentration Results by OES for Facility-Specific Modeling
of Annual Maximum Releases between 2013 and 2019 for the Maximum Operating
Days per Year 71
Table 2-6. OES-COU Crosswalk for Identified Facilities Releasing to Surface Water 72
Table 2-7. Summary by OES of Data Sources for Releases and Receiving Water Body Flow 73
Table 2-8. Hypothetical Mean Annual Concentrations (|ig/L) for a Range of Annual Release and
Flow Rate Combinations, for a Facility with 250 Days of Release per Year 74
Table 2-9. Occurrence of Facilities for Distributions of Maximum Annual 1,4-Dioxane Release
Amounts and Receiving Water Body Flow 74
Table 2-10. Estimated Surface Water Concentrations (|ig/L) Due to DTD Loading for a Range of
Populations and Hydrologic Flows 75
Table 2-11. Estimated Percent Occurrence of Combinations of Contributing Population to POTWs
and Receiving Water Body Flow, from Combined ICIS-NPDES and 2020 Census
Data 75
Table 2-12. Distribution of Potential Concentrations in Surface Water Resulting from Hydraulic
Fracturing Operations from a Single Site Reporting 1,4-Dioxane as an Ingredient 76
Table 2-13. Aggregate Probabilistic Results Showing Distribution of Total 1,4-Dioxane
Concentration in Surface Water (Release Plus Background) 77
Table 2-14. Potential Groundwater Concentrations (|ig/L) of 1,4-Dioxane Found in Wells within 1
Mile of a Disposal Facility Determined by Using the DRAS Model 87
Table 2-15. Total Annual Release Summary 90
Table 2-16. Summary of Select Statistics for the 95th Percentile Estimated Annual Average
Concentrations from the "Full-Screening" Analysis for 1,4-Dioxane Releases
Reported to TRI 94
Table 3-1. Estimated Dermal Absorbed Dose (mg/day) for Workers in Various Conditions of Use .... 104
Table 3-2. Summary of the Weight of Scientific Evidence for Occupational Exposure Estimates by
OES 106
Table 3-3. Adult and Infant Exposures Estimated from Facility-Specific Releases 112
Table 3-4. Adult LADD Exposures (mg/kg/day) Estimated from 1,4-Dioxane DTD Consumer and
Commercial Releases 114
Table 3-5. Adult ADR, ADD, and LADD Exposures Estimated from Disposal of Hydraulic
Fracturing Produced Waters to Surface Water 115
Table 3-6. Adult LADD Exposures from Aggregate Concentrations Estimated Downstream of
Release Sites (Including DTD Releases and Direct and Indirect Industrial Releases)... 115
Table 3-7. Adult LADD Exposures Estimated from Groundwater Contamination from Landfills
under Varying Landfill Conditions 117
Table 3-8. Estimated Exposures Resulting from Groundwater Contamination from Disposal of
Hydraulic Fracturing Produced Water 117
Table 3-9. Lifetime Average Daily Concentrations Estimated within 10 km of 1,4-Dioxane Releases
to Air 119
Table 3-10. Exposures from Fugitive Emissions Estimated within 1,000 m of Hydraulic Fracturing
Operations 121
Table 3-11. Exposures from Fugitive Emissions Estimated near Industrial and Institutional Laundry
Facilities 124
Table 4-1. Hazard Values Used for 1,4-Dioxane in this Supplement 139
Page 11 of 570
-------
Table 5-1. Use Scenarios, Populations of Interest, and Toxicological Endpoints Used for Acute and
Chronic Exposures 144
Table 5-2. Lifetime Cancer Risk Estimates for 1,4-Dioxane Concentrations Detected in Finished
Drinking Water 151
Table 5-3. Proximity of Nearest Downstream Drinking Water Intakes to Facilities Resulting in
Cancer Risk Greater than 1 x 1CT6 153
Table 5-4. Lifetime Cancer Riska Estimates from DTD Releases Alone (at the Point of Release)
under a Range of Population and Flow Rate Scenarios 157
Table 5-5. Lifetime Cancer Risks Estimated from Hydraulic Fracturing Produced Waters Disposed to
Surface Water under a Range of Scenarios 159
Table 5-6. Lifetime Cancer Risks Estimated for Modeled Groundwater Concentrations Estimated
under Varying Landfill Conditions 164
Table 5-7. Lifetime Cancer Risks Estimated for Modeled Groundwater Concentrations Resulting
from Disposal of Hydraulic Fracturing Produced Water 165
Table 5-8. Inhalation Lifetime Cancer Risks within 10 km of Industrial Air Releases Based on 95th
Percentile Modeled Exposure Concentrations 167
Table 5-9. Lifetime Cancer Risk Estimates for Fugitive Emissions from Hydraulic Fracturing 172
Table 5-10. Lifetime Cancer Risk Estimates for Fugitive Emissions from Industrial and Institutional
Laundry Facilitiesa 173
Table 5-11. Summary of PESS Considerations Incorporated throughout the Analysis and Remaining
Sources of Uncertainty 174
LIST OF FIGURES
Figure 1-1. 1,4-Dioxane Life Cycle Diagram 32
Figure 1-2. Production of 1,4-Dioxane as a Byproduct and Potential Exposure Pathways 33
Figure 1-3. Conceptual Model for Occupational Exposures from Industrial and Commercial
Activities 36
Figure 1-4. Conceptual Model for Environmental Releases and General Population Exposures 38
Figure 1-5. Overview of Analyses Included in this Supplement to the Risk Evaluation for 1,4-
Dioxane 42
Figure 2-1. Overview of EPA's Approach to Estimate Daily Releases for Each OES 46
Figure 2-2. 1,4-Dioxane Annual Water Releases as Reported to TRI and DMR, 2013-2019 51
Figure 2-3. 1,4-Dioxane Annual Releases to Land as Reported to TRI, 2013-2019 52
Figure 2-4. 1,4-Dioxane Annual Releases to Air as Reported by TRI, 2013-2019 53
Figure 2-5. Locations of Hydraulic Fracturing Operations that Report 1,4-Dioxane in Produced
Waters 54
Figure 2-6. Frequency of Nationwide Measured 1,4-Dioxane Surface Water Concentrations Retrieved
from the Water Quality Portal, 1997-2022 61
Figure 2-7. Frequency of Detection Limits for Nationwide Non-detect 1,4-Dioxane Surface Water
Samples Retrieved from the Water Quality Portal, 1997-2022 62
Figure 2-8. Detectable Concentrations of 1,4-Dioxane in Surface Water from the Water Quality
Portal, 1997-2022 62
Figure 2-9. Frequency of 1,4-Dioxane Concentrations Monitored in Raw (Untreated) Drinking Water
Derived from Surface Water 64
Figure 2-10. Frequency of 1,4-Dioxane Concentrations Monitored in Finished (Treated) Drinking
Water Derived from Surface Water 64
Figure 2-11. Map of Counties Containing PWSs that Reported Monitoring of Finished Drinking
Water Drawn from Surface Water for 1,4-Dioxane under UCMR3 65
Figure 2-12. Schematic of the EWISRD-XL Model Inputs and Outputs 67
Page 12 of 570
-------
Figure 2-13. Distributions of Surface Water Concentrations Estimated by Aggregate Probabilistic
Model for Each OES 78
Figure 2-14. Case Study Comparison of Modeled and Monitored Concentrations in Brunswick
County 80
Figure 2-15. Frequency of Nationwide Detected 1,4-Dioxane Groundwater Concentrations (n =
2,284) Retrieved from the Water Quality Portal, 1997-2022 82
Figure 2-16. Detectable Concentrations of 1,4-Dioxane in Groundwater from the Water Quality
Portal, 1997-2022 82
Figure 2-17. Groundwater Concentrations of 1,4-Dioxane vs. Sample Collection Date for Data
Collected between 1997 and 2022 83
Figure 2-18. Brief Description of Methodologies and Analyses Used to Estimate Ambient Air
Concentrations and Exposures 91
Figure 3-1. Potential Human Exposure Pathways to 1,4-Dioxane for the General Population 109
Figure 5-1. Distribution of Adult Lifetime Cancer Risk across all Facilities, Assuming No Additional
Dilution Occurs between the Point of Release and the Location of Drinking Water
Intakes 152
Figure 5-2. Distribution of Adult Lifetime Cancer Risk across Facilities with High Quality Release
Data, Assuming No Additional Dilution Occurs between the Point of Release and the
Location of Drinking Water Intakes 153
Figure 5-3. Distribution of Dilution of 1,4-Dioxane Concentrations at Downstream Drinking Water
Intakes 154
Figure 5-4. Distribution of Adult Lifetime Cancer Risk across all Facilities, Assuming Dilution to 1%
of Initial Concentrations in the Receiving Water Body 154
Figure 5-5. Histograms of Lifetime Cancer Risk Estimates for Aggregate Water Concentrations
Estimated Downstream of COUs with Vertical Lines Showing the Median and 95th
Percentile (P95) Values 161
LIST OF APPENDIX TABLES
TableApx C-l. Terminology Clarifications between the 2021 Draft Systematic Review Protocol and
the Systematic Review Protocol for the Supplement to the Risk Evaluation for 1,4-
Dioxane 200
Table_Apx C-2. Evaluation Criteria for Sources of Monitoring Data 213
Table Apx C-3. Evaluation Criteria for Sources of Experimental Data 221
Table Apx C-4. Evaluation Criteria for Sources of Database Data 228
Table Apx C-5. Considerations that Inform Evaluations of the Strength of the Evidence 233
Table Apx C-6. Evaluation of the Weight of Scientific Evidence for Exposure Assessments 234
TableApx D-l. Categories and Subcategories of Conditions of Use Included in the Scope of the
Risk Evaluation 241
Table Apx E-l. Summary of EPA's Estimates for the Number of Facilities for Each OES 244
Table Apx E-2. Summary of EPA's Estimates for Air and Water Release Days Expected for Each
OES 248
TableApx E-3. Summary of Daily Industrial and Commercial Water Release Estimates for 1,4-
Dioxane 256
Table Apx E-4. Summary of Weight of Scientific Evidence Conclusions in Water Release Estimates
by OES 262
Table Apx E-5. Summary of Daily Industrial and Commercial Land Release Estimates for 1,4-
Dioxane 278
Page 13 of 570
-------
TableApx E-6. Summary of Weight of Scientific Evidence Conclusions in Land Release Estimates
by OES 284
Table Apx E-7. Summary of Daily Industrial and Commercial Air Release Estimates for 1,4-
Dioxane 300
Table Apx E-8 Summary of Weight of Scientific Evidence Conclusions in Air Release Estimates by
OES 308
Table Apx E-9. Comparison of TRI/DMR Release Data to LCA Study for PET Byproduct 315
Table Apx E-10. Summary of Overall Weight of Scientific Evidence Conclusions for Environmental
Release Estimates by OES 321
Table_Apx E-l 1. TRI-CDR Use Code Crosswalk 324
Table Apx E-12. Summary of Parameter Values and Distributions Used in the Textile Release
Model 351
Table Apx E-13. Discrete Data Points on the Number of Operating Days at Textile Dye Sites 353
Table_Apx E-14. Triangular Distributions Ffixation 356
TableApx E-15. Summary of Parameter Values and Distributions Used in the Industrial and
Institutional Laundry Release Model 367
Table Apx E-16. Discrete Data Points on Mass Fraction of 1,4-Dioxane in Laundry Detergent 370
Table Apx E-17. Discrete Data Points on Daily Use Rate of Liquid Detergents 371
Table Apx E-18. Discrete Data Points on Daily Use Rate of Solid Detergents 371
Table Apx E-19. Data on the Fraction of Laundry Detergent Containing the Chemical of Interest 374
Table Apx E-20. Summary of Parameter Values and Distributions Used in the Hydraulic Fracturing
Release Model 384
TableApx E-21. Summary Statistics on Number of Operating Days at Hydraulic Fracturing Sites.... 386
Table Apx E-22. Summary Statistics on the Annual Use Rate of Fracturing Fluids at Hydraulic
Fracturing Sites 388
Table Apx E-23. Summary Statistics on the Mass Fractions of 1,4-Dioxane in Hydraulic Fracturing
Additives and Fluids 389
Table Apx E-24. Summary of Parameter Values and Distributions Used in the Industrial and
Commercial Use of Dish Soap and Dishwasher Detergent Release Model 398
Table Apx E-25. Discrete Data Points on Concentration of 1,4-Dioxane in Dish Soap 400
Table Apx E-26. Discrete Data Points on Concentration of 1,4-Dioxane in Dishwasher Detergent.... 401
TableApx F-l. Summary of Total Number of Workers and ONUs Potentially Exposed to 1,4-
Dioxane for Each Supplemental OES 407
Table Apx F-2. Glove Protection Factors for Different Dermal Protection Strategies from ECETOC
TRA v3 409
Table Apx F-3. Textile Dye Worker Exposure Data Evaluation 414
Table Apx F-4. Inhalation Exposures of Workers for the Use of Textile Dye Based on Monitoring
Data 416
Table Apx F-5. Occupational Inhalation Monitoring Data for Textile Dyes 417
Table_Apx F-6. Antifreeze Data Source Evaluation 422
Table Apx F-7. Modeled Occupational Inhalation Exposures for Antifreeze 423
Table Apx F-8. Inhalation Exposures of Workers for the Use of Antifreeze Based on Modeling 423
Table Apx F-9. Surface Cleaner Worker Exposure Data Evaluation 425
Table Apx F-10. Inhalation Exposures of Workers for the Use of Surface Cleaner Based on
Monitoring Data 426
Table Apx F-l 1. Occupational Inhalation Monitoring Data for Surface Cleaner 427
Table Apx F-12. Dish Soap Worker Exposure Data Evaluation 429
Table Apx F-13. Modeled Occupational Inhalation Exposures for Dish Soap 430
TableApx F-14. Inhalation Exposures of Workers for the Use of Dish Soaps Based on Modeling .... 430
Page 14 of 570
-------
TableApx F-15. Modeled Occupational Inhalation Exposures for Dishwasher Detergent 432
TableApx F-16. Inhalation Exposures of Workers for the Use of Dishwasher Detergents Based on
Modeling 433
Table Apx F-17. Laundry Detergent Worker Exposure Data Evaluation 437
Table Apx F-18. Modeled Occupational Inhalation Exposures for Industrial Laundries 437
Table Apx F-19. Modeled Occupational Inhalation Exposures for Institutional Laundries 438
Table Apx F-20. Inhalation Exposures of Workers for the Use of Laundry Detergent in Industrial
Laundries Based on Modeling 438
Table Apx F-21. Acute and Chronic Inhalation Exposures of Workers for the Use of Laundry
Detergent in Institutional Laundries Based on Modeling 439
Table Apx F-22. Paint and Floor Lacquer Worker Exposure Data Evaluation 442
TableApx F-23. Inhalation Exposures of Workers for the Use of Paint and Floor Lacquer Based on
Monitoring Data 443
Table Apx F-24. Occupational Inhalation Monitoring Data for Paint and Floor Lacquer 444
Table Apx F-25. Polyethylene Terephthalate (PET) Byproduct Worker Exposure Data Evaluation ... 449
TableApx F-26. Inhalation Exposures of Workers for PET Byproduct Based on Monitoring Data.... 450
Table Apx F-27. Occupational Inhalation Monitoring Data for Polyethylene Terephthalate (PET)
Byproduct 451
Table Apx F-28. Ethoxylation Process Byproduct Worker Exposure Data Evaluation 464
Table Apx F-29. Inhalation Exposures of Workers for the Ethoxylation Process Byproduct Based on
Monitoring Data 465
Table Apx F-30. Occupational Inhalation Monitoring Data for Ethoxylation Process Byproduct 466
Table Apx F-31. Hydraulic Fracturing Worker Exposure Data Evaluation 471
Table Apx F-32. Modeled Occupational Inhalation Exposures for Hydraulic Fracturing 472
Table Apx F-33. Inhalation Exposures of Workers for Hydraulic Fracturing Based on Modeling 472
Table Apx F-34. Estimated Inhalation Exposure (mg/m3) for Workers During Various Conditions of
Use 474
TableApx F-35. Summary of Weight of Scientific Evidence Conclusions in Inhalation Exposure
Estimates by OES 479
TableApx F-36. Summary of Parameter Values and Distributions Used in the Antifreeze Exposure
Modeling 486
TableApx F-37. Summary of Parameter Values and Distributions Used in the Laundry Detergent
Exposure Modeling 495
TableApx F-38. Summary of Parameter Values and Distributions Used in the Hydraulic Fracturing
Exposure Modeling 505
TableApx F-39. Summary of Parameter Values and Distributions Used in the Industrial and
Commercial Use of Dish Soap and Dishwasher Detergent Exposure Modeling 512
Table Apx G-l. Summary of Community Water Systems with Treatment Processes Capable of
Removing 1,4-Dioxane 517
Table Apx G-2. Summary of per Capita DTD Loading Estimates from SHEDS-HT Modeling 521
Table Apx G-3. Summary of Case Study Locations Including Modeled and Observed Surface Water
Concentrations 522
Table Apx G-4. Distribution of per Capita DTD Loading, in G/Day, by Product, for Non-commercial
Uses Modeled by SHEDS-HT 529
Table Apx G-5. Proportions of Population Expected to Contribute to DTD Loading through
Commercial Activities and Product Uses 529
Table Apx G-6. Summary of Proximity of Downstream Drinking Water Intakes to Releasing
Facilities Resulting in Modeled Risk above 1E-06 533
Page 15 of 570
-------
TableApx G-7. Ranges of Dilution and Diluted 1,4-Dioxane Concentrations Modeled at Drinking
Water Intakes Downstream of Industrial Releases 534
Table Apx G-8. Ranges of LADD and Adult Lifetime Cancer Risk Estimates for Diluted 1,4-
Dioxane Concentrations Modeled at Drinking Water Intakes Downstream of Industrial
Releases 535
Table Apx H-l. Release Year, TRI Facility ID, Facility Name, State, Registry Number, Disposal
Type, and Disposal Weight for On-Site Class I Underground Injection Wells
According to TRI 536
Table Apx H-2. Release Year, Source TRI Facility ID, Source State, Receiving Facility RCRA ID,
State, Disposal Type, and Disposal Weight for Off-Site Class I Underground Injection
Wells According to TRI and RCRAInfo Databases 537
Table Apx H-3. Release Year, TRI Facility ID, Facility Name, State, CERCLIS ID, Disposal Type,
and Disposal Weight for RCRA Subtitle C Landfills According to TRI 537
Table Apx H-4. Release Year, Source TRI Facility ID, Source State, Receiving Facility RCRA ID,
State, Disposal Type, and Disposal Weight for Off-Site Class I Underground Injection
Wells According to TRI and RCRAInfo Databases 538
Table_Apx H-5. Input Variables for Chemical of Concern 539
Table_Apx H-6. Waste Management Unit (WMU) Properties 540
Table Apx H-7. Potential Groundwater Concentrations (mg/L) Based on Disposal of 1,4-Dioxane to
Unlined and Clay-Lined Landfills as Assessed by Applying the EPACMTP Model.... 542
Table_Apx H-8. Input Variables for Chemical of Concern 544
Table_Apx H-9. Waste Management Unit 546
Table Apx J-l. Release Estimates from 2019 TRI Used for Ambient Air: Screening Methodology for
1,4-Dioxane 550
Table Apx J-2. Exposure and Risk Estimates from the Ambient Air: Screening Methodology for 1,4-
Dioxane Releases Reported to TRI 552
TableApx J-3. Exposure Scenarios and Inputs Utilized for Pre-screening Analysis of Hydraulic
Fracturing, Industrial Laundry, and Institutional Laundry COU 553
Table Apx J-4. Description of Daily or Period Average and Air Concentration Statistics 557
Table Apx J-5. Summary of Fenceline Community Exposures Expected near Facilities Where
Modeled Air Concentrations Indicated Risk for 1,4-Dioxane 560
Table Apx J-6. Summary of Groups of Facilities Considered in Aggregate Analysis 564
Table Apx K-l. Summary of Changes to Occupational Exposure and Risk Estimates 567
Table_Apx K-2. Summary of Revisions to Release Assessments 569
LIST OF APPENDIX FIGURES
FigureApx C-l. Overview of the TSCA Risk Evaluation Process with Identified Systematic Review
Steps 198
Figure Apx C-2. Literature Inventory Tree - Environmental Releases and Occupational Exposure
Search Results for 1,4-Dioxane 209
Figure Apx C-3. Literature Inventory Tree - General Population, Consumer, and Environmental
Exposure Search Results for 1,4-Dioxane 210
Figure_Apx D-l. COU and OES Mapping 240
Figure Apx E-l. Flowchart of a Monte Carlo Method Implemented in a Microsoft Excel-Based
Model Using a Monte Carlo Add-In Tool 343
Figure Apx E-2. Environmental Release Points (Numbered) and Occupational Exposure Points
(Lettered) During Textile Dying 348
Figure_Apx E-3. Container Cleaning (Daily Release Point 2) Sensitivity Chart 357
Page 16 of 570
-------
FigureApx E-4. Spent Dyebath and Equipment Cleaning (Daily Release Point 3) Sensitivity Chart. 357
FigureApx E-3. Environmental Release Points (Numbered) and Occupational Exposure Points
(Letterd) During Industrial/Institutional Laundering Operations 358
Figure Apx E-6. Sensitivity Chart for Fugitive Air Release During Unloading Liquid Detergents
(Daily Release Point 3) at Institutional Laundries 376
Figure Apx E-7. Sensitivity Chart for Release from Dust Generation During Unloading Solid
Detergents (Daily Release Point 4) at Industrial Laundries 377
Figure Apx E-4. Environmental Release Points (Numbered) and Occupational Exposure Points
(Lettered) During Hydraulic Fracturing 378
Figure Apx E-9. Sensitivity Chart for Fugitive Air Release During Unloading (Daily Release Point
1) at Hydraulic Fracturing Sites 391
Figure Apx E-10. Sensitivity Chart for Release from Flowback and Produced Water (Daily Release
Point 8) at Hydraulic Fracturing Sites 392
Figure Apx E-5. Environmental Release Points (Numbered) and Occupational Exposure Points
(Lettered) During Industrial and Commercial Use of Dish Soap and Dishwasher
Detergent 393
Figure Apx E-12. Sensitivity Chart for Container Disposal (Daily Release Point 2) at Dishwashing
Sites 405
Figure Apx E-13. Sensitivity Chart for Releases from Dishwashing (Daily Release Point 4) at
Dishwashing Sites 405
Figure Apx F-l. Environmental Release and Occupational Exposure Points During Textile Dying.... 413
Figure Apx F-2. Environmental Release and Occupational Exposure Points During
Industrial/Institutional Laundering Operation 434
Figure Apx F-3. Environmental Release and Occupational Exposure Points During Hydraulic
Fracturing 470
Figure Apx F-4. Sensitivity Chart for 8-Hour TWA Occupational Inhalation Exposure to 1,4-
Dioxane at Antifreeze Use Sites 489
Figure Apx F-4. Environmental Release Points (Numbered) and Occupational Exposure Points
(Lettered) During Industrial/Institutional Laundering Operations 490
Figure Apx F-5. Sensitivity Chart for 8-Hour TWA Occupational Inhalation Exposure to 1,4-
Dioxane Vapor at Institutional Laundries 497
Figure Apx F-6. Sensitivity Chart for 8-Hour TWA Occupational Inhalation Exposures to 1,4-
Dioxane Total Particulates at Industrial Laundries 498
Figure Apx F-5. Environmental Release Points (Numbered) and Occupational Exposure Points
(Lettered) During Hydraulic Fracturing 499
Figure Apx F-8. Sensitivity Chart for 8-Hour TWA Occupational Inhalation Exposure to 1,4-
Dioxane at Hydraulic Fracturing Sites 507
Figure Apx F-6. Environmental Release Points (Numbered) and Occupational Exposure Points
(Lettered) During Industrial and Commercial Use of Dish Soap & Dishwasher
Detergent 508
Figure Apx F-10. Sensitivity Chart for 8-Hour TWA Occupational Inhalation Exposure to 1,4-
Dioxane at Sites Using Dish Soap 514
Figure Apx F-l 1. Sensitivity Chart for 8-Hour TWA Occupational Inhalation Exposure to 1,4-
Dioxane at Sites Using Dishwasher Detergents 514
Figure Apx G-l. Example Raw and Finished Water Concentrations from a PWS Without Processes
to Remove 1,4-Dioxane 516
Figure Apx G-2. Schematic of the General Fit-for-Purpose EWISRD-XL Model 520
Figure_Apx G-3. Map of Brunswick County, NC Model Case Study 523
Page 17 of 570
-------
FigureApx G-4. Plot Comparing Results from Brunswick County Case Study Modeling with
Observed Concentrations 524
Figure_Apx G-5. Map of the Columbia, TN, Case Study 525
Figure_Apx G-6. Map of the East Liverpool, OH, Case Study 526
Figure Apx G-7. Schematic of the Flow of Data within the EWISRD-XL-R Probabilistic Model 528
Figure Apx G-8. Distribution of Mean Annual Modeled Flow Rates for NHDPlus V2.1 Reaches
Identified Within 5 km of Hydraulic Fracturing Wells Reporting 1,4-Dioxane 530
Figure Apx G-9. Distribution of Modeled Ranges of 1,4-Dioxane Concentrations in Streams near
Hydraulic Fracturing Wells Reporting 1,4-Dioxane 531
Figure Apx G-10. Generic Schematic of Hypothetical Release Point with Surface Water Intakes for
Drinking Water Systems Located Downstream 532
FigureApx G-l 1. Summary Distribution of Mean Annual Flow at Stream Reaches Matched with
Drinking Water Intakes 534
Figure Apx J-l. Summary of Methodologies Used to Estimate Ambient Air Concentrations and
Exposures 549
Figure Apx J-2. Exposure Scenarios Modeled for Max and Mean Release Using IIOAC Model for
Ambient Air: Screening Methodology 551
Figure Apx J-3. Modeled Receptor Locations for Finite Distance Rings 555
Figure_Apx J-4. Modeled Receptor Locations for Area Distance 556
Figure Apx J-5. Example of Group of Air Releasing Facilities with Overlapping 10 km Buffers for
Aggregate Air Risk Screening 561
Figure Apx J-6. Decision Tree for Characterizing Aggregate Air Risk for Multiple Facilities 563
Figure_Apx J-7. Map of Aggregated Air Facilities, Group 1 564
Figure_Apx J-8. Map of Aggregated Air Facilities, Group 2 565
Figure_Apx J-9. Map of Aggregated Air Facilities, Group 3 565
Figure_Apx J-10. Map of Aggregated Air Facilities, Group 4 566
Figure_Apx J-l 1. Map of Aggregated Air Facilities, Group 5 566
Page 18 of 570
-------
ACKNOWLEDGEMENTS
This report was developed by the United States Environmental Protection Agency (EPA or the Agency),
Office of Chemical Safety and Pollution Prevention (OCSPP), Office of Pollution Prevention and Toxics
(OPPT).
Acknowledgements
The Assessment Team gratefully acknowledges the participation, input, and review comments from
OPPT and OCSPP senior managers and advisors. Acknowledgement is also given for the contributions
of interagency reviewers that included multiple federal agencies and assistance provided from EPA
contractors ERG (Contract No. 68HERD20A0002), ICF (Contract No. EP-W-12-010), and Versar
(Contract No. EP-W-17-006). Special acknowledgement is given for the contributions of technical
experts from EPA's Office of Research and Development, including Daniel Dawson for initial
development of the 1,4-dioxane water model and Caroline Ring for input on probabilistic modeling
approaches for the water pathway.
Docket
Supporting information can be found in the public docket (Docket ID: EPA-HQ-OP] ).
Disclaimer
Reference herein to any specific commercial products, process, or service by trade name, trademark,
manufacturer, or otherwise does not constitute or imply its endorsement, recommendation, or favoring
by the United States Government.
Authors/Contributors
Jeffrey Morris (Division Director), Yvette Selby-Mohamadu (Deputy Division Director), Rochelle
Bohaty (Branch Chief/Management Lead), Susanna Wegner (Assessment Lead), Sarah Au, Rebecca
Feldman, Mark Gibson, Bryan Groza, Franklyn Hall, Lauren Knapp, Shannon Rebersak, Shawn
Shifflett, Adam Theising, Jason Todd, Kevin Vuilleumier, Cindy Wheeler, and Daniel Whitby
Executive Team
This supplement was reviewed and cleared for release by OPPT and OCSPP leadership, including senior
advisors Stan Barone, Jeff Dawson, Anna Lowit, and Ryan Schmit, as well as senior leaders Mark
Hartman (Deputy Office Director, OPPT) and Elissa Reaves (Office Director, OPPT).
Technical Support
Mark Gibson, S. Xiah Kragie, and Hillary Hollinger
Internal Review
This assessment was provided for review to scientists in EPA's Program and Region Offices, including
• Office of the Administrator/Office of Children's Health Protection
• Office of Air and Radiation
• Office of Chemical Safety and Pollution Prevention/Office of Pesticide Programs
• Office of General Council
• Office of Land and Emergency Management
• Office of Research and Development
• Office of Water
Page 19 of 570
-------
Sn 111111 :ir\ of Risk l-'imlings sind Support lor Kisk Dclcrminiition
1.4-l)ioxanc is a sokent used in a \aricl\ of commercial aiicl inckistrial iippliccilions in the I nited
Suites ll is alst> produced as a b\ prockicl in se\eral manufacturing processes aiicl may remain present
as a byproduct in consumer aiicl commercial products. including soaps, detergents. and cleaning
products I leallh effects of concern for l.4-dio\anc include cancer and effects in li\er and damage to
olfactory tissue (cells in\ol\ed in smell) People may be exposed to 1,4-dioxane through
occupational exposure, consumer products, or contact with water, land, or air where 1.4-dioxane has
been released to the en\ ironment from industrial and commercial sources or from consumer and
commercial products washed down the drain or disposed of in landlllls
The 2<>2<> risk e\alnation lor 1,4-dioxane e\ aluated risks from a range of occupational and consumer
uses, risks to aquatic species, and risks to the general population resulting from incidental
recreational contact with water, ll did not e\aluate ueneral population exposures to 1.4-dioxane in
drinking water or air and did not e\ aluate the lull ranue of exposures that might result IY0111 1,4-
dioxane produced as a byproduct
This document is a supplement to the 2<>2<> risk e\ aluation ll completes the Toxics Substances
Control Act (TSCA) risk e\ aluation for 1.4-dioxane by (I) more coniprehensk ely e\aluatinu risks
IV0111 1.4-dioxane present as a byproduct, and (2) e\aluatinu risks from ueneral population exposures
to 1.4-dioxane released to water, air. and land. This analysis identified cancer risk estimates higher
than I in |i).()oo(| |o ') lor a range of typical and high-end occupational exposures to 1.4-dioxane
produced as a byproduct ll also identified cancer risk estimates higher than I in I million (I in ")
for a ranue of ueneral population exposure scenarios associated with 1.4-dioxane in drinkinu water
son reed downstream of release sites and in air within I km of releasing facilities Although these risk
estimates include inherent uncertainties and the o\erall confidence in specific risk estimates \aries.
the analysis pro\ ides support I'or the Agency to make a determination about whether 1,4-dioxane
poses an unreasonable risk and to identify drkers of unreasonable risk among exposures for people
(I) with occupational exposure to 1.4-dioxane under some conditions of use. (2) who rely 011 sources
of drinkinu water located downstream of release sites, and (.1) breathing air near release sites.
I11 parallel to this supplement. I-PA is releasing an updated risk determination for 1.4-dioxane l-PA
has determined that 1.4-dioxane presents an unreasonable risk of injury to human health under the
conditions of use That determination is based 011 information presented in the 2<>2<> risk e\ aluation
for 1.4-dioxane as well as in this 2<>24 supplement Ik-cause the risk determination is based in part 011
information beyond the scope of this supplement, it is presented as a separate document The analysis
presented in this supplement supports findings of unreasonable risk to workers and the ueneral
population from drinkinu water exposure from some conditions of use.
EXECUTIVE SUMMARY
This document is a supplement to the Final Risk Evaluation for 1,4-Dioxane that was published
December 2020 (also referred to as the "2020 RE"). EPA conducted this supplemental analysis because
contrary to the law's requirement for TSCA risk evaluations to be carried out on the "chemical
substance" under the conditions of use (also referred to as COUs or TSCA COUs), the 2020 RE
excluded certain known human exposure pathways that are important to understanding the health
implications of exposure to 1,4-dioxane. This supplement completes EPA's risk evaluation on the
chemical substance and positions the Agency to comprehensively address identified unreasonable risks.
Page 20 of 570
-------
1,4-Dioxane is primarily used as a solvent in commercial and industrial applications. It can also be
produced as a byproduct of several common manufacturing processes, including but not limited to
ethoxylation processes used in the production of surfactants used in soaps and detergents and production
of polyethylene terephthalate (PET) plastics. Even though it is not intentionally added, 1,4-dioxane
produced as a byproduct might remain present in consumer and commercial products—including soaps
and detergents, cleaning products, antifreeze, textile dyes, and paints/lacquers. 1,4-Dioxane is released
to the environment from industrial and commercial releases and from consumer and commercial
products that are washed down the drain or disposed of in landfills. People may be exposed to 1,4-
dioxane through occupational exposure, consumer products, or contact with water, land, or air where
1,4-dioxane has been released to the environment. Health effects of concern for 1,4-dioxane include
cancer and adverse effects to the liver and nasal tissue.
The 2020 RE did not evaluate risks from two critical areas: (1) general population exposures to 1,4-
dioxane in drinking water or air, and (2) the full range of exposure that may result from 1,4-dioxane
produced as a byproduct. During review of the 2019 draft risk evaluation, peer reviewers and public
commenters raised concerns that failure to consider these exposure pathways could leave portions of the
population at risk. These concerns include the fact that 1,4-dioxane has been detected in drinking water
and is not readily removed through conventional water treatment. In addition, 1,4-dioxane produced as a
byproduct results in occupational exposures that were not evaluated in the 2020 RE. Finally, 1,4-dioxane
produced as a byproduct also contributes to 1,4-dioxane in drinking water through industrial releases to
water sources as well as down-the-drain (DTD) disposal of consumer and commercial products.
This supplement expands on the analysis of COUs in which 1,4-dioxane is present as a byproduct to
include additional COUs for which information is reasonably available and consider associated
occupational exposures, including PET manufacturing, ethoxylation processing, hydraulic fracturing,
industrial/commercial use of products containing 1,4-dioxane as a byproduct. This supplement also
evaluates risks to the general population—including potentially exposed or susceptible subpopulations
(PESS)—from exposure to 1,4-dioxane through drinking water or air resulting from all industrial
releases (including those resulting from 1,4-dioxane produced as a byproduct) as well as DTD releases
of consumer and commercial products.
EPA released a draft of this supplement in July 2023. The Agency's evaluation of additional human
exposure pathways included new methods and novel applications of existing methods that were subject
to peer review at a Science Advisory Committee on Chemicals (SACC) meeting in September 2023. In
addition, EPA received public comments on the 2023 draft supplement. This 2024 revised supplement
addresses public comments and SACC recommendations. Following public release of this supplement,
EPA will initiate steps to address unreasonable risks identified through its complete evaluation of 1,4-
dioxane.
For this supplement, EPA relied on the physical and chemical properties, chemical life cycle
information, environmental fate and transport information, and the hazard identification and dose-
response analysis presented in the 2020 RE. All hazard values used in this supplement were derived
from the points of departure (PODs) previously peer reviewed by the SACC and published in the 2020
RE. Some of the exposure scenarios evaluated in this analysis required duration adjustments to the
previously established hazard values; however, the underlying hazard endpoints and PODs remain the
same.
In this supplement, EPA evaluated cancer and non-cancer risks from occupational and general
population exposure scenarios using available modeling and/or monitoring information. The Agency
Page 21 of 570
-------
evaluated occupational exposures through inhalation and dermal contact under a range of industrial and
commercial uses, including scenarios where exposures are expected to be highest. EPA evaluated
general population exposures to 1,4-dioxane through drinking water and air that could result from
releases to surface water, groundwater, land, and air. To be protective of PES S and sentinel exposures,
EPA developed risk estimates for the scenarios, populations, and life stages with the highest levels of
potential exposure, including fenceline communities. The Agency also considered site-specific
exposures, such as combined or additive releases from multiple releasing facilities within a single air or
water exposure pathway.
Risks to Workers
EPA estimated cancer and non-cancer risks for a set of new occupational COUs where 1,4-dioxane is
present as a byproduct.
• Dermal Exposure: Dermal occupational exposure is expected to occur as a result of worker
activities such as transfer operations, application of 1,4-dioxane containing formulations, and the
cleaning of equipment. Cancer risk estimates for dermal exposures range from 8.1 x 1CT7 to
7.3x 1CT3 for central tendency exposure and from 5.0/10 6 to 2.8/10 2 for high-end exposures
across COUs. Overall confidence in risk estimates for occupational dermal exposures is medium
for all occupational exposure scenarios.
• Inhalation Exposure: Inhalation exposure to 1,4-dioxane is expected to occur based on scenario-
specific considerations described in the bulleted items below. Cancer risk estimates for
inhalation exposure range from 4.8/10 " to 1.9/10 4 for central tendency exposures and from
4.8x 10~10 to 7,4/ 10 3 for high-end exposures across COUs based on the distribution of exposure
estimates. Occupational exposure scenarios with the highest estimates of risk from inhalation
exposure are summarized below.
o PET Manufacturing: Workers may inhale 1,4-dioxane generated as a byproduct of PET
plastic manufacturing. Cancer risk estimates for inhalation exposure range from 2,8/10 4
for central tendency exposures to 2.9x10 3 for high-end exposures. There is uncertainty
regarding the risk estimates because the extent to which the monitoring data reflect
current practices is unknown. Overall confidence in risk estimates for PET plastic
manufacturing is medium to high.
o Hydraulic Fracturing Operations: 1,4-Dioxane inhalation exposures may occur during
hydraulic fracturing operations due to its documented presence in scale inhibitors,
additives, friction reducers, and surfactants used in fracturing fluid formulations. Cancer
risk estimates for inhalation exposure range from 2,2/ 10 6 for central tendency exposures
to2.5xl0~4 for high-end exposures. There is uncertainty regarding the model inputs used
to estimate exposures and the extent to which they reflect the actual distribution of
hydraulic fracturing occupational exposures and workplace practices. Overall confidence
in risk estimates for hydraulic fracturing operations is medium to high.
o Ethoxylation Processes: 1,4-Dioxane may be generated as a byproduct in ethoxylation
reactions during the manufacture of common surfactants that result in worker inhalation
exposure. Cancer risk estimates for inhalation exposure range from 2.1 x 10~4 for central
tendency exposures to 5.4xl0~4 for high-end exposures. There is uncertainty regarding
the risk estimates due to the low number of data points and age of certain data points.
There is also uncertainty in the worker activities covered by the monitoring data and
whether all foreseeable activities, corresponding exposures, and workplace operations are
represented. Overall confidence in risk estimates for ethoxylation processes is medium.
Page 22 of 570
-------
o Textile Dye: 1,4-Dioxane is present in textile dyes as an unintentional byproduct in
ethoxylated substances that may be used as a formulation component in textile dyes.
Cancer risk estimates for inhalation exposure range from 1,9x 10 4 for central tendency
exposures to 7.4x10 3 for high-end exposures. There is uncertainty regarding the risk
estimates due to the low number of data points and high number of non-detects. There is
also uncertainty in the worker activities covered by the monitoring data and whether all
foreseeable activities, corresponding exposures, and workplace operations are
represented. Overall confidence in risk estimates for textile dying is medium.
Risk to the General Population
Risks from Exposure through Drinking Water Sourcedfrom Surface Water: EPA estimated cancer and
non-cancer risks for a range of general population exposures to surface water used as drinking water.
1,4-Dioxane is not readily removed through typical wastewater or drinking water treatment processes.
Sources of 1,4-dioxane in surface water include direct and indirect industrial releases from COUs where
1,4-dioxane is manufactured, processed, or used; industrial COUs where 1,4-dioxane is present due to
production as a byproduct (including PET manufacturing, ethoxylation processes, and hydraulic
fracturing operations); and DTD releases of 1,4-dioxane present in consumer and commercial products.
Monitoring data demonstrate that 1,4-dioxane is present in some source water and finished drinking
water samples. Measured concentrations in finished drinking water samples resulted in cancer risk
estimates greater than 1 x 10~6 at the high-end of the distribution of monitoring samples. However,
available surface water monitoring datasets are not designed to reflect source water impacts of direct and
indirect releases into water bodies. Therefore, EPA estimated concentrations modeled for a range of
specific release scenarios. The Agency evaluated the performance of the models against monitoring data
from site-specific locations serving as cases studies. This evaluation demonstrated general agreement
between modeled concentrations and monitoring data, thereby increasing confidence in risk estimates
based on modeled concentrations.
EPA used modeled water concentrations to evaluate risks from a range of sources individually and in
aggregate {i.e., by evaluating risks from water concentrations resulting from multiple sources of 1,4-
dioxane releasing to the same water bodies). The Agency evaluated cancer risks for individuals exposed
through drinking water over 33 years as well as for individuals exposed for a full lifetime (78 years). For
each of the sources assessed, cancer risk estimates based on mean drinking water ingestion rates over 33
years of exposure to modeled concentrations in receiving water bodies at the point of release may
exceed 1x 10~6 or 1 x 10~4 under some conditions.
• Industrial Releases to Surface Water: Risk from individual facilities vary substantially within and
across COUs, with cancer risk estimates ranging from 5,4x10 13 to 2.5 x 10 2, Overall confidence
in risk estimates for specific facilities depends on confidence in facility-specific release data, but
confidence in the overall analysis is medium to high.
• Down-the-Drain Releases to Surface Water: EPA evaluated the conditions under which DTD
releases contribute to different levels of risk and identified plausible scenarios in which risks
from DTD releases result in risks greater than 1 in 1 million. Risk estimates from modeled DTD
releases are highest in locations where large populations contribute to these releases and where
they are discharged to streams with low flow. Overall confidence in this analysis is medium.
• Hydraulic Fracturing Releases to Surface Water: Cancer risk estimates from modeled hydraulic
fracturing waste releases to surface water are 3.85xl0~8 for median modeled releases and
l.SlxlO-6 for 95th percentile modeled releases. Overall confidence in this analysis is medium.
Page 23 of 570
-------
• Aggregate Releases to Surface Water: Probabilistic modeling provides a distribution of risk
estimates reflecting a range of drinking water scenarios that account for aggregate sources of 1,4-
dioxane in water. Confidence in risk estimates for specific facilities depends on confidence in
facility-specific release data used as model inputs, and overall confidence in this aggregate
analysis is medium to high.
The degree of 1,4-dioxane dilution that occurs between the point of release and the point of drinking
water intakes is highly variable and site-specific; therefore, it is a source of uncertainty in the analysis.
EPA used two different methods to estimate the impact of downstream dilution on risk estimates and
found that, under some circumstances, lifetime cancer risk remained above 10 6 at drinking water
intakes located downstream from industrial releases.
The impacts of longer exposure durations or higher drinking water ingestion rates were also assessed in
the revised supplement and result in greater exposure and therefore risk. Individuals exposed over a full
lifetime (78 years) could have exposure and risk approximately 2.3 times greater than those calculated
for 33 years of exposure. Because some people may live in a community near 1,4-dioxane releases for
longer durations, EPA agrees with the peer review recommendation to utilize a full lifetime of exposure
for assessing lifetime cancer risks for fenceline communities. Lifetime cancer risk estimates based on
95th percentile drinking water ingestion rates could result in 3 to 4 times higher exposures and risks than
those based on mean ingestion rates, depending on the age groups exposed. Although consideration of
alternate exposure factors such as lifetime and ingestion rates result in increased risks of less than an
order of magnitude, where the original estimates are close to the applicable benchmark, this could result
in changes to overall risk conclusions.
Risks from Exposure through Drinking Water Sourcedfrom Groundwater: EPA estimated cancer and
non-cancer risks for a range of general population exposures to groundwater used as drinking water.
Sources of 1,4-dioxane in groundwater include leachate from landfills and disposal of hydraulic
fracturing waste. DTD releases to septic fields from consumer and commercial products containing 1,4-
dioxane, as well as historical disposals of 1,4-dioxane, are other potential sources of groundwater
contamination; however, these were not considered in this assessment. Overall confidence in these risk
estimates is low to medium.
• 1,4-Dioxane in Groundwater from Hydraulic Fracturing: Cancer risk estimates for people
exposed to modeled groundwater concentrations over 33 years are 4.Ox 1CT7 for median modeled
releases and 8.6xl0~6for 95th percentile modeled releases.
• 1,4-Dioxane in Groundwater Resulting from Landfill Leachate: Cancer risk estimates increase
under scenarios with higher leachate concentrations and loading rates. Monitoring data for
groundwater contamination surrounding landfills were not readily available for comparison.
Risk from Exposure through Air: EPA estimated cancer and non-cancer risks for a range of general
population exposures to 1,4-dioxane in air. 1,4-Dioxane concentrations in air depend on the facility-
specific release amount, stack height(s), topography, and meteorological conditions—not on specific
COUs.
Potential sources of 1,4-dioxane in air include industrial releases reported to Toxics Release Inventory
(TRI), fugitive emissions from hydraulic fracturing, and emissions from institutional and industrial
laundries. The highest estimated risks occurred within 1,000 m of industrial release sites. EPA also
estimated risk from the aggregate exposures from multiple facilities releasing 1,4-dioxane in proximity
to fenceline communities. This aggregate analysis did not identify locations with aggregate cancer risk
Page 24 of 570
-------
greater than 1 x 10 6 that did not already have cancer risk above that level from an individual facility and
therefore did not have a substantial impact on the overall findings.
• Air Releases Reported to TRI: Cancer risk estimates based on 33- and 78-year exposure
durations and 95th percentile modeled air concentrations within 1,000 m (approximately 0.6
mile) of the highest risk facilities in each COU range from 1,0x 10~10 to 1.1x 10~4 for 33 years of
exposure and from 2,4/10 10 to 2,6/10 4 for 78 years of exposure. Cancer risk estimates based
on 33-year exposure duration and 50th percentile modeled exposure concentrations within 1,000
m of the highest risk facilities range from 2.5/10 " to 8.3/10 5 for 33 years of exposure and
from 5.9/10 " to 1.9/10 4 for 78 years of exposure. Although individual risk estimates for
specific locations should be interpreted with caution, most estimates are informed by moderate to
robust modeling approaches and input data. Overall confidence in risk estimates for inhalation
exposures resulting for air concentrations modeled for industrial releases ranges from low to
high, depending on the level of confidence in release information underlying risk estimates for
specific facilities and COUs.
• Fugitive Air Emissions from Hydraulic Fracturing Operations: Cancer risk estimates based on
33-year exposure duration within 1,000 m of hydraulic fracturing operations range from 2,2/ 10 8
to 7,1 / 10 5 for a range of air model scenarios across a range of high-end (95th percentile) and
central tendency release scenarios. Overall confidence in risk estimates for inhalation exposures
resulting for air concentrations modeled based on releases from hydraulic fracturing operations is
medium.
• Emissions from Industrial and Institutional Laundries: Cancer risk estimates based on 33 year-
exposure duration within 1,000 m of industrial and institutional laundries range from 1,5/10 "
to 3,8/10 8 across a range of high-end exposure scenarios. Overall confidence in risk estimates
from inhalation exposures resulting from industrial and institutional laundries is medium.
Unreasonable Risk Determination
In parallel to this supplement, EPA is releasing an updated risk determination for 1,4-dioxane. The
Agency has determined that 1,4-dioxane presents an unreasonable risk of injury to health under the
conditions of use. This determination is based on the information in the 2020 RE and this 2024
Supplement to the Risk Evaluation for 1,4-Dioxane, including the appendices and supporting documents
(see Appendix B). Because the risk determination is based in part on information beyond the scope of
this supplement, it is presented as a separate document. The analysis presented in this supplement
supports findings that the following COUs contribute to unreasonable risks for 1,4-dioxane:
• Manufacture (including domestic manufacture and import)
• Processing (including repackaging, recycling, non-incorporative, as a reactant, and as a
byproduct, including ethoxylation processing and polyethylene terephthalate [PET]
manufacturing)
• Industrial/commercial use: Intermediate
• Industrial/commercial use: Processing aid
• Industrial/commercial use: Other uses: Hydraulic fracturing
• Industrial/commercial use: Arts, crafts, and hobby materials: Textile dye
• Industrial/commercial use: Cleaning and furniture care products: Surface cleaner
• Industrial/commercial use: Laundry and dishwashing products: Dish soap
• Industrial/commercial use: Laundry and dishwashing products: Dishwasher detergent
• Industrial/commercial use: Laundry and dishwashing products: Laundry detergent
• Industrial/commercial use: Paints and coatings: Paint and floor lacquer
• Consumer use: Cleaning and furniture care products: Surface cleaner
Page 25 of 570
-------
• Consumer use: Laundry and dishwashing products: Dish soap
• Consumer use: Laundry and dishwashing products: Dishwasher detergent
• Consumer use: Laundry and dishwashing products: Laundry detergent
• Consumer use: Paints and coatings: Paint and floor lacquer
• Disposal
Analysis presented in the 2020 risk evaluation further supports the unreasonable risk conclusions for
some of the above COUs as well as other COUs not identified here.
Page 26 of 570
-------
1 INTRODUCTION
1,4-Dioxane was one of the first 10 chemicals to undergo the Toxics Substances Control Act (TSCA)
risk evaluation process following passage of the 2016 Frank R. Lautenberg Chemical Safety for the 21st
Century Act, which amended TSCA. 1,4-Dioxane is primarily used as a solvent in a variety of
commercial and industrial applications such as the manufacture of other chemicals (e.g., adhesives,
sealants) or as a processing aid or laboratory chemical. It is produced as a byproduct in several
manufacturing processes, including ethoxylation, sulfonation, sulfation, and esterification. Although
there are no direct consumer uses, 1,4-dioxane produced as a byproduct in the aforementioned processes
can be present in commercial and consumer products, including soaps, detergents, and cleaning
products. Use of these products may result in direct occupational and consumer exposures. Disposal of
these products down-the-drain (DTD) may contribute to general population exposure to 1,4-dioxane
present in some U.S. surface waters.
1.1 Regulatory Context
In the 2019 draft 1,4-dioxane risk evaluation, EPA reviewed the exposures, hazards, and risks of 1,4-
dioxane from occupational exposures and surface water exposures to environmental organisms. It also
included the physical and chemical properties, lifecycle information, environmental fate and transport
information, and hazard identification and dose-response analysis. However, the 2019 draft risk
evaluation excluded general population exposures through drinking water and air and conditions of use
(also referred to as COUs or TSCA COUs) in which 1,4-dioxane is present as a byproduct. These
exclusions were based in part on an interpretation that EPA had broad discretionary authority under
TSCA to categorically exclude conditions of use from the scope of its evaluations, and, as described in
the 2018 1,4-Dioxane Problem Formulation} that certain exposure pathways need not be considered if
they were under the jurisdiction of other EPA regulatory programs or analytical processes.
These analyses were reviewed by the Science Advisory Committee on Chemicals (SACC2) in 2019. The
SACC raised a number of concerns regarding the evaluation and approach, but particularly noted its
concerns about the Problem Formulation straying from "basic principles of risk assessment," the
omission of well-known exposure routes, and that general lack of comprehensiveness undermining
EPA's ability to protect against risks to human health and the environment. As stated in the meeting
minutes and final report3 from the July 2019 SACC meeting, "there was general dissatisfaction in the
Committee that the human health risk characterization did not extend to the general population since
there was no indication in the Evaluation that other offices in the EPA had plans to conduct such a
characterization." Furthermore, "several committee members also observed that failure to assess 1,4-
dioxane exposure in the general population may leave substantial portions of the population at risk. This
is particularly concerning for drinking water." The SACC also raised concerns about potential risks from
1,4-dioxane produced as a byproduct, recommending that "EPA should provide a detailed discussion of
the scientific basis for the exclusion of impurity or byproduct formation of 1,4-dioxane."
Public stakeholders also raised concerns about water monitoring data demonstrating the presence of 1,4-
dioxane in drinking water. Commenters also identified additional sources of 1,4-dioxane that had not
1 The 1,4-Dioxane Problem Formulation is available at https://www.epa.gov/assessing-and-managing-eheniieats-iiiider-
tsca/14-dioxane-problem-fonnnlation.
2 Additional information about SACC is available at https://www.epa. gov/tsca-peer-review/science-advisorv-cominittee-
chemicals-basic-information.
3 The SACC July 2019 meeting minutes and final report (Document ID EPA-HQ-OPPT-2019-0237-0064) are available at
https://www. regulations. gov/doeument/EP A-HO-OPPT-2019-0237-00o4.
Page 27 of 570
-------
been assessed, including COUs in which 1,4-dioxane is produced as a byproduct, and raised concerns
that some of these COUs may be an important source of exposure to 1,4-dioxane in water.
In November of 2020, EPA released a supplement to the draft 1,4-dioxane risk evaluation for public
comment. The November 2020 supplement to the draft assessed eight additional COUs of 1,4-dioxane
as a byproduct in consumer products and general population exposure from incidental contact with
surface water. The EPA determined that the additional analysis did not warrant SACC review as no
novel science was utilized.
In December 2020, the Agency published the Final Risk Evaluation for 1,4-Dioxane (also referred to as
the "2020 RE") v * * \ ,020c). The 2019 draft and 2020 supplement were both incorporated into
the 2020 RE, which assessed risks for
• worker and occupational non-user (ONU) exposures to 1,4-dioxane through 16 industrial and
commercial COUs;
• consumer and bystander exposures to 1,4-dioxane present as a byproduct4 in eight consumer
product categories;
• general population exposure via incidental/recreational contact with 1,4-dioxane present in
surface water from industrial releases; and
• aquatic species' exposures to 1,4-dioxane present in surface water.
In January 2021, the White House issued Executive Order 13990 instructing that the federal government
be guided by the best science and be protected by processes that ensure the integrity of federal decision-
making, and established the Administration's policy of, among other concerns, following the science,
improving public health and protecting the environment, limiting exposure to dangerous chemicals, and
prioritizing environmental justice when delivering on these concerns. Executive Order 13990 also
instructs agencies to (1) review actions issued between January 20, 2017, and January 20, 2021, which
may be inconsistent with or present obstacles to implementing the policy established in the order and;
(2) consider suspending, revising, or rescinding such actions.
Upon further review, EPA determined that the approach taken in the 2020 RE {i.e., the exclusion of
reasonably foreseeable exposures to workers, as well as exposures to the general population from air,
water, and disposal) was inconsistent with the plain language of TSCA section 6 and left potential
risks—including risks to potentially exposed and susceptible subpopulations (PESS)—unaccounted for.
The law's requirement that EPA conduct risk evaluation on a "chemical substance" under the COUs
requires the Agency to determine the chemical's COUs and to not otherwise exclude those COUs from
the scope of the risk evaluation.
In June of 2021, EPA announced that additional analysis was needed to consider critical exposure
pathways not assessed in the final risk evaluations for the first 10 chemicals (including, but not limited
to, ambient air, ambient water, and drinking water). For many of the first 10 risk evaluations, EPA
applied the Draft Screening Level Approach for Assessing Ambient Air and Water Exposures to
Fence line Communities Version 1.0. published in January 2022 to determine whether further analysis
was needed. For 1,4-dioxane, however, EPA determined that a more in-depth analysis was needed to
address concerns about known drinking water contamination (described in Section 1.3.1.3) and to more
fully evaluate COUs in which 1,4-dioxane is present as a byproduct (described in Section 1.3.1.1), and
signaled its intention to re-open and formally supplement the 1,4-dioxane risk evaluation.
4 Byproduct means a chemical substance produced without a separate commercial intent during the manufacture, processing,
use, or disposal of another chemical substance(s) or mixture(s).
Page 28 of 570
-------
This 2024 supplement to the 2020 RE is intended to complete the risk evaluation on the chemical
substance 1,4-dioxane as required under TSCA by (1) expanding the analysis of COUs in which 1,4-
dioxane is present as a byproduct to include additional COUs and consider associated occupational
exposures; and (2) evaluating risks from general population exposures to 1,4-dioxane released to surface
and groundwater, air, and land.
EPA's evaluation of these additional human exposure pathways included new methods and novel
applications of existing methods. This supplement is the first under amended TSCA to evaluate:
exposures and risks from a chemical produced as a byproduct, aggregate risks for communities relying
on drinking water sourced from surface water receiving a chemical from multiple sources, risks for
communities relying on drinking water sourced from groundwater, aggregate risks for communities
exposed through air near multiple release sites, and consideration of multiple years of environmental
release data.
In July 2023, EPA released a draft of this supplement and a draft update to the risk determination for
1,4-dioxane. The new methods and novel applications of existing methods included in this supplement
were subject to peer review at a SACC meeting in September 2023. In addition, EPA received public
comments on the draft of this supplement. The Agency considered all SACC recommendations and
public comments. EPA provides responses to major comments in a response to comment document
accompanying this revised supplement. Major revisions made to this revised supplement in response to
SACC and public comment include:
• Revisions to Occupational Exposure and Risk Estimates: As detailed in Appendix K, exposure
and risk estimates for some COUs were revised based on revisions to Monte Carlo models,
revised model input assumptions, and/or incorporation of additional data recommended by the
SACC or submitted through public comment. In some cases, these revisions increased or
decreased risk estimates by up to an order of magnitude. For other COUs, these revisions had no
quantitative impact on risk estimates.
• Revisions to Release Assessments: As detailed in Appendix K, EPA revised release estimates for
some COUs based on revised Monte Carlo models and alternate input assumptions. For
hydraulic fracturing releases to surface water, the revised release estimates were used to generate
revised exposure and risk estimates. For other revised release estimates, EPA did not revise the
corresponding exposure and risk estimates because the magnitude of the change was not
expected to be sufficient to alter overall risk conclusions.
• Consideration of Alternate Exposure Factors: Although EPA retained risk estimates based on
original exposure assumptions, the revised supplement discusses the extent to which alternate
assumptions about exposure amount and duration would increase risk estimates. For example,
while EPA originally assessed risks for the general population associated with 33 years of
exposure to 1,4-dioxane through air or water, the revised supplement includes consideration of
risks resulting from a full lifetime (78 years) of exposure.
• Consideration of Aggregate Risk across Routes: Although EPA retained risk estimates based on
individual routes, the revised supplement discusses the extent to which aggregation across routes
would alter risk conclusions.
• Clarifications on Methods: In an effort to improve clarity and transparency in response to
comments on a range of topics, EPA made revisions to the narrative throughout this revised
supplement by providing more detailed explanation of methodologies, approaches, and
assumptions.
Page 29 of 570
-------
In tandem with this revised supplement, EPA is also releasing a revised risk determination for 1,4-
dioxane. The revised risk determination takes into account revisions made to this revised supplement as
well as public comments received on the 2023 draft risk determination. Following release of this revised
supplement and the revised risk determination, EPA will initiate steps to address unreasonable risks
identified through its complete evaluation of 1,4-dioxane.
1.2 Scope
This supplement is intended to evaluate risks from exposure pathways and COUs for 1,4-dioxane that
were not assessed in the 2020 RE. Additional exposure pathways and new COUs included in this
supplement were identified based on information submitted in previous public comments and other
reasonably available information. For the current analysis, EPA is relying on the physical and chemical
properties, as well as lifecycle information, environmental fate and transport information, and hazard
identification and dose-response analysis presented in the 2020 RE (Sections 1.1, 1.4, 2.1, and 3.2 of the
2020 RE, respectively). Furthermore, this supplement does not re-evaluate the occupational, consumer,
or ecological exposure pathways and risks that were previously assessed in the 2020 RE.
This supplement more fully evaluates COUs in which 1,4-dioxane is present as a byproduct (described
in Section 1.3.1.1). Specifically, EPA considered 1,4-dioxane present as a byproduct in commercial
products (corresponding to the consumer products considered in the 2020 RE). The Agency also
identified a new set of COUs, based on reasonably available information, where 1,4-dioxane is produced
or present as a byproduct—including ethoxylation processing, polyethylene terephthalate (PET)
manufacturing, and hydraulic fracturing. A more detailed list of the new COUs and COU subcategories
considered in this supplement is presented in Section 2.1.1.
This supplement to the 2020 RE evaluates risks for the following exposure pathways:
• Occupational exposure to
o 1,4-dioxane present as a byproduct in commercial products (corresponding to consumer
products considered in the 2020 RE); and
o 1,4-dioxane produced or present as a byproduct in additional industrial COUs for which
information on the presence of 1,4-dioxane is reasonably available, including
ethoxylation processing, PET manufacturing, and hydraulic fracturing (Sections 3.1,
5.2.1).
• General population exposures to
o 1,4-dioxane present in drinking water sourced from surface water as a result of all direct
and indirect industrial releases and DTD releases of consumer and commercial products
(Sections 2.3.1, 3.2.2 and 5.2.2.1);
o 1,4-dioxane present in drinking water sourced from groundwater contaminated as a result
of disposals (Sections 2.3.2, 3.2.2.2 and 5.2.2.1.6); and,
o 1,4-dioxane released to air from industrial and commercial sources (Sections 2.3.3, 3.2.3,
and 5.2.2.3).
Many of the COUs assessed in this supplement contribute to more than one exposure pathway. For
example, 1,4-dioxane present as a byproduct of PET manufacturing may contribute to occupational
exposures during manufacturing as well as general population exposures through releases to water and
air. In addition, for many of the exposure pathways assessed, multiple COUs contribute to 1,4-dioxane
exposure. For example, many COUs can contribute to general population exposures to 1,4-dioxane in
surface water, including industrial releases from a range of COUs and DTD releases of consumer and
commercial products. In this supplement, EPA evaluated general population exposures resulting from
each type of known releases, including releases associated with COUs evaluated in the 2020 RE and
Page 30 of 570
-------
releases associated with new COUs introduced in this supplement due to the presence of 1,4-dioxane
produced as a byproduct.
1.3 Use Characterization
1.3.1 Conceptual Models
The life cycle diagram for 1,4-dioxane in Figure 1-1 summarizes the conditions of use that are within the
combined scope of the 2020 RE and the current supplement. The life cycle diagram has been updated
from the 2020 RE to highlight additional sources of 1,4-dioxane produced as a byproduct, including
commercial products and industrial uses, releases, and disposals (e.g., PET manufacturing, ethoxylation
byproducts, disposal of hydraulic fracturing produced waters).
Page 31 of 570
-------
MFG/IMPORT
PROCESSING
Manufacture
(includes import)
(i million lbs.)
Processing as a
Rea eta nt/lnter mediate
(Not reported in 2016 CDR)
Repackaging
(270,000 lbs.)
Non-incorporative
Activities
(270,00 lbs.)
Recycling
INDUSTRIAL, COMMERCIAL,
CONSUMER USES
RELEASES and
WASTE DISPOSAL
Industrial Uses
Consumer Uses
{Present only as Byproduct)
Processing Aids, Not
Otherwise Listed
(270,000 lbs.)
e.g., wood pulping, etching
of fluropolymers
Functional Fluids
(Open and Closed Systems)
(<150,000 lbs.)
e.g., hydraulic fluid.
Laboratory Chemicals
(<150,000 lbs.)
e.g., laboratory reagent
Adhesive Sealants
e.g., film cement
Other Industrial Uses
e.g., spray polyurethane
foam, printing and printing
compositions; dry film
lubricant
Paints and Coatings
e.g., Latex Wall Paint or
Floor Lacquer
Cleaning and Furniture
Care Products
e.g.. Surface Cleaner
Laundry and Dishwashing
Products
e.g., Dish Soap, Dishwasher
Detergent, Laundry
Pete rgent
Arts, Crafts and Hobby
Materials
e.g.. Textile Dye
Automotive Care Products
e.g., Antifreeze
Commercial Uses
(Present only as Byproduct)
Paints and Coatings
e.g.. Latex Wall Paint or
Floor Lacquer
Clea ning a nid Fu miture
Care Products
e.g., Surface Cleaner
Other Consumer Uses
e.g., Spray Polyurethane
Foam, Antifreeze
Laundry and Dishwashing
Products
e.g.. Dish Soap, Dishwasher
Detergent, Laundry
Detergent
Arts, Crafts and Hobby
Materials
e.g.. Textile Dye
Automotive Care Products
e.g.. Antifreeze
Other Consumer Uses
e.g.. Spray Polyurethane
Foam, Antifreeze
Industrial uses including 1,4-dioxane as a byproduct
(e.g., ethoxylation processes, PET manufacturing, and hydraulic fracturing operations)
Disposal
Byproduct
U ses from
Manufacturing
J
Assessed in the Supplement to
the Risk Evaluation
Figure 1-1. 1,4-Dioxane Life Cycle Diagram
Note: This life cycle diagram has been expanded from what was published in the 2020 RE to include additional sources of 1,4-dioxane produced as a
byproduct (indicated in blue boxes). See Appendix D for a complete table of COUs considered in the 2020 RE and this supplement.
Page 32 of 570
-------
1.3.1.1 1,4-Dioxane as a Byproduct
1,4-Dioxane produced as a byproduct may be a source of exposure for several of the occupational and
general population exposure pathways evaluated in this supplement. Byproduct refers to a chemical
substance produced without a separate commercial intent during the manufacture, processing, use, or
disposal of another chemical substance(s) or mixture(s). In this assessment, the term byproduct is used
to refer to 1,4-dioxane produced during manufacturing or industrial processes, including 1,4-dioxane
that remains present in downstream processes or in consumer and commercial products.
In the 2020 RE, EPA evaluated risks to consumers and bystanders from 1,4-dioxane present as a
byproduct in consumer products. In this supplement, EPA expanded on the previous evaluation to
consider risks from all other pathways of exposure to 1,4-dioxane produced as a byproduct for which
information is reasonably available. Figure 1-2 summarizes both what is known about the processes that
may result in 1,4-dioxane production and how it may contribute to human exposures through a range of
exposure pathways.
Source
chemicals
Occupational
exposure
Chemical
processes
..in many types of
manufacturing
facilities
(ethoxylation,
esterification,
sulfonation,
sulfation)
Direct and indirect*
industrial releases to
air and water
General population
exposure through air
or drinking water
1,4-dioxane
produced as a
byproduct of
manufacturing
'I
Removal of 1,4-dioxane
prior to product
formulation and releases
Consumer and
occupational
exposure ^
t
Consumer and
commercial products
Down-the-drain
releases*
*these releases may also go to wastewater treatment plants
Figure 1-2. Production of 1,4-Dioxane as a Byproduct and Potential Exposure Pathways
1,4-Dioxane is produced as a byproduct in several common manufacturing reactions, including in
manufacturing of PET plastics and in ethoxylation reactions during the manufacture of common
surfactants. In some facilities, additional processing steps may remove 1,4-dioxane produced as a
byproduct prior to product formulation and environmental releases, but the full extent of this practice
across industries is not known. Occupational exposure to 1,4-dioxane produced as a byproduct may
occur at manufacturing facilities and hydraulic fracturing operations. Releases of 1,4-dioxane from
manufacturing and industrial sites may also contribute to general population exposures through drinking
water and air.
1,4-Dioxane produced as a byproduct has also been detected in consumer and commercial products,
resulting in potential exposure to consumers and bystanders (evaluated in the 2020 RE) or workers and
ONUs (evaluated as described in Section 3.1.2). For example, dermal and/or inhalation exposures to 1,4-
dioxane are expected for workers during the use of dish soap and dishwashing detergent from unloading
and transferring detergent formulation, transport container cleaning, and washing operations due to the
presence of 1,4-dioxane as a surfactant byproduct. In addition, consumer and commercial products
Page 33 of 570
-------
containing 1,4-dioxane may contribute to general population exposures through drinking water when
released DTD.
In this supplement, EPA evaluated pathways of exposure to 1,4-dioxane produced as a byproduct that
were not previously assessed. Specifically, the Agency considered 1,4-dioxane present as a byproduct in
commercial products (corresponding to the consumer products considered in the 2020 RE). EPA
considered the direct occupational exposures that result from use of these commercial products as well
as the DTD releases of consumer and commercial products, which contribute to general population
exposures through surface water. EPA also identified a new set of COUs where 1,4-dioxane is produced
or present as a byproduct based on information submitted by public commenters and other reasonably
available information. For each of these new COUs, the Agency evaluated occupational exposure as
well as industrial releases that contribute to general population exposures via drinking water and air. The
available information supporting inclusion of each of the new COUs is described below. A more
detailed list of the new COUs and COU subcategories considered in this supplement is presented in
Section 2.1.1.
The following COUs are known to produce 1,4-dioxane as a byproduct based on reasonably available
information, but 1,4-dioxane produced as a byproduct may also be present in other industries that have
not yet been identified:
• Industrial/Commercial Use of Products Containing 1,4-Dioxane as a Byproduct: 1,4-Dioxane
is present in a range of commercial products (including textile dyes, antifreeze, surface cleaners,
dish soaps, laundry detergents, and paint and floor lacquer) because it is produced as a byproduct
during the manufacture of ingredients such as ethoxylated surfactants. While 1,4-dioxane present
as a byproduct in consumer products was previously assessed, evaluation of 1,4-dioxane in these
corresponding commercial products is new in this supplement. 1,4-Dioxane present in
commercial products can result in occupational exposure in commercial settings as well as DTD
releases that contribute to general population exposures via drinking water.
• PET Manufacturing: 1,4-Dioxane has been identified as a byproduct in the manufacture of PET
( ). EPA does not have information on the byproduct concentration of 1,4-
dioxane in PET, which is produced by the esterification of terephthalic acid to form
bishydroxyethyl terephthalate (BHET) (Forkner et at.. 2004). BHET polymerizes in a
transesterification reaction catalyzed by antimony oxide to form PET (Forkner et at.. 2004). In
2014, 20.6 million metric tons of PET were used in the United States (McDaniel and
DesLauriers. 2015). 1,4-Dioxane produced during PET manufacturing may result in occupational
exposures and may contribute to general population exposures via releases to water and air.
• Ethoxylation Processing: 1,4-Dioxane may be formed as a byproduct of reactions based on
condensing ethylene oxide or ethylene glycol during manufacture of detergents, shampoos,
surfactants, some food additives, and certain pharmaceuticals (HHS. 2016). In cosmetic
ethoxylated raw materials and ethoxylated alkyl sulfates, 1,4-dioxane has been detected at
concentrations of 0.48 to 1,410 ppm (\ c. « i1 \ J020c; Saraii and Shirvani. IVI ; Oavarani et
at.. 2012; Stack et at.. 2001). Polyethoxylated raw materials are widely used in cosmetic
products as emulsifiers, foaming agents, and dispersants (Stack et at.. 2001). They are produced
by polymerizing ethylene oxide, usually with a fatty alcohol, to form polyethoxylated alcohols
that may be used to synthesize other products such as sulfated surface-active agent. During the
ethoxylation process, 1,4-dioxane can be formed as a byproduct by the dimerization of ethylene
oxide (Stack et at.. 2001). The volume of 1,4-dioxane produced as a byproduct of ethoxylation
reactions is unknown. 1,4-Dioxane produced during ethoxylation processing may result in
Page 34 of 570
-------
occupational exposures and may contribute to general population exposures via releases to water
and air. It also contributes to the presence of 1,4-dioxane in consumer and commercial products.
• Hydraulic Fracturing: Hydraulic fracturing stimulates an existing oil or gas well by injecting a
pressurized fluid containing chemical additives into the well ( 2022c). 1.4-Dioxane is
measured in fracturing fluid, a water-based fluid that contains several chemical additives and in
waste fluid (produced waters). FracFocus 3.0 contains self-reported information indicating that
1,4-dioxane is present in hydraulic fracturing fluid additives, as scale inhibitors, additives,
biocides, friction reducers, and surfactants (GWPC and IOGCC. 2022). According to the
FracFocus 3.0 database, 1,4-dioxane is present in weight fractions ranging from 2.3 xlO~u to 0.05
within hydraulic fracturing additives and 1.00/10 12 to 4,30/10 6 in hydraulic fracturing fluids
(GWPC and IOGCC. 2022). 1,4-Dioxane has been documented to have a concentration of 60
|.ig/L in hydraulic fracturing produced waters (Lester et at.. 2015). The presence of 1,4-dioxane
in fracturing fluid may result in occupational exposures. It may also contribute to general
population exposures via discharge to surface water, groundwater, or fugitive air emissions from
fracturing operations.
1.3.1.2 Occupational Exposures
The conceptual model in Figure 1-3 presents the exposure pathways, exposure routes, and hazards to
people from industrial and commercial releases and uses of 1,4-dioxane. Blue shading highlights the
exposures evaluated in this supplement. Workers and ONUs may have acute (8-hour) or chronic (annual
to lifetime) exposures to 1,4-dioxane produced as a byproduct during PET manufacturing, ethoxylation
processes, or hydraulic fracturing operations. Workers and ONUs may also have acute or chronic
exposures to 1,4-dioxane present as a byproduct in commercial products, including detergents, cleaners,
and lacquers.
Page 35 of 570
-------
INDUSTRIAL AND COMMERCIAL ACTIVITIES/USE EXPOSURE PATHWAY
EXPOSURE ROUTE &
DURATION
RECEPTORS
HAZARDS
INDUSTRIAL USES
Manufacture
(Including Import)
Processing:
• Processing as a
reactant/i ntermediate
• Repackaging
• Non-incorporative
activities
Recycling
Processing Aid, Not
Otherwise Listed
Functional Fluids
(Open and Closed Systems)
Laboratory Chemicals
Other Industrial Uses
1,4-Dioxane as a Byproduct
COMMERCIAL USES
(Present only as Byproduct)
Paints and Coatings
e.g., Latex Wall Paint or
Floor Lacquer
Cleaning and Furniture
Care Products
e.g., Surface Cleaner
Laundry and Dishwashing
Products
e.g., Dish Soap, Dishwasher
Detergent, Laundry
Detergent
Arts, Crafts and Hobby
Materials
e.g., Textile Dye
Automotive Care Products
e.g., Antifreeze
Other Consumer Uses
e.g., Spray Polyurethane
Foam, Antifreeze
T
Waste Handling,
Treatment, and Disposal
-~ Liquid Contact
Dermal
(acute dose, chronic
ADD, LADD)
Hazards Potentially
Associated with
Acute and/or
Chronic Exposures
Vapor/Mist
i
L
l Fug
Emis
tive
sions
Inhalation
(8hrTWA, chronic
ADC, LA DC)
Wastewater, Liquid Wastes, Solid
Wastes
Assessed in the Supplement to
the Risk Evaluation
Figure 1-3. Conceptual Model for Occupational Exposures from Industrial and Commercial Activities
Page 36 of 570
-------
In this supplement, EPA evaluated acute and chronic cancer and non-cancer risks from occupational
inhalation and dermal exposures to 1,4-dioxane produced as a byproduct during PET manufacturing and
ethoxylation processes, hydraulic fracturing fluids and waste containing 1,4-dioxane, and commercial
products containing 1,4-dioxane.
1.3.1.3 General Population Exposures
The conceptual model in Figure 1-4 presents general population exposure pathways and hazards from
environmental releases and wastes associated with COUs (red, blue, and purple shading for each source
in the figure corresponds to the environmental media to which they release). The disposal and release
scenarios illustrated in Figure 1-1, Figure 1-2, and Figure 1-3 all contribute to the releases to air, water
and land that may result in the general population exposures illustrated in Figure 1-4. The general
population may be exposed to 1,4-dioxane released to surface water (blue shading), groundwater (red
shading), and air (purple shading). Drinking water exposures are evaluated based on releases to both
surface water and groundwater and these are each described in the drinking water subsection below.
EPA's evaluation of general population exposures considers potentially exposed or susceptible
subpopulations (PESS). Exposures to 1,4-dioxane through air and water could result in risk to fenceline
communities. As defined in the Draft Screening Level Approach for Assessing Ambient Air and Water
Exposures to Fenceline Communities Version 1.0 ( 2d) (also referred to as the "2022
Fenceline Report") fenceline communities are members of the general population that are in proximity
to air emitting facilities or a receiving water body, and who therefore may be disproportionately exposed
to a chemical undergoing risk evaluation under TSCA section 6. For the air pathway, proximity goes out
to 10,000 m from an air emitting source. For the water pathway, proximity does not refer to a specific
distance measured from a receiving water body, but rather to those members of the general population
that may interact with the receiving water body and thus may be exposed.
Page 37 of 570
-------
RELEASES AND WASTES FROM
INDUSTRIAL / COMMERCIAL / CONSUMER
USES
Industrial Pre-
Treatment or
Industrial WWT
EXPOSURE PATHWAYS
Indirect discharge
~
Wastewater or
Liquid Wastes
POTW
Down-the-Drain
Consumer Product Use
Hydraulic
Fracturing
Produced Waterl
Underground
Injection
Hazardous and
Municipal
Waste Landfill
Solid Wastes
Liquid Wastes
Hazardous an
Municipal
Waste
Off-site Waste
T ransfer
Recycling, Other
Treatment
Emissions to Air
EXPOSURE ROUTES
RECEPTORS
HAZARDS
Hazards Potentially
Associated with
Acute and/or Chronic
Exposures
| Land Pathway
| Air Pathway
H Water Pathway
Included in 2020 RE (No Revisit)
Figure 1-4. Conceptual Model for Environmental Releases and General Population Exposures
Page 38 of 570
-------
The 2020 RE included an assessment of acute incidental/recreational general population exposure to 1,4-
dioxane in surface water resulting from industrial releases. It did not evaluate risks from other general
population exposure pathways such as drinking water or air.
In this supplement, EPA evaluated additional general population exposure via air and drinking water.
EPA evaluated acute (24-hour) and chronic (annual to lifetime) cancer and non-cancer risks from these
exposure pathways. Where data were reasonably available, EPA incorporated releases of 1,4-dioxane
produced as a byproduct into these pathways. The Agency also considered aggregate 1,4-dioxane
exposures and risks from multiple releasing facilities or COUs for each pathway.
1.3.1.3.1 Drinking Water
1,4-Dioxane may enter surface water through direct and indirect industrial releases, DTD releases from
consumer and commercial products via wastewater treatment facilities, and releases of wastewater from
hydraulic fracturing sites. Similarly, 1,4-dioxane released or disposed of through various land pathways
may reach groundwater under some conditions. There is potential for general population exposures to it
if contaminated surface water or groundwater are used as drinking water. 1,4-Dioxane is mobile in water
and does not readily degrade in water. Available data indicate that typical wastewater treatment and
drinking water treatment methods are not effective at removing 1,4-dioxane. The subsections below
explain how general population exposures through surface and groundwater were considered.
Surface Water Pathway
1,4-Dioxane was included in the third Unregulated Contaminant Monitoring Rule (UCMR3) (
2017d) published in May of 2012 requiring community water systems to monitor for 1,4-dioxane
between 2013 and 2015. National and state water monitoring programs have detected 1,4-dioxane in
drinking water and drinking water sources (as described in Section 2.3.1.10). In the absence of
monitoring data, estimating 1,4-dioxane surface water concentrations can be complex because in
addition to direct and indirect industrial and commercial releases, upstream sources from releasing
facilities and DTD releases of consumer and commercial products contribute to surface water
contamination.
EPA evaluated surface water concentrations (Section 2.3.1) and drinking water exposures (Section
3.2.2.1) that could result from direct and indirect industrial releases, DTD releases of consumer and
commercial products, and disposal of wastewater from hydraulic fracturing sites. EPA used a novel
aggregate model to predict water concentrations of 1,4-dioxane that could result from multiple sources
that release 1,4-dioxane to the same receiving water bodies. Surface water concentrations estimated for
each source in isolation and from multiple sources in aggregation were used to evaluate potential
exposures (Section 3.2.2.1) and risks (Section 5.2.2.1) from general population oral exposure to 1,4-
dioxane in drinking water.
Land Pathway to Groundwater
Groundwater contamination with 1,4-dioxane presents a potential risk when the chemical substance is
released to landfills, underground injection wells, or surface impoundments. Due to its physical-
chemical properties (e.g., water solubility, Henry's Law constant) and fate characteristics (e.g.,
biodegradability, half-life in groundwater), 1,4-dioxane is anticipated to persist in groundwater for
months to years. This persistence has resulted in higher 1,4-dioxane concentrations in groundwater
relative to other media (ATSDR. ^ ). EPA considered potential for groundwater contamination
following disposal of waste containing 1,4-dioxane to landfills, underground injection of 1,4-dioxane
waste, and disposals of hydraulic fracturing waste containing 1,4-dioxane. Groundwater concentrations
Page 39 of 570
-------
estimated for each source (Section 2.3.2) were used to evaluate exposure (Section 3.2.2.2) and risks
(Section 5.2.2.1.6) for communities who rely on groundwater as a source of drinking water.
1.3.1.3.2 Air
Industrial releases to air include those from sites where 1,4-dioxane is manufactured intentionally as
well as those where it is produced or present as a byproduct. In this supplement, EPA evaluated
exposures and risks for communities located near release sites (fenceline communities) because they are
the members of the general population that are expected to be PESS due to their greater exposure. EPA
applied the methodology presented in the 2022 fenceline report ( 022d) to evaluate risks
from industrial air releases to fenceline communities. EPA expanded the fenceline methodology to
consider multiple years of release data in this supplement in response to SACC recommendations. In
addition to considering risks from individual facilities, EPA evaluated risks from aggregate exposures in
cases where multiple facilities reporting 1,4-dioxane releases to air were in proximity. The Agency also
evaluated potential risks to fenceline communities from air emissions of 1,4-dioxane modeled for
hydraulic fracturing operations and industrial and commercial laundries.
1.3.1.3.3 Aggregate Exposure
EPA has defined aggregate exposure as "the combined exposures to an individual from a single
chemical substance across multiple routes and across multiple pathways (40 CFR § 702.33)." In this
supplement, EPA considered the combined 1,4-dioxane exposure an individual may experience due to
releases to air or water from multiple sources. For general population drinking water exposure scenarios,
EPA evaluated combined exposure and risks from multiple sources of 1,4-dioxane in surface water,
including direct and indirect industrial releases, DTD releases, and upstream background contamination
(Section 5.2.2.1). For general population air exposure scenarios, EPA evaluated combined exposure and
risk across multiple facilities in proximity releasing to air (Section 5.2.2.3 and Appendix J.4). EPA
qualitatively considered aggregate exposures across exposure routes {i.e., across oral and inhalation) and
across exposure pathways {i.e., across air and water) but did not quantitatively aggregate these
exposures due to uncertainties around the additivity of effects across routes. The rationale for the scope
of aggregate analysis in this supplement and remaining sources of uncertainty are further discussed in
Section 5.2.4.
1.3.2 Potentially Exposed or Susceptible Subpopulations
TSCA section 6(b)(4)(A) requires that risk evaluations "determine whether a chemical substance
presents an unreasonable risk of injury to health or the environment, without consideration of costs or
other non-risk factors, including an unreasonable risk to a potentially exposed or susceptible
subpopulation identified as relevant to the risk evaluation by the Administrator, under the conditions of
use." TSCA section 3(12) states that "the term 'potentially exposed or susceptible subpopulation'
[PESS] means a group of individuals within the general population identified by the Administrator who,
due to either greater susceptibility or greater exposure, may be at greater risk than the general population
of adverse health effects from exposure to a chemical substance or mixture, such as infants, children,
pregnant women, workers, or the elderly."
Considerations related to PESS can influence the selection of relevant exposure pathways, the sensitivity
of derived hazard values, the inclusion of particular populations, and the discussion of uncertainties
throughout the assessment. Factors that may contribute to increased exposure or biological susceptibility
to a chemical include lifestage, pre-existing disease, lifestyle activities {e.g., smoking, physical activity),
occupational and consumer exposures (including workers and occupational non-users, consumers and
bystanders), geographic factors {e.g., fenceline communities), socio-demographic factors, nutrition,
Page 40 of 570
-------
genetics, unique activities (e.g., subsistence fishing), aggregate exposures, and other chemical and non-
chemical stressors.
This supplement considers PESS throughout the human health exposure assessment and risk
characterization. The hazard assessment and dose-response analysis used in this supplement incorporate
all PESS considerations described previously in the 2020 RE. Section 5.2.3 provides a summary of how
specific factors contributing to exposure and susceptibility were addressed in this assessment and
identifies remaining sources of uncertainty for PESS.
1.4 Systematic Review
EPA used the TSCA systematic review process described in the Draft Systematic Review Protocol
Supporting TSCA Risk Evaluations for Chemical Substances, Version 1.0: A Generic TSCA Systematic
Review Protocol with Chemical-Specific Methodologies. ( 1021a) (also referred to as "2021
Draft Systematic Review Protocol") to identify information needed to evaluate additional COUs and
exposure pathways considered in this supplement. Appendix C provides additional information on the
literature search strategy, data screening, evaluation, extraction, and evidence integration steps
performed in support of this assessment—including clarifications and updates made to the 2021 Draft
Systematic Review Protocol to better address assessment needs for this supplement.
1.5 Document Outline
This supplement to the 2020 risk evaluation for 1,4-dioxane comprises the following sections and
appendices:
• Section 1 presents information on the scope of the supplement. It also includes an overview of
the systematic review process used in this analysis. Appendix A provides a list of abbreviations
and acronyms used throughout this report while Appendix B provides the full name and links to
all supplemental documents associated with this supplement. A more detailed description of the
systematic review protocol for this assessment is presented in Appendix C, while Appendix D
provides a crosswalk of COUs with occupational exposure scenarios.
• Section 2 presents an overview of releases and concentrations of 1,4-dioxane in the environment.
A more detailed description of the industrial and commercial environmental release assessment
is presented in Appendix E. Methods for estimating environmental concentrations of 1,4-dioxane
are described in more detail in Appendix G (surface water), Appendix H (groundwater), and
Appendix J (air).
• Section 3 presents the human exposure assessment for occupational and general population
exposure pathways. Details of the occupational exposure assessment are presented in Appendix
F and details of the general population exposure assessment are presented in Appendix I and
Appendix J.
• Section 4 provides a summary of the human health hazard and dose-response assessment
previously published in the 2020 RE and describes duration adjustments made for the current
analysis.
• Section 5 presents risk characterization based on the conditions of use and exposure pathways
evaluated in this supplement. Details of risk calculations and risk estimates are provided in the
set of supplemental risk calculator files listed in Appendix B. Section 5 also includes a
discussion of PESS based on both greater exposure and susceptibility, a description of aggregate
and sentinel exposures, and a discussion of assumptions and uncertainties and the impact on the
supplemental risk evaluation.
Page 41 of 570
-------
Figure 1-5 provides an overview of how the analyses presented in each section are integrated into risk
characterization.
Release Assessment
Section 2
Releases to
surface water
Releasesto
land & groundwater
Releasesto air
Exposure Assessment
Section 3
x N
Occupational
Exposure Scenarios
Inhalation exposure
Dermal exposure
a
General Population
Exposure Scenarios
Oral exposures
through drinking
water
Inhalation
exposure
through air
Risk Characterization
Section 5
Occupational Risk
Characterization
Inhalation risk
Dermal risk
General Population
Risk Characterization
Drinking water risk
Inhalation risk
Hazard Assessment
Section 4
Occupational
hazard values for
acute and chronic
non-cancerand
cancer (consistent
with 2020 RE)
General population
hazard valuesfor
acute, chronic non-
cancer and cancer
hazard values
(derived from PODs
in the 2020 RE, with
some duration
adjustments)
Figure 1-5. Overview of Analyses Included in this Supplement to the Risk Evaluation for 1,4-
Dioxane
Page 42 of 570
-------
2 RELEASES AND CONCENTRATIONS
2.1 Approach and Methodology
2.1.1 Industrial and Commercial Releases
Releases to the environment are one component of potential exposure and may be derived from reported
data that are obtained through direct measurement via monitoring, calculations based on empirical data,
and/or assumptions and models.
The original COUs for 1,4-dioxane are summarized in Table 1-4 of the Final Risk Evaluation for 1,4-
Dioxane ( )20c). Additional COUs included in this supplement due to 1,4-dioxane produced
as a byproduct are presented in Table 2-1. For general population exposures, this supplement considers
releases from all COUs (including the original COUs included in the 2020 RE and the additional COUs
associated with 1,4-dioxane produced as a byproduct). For occupational exposures, this supplement
focuses on the additional COUs associated with 1,4-dioxane produced as a byproduct. For additional
information and context on the inclusion of these COUs in the supplement, refer to Sections 1.1 and 1.2.
For the full table of COUs, including those previously assessed in the Final Risk Evaluation for 1,4-
Dioxane ( )20c\ see Appendix D.
A COU is a combination of life cycle stage, category, and subcategory, as shown in Table 2-1. The COU
subcategory is the most granular description of the use. EPA mapped each COU to an occupational
exposure scenario (OES). The purpose of an OES is to group, where appropriate, COUs based on
similarity of the operations and data availability for each COU. For each OES, EPA estimated air, land,
and water releases and occupational dermal and inhalation exposures. The Agency mapped OESs to
COUs using professional judgment based on reasonably available5 data and information that describe
how releases and exposures take place within an occupational COU. EPA may group multiple COUs
into an OES if the release and exposure potential is similar across the COUs and there is insufficient
data to differentiate the COUs. This grouping minimized repetitive assessments. Alternatively, EPA may
assign multiple OESs to one COU if there are several ways in which release and exposure takes place
for the given COU and sufficient data exist to separately assess the OES. Appendix D.l shows mapping
between COUs and OESs. A crosswalk of the COUs with the OESs assessed is provided in Table 2-1.
As shown in Table 2-1, most COU life cycle stage, category, and subcategory combinations map to a
single OES with a similar or identical name to the COU subcategory. However, for the COU
subcategory of dish soap, dishwasher detergent, and laundry detergent, EPA assigned four OESs: (1)
dish soap, (2) dishwasher detergent, (3) laundry detergent (industrial), and (4) laundry detergent
(institutional). Institutional use of laundry detergent equates to commercial use.
EPA assessed environmental releases (air, water, and land) and occupational exposures (inhalation and
dermal) to 1,4-dioxane for each of the OESs listed in Table Apx D-l. EPA used the environmental
release estimates for each OES for subsequent environmental concentrations and general population
exposure calculations.
5 Reasonably available information is defined in TSCA at 40 CFR 702.33 as "information that EPA possesses, or can
reasonably obtain and synthesize for use in risk evaluations, considering the deadlines.. .for completing the evaluation.
Page 43 of 570
-------
Table 2-1. Additional Categories and Subcategories of COUs and Associated OESs Included in the Scope of the Supplement Due to
the Presence of 1,4-Dioxane Produced as a Byproduct"
Condition of Use
OES Mapped to COU
Life Cycle Stage
Category''
Subcategory'
Processing
Byproduct
Byproduct produced during the ethoxylation processes
Ethoxylation process byproduct
Byproduct produced during the production of polyethylene
terephthalate
Polyethylene terephthalate (PET)
byproduct
Industrial Use,
Commercial Use
Other uses
Hydraulic fracturing
Hydraulic fracturing
Consumer Use,
Commercial Use
Paints and coatings
Latex Wall Paint or Floor Lacquer
Paint and floor lacquer
Cleaning and furniture care
products
Surface Cleaner
Surface cleaner
Laundry and dishwashing
products
Dish soap
Dishwasher detergent
Laundry detergent
Dish soap
Dishwasher detergent
Laundry detergent (industrial)''
Laundry detergent (institutional)''
Arts, crafts, and hobby materials
Textile dye
Textile dye
Consumer Use,
Commercial Use
Automotive care products
Antifreeze
Antifreeze
Disposal
Disposal
Industrial pre-treatment
Disposal
Industrial wastewater treatment
Publicly owned treatment works (POTW)
Underground injection
Municipal landfill
Hazardous landfill
Other land disposal
Municipal waste incinerator
Hazardous waste incinerator
Off-site waste transfer
Page 44 of 570
-------
Condition of Use
OES Mapped to COU
Life Cycle Stage
Category''
Subcategory1
COU = condition of use; OES = occupational exposure scenario
aNew COUs and associated OESs where 1,4-dioxane is produced as a byproduct.
b These categories of COU reflect CDR codes and broadly represent conditions of use for 1,4-dioxane in industrial and/or commercial settings.
c These subcategories reflect more specific uses of 1,4-dioxane.
d Laundry detergent use may occur in industrial, commercial, or consumer settings. Sufficient information was available to separately assess each use setting;
thus, there are two OESs—one for industrial and one for institutional, which equates to commercial use.
Page 45 of 570
-------
2.1.1.1 General Approach and Methodology for Environmental Releases
For each OES, a daily air, land, and water release was estimated based on annual releases, release days,
and the number of facilities (Figure 2-1). The blue boxes represent primary sources of release data that
are used to develop annual releases, release days, and number of facilities. The information in the green
boxes is aggregated by OES to provide a daily release estimate.
Data reported to the Toxics Release Inventory (TRI) and discharge monitoring reports (DMRs) are the
primary sources of release data that EPA used for the release assessments. Under the Emergency
Planning and Community Right-to-Know Act (EPCRA) section 313, 1,4-dioxane has been a TRI-
reportable substance since 1987. The TRI database includes information on releases of 1,4-dioxane to
air, water, and land—in addition to how it is being managed through recycling, treatment, and burning
for energy recovery. Under the Clean Water Act (CWA), surface water discharges reported in DMRs are
based on required monitoring as part of a facility's National Pollutant Discharge Elimination System
(NPDES) permit. Where releases are expected but TRI and DMR data were not available, EPA
estimated releases using data from literature, process information, relevant emission scenario documents
(ESDs), or generic scenarios (GSs), or existing EPA models.
Figure 2-1. Overview of EPA's Approach to Estimate Daily Releases for Each OES
TRI = Toxics Release Inventory; DMR = discharge monitoring report; ESD = emission scenario
document; GS = generic scenario
2.1.1.2 Water Release Estimates
EPA followed a similar approach for estimating industrial and commercial water releases as it did in the
2020 RE, with one key difference. Here, the Agency evaluated multiple years of data using data from
2013 to 2019 TRI (U.S. EPA 2022h) and 2013 to 2019 DMR (U.S. EPA 2022c). as opposed to
utilizing 1 year of data.
Where water releases are expected for an OES but TRI and DMR data were not available, EPA
estimated industrial, and commercial water releases using two approaches. If available, the Agency used
data from literature, ESDs, and GSs in conjunction with Monte Carlo simulation where sufficient data
were available to vary calculation input parameters to estimate industrial and commercial water releases
(see Appendix E.3 for additional information). If no data from literature, ESDs, or GSs were available,
Page 46 of 570
-------
EPA used Stochastic Human Exposure and Dose Simulation for High Throughput (SHEDS-HI) DTD
modeling. SHEDS-HT predicts a per capita DTD loading of 1,4-dioxane, which is combined with an
estimation of the population contributing to publicly owned treatment works (POTW) effluent on the
modeled water body stream to produce an estimated DTD loading. See Section 2.3.1 for additional
explanation of the DTD release modeling. Note that EPA only used SHEDS-HT DTD modeling to
estimate commercial releases when no other reasonable information was available, which was only the
case for the Surface cleaner OES.
For the following OESs, EPA either could not estimate water releases due to lack of reasonably
available data or information or did not expect water releases based on volatility and use patterns:
• Functional Fluids (Closed-Systems): Water release data were not available for this OES.
However, EPA expects that the sources of release for this OES to be similar to those for the
Industrial Uses OES (per process information in the 2020 RE, Appendix G.6.4). Therefore, EPA
grouped the water release assessment for Functional Fluids (Closed-Systems) into that for
Industrial Uses. However, there is uncertainty in this assumption of similar release sources
between these OESs.
• Laboratory Chemical, Film Cement, and Dry Film Lubricant: Wastewater discharges
containing 1,4-dioxane were not expected for these OESs; releases from these OESs are expected
to be to air from volatilizations and landfill/incineration from disposal of empty containers and
other waste (see 2020 RE, Appendix G).
• Antifreeze: Wastewater discharges containing 1,4-dioxane were not expected for this OES;
releases from this OES are expected to be to air from volatilizations during antifreeze changeouts
and to landfill/incineration from disposal of empty antifreeze containers and spent antifreeze.
• Paints and Floor Lacquer: Wastewater discharges containing 1,4-dioxane were not expected for
this OES; releases from this OES are expected to be to air from volatilizations during
painting/drying and to landfill/incineration from disposal of empty paint containers, used paint
brushes/rollers, or solvent washes of paint brushes/rollers.
2.1.1.3 Land Release Estimates
EPA used data from 2019 TRI ( ) to estimate industrial and commercial land releases
that were mapped to each OES with the exception of the Disposal OES. For that OES, EPA performed a
more detailed analysis using data from 2013 to 2019 TRI ( 2022h). Where land releases are
expected for an OES, but TRI data were not available, releases were estimated using reasonably
available data from literature, ESDs, and GSs in conjunction with Monte Carlo simulation (Palisade.
2022a) to allow for variability in calculation input parameters where sufficient data were available to
inform such variability.
EPA did not estimate daily land releases due to the high level of uncertainty in the number of release
days. This uncertainty is because facility operating days does not correlate directly to releases. For
example, a facility may wait until a dumpster or other waste receptacle is full before disposing of it.
Because these releases may occur on a daily, weekly, and even monthly scale, EPA used the annual land
releases reported in TRI data or modeled without estimating land releases for a different frequency
(daily, weekly, monthly). See Appendix E.4 for additional information. Annual land release estimates
were used to estimate potential groundwater contamination from landfill releases as described in Section
2.3.2.
Page 47 of 570
-------
For the following OESs, EPA was not able to estimate land releases due to lack of data or information or
did not expect land releases due to physical form, use patterns, and lack of data:
• Manufacturing, Import and Repackaging, and Functional Fluids (Open-Systems): Data from
2019 TRI (U.S. EPA. 2022h) indicated that there were no releases of 1,4-dioxane to land from
facilities that EPA mapped to these OESs. EPA did not have additional reasonably available
information to estimate land releases from these OESs.
• Functional Fluids (Closed-Systems): See explanation in the preceding section, "Water Release
Estimates."
• 3D Printing: Industrial applications of this OES are expected to be accounted for in the
Industrial Uses TRI data. Per Appendix G.6.8 of the 2020 RE, 3D printing ink containing 1,4-
dioxane is used in research labs to print biomedical products. Because the 2019 TRI data for the
Industrial uses OES include medicinal and pharmaceutical manufacturing NAICS codes, medical
research labs that conduct 3D printing with 1,4-dioxane inks may be captured in that OES.
Therefore, EPA grouped the land release assessment for 3D Printing into that OES for Industrial
uses. However, the extent to which all potential 3D printing sites that use 1,4-dioxane are
captured in the Industrial Uses TRI data is unknown.
EPA also notes that the Hydraulic fracturing OES is associated with certain specific land releases that
may not apply to other OESs, such as the releases of wastewater containing 1,4-dioxane to deep well
injection or surface impoundments, which are considered land releases in this assessment. The Agency
estimated these deep well injection and surface impoundment releases, which were used in addition to
landfill releases, to estimate potential groundwater contamination from hydraulic fracturing described in
Section 2.3.2.
2.1.1.4 Air Release Estimates
EPA applied the following tiered approach to developing air release, exposure, and risk estimates:
1. Pre-screening analysis,
2. Single-year fenceline analysis, and
3. Multi-year fenceline analysis.
2.1.1.4.1 Pre-screening Analysis
This analysis is described in the Draft TSCA Screening Level Approach for Assessing Ambient Air and
Water Exposures to Fenceline Communities and consisted of extracting data for all facilities reporting
1,4-dioxane air releases to the 2019 TRI (U.S. EPA. 2022h). The extracted data were reviewed to
identify the maximum single facility release reported across all reporting facilities. Additionally, the
arithmetic average (mean) value of all reported releases was calculated. These two release values were
used for further analysis to estimate exposure concentrations at select distances from the releasing
facility.
2.1.1.4.2 Single-Year Fenceline Analysis
Where available, EPA used data from 2019 TRI to estimate industrial and commercial air releases in
accordance with the Draft TSCA Screening Level Approach for Assessing Ambient Air and Water
Exposures to Fenceline Communities (U, 2022d). Facilities are only required to report to TRI if
the facility has 10 or more full-time employees; is included in an applicable North American Industry
Classification System (NAICS) code; and manufactures, processes, or uses the chemical in quantities
greater than a certain threshold. Due to these limitations, some sites that manufacture, process, or use
1,4-dioxane may not report to TRI and are therefore not included in these datasets.
Page 48 of 570
-------
Where air releases are expected for an OES, but TRI data were not available, industrial, and commercial
air releases were estimated using data from literature, ESDs, and GSs in conjunction with Monte Carlo
simulation (Palisade. 2022a) to allow for variability in calculation input parameters where sufficient data
were available to inform such variability. See Appendix E.5 for additional information.
For the following OESs, EPA was not able to estimate industrial and commercial air releases due to lack
of data or information:
• Functional Fluids (Closed-Systems): See previous explanation in the "Water Release
Estimates" section above.
• 3D Printing: Industrial applications of this OES are expected to be accounted for in the
Industrial Uses TRI data. Per Appendix G.6.8 of the 2020 RE, 3D printing ink containing 1,4-
dioxane is used in research labs to print biomedical products. Because the 2019 TRI data for the
Industrial Uses OES include medicinal and pharmaceutical manufacturing NAICS codes,
medical research labs that conduct 3D printing with 1,4-dioxane inks may be captured in that
OES. Therefore, EPA grouped the air release assessment for 3D Printing into that OES for
Industrial Uses. However, the extent to which all potential 3D printing sites that use 1,4-dioxane
are captured in the Industrial Uses TRI data is unknown.
• Textile Dyes: EPA did not find relevant reasonably available 1,4-dioxane or surrogate TRI data,
literature sources, sufficient process information, nor ESD or GS with air release estimation
approaches to estimate air releases for this OES. Therefore, EPA was not able to estimate air
releases for this OES.
2.1.1.4.3 Multi-year Analysis
The multi-year analysis incorporates (SACC) recommendations on the Draft TSCA Screening Level
Approach for Assessing Ambient Air and Water Exposures to Fenceline Communities (
2022d) to evaluate multiple years of chemical release data to estimate exposures and associated risks to
fenceline communities. This is achieved by conducting a facility-by-facility evaluation of all 1,4-
dioxane releases reported to TRI from 2015 through 2020. Data for these 6 years were obtained from the
TRI database (TRI basic plus files downloaded on August 5, 2022). Annual release data for 1,4-dioxane
were extracted from the entire TRI data set for all facilities reporting air releases of 1,4-dioxane for one
or more years between 2015 and 2020. Facilities were categorized into OESs for exposure modeling
purposes and later cross-walked to COUs for risk management purposes.
2.2 Environmental Releases
2.2.1 Industrial and Commercial Releases
This section summarizes the estimated air, water, and land releases for each OES; the weight of
scientific evidence conclusions for these estimates; and the strengths, limitations, assumptions, and key
sources of uncertainty for these estimates.
2.2.1.1 Release Estimates Summary
EPA estimated air, water, and land releases of 1,4-dioxane using various methods and information
sources, including
• TRI and DMR data for Manufacturing, import, and repackaging, Industrial uses, Functional
fluids, 3D Printing, Disposal, PET byproduct, and Ethoxylation byproduct,
Page 49 of 570
-------
• GSs and ESDs modeling with and without Monte Carlo for Laboratory chemicals, Spray
polyurethane foam, Textile dye, Antifreeze, Dish soap, Dishwasher detergent, Laundry
detergent, Paint and floor ;acquer, and Hydraulic fracturing,
• Process information for Film cement and Dry film lubricant, and
• SHEDS-HT DTD Modeling for Surface Cleaner.
Note that SHEDS-HT DTD modeling was conducted for multiple additional COUs/OESs, as described
in Section 2.3.1 and Appendix G; however, commercial releases were assessed using alternate methods
as described above for all OESs other than Surface cleaner.
EPA combined its estimates for annual releases, release days, and number of facilities to estimate a
range of daily releases for all OESs, including those presented in the December 2020 RE. The COUs
associated with each OES are summarized in TableApx D-l. A summary of these industrial and
commercial releases for air, water, and land are presented in Table Apx E-3, Table Apx E-5, and
Table Apx E-7, respectively. These release estimates are for total releases from a facility and may
include multiple points of release, such as multiple outfalls for discharges to surface water or multiple
point sources for air emissions. Note that for some release estimates, there is uncertainty and variability
in the potential media of release. In such cases, EPA did not have sufficient information to partition the
release estimates between all potential media of release and they are replicated between the air, land, and
water subsections if there is overlap in the potential media of release.
EPA mapped these releases by media, state, and tribal territory for the conterminous United States.
Alaska, American Samoa, Guam, Hawaii, N. Mariana Islands, Puerto Rico, and the U.S. Virgin Islands
are not mapped since no estimated releases are known.
Surface water releases as reported by TRI and DMR are presented in Figure 2-2. For surface water
releases, the data are divided based on the source of data {i.e., DMR, TRI) and whether the release is
from a direct (on-site) source, including on-site wastewater treatment systems, or indirect (offsite)
source where the chemical substance was taken to a different location for potential release, such as a
POTW. The largest releases have been from PET manufacturing in South Carolina (2,512,434 kg in
2019), Alabama (170,526 kg in 2015; 125,903 kg in 2014; and 111,924 kg in 2017), Tennessee (15,168
kg in 2018), and West Virginia (14,134 kg in 2016 and 12,229 kg in 2014).
Page 50 of 570
-------
Figure 2-2.1,4-Dioxane Annual Water Releases as Reported to TRI and DMR, 2013-2019
Note: Some symbols for individual years may overlap and obscure annual releases at each site.
Alaska, American Samoa, Guam, Hawaii, N. Mariana Islands, Puerto Rico, and the U.S. Virgin Islands are not
shown due to no known modeled or estimated releases.
Land disposals as reported by TRI are available in Figure 2-3. The largest disposals have been to on-site
Class I Underground Injection Wells in Texas (169,035 kg in 2013; 42,865 kg in 2015; 10,729 kg in
2018), On-site Subtitle C Landfills in Oregon (7,321 kg in 2014; 7,000 kg in 2013; and 6,076 kg in
2015), and Offsite Other Landfills in Indiana (862 kg in 2019; 603 kg in 2018; and 354 kg in 2017). Air
release as reported by TRI are available in Figure 2-4.
Page 51 of 570
-------
Tribal Lands
On-site Disposal
• Disposal to RCRASubtitle C Landfill
¦ Underground Injection to Class 1 Wells
Offsite Disposal
a Disposal to Other Landfills
© Disposal to RCRASubtitle C Landfills
¦ Underground I njection to Class 1 Wells
Mass of Disposal (kg)
0.001 - 1.00
o >1.00-10.0
O >10.0-100
O >100-862
200 400 800 1,200 1,600
l Kilometers
Figure 2-3.1,4-Dioxane Annual Releases to Land as Reported to TRI, 2013-2019
Note: Some symbols for individual years may overlap and obscure annual releases at each site.
Alaska, American Samoa, Guam, Hawaii, N. Mariana Islands, Puerto Rico, and the U.S. Virgin Islands are not
shown as there are no known modeled or estimated releases.
For air releases, the largest emissions have been in Illinois (9,943 kg/year), South Carolina (3,495
kg/year), and Texas (2,097 kg/year). Collectively, these figures give insight into the spatial distribution
of releases and corresponding amount across the contiguous United States. A full summary of these
estimates can be found in Appendix E.
Page 52 of 570
-------
Tribal Lands
Total Air Emissions (kg/yr)
O 0.00-10.0
• >10.0-100
• >100 - 1,000
• >1,000 - 10,000
• >10,000 - 100,000
200 400 800 1,200 1,600
I Kilometers
Figure 2-4. 1,4-Dioxane Annual Releases to Air as Reported by TRI, 2013-2019
Note: Some symbols for individual years may overlap and obscure annual releases at each site.
Alaska, American Samoa, Guam, Hawaii, N. Mariana Islands, Puerto Rico, and the U.S. Virgin Islands are not
shown as there are no known modeled or estimated releases.
In addition to mapping releases from TRI and DMR, EPA also mapped hydraulic fracturing sites
reporting the presence of 1,4-dioxane in hydraulic fracturing operation fluids according to FracFocus 3.0
(GWPC and IOGCC. 2022). These operations are primary sited in a wide range of shale plays across the
country (as indicated by the multi-colored plays mapped in Figure 2-5). The Delaware play in Texas has
the largest number of operations (n = 158) followed by the Niobrara in Colorado (n = 86) and the Utica
play that spreads across Pennsylvania and Ohio (n = 70).
Page 53 of 570
-------
Figure 2-5. Locations of Hydraulic Fracturing Operations that Report 1,4-Dioxane in
Produced Waters
Note: Some symbols for individual years may overlap and obscure annual releases at each site.
Alaska, American Samoa, Guam, Hawaii, N. Mariana Islands, Puerto Rico, and the U.S. Virgin Islands are
not shown as there are no known modeled or estimated releases.
The basis for overall data quality determinations is also included in the water, air, and land summary
subsections below. Each source is evaluated on multiple metrics based on defined criteria. For air, water,
and land releases, all monitoring data had data quality ratings of medium/high. Modeled data had data
quality ratings of medium or high.
2.2.1.2 Weight of Scientific Evidence Conclusions for Environmental Releases
EPA's judgment on the weight of scientific evidence is based on the strengths, limitations, and
uncertainties associated with the release estimates. The Agency considers factors that increase or
decrease the strength of the evidence supporting the release estimate—including quality of the
data/information, applicability of the release data to the COU (including considerations of temporal
relevance, locational relevance), and the representativeness of the estimate for the whole industry. In
general, the use of Monte Carlo modeling improves the weight of scientific evidence due to the
incorporation of variability; however, the weight of scientific evidence is largely tied to the strengths
and limitations of the underlying model equations and input parameter datasets. The weight of scientific
evidence is summarized using the descriptors of robust, moderate, slight, or indeterminant, according to
EPA's Application of Systematic Review in TSCA Risk Evaluations (U.S. EPA. 2018c). For example, a
conclusion of moderate weight of scientific evidence is appropriate where there is measured release data
Page 54 of 570
-------
from a limited number of sources such that there is a limited number of data points that may not cover
most or all of the sites within the COU. A conclusion of slight weight of scientific evidence is
appropriate where there is limited information that does not sufficiently cover all sites within the COU,
and the assumptions and uncertainties are not fully known or documented. See EPA's Application of
Systematic Review in TSCA Risk Evaluations (U. 2018c) for additional information on weight of
scientific evidence conclusions.
A summary of air, land, and water release estimation approaches with the associated weight of scientific
evidence conclusion is compiled for each OES in Table 2-2. In summary, all TRI/DMR monitoring data
had data quality ratings of medium/high. For supplemental releases assessed with TRI/DMR (PET
byproduct, Ethoxylation byproduct, Disposal), the weight of scientific evidence conclusion was
moderate to robust because the reasonably available information relevant for the conditions of use of
1,4-dioxane at facilities in TRI and DMR is limited. The underlying data used in modeled release
estimates had data quality ratings of medium or high. For releases that used SHEDS-HT modeling
(Surface cleaner), the weight of scientific conclusion was slight since there is uncertainty in the
application of this modeling for a commercial setting, and this case study does not represent all sites in
this OES. For supplemental releases that used GS/ESDs or other data sources, the weight of scientific
conclusion was moderate when used in tandem with Monte Carlo modeling (Textile dye, Laundries,
Dish soap, Dishwasher detergent), and slight/moderate when used alone (Antifreeze, Paint and floor
lacquer). For Hydraulic fracturing, the weight of scientific conclusion was moderate to robust since
FracFocus 3.0, an ESD, and Monte Carlo modeling were used. See Appendix E.8 for a summary of
EPA's overall weight of scientific evidence conclusions for its release estimates for each of the assessed
OESs.
Page 55 of 570
-------
Table 2-2. Summary of the Weight of Scientific Evidence for Environmental Release Estimates by PES
Water
Land
Air
OES
Approach
Data
Quality
Weight of
Scientific
Approach
Data
Quality
Weight of
Scientific
Approach
Data
Quality
Weight of
Scientific
Rating"
Evidence
Rating"
Evidence
Rating"
Evidence
Manufacturing
TRI and DMR
Medium
Moderate
to robust
TRI
Medium
Moderate
to robust
TRI
Medium
Moderate to
robust
Import and
TRI and DMR
Medium
Moderate
TRI
Medium
Moderate
TRI
Medium
Moderate to
repackaging
to robust
to robust
robust
Industrial uses
TRI and DMR
Medium
Moderate
to robust
TRI
Medium
Moderate
to robust
TRI
Medium
Moderate to
robust
Functional fluids
TRI and DMR
Medium
Moderate
TRI
Medium
Moderate
TRI
Medium
Moderate to
(open-system)
to robust
to robust
robust
Functional fluids
Assessed as a part
N/A
Slight
Assessed as a part
N/A
Slight
Assessed as a part
N/A
Slight
(closed-system)
of Industrial Uses
OES
of Industrial Uses
OES
of Industrial Uses
OES
Laboratory
chemical
GS indicates no
water releases
High
Slight to
moderate
GS modeling
High
Slight to
moderate
GS modeling
High
Slight to
moderate
Film cement
Process
information
indicates no water
releases
High
Slight to
moderate
Modeling with
process
information
High
Slight to
moderate
Modeling with
process
information
High
Slight to
moderate
Spray foam
application
GS modeling
Medium
Slight to
moderate
GS modeling
Medium
Slight to
moderate
GS modeling
Medium
Slight to
moderate
Printing inks (3D)
DMR
Medium
Moderate
to robust
Assessed as a part
of Industrial Uses
OES
N/A
Slight
Assessed as a part
of Industrial Uses
OES
N/A
Slight
Dry film lubricant
Process
information
indicates no water
releases
High
Slight to
moderate
Modeling with
process
information
High
Slight to
moderate
Modeling with
process
information
High
Slight to
moderate
Disposal
TRI and DMR
Medium
Moderate
to robust
TRI
Medium
Moderate
to robust
TRI
Medium
Moderate to
robust
Page 56 of 570
-------
Water
Land
Air
OES
Approach
Data
Quality
Weight of
Scientific
Approach
Data
Quality
Weight of
Scientific
Approach
Data
Quality
Weight of
Scientific
Rating"
Evidence
Rating"
Evidence
Rating"
Evidence
Textile dye
ESD modeling
with Monte Carlo
Medium
Moderate
ESD modeling
with Monte Carlo
Medium
Moderate
Not assessed due
to lack of
information
N/A
Indeterminant
Antifreeze
Process
information
indicates no water
releases
High
Slight to
moderate
Modeling with
process
information
High
Slight to
moderate
Modeling with
process
information
High
Slight to
moderate
Surface cleaner
SHEDS-HT and
generic modeling
with process
information
High
Slight
SHEDS-HT and
generic modeling
with process
information
High
Slight
SHEDS-HT and
generic modeling
with process
information
High
Slight
Dish soap
Process
information with
Monte Carlo
modeling
High
Moderate
Process
information with
Monte Carlo
modeling
High
Moderate
Process
information with
Monte Carlo
modeling
High
Moderate
Dishwasher
Process
High
Moderate
Process
High
Moderate
Process
High
Moderate
detergent
information with
Monte Carlo
modeling
information with
Monte Carlo
modeling
information with
Monte Carlo
modeling
Laundry detergent
(industrial and
ESD modeling
with Monte Carlo
Medium
Moderate
ESD modeling
with Monte Carlo
Medium
Moderate
ESD modeling
with Monte Carlo
Medium
Moderate
institutional)
Paint and floor
lacquer
ESD and process
information
indicates no water
releases
Medium
Slight to
Moderate
ESD modeling
Medium
Slight to
Moderate
ESD modeling
Medium
Slight to
Moderate
PET byproduct
TRI and DMR
Medium
Moderate
to robust
TRI
Medium
Moderate
to robust
TRI
Medium
Moderate to
robust
Ethoxylation
TRI and DMR
Medium
Moderate
TRI
Medium
Moderate
TRI
Medium
Moderate to
process byproduct
to robust
to robust
robust
Hydraulic
factoring
ESD modeling
with Monte Carlo
High
Moderate
to robust
ESD modeling
with Monte Carlo
High
Moderate
to robust
ESD modeling
with Monte Carlo
High
Moderate to
robust
a Data quality ratings of modeling approaches are based on the GS/ESD or process information.
Page 57 of 570
-------
2.2.1.3 Strengths, Limitations, Assumptions, and Key Sources of Uncertainty for the
Environmental Release Assessment
EPA estimated air, water, and land releases of 1,4-dioxane using various methods and information
sources, including TRI and DMR data, GSs and ESDs modeling with and without Monte Carlo, process
information, and SHEDS-HT DTD Modeling.
TRI and DMR were determined to have the overall data quality determination of medium through
EPA's systematic review process. Uncertainties for using TRI and DMR data are discussed in the Final
Risk Evaluation for 1,4-Dioxane (U.S. EPA. 2020c). In summary, these uncertainties include
underestimation of the number of sites for a given OES due to reporting thresholds in TRI, the accuracy
of EPA's mapping of sites reporting to TRI and DMR to a specific OES, quality of the data reported to
TRI and DMR, and the representativeness of past years data toward current conditions.
Due to TRI reporting thresholds, estimated releases using TRI data may not be representative of sites
that handle 1,4-dioxane at quantities below the TRI reporting threshold. There is additional uncertainty
for sites that report to TRI with Form A because these sites do not report release quantities if the
quantity did not exceed 500 lb for the total annual reportable release amount. For these sites, EPA
assessed a "what-if' scenario, which assumes the entire 500 lb going to single media of release, noting
that the 500 lb should not be added over all release media. Additional information on TRI uncertainties
is provided in Appendix E.7. In addition, as discussed in Section 2.2.1, EPA used data from the 2019
reporting year to estimate air and land releases. A key source of uncertainty in the assessment of air and
land releases is whether 2019 TRI data are representative of releases from other reporting years. This
does not apply to the water release estimates because EPA used data from reporting years 2013 to 2019.
A strength of using TRI is that it compiles reasonably available release data for all facilities that reported
to EPA. However, not all facilities are required to report to TRI.
Some uncertainties of using DMR data include the accuracy of EPA's mapping of sites reporting to
DMR to a specific OES, and quality of the data reported to DMR. Also, an uncertainty of using the
ECHO Pollutant Loading Tool Advanced Search option is that average measurements may be reported
as a quantity (kg/day) or a concentration (mg/L). Calculating annual loads from concentrations requires
adding wastewater flow to the equation, which increases the uncertainty of the calculated annual load. In
addition, for facilities that reported having zero pollutant loads to DMR, the EZ Search Load Module
uses a combination of setting non-detects equal to zero (if all data from the facility over the year were
non-detect) and as one-half the detection limit (if some data were detect and other data were non-detect
over the year, the non-detect values are set at half the detection limit) to calculate the annual pollutant
loadings. This method could cause overestimation or underestimation of annual and Daily pollutant
loads; however, EPA uses this method for handling non-detects as it is consistent with the established
procedures for the EZ Search Load Module. A strength of using DMR data and the Pollutant Loading
Tool is that the tool calculates an annual pollutant load by integrating monitoring period release reports
provided to the EPA and extrapolating over the course of the year. However, this approach assumes
average quantities, concentrations, and hydrologic flows for a given period are representative of other
times of the year.
Additionally, there is uncertainty when the reported surface water discharges for a given site differs
between DMR and TRI for the same year. In these instances, EPA uses the higher of the reported
discharge quantities. These differences are due to TRI annual release totals being directly reported by a
facility, while DMR annual release totals are estimated from DMR monitoring data. While differences
Page 58 of 570
-------
between annual estimates between the two systems are common, the magnitudes of differences vary by
facility and reporting year.
Where TRI and DMR data were not reasonably available, EPA used GS and ESDs. One uncertainty for
this method is lack of specific 1,4-dioxane data. Because GS/ESDs are generic, assessed parameter
values may not always be representative of applications specific to 1,4-dioxane use in each OES.
Another uncertainty is lack of consideration for release controls. The GS/ESDs assume that all activities
occur without any release controls, and in an open-system environment where vapor and particulates
freely escape (U.S. EPA. 2022e; QEC n 1 , 201 la. b). Actual releases may be less than estimated if
facilities utilize pollution control methods. Although 1,4-dioxane monitoring data are preferred to
modeled data, EPA strengthened modeled estimates by using Monte Carlo modeling to allow for
variation in environmental release calculation input parameters according to the GS/ESD and other
literature sources. However, EPA did not utilize Monte Carlo modeling for all GS/ESD, which is a
limitation of this assessment. Table Apx E-8 includes information on which GS/ESDs were used in
tandem with Monte Carlo modeling.
EPA used process information to quantify environmental releases for the film cement and dry film
lubricant OESs. This process information is from the 2020 RE ( >20c) and the underlying
sources were determined to have high overall data quality determinations through EPA's systematic
review process. To develop these release estimates, EPA made assumptions on the likely media of
release for various releases sources and, in some cases, used standard EPA models in conjunction with
process information to estimate the release quantity. A source of uncertainty in this approach is the
representativeness of these estimates regarding all sites that use 1,4-dioxane for this OES.
EPA used SHEDS-HT DTD modeling to estimate environmental releases to surface water or land for
the surface cleaner OES because no other data or information were reasonably available. The main
source of uncertainty is that the SHEDS-HT DTD modeling is for a single case study location,
Liverpool, OH. It is uncertain whether the release estimates generated from this case study are
applicable to other areas of the country. Additionally, EPA is unsure whether the use of SHEDS-HT
results in a high-end or typical exposure scenario, so the use of these data may lead to over or
underestimates of releases. Additional uncertainties associated with using SHEDS-HT to estimate
commercial releases for the surface cleaner OES is provided in Appendix E.7.
To assess daily air and water discharges, EPA divided annual release loads by the number of facility
release days to estimate the daily release load for the facility. There is uncertainty if the assumed release
duration is applicable to all sites for a given OES; therefore, the average daily releases may be higher if
sites have fewer release days or lower if they have greater release days. Furthermore, 1,4-dioxane
concentrations in air emissions and wastewater release to receiving water bodies at each facility may
vary from day-to-day such that on any given day the actual daily releases may be higher or lower than
the estimated average daily discharge. Thus, this approach minimizes variations in emissions and
discharges from day to day. EPA did not estimate daily land releases due to the high level of uncertainty
in the number of release days associated with land releases. The Agency expects that sites may not send
waste to landfills every day and are more likely to accumulate waste for periodic shipments to landfills.
However, sites that release to municipal landfills may have more frequent release days based on the
frequency of shipments.
Spills and leaks may occur in multiple OES. Generally, releases and exposures from spills and leaks are
assessed within the OES where they occur, as TRI data includes releases from accidental releases such
as spills and GS/ESD typically include assessment approaches for spills when supported by data.
Page 59 of 570
-------
However, due to the highly variable nature of spills, there is uncertainty in the representativeness of any
data on spills toward all potential accidental releases for a given OES.
2.3 1,4-Dioxane Environmental Concentrations
2,3,1 Surface Water Pathway
Surface water contamination from 1,4-dioxane can occur from direct releases of wastewater from
industrial operations, discharges from wastewater treatment plants containing DTD releases of 1,4-
dioxane from consumer and commercial product usage {i.e., dish soap, laundry detergent, etc.), and
other activities where 1,4-dioxane may be present as a byproduct, such as in hydraulic fracturing
operations. To understand possible exposure scenarios from these practices, EPA assessed exposures to
the general population from ambient surface and drinking water. These exposures are due to 1,4-dioxane
being directly or indirectly discharged to receiving water bodies.
The evaluation of these exposures considered both the review of reasonably available monitoring data to
both ambient surface water and drinking water as well as the modeling of estimated exposures due to
releases. Although EPA identified a robust set of surface and drinking water monitoring data (Section
2.3.1.1) indicating the presence of 1,4-dioxane in these pathways, it was collected independent of release
data, and cannot be attributed to specific sources (Section 2.2). Therefore, EPA relied primarily on a
series of modeling approaches to estimate concentrations of 1,4-dioxane in surface water near known
release locations (Sections 2.3.1.2 and 2.3.1.3). For this assessment, EPA modeled concentrations
resulting from industrial releases for all COUs releasing to surface water, including those assessed in the
2020 RE, as well as those producing 1,4-dioxane as a byproduct. To the degree possible, the relationship
between monitoring and modeled data is further evaluated in Section 2.3.1.4.
2.3.1.1 Monitoring Data
Environmental concentration data for 1,4-dioxane in ambient surface water {i.e., measured in rivers,
streams, lakes, and ponds, rather than within industrial operations or drinking water systems) across the
country, as well as routine monitoring conducted by public water systems (PWSs) of raw (untreated)
source water and finished (treated) drinking water were collected from readily available public databases
and publications. The methods for retrieving and processing ambient surface water and PWS data are
described in detail in Appendix G. 1.
Ambient Surface Water Monitoring
Data were retrieved from the Water Quality Portal (WQP) to characterize observed concentrations in
ambient surface water (NWQMC. 2022). These monitored values were not assessed for proximity to
sources of drinking water and are instead analyzed to generally characterize the observed ranges of 1,4-
dioxane concentrations in ambient surface water—irrespective of the reasons for sample collection—and
to provide context for the modeled surface water concentrations presented in Section 2.3.1.3. Data
retrieved in July 2022 included sampling dates from 1997 to 2022 and resulted in 12,471 available
sample results. Full details of the retrieval and processing of ambient surface water monitoring data from
the WQP are presented in Appendix G. Table 2-3 shows the range of 1,4-dioxane concentrations
detected in surface water samples. Most {i.e., 92.3%) of the sample records available had no level of 1,4-
dioxane detected above the reported detection limit for the analysis (referred to as "non-detects"), with
limits of detection ranging from 0.001 to 28,000 |ig/L across all samples. The 105 detected values
ranged from 0.016 to 470 |ig/L, with a median of 1.10 |ig/L. Since the range of detected concentrations
fall within the range of detection limits, it is possible that there are additional instances of 1,4-dioxane
occurrence that were not able to be reported due to analytical limitations. Figure 2-6 and Figure 2-7
show the distribution of detected concentrations and reported detection limits of non-detect samples,
Page 60 of 570
-------
respectively. The highest concentrations reported in this dataset are noted in the metadata to have been
collected at the point of discharge from an industrial facility, while for most samples, the reason for
sampling, or sample location in relation to expected releases is not included in the metadata. Figure 2-8
shows the spatial distribution of detected samples. For the entire dataset, including non-detects,
approximately 70 percent of the samples were collected from the states of North Carolina, New Mexico,
and New Jersey. Of the 105 detected values, 46 percent are in Pennsylvania, 21 percent in North
Carolina, and 14 percent in Illinois. In the absence of a national standardized study of 1,4-dioxane in
ambient surface water (analogous to the UCMR monitoring in drinking water), and without more
national coverage and metadata, it is difficult to characterize the national occurrence of 1,4-dioxane in
surface water. It is apparent from the available monitoring data that certain areas may be more likely to
have higher concentrations, while many others have little or no detected 1,4-dioxane. Over-
representation of certain states or regions may reflect targeted sampling campaigns of specific locations
expected to have higher concentrations, and conclusions about areas without monitoring data cannot be
drawn without further exploration through modeling.
0.01 0.02 0,05 0.10 0.20 0,50 1 2 5 10 20 50 100 200 500
Detected 1,4-Dioxane Concentration (|jg/L)
Figure 2-6. Frequency of Nationwide Measured 1,4-Dioxane Surface Water Concentrations
Retrieved from the Water Quality Portal, 1997-2022
Note: Detectable levels of 1,4-dioxane may vary by location.
Page 61 of 570
-------
c
CD
250-
c
o
OJ
?no-
1—
c
CI)
o
c
150 -
o
o
4—
o
100-
>N
o
c
CI)
50-
cr
0
LL
I I 11 III!" I I 1111 III I I I I mil I I I111 III I I I I Hill I I I 11 Mil ' I I I 11111^ I III 111 ll I ^ I III
10~3 10~2 1CT1 10° 101 102 103 104
Detection Limit for 1,4-Dioxane Non-Detect (pg/L)
Figure 2-7. Frequency of Detection Limits for Nationwide Non-detect 1,4-Dioxane
Surface Water Samples Retrieved from the Water Quality Portal, 1997-2022
Figure 2-8. Detectable Concentrations of 1,4-Dioxane in Surface Water from the Water
Quality Portal, 1997-2022
Note: Detectable levels of 1,4-dioxane may vary by sampling location.
Alaska, American Samoa, Guam, Hawaii, N. Mariana Islands, Puerto Rico, and the U.S. Virgin
Islands are not shown as there are no known monitoring data above detection limits.
Page 62 of 570
-------
Drinking Water Monitoring Data
The Safe Drinking Water Act (SDWA) authorizes the EPA to set national health-based standards for
drinking water to protect against both naturally occurring and man-made contaminants that may be
found in drinking water. The National Primary Drinking Water Regulations (NPDWRs) are legally
enforceable primary standards and treatment techniques that apply to PWSs. Although states, tribes or
territories that have been approved as the primary implementation authority for drinking water may
require monitoring or impose limits for contaminants beyond those regulated under SDWA, there are
not currently national requirements to routinely monitor or limit 1,4-dioxane in finished water from
PWSs. In support of the SDWA, EPA often relies on data from the UCMR program as the best available
occurrence information to support its regulatory determinations {i.e., to judge whether a particular
contaminant is known to occur or there is substantial likelihood the contaminant will occur in public
water systems with a frequency and at levels of public health concern). UCMR monitoring is designed
to produce a data set that is nationally representative of public water systems (PWSs) across the country,
but its focus is on finished water (rather than source water), and it may not capture worst-case
conditions. PWS monitoring data of finished drinking water were collected for 1,4-dioxane via EPA's
published UCMR3 dataset from 2013 to 2015, as well as raw and finished drinking water monitoring
from additional individual state databases (CA, MA, and NY) from 2008 to 2022 (CA Water Board.
2022; NY DOR 2022; Commonwealth of Massachusetts. , ) UCMR3 data were
filtered to only include facilities flagged as using surface water, while the individual state data were
filtered down to only those systems with surface water listed as the primary source in SDWIS. Datasets
were processed to ensure that no samples were repeated in multiple datasets. These PWS monitoring
data were collected to assess possible exposures to the general population through drinking water.
Descriptions of the data retrieval and processing methods are presented in Appendix G.2.
The combined datasets resulted in 16,972 samples from 2,847 PWSs across 50 states (Table 2-3).
Reported detection limits across the PWS datasets ranged from 0.0001 to 3 |ig/L, with 81 percent of
samples reporting a detection limit of 0.07 |ig/L. To the extent that it could be determined from the
database records, samples were separated into raw (untreated) water from the PWS intake or finished
(treated) water being sent to the distribution system. The distribution of raw water monitoring
concentrations is presented in Figure 2-9, and the distribution of finished drinking water concentrations
is presented in Figure 2-10.
Table 2-3. Summary of PWS Monitoring Datasets of 1,4-Dioxane Monitoring in PWSs Using
Surface Water as a Source
Dataset of
Origin
Number of
Samples
Minimum
Concentration
(jug/L)
Median
Concentration
(Hg/L)
Maximum
Concentration
fag/L)
Start
Year
End
Year
CA
1,797
0.25
0.5
1.5
2011
2022
MA
949
0.049
0.22
3.8
2008
2022
NY
615
7.20E-05
0.035
1
2015
2022
UCMR3
13,611
0.035
0.035
13.3
2013
2016
Note: for the summary presented in this table, results reported as below their respective detection limit were
assigned a value of half of the detection limit.
Page 63 of 570
-------
Detected 1,4-Dioxane Concentration in Raw Water (ug/L)
Figure 2-9. Frequency of 1,4-Dioxane Concentrations Monitored in Raw (Untreated)
Drinking Water Derived from Surface Water
Data retrieved from state databases (CA, MA, and NY) between 2008-2022.
Detected 1,4-Dioxane Concentration in Finished Drinking Water (ug/L)
Figure 2-10. Frequency of 1,4-Dioxane Concentrations Monitored in Finished (Treated)
Drinking Water Derived from Surface Water
Data retrieved from the UCMR3 and state databases (CA, MA, and NY) between 2008-2022
Note: the detection limit for the method used in UCMR3, and the most common detection limit reported
in state databases is 0.07 (ig/L.
Water treatment systems may vary widely across the country based on available and utilized water
treatment processes that depend on whether source water is groundwater or surface water. These
processes typically include disinfection, coagulation/flocculation, sedimentation, and filtration (U.S.
EPA. 2006a). In assessing drinking water exposures, the ability to treat and remove or transform
chemicals in possible drinking water supplies should be considered. Typical treatment processes do not
remove 1,4-dioxane from ambient surface water and groundwater prior to possible general population
consumption as drinking water and treatment processes that do effectively remove 1,4-dioxane are
uncommon. EPA therefore assumes zero removal in the following analyses to provide a conservative
estimate of general population drinking water exposures. Even without treatment processes that remove
1,4-dioxane, multiple sources of water may be mixed within the same drinking water system which may
Page 64 of 570
-------
result in finished water with lower concentrations than one of the higher contributing source water
concentrations. EPA acknowledges that the surface water concentration at a single intake location may
be higher than the finished drinking water once mixed with other sources. Further discussion of the
prevalence of treatment processes across water systems, and the methodology for identifying raw and
finished drinking water monitoring samples is presented in Appendix G. 1.2.
Figure 2-11 shows the spatial distribution of UCMR3 samples at the county level, with 1,4-dioxane
detected in 25 percent or 240 of 943 counties with participating water systems.
Maximum Reported 1,4-dioxane
Concentration
(M9/L)
6,21 - 13.30
2.49 - 6.20
1.21 - 2.48
0.62 - 1.20
I 1 0.07 - 0.61
I 1 < 0.07 (Not Detected)
I I No Data
0 125 250
500
750
1,000
I Miles
Figure 2-11. Map of Counties Containing PWSs that Reported Monitoring of Finished
Drinking Water Drawn from Surface Water for 1,4-Dioxane under UCMR3
Note: UCMR3 monitoring of 1,4-dioxane required four sampling events, one for each season, to capture
temporal variability. Each county highlighted may include one or multiple PWSs reporting data.
Monitored drinking water data were also included in exposure and risk estimates to assess the human
health implications of drinking water concentrations in this range. Since the UCMR program and state
monitoring datasets are not designed to reflect source water impacts of direct and indirect releases into
water bodies, EPA's TSCA program relied on estimated concentrations modeled for a range of specific
release scenarios to characterize risks from the water pathway. The Agency evaluated the performance
of the models used to estimate water concentrations with monitoring data from site-specific locations
serving as cases studies. These case study comparisons demonstrated general consistency between
modeled concentrations and monitoring data, thereby increasing confidence in risk estimates based on
modeled concentrations.
Page 65 of 570
-------
Although monitoring data confirm that 1,4-dioxane is present in drinking water in some locations,
samples collected under the UCMR program are designed to be nationally representative of drinking
water occurrence and not specifically associated with industrial releases of 1,4-dioxane. Since these
monitoring data may not reflect the 1,4-dioxane concentrations that result from industrial releases, EPA
relied on modeling to estimate 1,4-dioxane concentrations that occur near release sites.
2.3.1.2 Surface Water and Drinking Water Modeling
To assess possible general population exposures to 1,4-dioxane via industrial releases to surface water,
concentrations of 1,4-dioxane in surface water were modeled using two separate approaches. First, a
facility-specific approach aimed to quantify the maximum expected aqueous concentrations resulting
from reported 1,4-dioxane discharges from individual facilities in isolation. Second, a probabilistic
model was applied to assess the range of expected aqueous concentrations resulting from reported 1,4-
dioxane discharges across a COU, with consideration of expected ranges of background concentrations
of 1,4-dioxane from DTD loading and other unreported releases.
2.3.1.2.1 Modeling Methodology
A detailed description of modeling methods is presented in Appendix G.2.
As described in Section 2.2, annual releases of 1,4-dioxane to surface water from regulated dischargers
were retrieved from TRI and DMR. To the extent possible, modeled hydrologic flow data {i.e., stream
flow) associated with the receiving water body to which each facility released was retrieved from the
NHDPlus V2.1 dataset ( ). The receiving water body was identified either through
NPDES permit information for the releasing facility, or the nearest identified NHDPlus V2.1 flowline.
Detailed methods for the retrieval and processing of flow data are presented in Appendix G.2.1.
Facility-Specific Modeling
Facility-specific modeling was conducted to estimate concentrations in receiving water bodies resulting
from the greatest facility-specific annual release reported between 2013 through 2019. This modeling
approach employed the equations used to model releases from facilities in the E-FAST 2014 model
( 2014) and is described in Appendix G.2.2. For each facility and annual release amount, three
different scenarios for days of release per year were considered: 1 day, 30 days, and expected number of
days of operation reported in Table Apx E-2 (referred to as the "maximum" number of days and ranges
from 250 to 365 days depending on OES). These additional scenarios with lower numbers of days of
operation provide more conservative estimates of resulting surface water concentrations and are
intended to evaluate the full range of possible facility release patterns based on the best available
information. Two flow metrics were evaluated: the lowest monthly average flow from NHDPlus, and the
harmonic mean flow derived from E-FAST 2014 methodology. The resulting concentrations from the
facility-specific modeling are used in calculations of general population exposure and human health
outcomes.
Probabilistic Modeling
The probabilistic modeling approach was conducted to consider multiple years of release data per
facility and multiple modeled flow metrics from NHDPlus V2.1 (U.S. EPA. 2016c) per facility to
generate a distribution of potential surface water concentrations resulting from releases across each
COU. The underlying model for the probabilistic approach is a fit-for-purpose model developed by EPA
in Microsoft Excel, the EWISRD-XL model (Estimating Water Industrial Surface Release and Down
the Drain in Excel). The EWISRD-XL model was designed to model 1,4-dioxane inputs to a stream
segment, including existing in-stream concentrations (including from unregulated sources), DTD
loading from consumer and commercial products, and industrial releases, as a steady-state snapshot of a
Page 66 of 570
-------
single point in time (Figure 2-12). Outputs from the model include the resulting downstream
concentration and the relative contribution from each input source to that concentration. Examples of the
EWISRD-XL model applied to three specific case study locations (Brunswick County, NC, Columbia,
TN, and Liverpool, OH) are included in three Supplemental Information Files (U.S. EPA. 2024p. g, r).
POTW Loading:
Down-the-drain
from upstream
population
Indirect industrial
releases
Inputs: -
Figure 2-12. Schematic of the EWISRD-XL Model Inputs and Outputs
For the probabilistic 1,4-dioxane COU modeling, an R script (R Core Team. 2022) was developed to
rapidly run multiple iterations of the EWISRD-XL model. In this configuration, called the EWISRD-
XL-R model, the underlying calculations were performed by EWISRD-XL model, and an R script
wrapper managed the processing of input and output data. For the probabilistic COU modeling, the
EWISRD-XL-R model developed to calculate the receiving water body concentrations at the point-of-
release by a facility. The EWISRD-XL-R results include the concentrations due only to releases from
facilities, as well as an estimated background concentration of DTD and unmonitored releases. The full
details of the underlying EWISRD-XL model and the probabilistic implementation are presented in
Appendix G.2.3. Distributions of total concentrations {i.e., the sum of resulting facility releases and
background concentrations) estimated by the probabilistic model were used for additional calculations of
general population exposure and human health outcomes.
A series of case studies was developed with the EWISRD-XL model to evaluate its performance across
various 1,4-dioxane release settings. These cases are presented in Appendix G.2.3.2.
2.3.1.2.2 Estimating Down-the-Drain Releases
To evaluate the anticipated ranges of DTD contributions of 1,4-dioxane to water bodies receiving
POTW effluent, a range of combinations of hydrologic flows and populations served by a POTW were
evaluated using the EWISRD-XL-R model. For this modeling exercise, only contributions from the
DTD component were used to calculate resulting surface water contributions (i.e., no facility releases or
existing background concentrations were included). Hypothetical combinations of hydrologic flows and
Page 67 of 570
-------
populations contributing to wastewater loading derived from the national distribution of hydrologic
flows and populated places were selected to represent a range of results, which were then compared with
concentrations expected from industrial releases and used to calculate ranges of human exposure and
risk. More detailed methodology for this calculation is presented in Appendix G.2.3.4.
2.3.1.2.3 Hydraulic Fracturing
Hydraulic fracturing is a process used to extract oil and gas from shale plays. After hydraulic fracturing
operations inject fluids to extract oil and gas, a substantial volume of water may be produced through
flowback. The composition of these produced waters depends both on the geochemistry of the injected
area and the injected fluids. 1,4-Dioxane has been reported to EPA as one of the chemicals present in
these produced waters by 411 facilities via FracFocus 3.0 ("GWPC and IOGCC. 2022). Estimated 1,4-
dioxane loadings of produced water to surface water from hydraulic fracturing activities (described in
Appendix E.9) were evaluated for expected ranges of resulting concentrations in receiving water bodies
using the EWISRD-XL-R model. Hydraulic fracturing wells reporting 1,4-dioxane use by FracFocus 3.0
were mapped, and flow data from nearby water bodies were collected from NHDPlus V2.1. A Monte
Carlo analysis was used to generate loadings to receiving water bodies from the distribution of modeled
releases and to pair them with hydrologic flows, resulting in a distribution of possible surface water
concentrations. Methodology for this analysis is presented in Appendix G.
2.3.1.2.4 Proximity to Drinking Water Sources
Drinking water exposures from facility-specific results assumed that the exposure occurs at the receiving
water body to provide a conservative estimation of drinking water exposures. However, the evaluated
water bodies may not be used as, or proximate to, actual drinking water sources and intakes. To give a
more robust characterization of possible drinking water exposures, known facility-specific releases were
mapped to drinking water sources using public water systems data stored in EPA's Safe Drinking Water
Information System Federal Data Warehouse ( 022g). This dataset is updated quarterly, and
the 2nd quarter 2022 version was used for this analysis. Following mapping, the colocation of and
proximity of releases to drinking water sources were evaluated. Locations of raw water intakes for
PWSs are considered sensitive by EPA Office of Water due to public safety concerns. Geospatial
analysis and the NHDPlus V2.1 flowline network were used to assess whether any known drinking
water intakes are located downstream of 1,4-dioxane releasing facilities. Methodology for this analysis
is presented in Appendix G.2.4.
2.3.1.3 Modeling Results
2.3.1.3.1 Facility-Specific Results
The facility-specific results show the expected concentration at the point of release from the facility
discharging 1,4-dioxane to receiving water bodies, without consideration of the contribution from other
sources. The total number of modeled releases within a given OES may be greater than the number of
1,4-dioxane releasing facilities in cases where facilities indirectly dispose of 1,4-dioxane by transferring
to another facility in addition to directly discharging 1,4-dioxane. Surface water concentrations resulting
from facility-specific modeling for one day of release are summarized in Table 2-4 and represent the
highest expected concentrations in receiving water bodies, due to the annual release amount being
discharged in a single day. Surface water concentrations resulting from facility-specific modeling for
maximum days of release are summarized in Table 2-5, and represent the lowest expected
concentrations in receiving water bodies due to the annual release spread out over the most days. The
single day release scenario allows consideration of a "worst-case scenario" given the available annual
release information and can inform an upper limit of concentrations resulting from releases. The
maximum days release scenario can inform a lower limit of expected concentrations from the available
Page 68 of 570
-------
annual release data. Full discussion on the evaluation of multiple release days is given in Section
2.3.1.2.1, but the range of evaluated release days is intended to provide to full range of expected surface
water concentrations resulting from possible facility release patterns and available information. As
described in Section 5, these variations in concentration due to days of release do not affect chronic
cancer risk estimates resulting from a particular releasing facility, due to annual averaging of exposure.
Resulting concentrations varied widely, both across and within OESs, due to variability in facility
release amounts as well as receiving water body flow magnitudes. Facility-specific releases are
organized around their identified OES as fully described in Section 2.1 and Appendix D.
Page 69 of 570
-------
Table 2-4. Summary of Surface Water Concentration Results by OES from Facility-Specific Modeling of Annual Maximum Releases
between 2013 and 2019 for 1 Operating Day per Year
OES
No. of
Releases
Modeled
Sum of
Annual
Releases
Modeled
(kg/year)
Annual Release by Facility
(kg/site-year)
Surface Water Concentration
(Lowest Monthly Flow)
(Hg/L)
Surface Water Concentration
(Harmonic Mean Flow)
(Hg/L)
Min
Mean
Max
Min
Mean
Max
Min
Mean
Max
Disposal
25
16,997
1.36E-04
6.80E02
7.95E03
1.50E-02
6.45E05
9.52E06
1.50E-02
4.77E05
7.34E06
Ethoxylation byproduct
8
112,076
4.54E-01
1.40E04
1.12E05
5.39E-03
2.58E06
2.07E07
3.01E-03
1.22E06
9.73E06
Functional fluids (open-
system)
6
17,711
3.80E-01
2.95E03
1.75E04
1.39E01
1.57E03
4.78E03
6.07E00
7.40E02
2.21E03
Import and repackaging
12
2,722
2.27E02
2.27E02
2.27E02
1.08E01
8.15E06
9.28E07
4.39E00
1.01E06
7.40E06
Industrial uses
31
70,343
2.07E-01
2.27E03
2.62E04
1.33E-02
5.11E05
4.64E06
6.52E-03
4.53E05
5.15E06
Manufacture
2
7,034
1.67E03
3.52E03
5.36E03
8.31E04
1.63E06
3.18E06
8.31E04
1.63E06
3.18E06
PET manufacturing
19
2,773,355
3.40E-01
1.46E05
2.51E06
2.77E00
1.07E06
1.66E07
1.28E00
1.05E06
1.66E07
Printing inks
1
5
5.45E00
5.45E00
5.45E00
2.05E03
2.05E03
2.05E03
2.05E03
2.05E03
2.05E03
Remediation
16
46
3.40E-05
2.91E00
2.39E01
1.50E-03
1.83E03
1.79E04
3.54E-04
1.52E03
1.37E04
Overall
120
3,000,290
3.40E-05
2.50E04
2.51E06
1.50E-03
1.45E06
9.28E07
3.54E-04
5.92E05
1.66E07
Page 70 of 570
-------
Table 2-5. Summary of Surface Water Concentration Results by OES for Facility-Specific Modeling of Annual Maximum Releases
OES
No. of
Releases
Modeled
Sum of
Annual
Releases
Modeled
(kg/year)
Annual Release by Facility
(kg/site-vear)
Surface Water Concentration
(Lowest Monthly Flow)
(Hg/L)
Surface Water Concentration
(Harmonic Mean Flow)
(jig/L)
Min
Mean
Max
Min
Mean
Max
Min
Mean
Max
Disposal
25
16,997
1.36E-04
6.80E02
7.95E03
5.99E-05
2.57E03
3.81E04
5.99E-05
1.90E03
2.94E04
Ethoxylation byproduct
8
112,076
4.54E-01
1.40E04
1.12E05
2.16E-05
1.03E04
8.26E04
1.20E-05
4.87E03
3.89E04
Functional fluids (open-system)
6
17,711
3.80E-01
2.95E03
1.75E04
5.63E-02
6.37E00
1.93E01
2.46E-02
3.00E00
8.95E00
Import and repackaging
12
2,722
2.27E02
2.27E02
2.27E02
4.32E-02
3.26E04
3.71E05
1.76E-02
4.04E03
2.96E04
Industrial uses
31
70,343
2.07E-01
2.27E03
2.62E04
5.31E-05
2.04E03
1.86E04
2.61E-05
1.81E03
2.06E04
Manufacture
2
7,034
1.67E03
3.52E03
5.36E03
3.32E02
6.52E03
1.27E04
3.32E02
6.52E03
1.27E04
PET manufacturing
19
2,773,355
3.40E-01
1.46E05
2.51E06
1.11E-02
4.29E03
6.63E04
5.12E-03
4.20E03
6.63E04
Printing inks
1
5
5.45E00
5.45E00
5.45E00
8.21E00
8.21E00
8.21E00
8.21E00
8.21E00
8.21E00
Remediation
16
46
3.40E-05
2.91E00
2.39E01
4.11E-06
5.01E00
4.90E01
9.69E-07
4.15E00
3.75E01
Overall
120
3,000,290
3.40E-05
2.50E04
2.51E06
4.11E-06
5.80E03
3.71E05
9.69E-07
2.37E03
6.63E04
Page 71 of 570
-------
Table 2-6 describes the crosswalk between identified OESs and relevant COUs under each for the
identified facility releases to surface water. The full facility-specific analysis is included in 1,4-Dioxane
Supplemental Information File: Drinking Water Exposure and Risk Estimates for 1,4-Dioxane Release
to Surface Water from Individual Facilities ('U.S. EPA. 2024b).
Table 2-6. OES-COU Crosswalk for Identified Facilities Releasing to Sur
'ace Water"
COU*
OES'
Life Cycle Stage
Category
Subcategory'
Manufacturing
Domestic manufacture
Domestic manufacture
Manufacturing
Import
Import
Repackaging
Import and repackaging
Processing as a reactant
Polymerization catalyst
Industrial uses
Non-incorporative
Basic organic chemical manufacturing
(process solvent)
Processing
Byproduct
Byproduct produced during the
ethoxylation process to make
ethoxylated ingredients for personal
care products
Ethoxylation process
byproduct
Byproduct produced during the
production of polyethlene terephtalate
PET byproduct
Intermediate use
Plasticizer intermediate
Catalysts and reagents for anhydrous
acid reactions, brominations, and
sulfonations
Industrial uses
Industrial Usea
Processing aids, not
otherwise listed
Wood pulping
Extraction of animal and vegetable
oils
Wetting and dispersing agent in
textile processing
Polymerization catalyst
Purification of process intermediates
Etching of fluoropolymers
Industrial uses
Functional fluids (open
and closed systems)
Polyalkylene glycol lubricant
Synthetic metalworking fluid
Cutting and tapping fluid
Functional fluids (open
system)
Industrial Use,
Commercial Use
Other Uses
Spray polyurethane foam
Printing and printing compositions,
including 3D printing
Dry film lubricant
Hydraulic fracturing
Printing inks (3D)
Disposal
Disposal
Remediation
Remediation
Disposal
Disposal
Industrial pre-treatment
Industrial wastewater treatment
Publicly owned treatment works
(POTW)
Underground injection
Municipal landfill
Disposal
Page 72 of 570
-------
COU''
OESrf
Life Cycle Stage
Category
Subcategory'
Hazardous landfill
Other land disposal
Municipal waste incinerator
Hazardous waste incinerator
Off-site waste transfer
" Although EPA has identified both industrial and commercial uses here for purposes of distinguishing scenarios in this
document, the Agency interprets the authority over "any manner or method of commercial use" under TSCA section 6(a)(5)
to reach both.
b As mapped to COU Life Cycle Stage, Category, and Subcategory in Table Apx D-l.
c Evaluated facilities within an OES may not encompass all listed COU subcategories.
d Note that identified OESs can encompass multiple COUs across different life cycle stages and categories.
To put the modeled releases in the context of the underlying data sources for release amounts and
receiving water body flow, Table 2-7 presents the results of the process of assigning the receiving water
body (by reach code in the NHDPlus 2.1 dataset) to each releasing facility. Those facilities with reach
code information in their NPDES permit were regarded as the highest confidence in an accurate match
to the actual discharging water body, followed by facilities matched geospatially to the nearest reach
code within 1 km of the facility. Facilities matched to reaches beyond 1 km from the facility but within 2
km provided lower confidence, and those without reach code matches were substituted with the lowest
non-zero flow within the OES as a conservative estimate. The full details of the flow matching process
are presented in Appendix G.2.1. Due to the assumptions described in Section 0 required to model
releases from facilities reporting only via TRI Form A, the percent of facilities within an OES using
Form A is also reported.
Table 2-7. Summary by OES of Data Sources for Releases and Receiving Water Body Flow
Method of Matching to Receiving Water Body
OES
Total
Number of
Releases
NPDES
Permit
Contains
Reach Code
Nearest
Reach
(within 1
km)
Nearest
Reach
within 2 km
Lowest Non-
zero Flow
within OES
Substituted
%of
Releases
Estimated
from TRI
Form A
Disposal
25
22
0
1
1
8
Ethoxylation
byproduct
8
1
1
1
0
0
Functional fluids
(open-system)
6
5
0
0
0
0
Import and
repackaging
12
1
2
2
7
100
Industrial uses
31
11
3
5
7
45
Manufacture
2
1
0
0
0
0
PET manufacturing
19
11
0
1
0
0
Printing inks
1
1
0
0
0
0
Remediation
16
14
2
0
0
0
Total
120
67
8
10
15
23
A generic table of annual facility release and average flow rates for the receiving water body is
presented in Table 2-8, which demonstrates the relationship between the facility and water body
Page 73 of 570
-------
characteristics regarding the resulting surface water concentrations. Table 2-9 shows the relative
occurrence of each of the releases modeled for this assessment within binned ranges of releases and
flows. Combined, these tables demonstrate that most facilities releasing 1,4-dioxane are initially
discharging to smaller water bodies, even in some cases where large annual release amounts result in
very high modeled concentrations.
Table 2-8. Hypothetical Mean Annual Concentrations (jig/L) for a Range of Annual Release and
Flow Rate Combinations, for a Facility with 250 Days of Release per Year
Annual Release Amount (kg)
1
10
100
1,000
10,000
100,000
1,000,000
Mean
Annual
Receiving
Water Body
Flow (cfs)
1
1.6
16
160
1.6E03
1.6E04
1.6E05
1.6E06
10
0.16
1.6
16
160
1.6E03
1.6E04
1.6E05
100
0.016
0.16
1.6
16
160
1.6E03
1.6E04
1,000
1.6E-03
0.016
0.16
1.6
16
160
1.6E03
10,000
1.6E-04
1.6E-03
0.016
0.16
1.6
16
160
100,000
1.6E-05
1.6E-04
1.6E-03
0.016
0.16
1.6
16
Table 2-9. Occurrence of Facilities for Distributions of Maximum Annual 1,4-Dioxane Release
Amounts and Receiving Water Body Flow
Annual Release Amount (kg)
<10
10 to 100
100 to
1,000
1,000 to
10,000
10,000 to
100,000
>100,000
Mean Annual
Receiving Water
Body Flow (cfs)
<10
14%
8%
11%
6%
2%
<1%
10 to 100
9%
7%
2%
<1%
1%
3%
100 to 1,000
3%
6%
2%
1%
<1%
<1%
1,000 to 10,000
1%
2%
2%
2%
2%
<1%
10,000 to 100,000
3%
1%
3%
4%
2%
<1%
2.3.1.3.2 Concentrations from Down-the-Drain Loading
Water concentrations of 1,4-dioxane resulting from DTD releases depend on the population size (an
indicator of the number of people using products and contributing to the releases) and the stream flows
of the receiving water bodies. The representative per capita DTD loading developed from modeling
results from SHEDS-HT was applied to a range of population sizes (100 to 1,000,000 people) and
stream flows (300 to 30,000 cfs) to develop a distribution of potential surface water concentrations.
Estimated surface water at the point of discharge by POTWs resulting from DTD releases ranged from
less than 0.0001 to 110 |ig/L (Table 2-10). The typical ranges of results from this analysis, representing
only the concentrations due to DTD loading, are comparable to the range of minimum to mean
concentrations calculated from individual facility releases in Section 2.3.1.3.1.
Page 74 of 570
-------
Table 2-10. Estimated Surface Water Concentrations (jig/L) Due to DTD Loading
Population Contributing to DTD Loading
100
1,000
10,000
100,000
1,000,000
Receiving
Water Body
Flow (cfs)
100
0.011
0.11
1.1
11
110
300
3.6E-03
0.036
0.36
3.6
36
1,000
1.1E-03
0.011
0.11
1.1
11
3,000
3.6E-04
3.6E-03
0.036
0.36
3.6
10,000
1.1E-04
1.1E-03
0.011
0.11
1.1
The occurrence of POTWs processing wastewater from various populations and the associated flows of
the receiving water bodies were investigated using data from the ICIS-NPDES database (
2013). to inform the interpretation of the above ranges of DTD loading concentrations. For communities
with a single POTW treating wastewater, most fell into the range of 100 to 10,000 people, with the
annual average flow of the receiving water body less than 300 cfs (Table 2-11).
Table 2-11. Estimated Percent Occurrence of Combinations of Contributing Population to
Population Contributing to DTD Loading
<100
100 to 1,000
1,000 to
10,000
10,000 to
100,000
100,000 to
1,000,000
Mean Annual
Receiving
Water Body
Flow (cfs)
<100
5%
44%
26%
4%
<1%
100 to 300
<1%
3%
4%
1%
<1%
300 to 1,000
<1%
2%
2%
1%
<1%
1,000 to 3,000
<1%
1%
2%
<1%
<1%
3,000 to 10,000
<1%
<1%
1%
<1%
<1%
>10,000
<1%
1%
1%
1%
<1%
2.3.1.3.3 Concentrations from Hydraulic Fracturing
The Monte Carlo distribution of potential surface water concentrations resulting from hydraulic
fracturing operations is presented in Table 2-12. Hydrologic flows in water bodies near hydraulic
fracturing wells reporting 1,4-dioxane as a constituent of wastewater ranged from less than 10 to 44,300
cfs. Due to the very low flows in many nearby streams, resulting concentrations were sensitive to the
receiving water body flow rate. The distribution of loading to surface water from hydraulic fracturing
represents the loading from a single site of hydraulic fracturing operations (described in Appendix
G.2.3.5) at the immediate point of discharge to the receiving water body. Concentrations estimated at the
highest end are comparable to mean to high-end facility releases presented in Section 2.3.1.3.1. More
than half of the modeled concentrations fell below the typical detection limit in drinking water of 0.07
Hg/L-
Page 75 of 570
-------
Table 2-12. Distribution of Potential Concentrations in Surface
Water Resulting from Hydraulic Fracturing Operations from a
Single Site Reporting 1,4-Dioxane as an Ingredient
Monte Carlo Distribution
Concentration (jig/L)
Maximum
157
99th Percentile
7.55
95th Percentile
2.73
Median
0.069
5th Percentile
3.38E-04
Minimum
2.79E-10
2.3.1.3.4 Aggregate Probabilistic Results
The aggregate probabilistic model predicts surface water concentrations at the point of facility releases
when incorporating potential contributions from DTD and other unmonitored sources. The model
incorporates multiple years of release data and was run with 10,000 iterations for each OES using
different combinations of direct and indirect facility releases, DTD releases, flows, and background
concentrations. This results in a more descriptive distribution of the potential releases. At the highest
end, the results of the aggregate probabilistic model are similar to those from the facility-specific
modeling. This is due to both the facility-specific modeling and the highest end of the probabilistic
modeling being based on the maximum reported releases from the modeled facilities. Additionally, the
loading from facilities far outweighs the contribution from background sources at the higher end. The
shape of the resulting distribution can be informative in its representation of the frequency of
concentrations exceeding a certain threshold.
Resulting surface water concentrations ranged from 1,45 x ] 0 4 to 7.34><103 |ig/L. Summaries of the
resulting concentrations by OES are presented in Table 2-13 and Figure 2-13. Overall, releases from
facilities tended to result in greater 1,4-dioxane concentrations in surface water than the expected ranges
of background concentrations. Background concentrations were derived from values of 1,4-dioxane
measured by drinking water systems using surface water as a source that were not downstream of known
1,4-dioxane releases (Figure 2-11). The "% of Releases Greater than Background" column in Table 2-13
refers to the frequency of model runs (out of the 10,000 per OES) in which the resulting concentration
from the facility release was greater than the generated background concentration resulting from DTD
and other unregulated surface water loading. A low percentage for this metric may suggest that releases
by a particular OES are typically outweighed by these other unreported releases with respect to their
contribution to surface water concentrations.
Page 76 of 570
-------
Table 2-13. Aggregate Probabilistic Results Showing Distribution of Total 1,4-Dioxane Concentration in Surface Water (Release
Plus Background)
OES
Min
(^g/L)
5th
Percentile
(jig/L)
25th
Percentile
(jLLg/L)
Median
(jig/L)
75th
Percentile
(jig/L)
95th
Percentile
(^g/L)
Max
(^g/L)
% of Releases
Greater than
Background
Disposal
1.88E-03
1.51E-01
1.98E-01
3.50E-01
8.52E-01
1.94E00
2.02E01
81
Ethoxylation process byproduct
2.25E-03
9.81E-02
1.35E-01
2.74E-01
4.65E-01
2.55E00
1.46E01
72
Functional fluids (open-system)
1.78E-04
8.20E-02
1.11 E—01
1.58E-01
2.79E-01
1.60E00
6.10E00
48
Import and repackaging
5.70E-03
1.32E-01
2.83E-01
6.60E01
3.25E02
1.42E03
2.12E03
90
Industrial uses
1.45E-04
5.15E-02
8.95E-02
1.22E-01
2.52E-01
1.33E01
2.26E02
44
Manufacture
1.10E02
3.32E02
3.32E02
7.19E02
2.32E03
5.48E03
7.34E03
100
PET byproduct
5.44E-03
1.42E-01
2.13E-01
4.65E-01
4.36E00
1.30E02
2.23E03
84
Printing inks (3D)
1.51E-01
3.17E-01
1.10E00
4.00E00
6.41E00
7.48E00
8.26E00
96
Remediation
5.34E-04
5.89E-02
9.04E-02
1.35E-01
2.73E-01
6.66E00
1.46E01
47
Page 77 of 570
-------
Disposal
CO
¦O
©
J
Import and Repackaging
c
CD
CD
-JH
J:
PET Manufacturing
c:
CU
"O
0
J
1-^-riTK^rT,
Ethoxylation byproduct
"O
CD
mk.
Industrial Uses
c
cu
1ZJ
CD
cdj
¦
Printing Inks
E
c
cu
~o
CD
Id
Functional Fluids (Open-System)
c
2
XJ
(D
-d
Manufacture
c
CU
T3
CD
Remediation
c
cu
~o
CD
ttd
i
10"' 10"3 10"
10"
10'
10°
Total (Facility Release + Background) Modeled 1,4-Dioxane Concentration (pg/L)
Figure 2-13. Distributions of Surface Water Concentrations Estimated by Aggregate Probabilistic Model for Each OES
Vertical lines indicate the median and 95th percentile (P95) surface water concentrations.
Page 78 of 570
-------
2.3.1.4 Comparison of Modeled and Monitored Surface Water Concentrations
At the higher end, the modeled concentrations from facility releases are several orders of magnitude
greater than those observed in the 1,4-dioxane monitoring data (Figure 2-6 and Table 2-4). This
difference in concentrations may be due to monitoring data being collected further downstream
(allowing for additional dilution), or on reaches that are not impacted by releasing facilities. Many of the
direct releasing facilities, and POTWs assessed for the DTD component, had a receiving water body
specified on their NPDES permits that was associated with a very small stream or industrial canal. These
small receiving water bodies, combined with larger loading values from the releases, resulted in high
modeled concentrations in surface water at the point of release. As this water travels downstream, it is
expected to eventually join with larger water bodies, where some decrease in concentration due to
dilution would occur.
Because most of the reasonably available monitoring data were generally not co-located with 1,4-
dioxane release sites, EPA relied primarily on modeling to estimate water concentrations that could
result from releases. Where co-located monitoring data were available, EPA compared modeled
concentrations to reasonably available monitoring data in the limited set of specific locations to evaluate
the performance of the model. Comparisons of modeled vs. monitoring water concentrations for this
limited set of "case study" locations demonstrate that modeled mean concentrations are generally
consistent with mean concentrations reported in monitoring data. For example, the Cape Fear River
upstream of the Brunswick County, NC drinking water intake was selected as a case study to test the
model due to abundant monitoring data in the region. Water concentrations modeled based on upstream
releases from an industrial facility in Fayetteville in combination with other upstream sources. As
illustrated in Figure 2-14, modeled surface water concentrations generally fell within the ranges reported
from monitored concentrations. Wide ranges of both monitored and modeled values were noted,
indicating variability among inputs to the system. Details of the case study comparisons for Brunswick
County and other locations are described in Appendix G.2.3.2. The agreement between monitoring and
modeled concentrations increases confidence in the model used to estimate water concentrations from
DTD releases and hydraulic fracturing, and to perform probabilistic modeling of aggregate
concentrations from multiple sources.
Page 79 of 570
-------
Date
Figure 2-14. Case Study Comparison of Modeled and Monitored Concentrations in
Brunswick County
2.3.1.5 Strengths, Limitations, and Sources of Uncertainty in Assessment Results for
Monitored and Modeled Drinking Water and Surface Water Concentrations
The evaluation of general population drinking water exposure scenarios are impacted by uncertainties
and assumptions surrounding inputs and the approaches used for modeling surface water concentrations
and estimation of the drinking water doses. In Appendix E.8, EPA assesses the overall confidence of
estimated releases for various OESs. For those OESs releasing to surface water, confidence is rated as
medium to high depending on an individual OES.
The modeling used and the associated default and user-selected inputs have the ability to affect overall
strength in evaluated general population exposures. The facility-specific releases methodology described
in Section 2.3.1.2.1, and the results in 2.3.1.3.1, rely on a modeling framework that does not consider
downstream fate or transport. However, the physical-chemical properties of 1,4-dioxane are expected to
moderate this limitation due to its likelihood to stay in the water column, and due to the lack of removal
during typical drinking water treatment process. To reduce uncertainties, EPA incorporated an updated
flow network and flow data into this assessment that allowed a more site-specific consideration of
release location and associated receiving water body flows. These facility-specific releases are also
evaluated on a per facility basis that does not account for additional sources of 1,4-dioxane that may be
present in the evaluated waterways. To help address these limitations in this risk evaluation, EPA
conducted additional aggregate and probabilistic approaches, evaluated in Section 2.3.1.2.1 and Section
2.3.1.3.4, that give a more complete overall estimation of possible 1,4-dioxane concentrations. EPA
acknowledges some uncertainty in the modeled flows represented for each reach of the NHDPlus V2.1
database, including the consideration that modeled flows are based on flow data collected from 1971 to
2000. Some variation in flow statistics may be expected for current and future flow conditions. Finally,
drinking water exposures from facility-specific results assume that the exposure occurs at the receiving
water body. The water bodies evaluated may or may not be used as drinking water sources. To address
this limitation, EPA evaluated the proximity of known 1,4-dioxane releases to known drinking water
sources as well as known drinking water intakes as described in Section 2.3.1.2.4.
To evaluate the accuracy of the aggregate model, case studies described in Appendix G.2.3.2 compared
modeled results to observed monitored concentrations. The three evaluated case studies give good
Page 80 of 570
-------
general agreement between available monitoring with modeled values. Overall, this gives strength to the
modeling assumptions, inputs and output calculations for areas that are lacking robust monitoring data.
The model is able to effectively capture the general influences of both DTD loading, facility loading and
upstream contributions to create an aggregation of possible ambient surface water concentrations of 1,4-
dioxane. The monitored data encompassed both ambient surface water monitoring as well as drinking
water system monitoring data. For the ambient surface water data, data is limited geographically and
temporally with many states having no reported data and even those areas reporting measured values
having limited samples over time. Monitored concentrations in close proximity to modeled releases were
rare, often making direct comparisons of modeled results unavailable. In most cases, monitoring data
represented water bodies without identified releases of 1,4-dioxane nearby.
The hydraulic fracturing analysis relies on a Monte Carlo distribution of loading values with some level
of uncertainty and is itself a Monte Carlo simulation with potential receiving water body flows. The
precision of such an analysis is lower at the most extreme (minimum and maximum) values.
2,3,2 Land Pathway (Groundwater)
Any activities where chemicals or wastes might be released to the environment has the potential to
pollute groundwater. To understand possible exposure scenarios from these practices, EPA assessed
drinking water exposure resulting from use of 1,4-dioxane contaminated groundwater due to chemical
injection to Underground Class I Wells, leaching from landfills where 1,4-dioxane or products
containing 1,4-dioxane have been disposed, and disposal of hydraulic fracturing produced water to
surface impoundments. Sections 2.3.2.1 through 2.3.2.4 provide a description and an assessment of each
disposal practice. Figure 2-15 and Figure 2-16 provide a visual summary of groundwater monitoring
data available through the WQP fNWOMC. 20221
2.3.2.1 Groundwater Monitoring Data
Measured, field-collected, data from environmental samples representing groundwater 1,4-dioxane
concentrations across the country were collected as direct groundwater monitoring results. These results
are collated by the National Water Quality Monitoring Council and stored in the WQP fNWOMC.
2022). Some monitoring results reported to the WQP included locations expected to be directly
impacted by 1,4-dioxane releases. Data were available from 1997 to 2022, resulting in 8,110 available
sample results. The distribution and detection percentages are presented in Figure 2-15 and mapped in
Figure 2-16. The process for identifying this data is provided in Appendix H.l. This analysis is intended
to characterize the observed ranges of 1,4-dioxane concentrations in groundwater, irrespective of the
reasons for sample collection, and to provide context for the modeled groundwater concentrations
presented in Sections 2.3.2.1 through 2.3.2.4. In order to better understand where highest groundwater
concentrations are occuring, EPA arbitrarily portioned the data based on order of magnitude differences
to best describe where and when data differences could be observed.
Page 81 of 570
-------
Detected 1,4-Dioxane Concentration ((jg/L)
Figure 2-15. Frequency of Nationwide Detected 1,4-Dioxane Groundwater
Concentrations (n = 2,284) Retrieved from the Water Quality Portal, 1997-2022
Figure 2-16. Detectable Concentrations of 1,4-Dioxane in Groundwater from the Water
Quality Portal, 1997-2022
Note: Alaska, American Samoa, Guam, Hawaii, N. Mariana Islands, Puerto Rico, and the U.S. Virgin Islands
are not shown as there are no known monitoring data above detection limits.
Page 82 of 570
-------
Figure 2-15 shows the range of detected concentrations of 1,4-dioxane in groundwater from 1997 to
2022. During this period the detection limits ranged from 0.028 |ig/L to 320 |ig/L. The maximum
detected concentration (31,000 |ig/L) occurred in Westville, IN, in 1997 at a former waste-oil refinery.
This site and many others identified in this monitoring data have ongoing remediation projects to
address these contamination plumes.
Recent changes in industrial activities and disposal may have largely reduced groundwater
contamination with 1,4-dioxane. As shown in Figure 2-17, samples collected prior to 2000 tended to be
substantially higher in concentration relative to those collected after 2003. This finding may be an
artifact of historical uses and industrial practices related to 1,4-dioxane. Although several samples are
still above 10 |ig/L, particularly in 2007, the bulk of data tend to fall between 1 and 10 |ig/L. Without a
thorough investigation of what practices have changed in industry, it is difficult to attribute this decline
to a single event but indicates continued work to prevent groundwater contamination.
Figure 2-17. Groundwater Concentrations of 1,4-Dioxane vs. Sample
Collection Date for Data Collected between 1997 and 2022
Figure 2-16 shows the spatial distribution of detected 1,4-dioxane concentration across the contiguous
states. This map shows nine locations with concentrations of 1,4-dioxane greater than 10 |ig/L. These
tend to be attributed to past industrial activities causing extensive groundwater contamination. In
addition to this monitoring data, groundwater contamination from disposing 1,4-dioxane to landfills has
been documented in Alaska (Li et al.. 2013). California (Li et al.. 2015; Adamson et al.. 20141
Michigan (Mohr and DiGuiseppi. 20101 New York (Lee et al.. 2020). and recently in Ohio
(https://cumulis.epa.gov/supercpad/cursites/csitinfo.cfm?id=0504014). EPA was not able to identify
reasonably available information specific to groundwater concentrations near or around underground
injection sites, landfills, or surface impoundments that received hydraulic fracturing produced water.
2.3.2.2 Disposal via Underground Injection
Underground injection is a method of disposal for hazardous wastes.6 There are generally six different
classes of underground wells, and only Class I Wells may be permitted to receive hazardous waste.
Oversight of these wells requires that they are designed and constructed to prevent the movement of
6 Additional information about underground injection can be found at https://www.epa.gov/uic.
Page 83 of 570
-------
injected waste streams into drinking water systems. Wells typically consist of three or more concentric
layers of pipe including surface casing, long string casing, and injection tubing. In addition, wells must
be sited at locations with geologies that mitigate any movement of contaminants outside of a confined
layer in case of a well failure. Extensive pre-siting geological tests confirm that the injection zone is of
sufficient lateral extent and thickness and is sufficiently porous so that fluids injected through the well
can enter the rock formation without extensive buildup of pressure or possible displacement of injected
fluids outside of the intended zone.
Potential pathways through which injected fluids can migrate to underground sources of drinking water
include failure of the well or improperly plugged or completed wells near the well. Well failures can be
detected by continuous monitoring systems or mechanical integrity tests, at which point the wells would
be shut-in until they are repaired. EPA's extensive technical requirements for Class I wells (40 CFR
148) are designed to prevent contamination of underground sources of drinking water through these
pathways. Operators must conduct appropriate mechanical integrity tests yearly for hazardous wells and
every 5 years for nonhazardous wells to ensure wells are fit for operation. Note that the loss or failure of
mechanical integrity does not necessarily mean that wastewater will escape the injection zone. This
added security can be attributed to redundant safety systems to protect against loss of waste
confinement.
2.3.2.2.1 Summary of Assessment for Disposal to Underground Injection
According to EPA's TRI database, there are two locations where 1,4-dioxane has been disposed of via
underground injection to Class I Wells. On-site disposals to Class I underground injection wells are
provided in TableApx H-l. On-site Class I underground injection wells may be owned and operated by
the producer of the waste. Off-site disposals to Class I underground injection wells are provided in
Table Apx H-2. Offsite Class I underground injection wells may be secondary entities that own and
operate the well. Both on-site and offsite underground injection wells must be permitted and regularly
inspected. Careful review of the permits and state databases corroborates that both sites are permitted
and compliant. These sites have implemented groundwater migration controls and the Enforcement and
Compliance History Online (ECHO) database ( I022f) indicates the site is currently in
compliance.
In addition to reviewing these permits, EPA reviewed reasonably available groundwater monitoring data
available via state databases as well as via the WQP (see Figure 2-15 and Figure 2-16) and found no
evidence of groundwater contamination near the facilities. Because underground injection is not
expected to result in groundwater contamination based on the reasonably available information, EPA did
not quantitatively estimate groundwater concentrations, exposures, or risks from underground injection.
2.3.2.2.2 Strengths, Limitations, and Sources of Uncertainty in Assessment of
Disposal to Underground Injection Wells
Because EPA did not quantitatively evaluate the potential exposure from disposing 1,4-dioxane via
underground injection, the major source of uncertainty is limited to the accuracy of state databases
providing monitoring data surrounding these wells. EPA believes these databases are reporting
accurately where contaminations are known, but only explored states where the TRI database indicated
there were disposals via underground injection. Disposals below the reporting requirement for TRI may
not be captured.
2.3.2.3 Disposal to Landfills
Landfills may have various levels of engineering controls to prevent groundwater contamination. These
can include industrial liners, leachate capturing systems, and routine integration of waste. However,
Page 84 of 570
-------
groundwater contamination from disposal of consumer, commercial, and industrial waste streams
continues to be a prominent issue for many landfills throughout the United States (Li et al.. JO I \ Li et
at.. 2013; Mohr and DiGuiseppi. 2010). These contaminations may be attributed to perforations in the
liners, failure of the leachate capturing system, or improper management of the landfills. 1,4-Dioxane
persists in groundwater and can migrate away from landfills into nearby communities at the same rate as
hydraulic flow (Mohr a iuiseppi. 2010). When these communities rely on groundwater as their
primary drinking water source, there is a potential for exposure via oral ingestion if that water is
contaminated with 1,4-dioxane and does not undergo treatment. Depending on the distance between the
landfill and a drinking water well, as well as the potential rate of release of landfill leachate into
groundwater, the concentration of this exposure can vary substantially.
Landfills are generally regulated under the Resource Conservation and Recovery Act (RCRA). RCRA
landfills can be classified as Subtitle C (hazardous waste landfills) or Subtitle D (municipal solid
nonhazardous waste landfills). Subtitle C establishes a federal program to manage hazardous wastes
from cradle to grave. The objective of the Subtitle C program is to ensure that hazardous waste is
handled in a manner that protects human health and the environment. When waste generators produce
greater than 100 kg per month of non-acutely hazardous waste, those hazardous wastes, including 1,4-
dioxane, meeting the U108 waste code description in 40 CFR 261.33, must be treated to meet the land
disposal restriction levels in 40 CFR part 268 and be disposed in RCRA subtitle C landfills. These
disposals are captured partially through the Toxics Release Inventory and are reported for onsite
facilities (TableApx H-3) and offsite facilities (TableApx H-4). Recent violations of permits are
reported in the footnotes of each table.
Review of state databases does not suggest any readily available evidence of groundwater contamination
near or coinciding with Subtitle C operations that could affect a drinking water supply. Similar review of
the data available via the WQP suggests that there are no known contaminations from RCRA Subtitle C
Landfills as reported to the TRI program (see Figure 2-14 and Figure 2-15). The absence of groundwater
contamination near RCRA Subtitle C Landfills may be attributed to many of the ongoing engineering
controls built into these facilities as well as active monitoring of groundwater wells around facilities. As
a result, EPA did not assess Subtitle C landfills further than understanding their permit violations.
Regulations established under Subtitle D ban open dumping of waste and set minimum federal criteria
for the operation of municipal waste and industrial waste landfills, including design criteria, location
restrictions, financial assurance, corrective action (clean up), and closure requirements. States play a
lead role in implementing these regulations and may set more stringent requirements. National
requirements for Subtitle D landfills are most specific for MSW landfills. MSW landfills built after 1990
must be constructed with composite liner systems and leachate collection systems in place. Composite
landfill liners consist of a minimum of two feet of compacted soil covered by a flexible membrane liner,
which work in concert to create a low hydraulic conductivity barrier and prevent leachate from being
released from the landfill and infiltrating to groundwater. A leachate collection system typically consists
of a layer of higher conductivity material above the composite liner that funnels leachate to centralized
collection points where it is removed from the landfill for treatment and disposal. Despite these controls,
releases may still occur due to imperfections introduced during construction or that form over time (Li et
al., 2015; Li et al., 2013; Mohr and DiGuiseppi, 2010); thus, groundwater monitoring is required to
identify and address any releases before there can be harm to human health and the environment. RCRA
Subtitle D requirements for non-MSW landfills are less stringent. In particular, nonhazardous industrial
landfills and C&D debris landfills do not have specified national requirements for construction and
operation and certain landfills are entirely exempt from RCRA criteria. Under the Land Disposal
Program Flexibility Act of 1996 (Pub.L. 104-119), some villages in Alaska that dispose of less than 20
Page 85 of 570
-------
tons of municipal solid waste daily (based on an annual average) may dispose of waste in unlined or
clay-lined landfills or waste piles for open burning or incineration.
There are a several potential sources of 1,4-dioxane to Subtitle D landfills. Waste generators that
produce less than 100 kg per month of non-acutely hazardous waste, including 1,4-dioxane meeting the
U108 waste code, may dispose of this waste in these landfills. Nonhazardous industrial wastes also have
the potential to contain 1,4-dioxane at variable concentrations. Consumer and commercial products may
also contain 1,4-dioxane in relatively low amounts. The greatest potential for release of disposed 1,4-
dioxane to groundwater is from landfills that do not have an adequate liner system. Thus, an objective of
this assessment is to evaluate the potential for groundwater contamination in the absence of landfill
controls.
This assessment was completed using the Hazardous Waste Delisting Risk Assessment Software
(DRAS). DRAS was specifically designed to address the Criteria for Listing Hazardous Waste identified
in Title 40 Code of Federal Regulations (40 CFR) Section 261.11(a)(3), a requirement for evaluating
proposed hazardous waste delistings. In this assessment, DRAS is being utilized to determine potential
groundwater concentrations of 1,4-dioxane after they have been disposed of into a non-hazardous waste
landfill. The results of this assessment are found in Table 2-14. This assessment relied on the default
waste loading rates for RCRA Subtitle C Landfills available in DRAS. Similarly, the assessment relied
on the default values for 1,4-dioxane as the chemical of concern. Lastly, leachate concentrations were
estimated for a range of possibilities until no risk could be identified at the lower end of those
concentrations. Because DRAS calculates a weight adjusted dilution attenuation factor (DAF) rather
than a groundwater concentration, a back of the envelop computation was used to convert the DAF to a
potential concentration that people living within 1 mile of a landfill might be exposed if the release were
not identified and remediated.
Page 86 of 570
-------
Table 2-14. Potential Groundwater Concentrations (jig/L) of 1,4-Dioxane Found in Wells within 1 Mile of a Disposal Facility
Determined by Using the DRAS Model
Leachate Concentration
(jig/L)
Loading Rate (kg)
4.55 E-04
4.55E-03
4.55E-02
4.55E-01
4.55E00
4.55 E01
4.55E02
4.55E03
4.55E04
4.55E05
1.00E-07
7.81E-13
7.46E-12
5.46E-11
5.21E-10
6.49E-09
6.17E-08
5.88E-07
5.62E-06
5.38E-05
5.13E-04
1.00E-06
7.81E-12
7.46E-11
5.46E-10
5.21E-09
6.49E-08
6.17E-07
5.88E-06
5.62E-05
5.38E-04
5.13E-03
1.00E-05
7.81E-11
7.46E-10
5.46E-09
5.21E-08
6.49E-07
6.17E-06
5.88E-05
5.62E-04
5.38E-03
5.13E-02
1.00E-04
7.81E-10
7.46E-09
5.46E-08
5.21E-07
6.49E-06
6.17E-05
5.88E-04
5.62E-03
5.38E-02
5.13E-01
1.00E-03
7.81E-09
7.46E-08
5.46E-07
5.21E-06
6.49E-05
6.17E-04
5.88E-03
5.62E-02
5.38E-01
5.13E00
1.00E-02
7.81E-08
7.46E-07
5.46E-06
5.21E-05
6.49E-04
6.17E-03
5.88E-02
5.62E-01
5.38E00
5.13E01
1.00E-01
7.81E-07
7.46E-06
5.46E-05
5.21 E-04
6.49E-03
6.17E-02
5.88E-01
5.62E00
5.38E01
5.13E02
1.00E00
7.81E-06
7.46E-05
5.46E-04
5.21E-03
6.49E-02
6.17E-01
5.88E00
5.62E01
5.38E02
5.13E03
1.00E01
7.81E-05
7.46E-04
5.46E-03
5.21E-02
6.49E-01
6.17E00
5.88E01
5.62E02
5.38E03
5.13E04
Concentrations organized by potential loading rates (kg) and potential leachate concentrations (ng /L).
Page 87 of 570
-------
2.3.2.3.1 Summary of Assessment for Disposal to Landfills
EPA determined through modeling that groundwater concentration of 1,4-dioxane increased with
increasing landfill load rate and increasing leachate concentration. With each progressive iteration of
loading rate or leachate concentration, potential groundwater concentrations increase by an order of
magnitude. When both loading rate and leachate increase by one order of magnitude, potential
groundwater concentration increase by two orders of magnitude. These increases can largely be
attributed to the increasing weight adjusted dilution attenuation factor and are what would be expected
for a chemical substances with 1,4-dioxane's physical-chemical properties (water solubility, Henry's
law constant) and fate characteristics (biodegradability, half-life in groundwater). 1,4-Dioxane migrates
in groundwater at the rate of hydraulic flow and can persist for greater than 30 days in anaerobic
environments (Adamson et al.. 2014; Mohr a iuiseppi. 2010) as described in the 2020 RE. Thus,
these concentrations are likely to represent the range of potential groundwater concentrations for PESS
living within a 1-mile radius of a RCRA Subtitle D landfills and other non-Subtitle C landfills.
EPA also determined that the modeled concentrations are within the range of concentrations of 1,4-
dioxane found in groundwater monitoring studies. A survey of monitoring studies in California has
demonstrated that 1,4-dioxane concentrations in groundwater can range from 9 |ig/L at 10th percentile
to 13,460 |.ig/L at the 90th percentile (Adamson et al.. 2014). Monitoring data from EPA's Third
Unregulated Contaminant Monitoring Rule (UCMR3) reported 1,4-dioxane concentrations in
groundwater ranging from 0.07 to 34 |ig/L (Adamson et al. JO I ; 1 c. « i1 \ JO I i) and local scales (Li
et al.. 2015; Li et al.. 2013; Mohr and DiGuiseppi. 2010). However, the modeled results may not
represent current conditions of waste management units in the United States. Both the DRAS model and
Page 88 of 570
-------
EPACMTP are based on a survey of drinking water wells located downgradient from a waste
management unit ( 8). Due to the age of the survey, it is unclear how the survey represents
current conditions and proximity of drinking water wells to disposal units. Similarly, it is not clear if the
surveyed waste management units are representative of current waste management practices.
2.3.2.4 Disposal of Hydraulic Fracturing Produced Water to Surface Impoundments
After hydraulic fracturing operations inject fluids to extract oil and gas, a substantial volume of water
may be produced through flowback. Otherwise known as produced waters, the composition of this water
depends both on the geochemistry of the injected area and the injected fluids ( .). 1,4-
Dioxane has been reported to EPA as one of the chemicals present in produced waters by 411 facilities
via FracFocus 3.0 (GWPC and IOGCC. 2022) (Table 2-15). A variety of options exist for these
produced waters after use in hydraulic fracturing operations ranging from underground injection,
treatment and subsequent use, treatment and discharge, or evaporation in surface impoundments. Each
of these options are subject to state and federal regulations ( .). When produced waters
are released to unlined surface impoundments, there is potential for groundwater contamination and
subsequent human exposure via drinking water. Thus, EPA conducted an assessment to determine the
range of groundwater concentrations within a 1-mile radius of surface impoundments receiving
produced water from hydraulic fracturing operations.
Based on the results of Monte Carlo analysis presented in Table Apx E-5, disposal to these surface
impoundments could account for up to 3 percent of all produced waters. 1,4-Dioxane has been
documented to have a concentration of 60 |ig/L in these produced waters (Lester et at.. 2015). Thus,
EPA assessed the potential for disposing of hydraulic fracturing produced water at the 5th, 50th, 95th,
and 99th percentiles as well as at the min, mean, and max to a managed surface impoundment assuming
these loading rates and concentration using DRAS. The results are presented in Table 2-15.
2.3.2.4.1 Summary of Assessment for Disposal of Hydraulic Fracturing Produced
Water
In general, EPA determined that groundwater concentrations of 1,4-dioxane would increase as more
produced water was released to surface impoundments. The values presented in Table 2-15 represent the
maximum 3 3-year receptor well concentration within a 1-mile radius of a hypothetical surface
impoundment that leaches into groundwater. With each progressive iteration of summary statistic for
loading rate, potential groundwater concentrations increase accordingly. This increase can be attributed
to the decrease in the weight-adjusted dilution attenuation factor. As the mass of 1,4-dioxane entering an
aquifer increases, it is less diluted and higher concentrations will be found downgradient. Due to its
physical-chemical properties (e.g., water solubility, Henry's Law constant) and fate characteristics (e.g.,
biodegradability, half-life in groundwater), 1,4-dioxane migrates in groundwater and can persist for
greater than 30 days in anaerobic environments (Adamson et at.. 2014; Mohr and DiGuiseppi. , ) as
described in the 2020 RE. Thus, these concentrations are likely to represent the range of potential
groundwater concentrations for people living within a 1-mile radius of a surface impoundment.
Page 89 of 570
-------
Table 2-15. Total Annua
Release Summary
Total Annual Release
(kg/site-year)
Monte Carlo
Analysis Summary
Statistic
Weight Adjusted
Dilution Attenuation
Factor
Potential
Groundwater
Concentration (jig/L)
1.68E00
Max
3.18
1.89E-05
01.87E-01
99th Percentile
3.91
1.54E-05
6.52E-02
95th Percentile
3.91
1.54E-05
1.47E-02
Mean
84
7.10E-07
3.83E-03
50th Percentile
495
1.20E-07
3.24E-05
5th Percentile
495
1.20E-07
1.06E-11
Min
135,000
0.00E00
2.3.2.4.2 Strengths, Limitations, and Sources of Uncertainty in Assessment Results
for Disposal from Hydraulic Fracturing Operations
Although it is well understood that 1,4-dioxane is present in produced waters from hydraulic fracturing
as reported in FracFocus (GWPC and IOGCC. 2022). the number of studies reporting the concentration
of the chemical substances in produced waters is limited (Lester et at.. 2015). FracFocus is generally
considered a moderately reliable source of information as it is based on data from thousands of fracking
wells across the United States. Further, both the release assessment (as discussed in Section 2.2.1.2; see
also Table Apx E-7) and the groundwater concentration assessment (Table 2-15) are modeled using a
Monte Carlo simulation. These conditions lower the confidence in the overall assessment.
2.3.3 Ambient Air Pathway
EPA developed and applied tiered methodologies and analyses to estimate ambient air concentrations
and exposures to members of the general population. These methodologies and analyses focus on
inhalation exposures to a sub-set of the general population referred to as fenceline communities.
Fenceline communities are defined as a subset of the general population that are in proximity to air
emitting facilities or a receiving water body, and who therefore may be disproportionately exposed to a
chemical undergoing risk evaluation under TSCA section 6(b). For the air pathway, proximity goes out
to 10,000 m from an air emitting source. The methodology and analyses were first presented in the 2022
Draft TSCA Screening Level Approach for Assessing Ambient Air and Water Exposures to Fenceline
Communities Version 1.0 (U.S. EPA. 2022d) and included the screening methodology and single-year
analysis. In response to SACC recommendations on the 2022 fenceline report to consider multiple years
of release data, EPA added the multi-year analysis to this supplemental risk evaluation. However, the
order of these analyses caused some confusion when the draft 1,4-dioxane supplemental risk evaluation
went through public comment and peer review because the multi-year analysis uses a lower tier model
(IIOAC7) after the single-year analysis used a more complex, higher tier model (American
Meteorological Society/Environmental Protection Agency Regulatory Model or AERMOD).
Nonetheless, since the multi-year analysis was intended to identify if consideration of multiple years of
release data resulted in different exposure characterization and risk conclusions using a lower tier model
to screen any differences in exposure/risks is a logical first step. Ultimately, EPA did not identify
differences in either exposure characterizations or risk conclusions when considering multiple years of
release data and therefore did not pursue additional analysis using both multiple years of data and the
higher tier model (AERMOD). The specific methodologies used in this assessment to evaluate general
7 The 110AC website is available at https://www.epa.gov/tsea~screening~tooIs/iioac~integrated~iiMloor-oiitdoor~air~caIeiilator.
Page 90 of 570
-------
population exposures to 1,4-dioxane in air are briefly described in Figure 2-18. Additional details on the
methodologies are provided in Appendix J.
Figure 2-18. Brief Description of Methodologies and Analyses Used to Estimate Ambient
Air Concentrations and Exposures
EPA used the air release estimates obtained using the methodology described in Section 2.1.1.3 as direct
inputs for the models used to estimate exposure concentrations at various distances from a releasing
facility. EPA expanded upon the methods described in the 2022 Draft TSCA Screening Level Approach
for Assessing Ambient Air and Water Exposures to Fenceline Communities Version 1.0 (U.S. EPA.
2022d) in response to SACC comments/recommendations by evaluating potential aggregate
concentrations from multiple facilities.
2.3.3.1 Measured Concentrations in Air
EPA did not identify quantitative outdoor air monitoring data for 1,4-dioxane.
2.3.3.2 Modeled Concentrations in Air
Because there is no air monitoring data for 1,4-dioxane, the Agency relied upon modeling to estimate
exposure concentrations to fenceline communities at various distances from a releasing facility.
Modeling was used for each analysis described in Figure 2-18 for 1,4-dioxane. For scenarios where the
screening methodology indicated a need for further analysis, EPA performed a full analysis using the
AERMOD and/or IIOAC. IIOAC analysis was performed for three COUs where no site-specific data
were available (Hydraulic fracturing, Industrial laundry facilities, Institutional laundry facilities) and is
briefly described in Section 2.3.3.2.4 with results presented and discussed in Sections 5.2.2.3.2 and
5.2.2.3.3. An expanded analysis to consider aggregate exposures was performed for 1,4-dioxane in
response to SACC comments/recommendations on the 2022 Draft TSCA Screening Level Approach for
Assessing Ambient Air and Water Exposures to Fenceline Communities Version 1.0 ( )22d).
Page 91 of 570
-------
2.3.3.2.1 Ambient Air: Screening Methodology
The Ambient Air: Screening Methodology utilizes EPA's IIOAC model to estimate high-end and central
tendency (mean) 1,4-dioxane exposure concentrations in ambient air at three distances from an emitting
facility: 100, 100 to 1,000, and 1,000 m. EPA developed and evaluated a range of exposure scenarios for
each of two categorical release amounts8 designed to capture a variety of release types, topography,
meteorological conditions, and release scenarios. A diagram of these exposure scenarios is provided in
Appendix J. Findings from the Ambient Air: Screening Methodology were used to inform the need for a
higher-tier analysis as well as provide insight into whether risk estimates above the benchmarks are or
are not expected for 1,4-dioxane.
The Ambient Air: Screening Methodology design inherently includes both estimates of exposures as
well as estimates of risks to inform the need, or potential need, for further analysis. If findings from the
Ambient Air: Screening Methodology estimate risk (acute non-cancer, chronic non-cancer, or cancer)
for a given chemical above (or below as applicable) typical Agency benchmarks, EPA generally will
conduct a higher-tier analysis of exposures and associated risks for that chemical. If findings from the
Ambient Air: Screening Methodology estimate risks that do not exceed (or fall below as applicable)
benchmarks, EPA may still conduct a limited higher-tier analysis at distances very near a releasing
facility (less than 100 m) to ensure potential risks are not missed.
A more detailed description of the Ambient Air: Screening Methodology for 1,4-dioxane is provided in
Appendix J, along with summarized results. In general, for 1,4-dioxane, the results of this analysis
identified risk estimates above screening benchmarks for cancer at multiple distances and for multiple
releases (max and mean). In accordance with the tiered methodology presented to the SACC in the 2022
Draft TSCA Screening Level Approach for Assessing Ambient Air and Water Exposures to Fenceline
Communities Version 1.0 (U.S. EPA. 2022d). because risk estimates exceeded the benchmark for cancer
based on the Ambient Air: Screening Methodology, EPA conducted multiple higher-tier analyses of all
facilities reporting releases of 1,4-dioxane to TRI.
2.3.3.2.2 Ambient Air: Single Year Methodology (AERMOD)
The Ambient Air: Single Year Methodology (AERMOD) utilizes the EPA's American Meteorological
Society/Environmental Protection Agency Regulatory Model (AERMOD)9 to estimate 1,4-dioxane
concentrations in ambient air at eight finite distances (5, 10, 30, 60, 100, 2,500, 5,000, and 10,000 m)
and one area distance from an emitting facility.10 The single year modeling analysis was conducted as
part of the 2022 fenceline work, and therefore completed prior to consideration of multiple years of
release data (multi-year analysis). EPA modeled two different types of release estimates, as applicable,
for 1,4-dioxane: (1) facility-specific chemical releases with source attribution when 2019 TRI data was
available, and (2) alternative release estimates representing a generic facility when 2019 TRI data was
not available for an OES. Daily and period average outputs were obtained via modeling, and post-
processing scripts were used to extract a variety of statistics from the modeled concentration
distribution, including the 95th (high-end), 50th (central tendency), and 10th (low-end) percentile 1,4-
dioxane concentrations at each distance modeled.
8 The pre-screening methodology from the 2022 fenceline analysis evaluated two categorical release values across all
facilities reporting releases to the 2019 TRI. The first is the maximum single facility release reported across all facilities; the
second is the mean (arithmetic average) of all releases reported across all facilities reporting.
9 See https://www.epa.gOv/scram/air-qnality-dispersion-modeiing-preferred-and-recommended-modeis#aermod for more
information.
10 For the one "area distance" evaluated, receptors are placed in a cartesian grid between approximately 200 and 900 m, at
100 m spacing. This results in a total of 456 receptors. The exposure estimates for the area distance represent the arithmetic
average (mean) exposure concentration across all 456 receptors within the "area distance" for each day.
Page 92 of 570
-------
A summary of the concentration ranges estimated using the Ambient Air: Single Year Methodology
(AERMOD) is provided in Table 2-16. The summary includes 11 OESs and select statistics (maximum,
mean, median, and minimum) calculated from the modeled concentration distributions within each OES
at each distance modeled. The associated range of estimated concentrations is based on 33 years
exposure duration and the maximum 95th percentile annual average exposure concentrations for each
distance. Although the range of concentrations are provided, there are many instances where the range
extends as many as 12 orders of magnitude from minimum to maximum concentration. This occurs
because within each OES there are several individual facilities evaluated and, in most cases, the reported
release values from each individual facility can vary widely (from 500 lb to several hundred thousand
pounds), which in turn affects the range of estimated exposure concentrations at a given distance.
Therefore, in trying to summarize the wide variety of releases into a single range, the variation in
estimated concentrations will also appear extensive. This is not indicative of an inadequate analysis or
methodology, but solely based on the variability of releases across facilities within a given OES.
Page 93 of 570
-------
Table 2-16. Summary of Select Statistics for the 95th Percentile Estimated Annual Average Concentrations from the "Full-
Screening" Analysis for 1,4-Dioxane Releases Reported to TRI
OES
# Facilities
Evaluated
in OES
Statistic
Annual Average Concentration (ppm) Estimated within 5 to 10,000 m of Releasing Facilities
5
10
30
60
100
100 to
1,000
2,500
5,000
10,000
Disposal
15
Max
4.26E-03
5.05E-03
1.80E-03
6.90E-04
3.15E-04
2.95E-05
1.89E-06
6.28E-07
2.09E-07
Mean
4.76E-04
6.92E-04
2.79E-04
1.12E-04
5.39E-05
6.00E-06
3.46E-07
1.18E-07
4.12E-08
Median
8.44E-06
1.65E-05
9.35E-06
8.70E-06
5.81E-06
7.64E-07
4.53E-08
1.48E-08
4.81E-09
Min
3.31E-15
9.85E-14
5.17E-11
9.72E-10
2.03E-09
1.10E-09
1.21E-10
3.81E-11
1.22E-11
Dry film
lubricant
8
Max
1.61E-10
7.14E-09
5.10E-07
3.88E-06
6.29E-06
9.92E-07
2.79E-08
8.44E-09
3.68E-09
Mean
2.06E-11
9.46E-10
1.90E-07
2.28E-06
4.05E-06
8.14E-07
1.95E-08
5.94E-09
2.45E-09
Median
2.46E-13
3.58E-11
1.59E-07
2.21E-06
4.00E-06
7.75E-07
1.88E-08
6.02E-09
2.66E-09
Min
4.05E-18
2.19E-13
5.64E-08
9.23E-07
2.39E-06
7.39E-07
1.36E-08
4.02E-09
1.40E-09
Ethoxylation
byproduct
6
Max
6.53E-03
1.36E-02
7.33E-03
3.09E-03
1.64E-03
3.81E-04
2.20E-05
9.00E-06
3.45E-06
Mean
1.74E-03
3.05E-03
1.49E-03
6.18E-04
3.23E-04
7.03E-05
4.02E-06
1.62E-06
6.13E-07
Median
2.44E-04
4.40E-04
2.08E-04
8.70E-05
4.40E-05
6.36E-06
3.09E-07
1.03E-07
3.40E-08
Min
4.08E-14
6.32E-13
4.29E-10
5.22E-09
1.15E-08
4.99E-09
7.17E-10
3.39E-10
1.40E-10
Film cement
1
Max
1.25E-04
1.31E-04
4.41E-05
2.28E-05
1.25E-05
2.29E-06
1.38E-07
4.60E-08
1.52E-08
Mean
3.90E-05
4.87E-05
2.04E-05
9.36E-06
4.82E-06
7.95E-07
4.44E-08
1.48E-08
4.87E-09
Median
2.02E-05
2.93E-05
1.74E-05
7.33E-06
3.57E-06
5.68E-07
2.48E-08
8.10E-09
2.65E-09
Min
3.17E-06
6.36E-06
4.97E-06
1.95E-06
8.99E-07
1.32E-07
4.44E-09
1.42E-09
4.61E-10
Functional
fluids (open-
system)
2
Max
1.28E-05
2.36E-05
1.03E-05
1.08E-05
1.82E-05
7.42E-06
6.78E-07
2.47E-07
8.81E-08
Mean
6.40E-06
1.18E-05
5.74E-06
7.71E-06
1.08E-05
4.24E-06
3.88E-07
1.45E-07
5.28E-08
Median
6.40E-06
1.18E-05
5.74E-06
7.71E-06
1.08E-05
4.24E-06
3.88E-07
1.45E-07
5.28E-08
Min
1.66E-11
1.93E-10
1.18E-06
4.61E-06
3.37E-06
1.06E-06
9.70E-08
4.30E-08
1.74E-08
Import and
repackaging
1
Single Facility
2.70E-11
5.57E-10
5.52E-08
4.17E-07
8.70E-07
3.21E-07
6.72E-08
4.12E-08
2.23E-08
Page 94 of 570
-------
OES
# Facilities
Evaluated
in OES
Statistic
Annual Average Concentration (ppm) Estimated within 5 to 10,(MM) m of Releasing Facilities
5
10
30
60
100
100 to
1,000
2,500
5,000
10,000
Industrial uses
12
Max
4.19E-03
4.78E-03
1.54E-03
5.67E-04
2.80E-04
7.15E-05
8.79E-06
3.22E-06
1.13E-06
Mean
8.76E-04
1.14E-03
4.08E-04
1.70E-04
9.63E-05
1.69E-05
1.48E-06
5.45E-07
1.94E-07
Median
8.76E-05
1.14E-04
3.83E-05
1.65E-05
9.94E-06
2.23E-06
3.19E-07
1.17E-07
4.04E-08
Min
7.75E-13
1.69E-12
2.40E-09
2.50E-08
1.23E-08
1.10E-09
6.36E-11
1.97E-11
6.14E-12
Laboratory
chemical
1
Max
2.06E-03
2.15E-03
7.26E-04
3.75E-04
2.06E-04
3.76E-05
2.27E-06
7.57E-07
2.50E-07
Mean
6.84E-04
8.52E-04
3.58E-04
1.64E-04
8.46E-05
1.39E-05
7.77E-07
2.59E-07
8.55E-08
Median
4.30E-04
5.65E-04
3.15E-04
1.36E-04
6.68E-05
1.08E-05
4.82E-07
1.59E-07
5.24E-08
Min
7.39E-05
1.48E-04
1.16E-04
4.55E-05
2.09E-05
3.08E-06
1.03E-07
3.30E-08
1.07E-08
Manufacturing
1
Single Facility
8.73E-03
1.63E-02
7.69E-03
3.22E-03
1.59E-03
1.42E-04
8.21E-06
2.54E-06
7.92E-07
PET
manufacturing
13
Max
8.01E-03
9.57E-03
3.50E-03
1.40E-03
6.43E-04
1.07E-04
2.07E-05
1.24E-05
6.58E-06
Mean
1.41E-03
1.89E-03
7.83E-04
3.36E-04
1.85E-04
3.31E-05
4.23E-06
2.08E-06
9.60E-07
Median
8.00E-04
1.64E-03
5.21E-04
2.27E-04
1.42E-04
2.64E-05
2.48E-06
1.09E-06
3.94E-07
Min
6.04E-12
8.54E-11
3.01E-08
2.43E-07
5.56E-07
3.02E-07
4.33E-08
2.07E-08
9.30E-09
Spray foam
application
1
Max
7.79E-07
8.40E-07
2.85E-07
1.50E-07
8.55E-08
1.55E-08
1.72E-09
6.30E-10
2.45E-10
Mean
2.68E-07
3.30E-07
1.34E-07
6.21E-08
3.28E-08
5.29E-09
4.78E-10
1.67E-10
5.97E-11
Median
1.41E-07
1.95E-07
1.14E-07
4.88E-08
2.36E-08
3.64E-09
2.25E-10
7.40E-11
2.40E-11
Min
2.51E-08
4.43E-08
3.45E-08
1.36E-08
6.07E-09
8.42E-10
3.26E-11
1.10E-11
3.74E-12
Page 95 of 570
-------
Based on the air concentrations estimated through the Ambient Air: Single Year Methodology, EPA also
estimated potential aggregate air concentrations resulting from the combined releases of multiple
facilities in proximity to each other. Details of the methods used to aggregate exposure and
corresponding risk are presented in Appendix J.4.
2.3.3.2.3 Ambient Air: Multi-Year Analysis (IIOAC)
The multi-year analysis utilizes EPA's IIOAC model to estimate high-end and central tendency (mean)
1,4-dioxane concentrations in ambient air at three distances from an emitting facility: 100, 100 to 1,000,
and 1,000 m. The multi-year analysis incorporates SACC recommendations on the 2022 Draft TSCA
Screening Level Approach for Assessing Ambient Air and Water Exposures to Fenceline Communities
Version 1.0 (U .S. EPA. 2022d) by evaluating multiple years of chemical release data to estimate
exposures and associated risks to fenceline communities including an average release across all years of
reported data for each facility evaluated. This is achieved for 1,4-dioxane by expanding upon
methodologies described in the 2022 fenceline report and conducting a facility-by-facility evaluation of
all 1,4-dioxane releases reported to TRI (2015 through 2020). Data for these 6 years were obtained from
the TRI database (TRI basic plus files downloaded on August 5, 2022). Annual release data for 1,4-
dioxane were extracted from the entire TRI data set for all facilities reporting air releases of 1,4-dioxane
for one or more years between 2015 and 2020. Facilities were categorized into occupational exposure
scenarios for modeling purposes to inform the release scenarios evaluated.
The multi-year analysis provides highlights of the year-to-year variability that exists in the release data
and illustrates the potential impact of considering multiple years of TRI data on exposure and risk
estimates. The findings from the multi-year analysis can also be used in a comparative manner to
determine how representative the single year of data used for the Ambient Air: Single Year
Methodology (AERMOD) presented in the 2022 fenceline report is or to provide additional confidence
in the findings from the Ambient Air: Single Year Methodology (AERMOD) described in the 2022
fenceline report for purposes of estimating exposures and associated risks to fenceline communities. In
broader terms, the multi-year analysis provides both a broad analysis of multiple years of release data
and enables a general comparison to the Ambient Air: Single Year Methodology (AERMOD) results
described above and in the 2022 fenceline report.
2.3.3.2.4 Ambient Air: IIOAC Methodology for COUs Without Site-Specific Data
(Hydraulic Fracturing, Industrial, and Institutional Laundry Facilities)
For COUs without site-specific data, EPA's IIOAC model was used to estimate high-end and central
tendency (mean) 1,4-dioxane concentrations in ambient air at three distances from an emitting facility
(100, 100 to 1,000, and 1,000 m). This methodology was applied for three unique COUs (hydraulic
fracturing, and industrial, and institutional laundry facilities) where there was no site-specific data
available for modeling in the 2019 1,4-dioxane risk evaluation. Environmental releases (fugitive and
stack) along with other data (like days of release) for these COUs were estimated using Monte Carlo
modeling. As such, the Ambient Air: IIOAC Methodology for COUs without site-specific data was
developed to allow modeling all possible iterations of releases provided across eight different exposure
scenarios, including consideration of source attribution as well as actual days of release. Additionally,
the product form for laundry detergent was provided, allowing for analysis of releases associated with
detergent in vapor only form, as well as solid form (particulate) either coarse (PM10) or fine (PM2.5). A
description of this methodology is provided in Appendix J along with a summary of the model inputs
and exposure scenarios evaluated. A full list of the inputs, exposure scenarios, and results is provided in
1,4-Dioxane Supplemental Information File: Air Exposure and Risk Estimates for 1,4-Dioxane
Emissions from Hydraulic Fracturing Operations ( 2024b) and 1,4-Dioxane Supplemental
Information File: Air Exposures and Risk Estimates for Industrial Laundry ( 2024c).
Page 96 of 570
-------
In response to SACC recommendations, EPA made revisions to the release assessments for hydraulic
fracturing and industrial and institutional laundries. The Agency considered the magnitude of impact of
these revisions to estimated releases to determine whether these revisions should be carried through
corresponding modeled air concentrations. For both hydraulic fracturing and for laundries, EPA
concluded that the shift in release estimates based on alternate assumptions, inputs or model
distributions are not expected to shift exposure and risk estimates sufficiently to alter the overall risk
conclusion. For those COUs, EPA has therefore retained the original air concentration modeling,
exposure and risk estimates based on the original release assessments published in the draft supplement.
2.3.3.3 Strengths, Limitations, and Sources of Uncertainty for Modeled Air
Concentrations
EPA has medium to high confidence in the air concentrations estimated from TRI release data using
IIOAC and AERMOD.
I 10AC
IIOAC estimates air concentrations at three pre-defined distances (100, 100 to 1,000, and 1,000 m). The
inherent distance limitations of IIOAC do not allow estimation of exposures closer to a facility (less than
100 m from the facility) where we expect to see higher exposures from fugitive releases.
IIOAC uses meteorological data from 14 pre-defined meteorological stations representing large regions
across the United States. This generalizes the meteorological data used to estimate exposure
concentrations where competing conditions can influence the exposure concentrations modeled upwind
and downwind of a releasing facility. To reduce the uncertainties associated with using regional
meteorological data, EPA conducted a sensitivity analysis of all 14 pre-defined meteorological stations
to identify which two within IIOAC tended to result in a high-end and central tendency estimate of
exposure concentrations. This maintained a more conservative exposure concentration estimate which is
then used in calculations to estimate risks. This approach adds confidence to the findings by ensuring,
potential risks would be captured under a high-end exposure scenario, while also providing insight into
potential risks under a less conservative exposure scenario (central tendency).
AERMOD
AERMOD is an EPA regulatory model and has been thoroughly peer reviewed; therefore, the general
confidence in results from the model is high but relies on the integrity and quality of the inputs used and
interpretation of the results. For the full analysis, EPA used releases reported to the 2019 TRI as direct
inputs to AERMOD. Although there is some uncertainty around the representativeness of using only a
single year of data, AERMOD successfully estimated exposure concentrations to fenceline communities.
Furthermore, in response to SACC recommendations to use multiple years of data to estimate exposures
and associated risks, EPA developed the IIOAC and conducted a multi-year analysis using 6 years of
TRI data and compared the results to those of AERMOD and found exposure concentration estimates
from the 2019 data is generally representative of other years.
AERMOD relied upon the latitude/longitude information reported by each facility to TRI as the location
for the point of release. Although this may generally be a close approximation of the release point for a
small facility (for example a single building), it may not represent the release point within a much larger
facility. Therefore, there is some uncertainty associated with the modeled distances from each release
point and the associated exposure concentrations to which fenceline communities may be exposed. For
small facilities where the latitude/longitude may closely approximate the release point, there is a less
uncertainty that the estimated exposure at the associated distance is representative of exposure to
Page 97 of 570
-------
fenceline communities at that distance. For larger facilities, where the latitude/longitude may be several
hundred meters away from the actual release point, there is a higher uncertainty that the estimated
exposure at the associated distance is representative of exposure to fenceline communities at that
distance.
The TRI datasets used for both AERMOD and IIOAC do not include source specific stack parameters
that can affect plume characteristics and associated dispersion of the plume. Therefore, EPA used pre-
defined stack parameters within IIOAC to represent stack parameters of all facilities modeled using each
of these methodologies. Those stack parameters include a stack height 10 m above ground with a 2-
meter inside diameter, an exit gas temperature of 300° Kelvin, and an exit gas velocity of 5 m per
second (see Table 6 of the IIOAC User Guide). These parameters were selected since they represent a
slow-moving, low-to-the-ground plume with limited dispersion which results in a more conservative
estimate of exposure concentrations at the distances evaluated. As such, these parameters may result in
some overestimation of emissions for certain facilities modeled.
Additionally, the assumption of a 10x 10 area source for fugitive releases may impact the exposure
estimates very near a releasing facility (5 and 10 m from a fugitive release). This assumption places the
receptor at 5 m directly on top of the release point which may result in an over or underestimation of
exposure. This assumption places the 10-meter receptor just off the release point that may again result in
either an over or underestimation of exposure depending on other factors like meteorological data,
release heights, and plume characteristics.
For facilities reporting releases to TRI via a TRI Form A (which is allowed for use by those facilities
releasing less than 500 lb of the chemical reported), EPA assumed the maximum release value of 500 lb
for exposure modeling purposes. TRI Form A reporters do not provide source attribution (fugitive or
stack releases) so EPA modeled each facility associated with a Form A submittal twice—once assuming
all 500 lb of the reporting threshold was fugitive and once assuming all 500 lb of the reporting threshold
was stack. There is no way to attribute a certain portion of the releases to each release type, so this
modeling approach represents a conservative estimate, in terms of total release, but may overestimate
exposure concentrations associated with each release type if a facility did not actually release all 500 lb
via a single release type or even combined release type. To avoid the potential double counting of
facility releases for TRI Form A reporters, when presenting potential exposures EPA presented only the
highest (more conservative) exposure concentration estimated for either of the two release types for
purposes of evaluating potential risks to fenceline communities. Given the exposure scenarios modeled,
this tended to result from the exposure scenario which assumed all 500 lb of the release were fugitive
releases.
Page 98 of 570
-------
3 HUMAN EXPOSURES
1,4-Dioxane - Human Exposures (Section 3):
Key Points
EPA evaluated reasonably available information for occupational exposures to 1,4-dioxane present
as a byproduct, and general population, including PESS, exposures to 1,4-dioxane present in the
environment.
• Occupational exposures for workers and ONUs to 1,4-dioxane present as a byproduct occur
during manufacturing, through use of commercial products, or in hydraulic fracturing
operations.
• General population exposures to 1,4-dioxane occur when 1,4-dioxane is present in potential
drinking water sources or ambient air, particularly in fenceline communities.
• EPA considered the potential for increased exposures across PESS factors throughout the
exposure assessment. PESS categories incorporated into this supplemental exposure
assessment include
o Lifestage (including formula-fed infant exposures),
o Occupational exposures (including high-end exposure scenarios), and
o Geography/site-specific factors (i.e., fenceline community exposures)
3.1 Occupational Exposures
1,4-Dioxane - Occupational Exposures (Section 3.1):
Key Points
EPA considered the reasonably available information to evaluate occupational exposures.
• EPA estimated occupational exposures to 1,4-dioxane through air and skin. The Agency
estimated both high-end and central tendency exposures for occupational exposure scenarios
associated with each COU.
• Exposure for most COUs was estimated based on monitoring data. For COUs without
monitoring data, EPA applied Monte Carlo statistical modeling approaches to estimate
exposures.
The following sections describe EPA's approach to assessing occupational exposures for OESs
involving industrial and commercial products containing 1,4-dioxane as a byproduct. The assessed OESs
include textile dye, antifreeze, surface cleaner, dish soap, dishwasher detergent, institutional and
industrial laundries, paints and floor lacquer, PET byproducts, ethoxylation process byproducts, and
hydraulic fracturing. For a crosswalk linking COUs to OESs, see Table 2-1. The remaining OESs have
occupational exposure assessments in Section 2.4.1 of the Final Risk Evaluation for 1,4-Dioxane (U.S.
EPA. 2020c).
EPA distinguishes between exposures to workers and exposures to ONUs. Normally, workers may
handle 1,4-dioxane and have direct contact with the chemical, such as operators, applicators (e.g., for
Page 99 of 570
-------
paints, cleaners), and maintenance workers. ONUs work in the general vicinity of workers but do not
handle 1,4-dioxane and do not have direct contact with 1,4-dioxane, such as supervisors and managers.
EPA evaluated acute and chronic inhalation exposures to workers and ONUs, and dermal exposures to
workers. The Agency did not assess dermal exposures to ONUs as EPA does not expect ONUs to have
routine dermal exposures in the course of their work.
The occupational exposure assessment for each COU comprises the following components:
• Process Description of the COU, including the role of the chemical in the use; process vessels,
equipment, and tools used during the COU; and descriptions of the worker activities, including
an assessment for potential points of worker exposure.
• Number of Sites that use the chemical for the given COU.
• Number of Workers and ONUs potentially exposed to the chemical for the given COU. Unless
mentioned otherwise in this report, the total number of workers and ONUs are number of
personnel per site per day. The details on estimation of the number of workers and ONUs are
discussed below for each COU.
• Central Tendency and High-End Estimates of Inhalation Exposure to workers and
occupational non-users. See "General Approach and Methodology for Environmental Releases"
for a discussion of EPA's statistical analysis approach for assessing inhalation exposure.
• Dermal Exposure estimates for multiple scenarios, accounting for simultaneous absorption and
evaporation, and different protection factors of glove use.
• Users include adult workers (>16 years old) exposed to 1,4-dioxane for 8-hour exposure.
• ONUs include adult workers (>16 years old) exposed to 1,4-dioxane indirectly by being in the
same work area of the building.
3.1.1 Approach and Methodology
EPA developed occupational exposure values representative of central tendency (50th percentile, mean)
conditions and high-end (90th and 99.9th percentiles). Additional explanation of central tendency and
high-end conditions are described in the Final Risk Evaluation for 1,4-Dioxane ( 2020c).
3.1.1.1 Process Description, Number of Sites, Number of Workers, and ONUs
EPA performed a literature search to find descriptions of processes involving 1,4-dioxane and
worker activities that could potentially result in occupational exposures. This literature search was
specific to the scope of this supplement and is described in Section 1.4. A summary of the data quality
evaluation results for the 1,4-dioxane occupational exposure sources are presented in the attachment
Systematic Review Supplemental File: Data Quality Evaluation and Data Extraction Information for
Environmental Release and Occupational Exposure ( 324x).
EPA used a variety of sources to supplement the data found through the Systematic Review
process. The additional sources included relevant NIOSH Health Hazard Evaluations, Generic
Scenarios, and ESDs. These sources were sometimes used to provide process descriptions of the COUs
as well as estimates for the number of sites and number of workers. Because CDR data were not
available for the COUs included in this occupational exposure assessment, EPA used data from the
Bureau of Labor Statistics (BLS) and the U.S. Census' Statistics of US Businesses (SUSB) to estimate
the number of sites, workers, and ONUs for each OES. This approach involved the identification of
relevant Standard Occupational Classification (SOC) codes within the BLS data for the identified
NAICS codes for each OES. First, EPA identified the affected NAICS codes. Then, EPA reviewed
occupation descriptions to designate which SOC codes contained potentially exposed workers and
Page 100 of 570
-------
ONUs. The Agency refined the estimates by using U.S. Census Bureau data. Next, EPA estimated the
percentage of workers using 1,4-dioxane instead of other chemicals to calculate number of workers per
site. Finally, this data was separated by COU. Additional details on this approach can be found in
Appendix G.5 of the Final Risk Evaluation for 1,4-Dioxane (U.S. EPA. 2020c). These estimates were
utilized in Monte Carlo modeling of environmental releases and occupational exposures.
3.1.1.2 Inhalation Exposures Approach and Methodology
EPA used inhalation monitoring data from literature sources having high or medium data quality ratings
during data evaluation. EPA used modeling approaches to estimate potential inhalation exposures where
inhalation monitoring data were not available.
The Agency reviewed workplace inhalation monitoring data collected by government agencies such as
OSHA and NIOSH, and monitoring data found in published literature {i.e., personal exposure
monitoring data and area monitoring data). Central tendency and high-end exposure values were
calculated from the monitoring data provided in the sources depending on the size of the dataset {i.e.,
number of data points). Where discrete sampling points were not provided in the source and EPA was
unable to calculate central tendency and high-end values, the Agency used values of central tendency
and high-end that were provided in the source. EPA's approach for evaluating central tendency and
high-end estimates from inhalation monitoring data is further discussed in the Final Risk Evaluation for
1,4-Dioxane ( 320c). EPA used the following types of monitoring data of 1,4-dioxane from
various sources to estimate occupational inhalation exposure:
• Personal sample monitoring data from directly applicable scenarios {e.g., personal breathing
zone [PBZ]). This type of monitoring data was used for the textile dye, surface cleaner, dish
soap, paint and floor lacquer, PET byproduct, and the Ethoxylation process byproduct OESs.
• Personal sample monitoring data from potentially applicable or similar scenarios. Specifically,
PBZ data from the dish soap OES was also used for the dishwasher detergent OES because these
OESs are expected to be similar.
EPA used the following models and modeling approaches to estimate occupational inhalation exposure
where no monitoring data were found:
• Monte-Carlo statistical modeling approaches, which was used for the antifreeze, laundry
detergent, and hydraulic fracturing OES. EPA developed these models for the purposes of this
assessment. The models and the associated sources of data used in the modeling are described in
detail in Appendices F.7, F.8, and F.9, respectively.
• Additional modeling approaches, including the use of surrogate data and fundamental modeling
approaches for the spray polyurethane foam OES in the Final Risk Evaluation for 1,4-Dioxane
(U.S. EPA. 2020c). Although this OES is included in the scope of this supplement, EPA
evaluated occupational exposure estimates for this OES in the published risk evaluation and
these estimates remain unchanged in this supplement.
• EPA AP-42 Loading Model estimates vapor releases that occur when vapor is displaced by
liquid during container loading. It calculates a vapor generation rate (G) using the physio-
chemical properties of the chemical.
• EPA Mass Balance Inhalation Model estimates occupational inhalation exposures assuming the
air immediately around the source of exposure behaves as a well-mixed zone. The Agency used
the vapor generation rate (G), calculated using the EPA AP-42 Loading Model, in conjunction
with this model to develop estimates of inhalation exposure.
Page 101 of 570
-------
• Generic Model for Central Tendency and High-End Inhalation Exposure to Total and Respirable
Particulates Not Otherwise Regulated (PNOR) estimates occupational inhalation exposures to
particulates containing the chemical using OSHA PNOR data.
EPA did not utilize occupational exposure limits to estimate occupational inhalation exposures in this
assessment because sufficient monitoring data or modeling approaches were available for all OES.
The Agency then used measured or modeled air concentrations to calculate exposure concentration
metrics essential for risk assessment. These exposures are presented as 8-hour time weighted averages
(TWAs) and used to calculate average daily concentrations (ADCs) and lifetime average daily
concentrations (LADCs). The ADC is used to estimate chronic, non-cancer risks and the LADC is used
to estimate chronic, cancer risks. These calculations required additional parameter inputs, such as years
of exposure, exposure duration and frequency, and lifetime years. See Appendix F.l for more
information about parameters and equations used to calculate acute and chronic exposures.
3.1.1.3 Dermal Exposures Approach and Methodology
EPA modeled dermal doses using the EPA Dermal Exposure to Volatile Liquids Model. This model
determines a dermal potential dose rate based on an assumed amount of liquid on skin during one
contact event per day and the steady-state fractional absorption for 1,4-dioxane. The amount of liquid on
the skin is adjusted by the weight fraction of 1,4-dioxane in the liquid to which the worker is exposed.
This is the same approach that EPA used in the Final Risk Evaluation for 1,4-Dioxane (U.S. EPA.
2020c).
3.1.1.4 Engineering Controls and Personal Protective Equipment
EPA's occupational exposure estimates do not assume the use of engineering controls or PPE.
Additionally, as discussed in Section 5.2.1, the risk estimates are based on exposures to workers in the
absence of PPE such as gloves or respirators. Reasonably available monitoring data or information on
effectiveness of engineering control and PPE for reducing occupational exposures to 1,4-dioxane during
the assessed OESs were not available. This section presents a general discussion on engineering controls
and PPE for informative purposes only.
OSHA recommends employers utilize the hierarchy of controls for reducing or removing hazardous
exposures. The most effective controls are elimination, substitution, or engineering controls. Respirators,
and any other personal protective equipment (PPE), are the last means of worker protection in the
hierarchy of controls and should only be considered when process design and engineering controls
cannot reduce workplace exposure to acceptable levels. OSHA's Respiratory Protection Standard (29
CFR 1910.134) provides a summary of respirator types by their assigned protection factor (APF). OSHA
defines the APF to mean the workplace level of respiratory protection that a respirator or class of
respirators is expected to provide to employees when the employer implements a continuing, effective
respiratory protection program according to the requirements of the OSHA Respiratory Protection
Standard. Exposure limits, respirator requirements, worker respirator use rates, and a table of APFs for
different types of respirators are provided in the 2020 RE ( 2020c).
OSHA's hand protection standard (29 CFR 1910.138) states that employers must select and require
employees to use appropriate hand protection when employees are expected to be exposed to hazards
such as those from skin absorption of harmful substances; severe cuts or lacerations; severe abrasions;
punctures; chemical burns; thermal burns; and harmful temperature extremes. Dermal protection
selection provisions are provided in § 1910.138(b) and require that appropriate hand protection is
selected based on the performance characteristics of the hand protection relative to the task(s) to be
performed, conditions present, duration of use, and the hazards to which employees will be exposed.
Page 102 of 570
-------
Unlike respiratory protection, OSHA standards do not provide protection factors (PFs) associated with
various hand protections. Data regarding the frequency of effective glove use in industrial and
commercial settings is limited. Although there are no regulatory standards containing dermal protection
factors, the European Center for Ecotoxicity and Toxicology of Chemicals (ECETOC) targeted risk
assessment (TRA) tool includes fixed, assigned protection factors equal to 5, 10, or 20 for various
dermal protection strategies. These are discussed in Appendix F.3 and further explained in the 2020 RE
( 2020c).
3.1.2 Occupational Exposure Estimates
In this section, EPA provides a summary of the exposure estimates for each OES, including estimates
for number of workers and ONUs, inhalation exposures, and dermal exposures. For the crosswalk
linking COUs to OESs, see Table 2-1. Note that EPA assessed dermal exposures for all OESs with the
same methodology, which is described at the end of this section.
3.1.2.1 Summary of Inhalation Exposure Assessment
EPA estimated central tendency and high-end occupational inhalation exposures using various methods
and information sources—including OSHA data, NIOSH health hazard evaluation data, and GSs and
ESDs with Monte Carlo modeling. EPA estimated inhalation exposures as 8-hour TWA values for the
COUs included in this supplement per Table 2-1. Using the estimated central tendency and high-end
inhalation exposures with the estimated exposure frequency, EPA then calculated the cancer and non-
cancer exposures using the calculations described in Appendix F.l.
A summary of the occupational inhalation exposures is presented Table Apx F-34. EPA used
monitoring data to estimate occupational inhalation exposures to workers for the textile dye, surface
cleaner, dish soap, dishwasher detergent, paint and floor lacquer, PET byproduct, and ethoxylation
byproduct conditions of use. This monitoring data was found to be relevant to these scenarios and based
on medium to high data quality. However, several of the scenarios had a low number of samples and
may have preceded changes in current industry practices. Additionally, sufficient representation of the
entire industry is uncertain due to the limited number of sites. For the remaining conditions of use
included in this supplement, which are antifreeze, laundry detergent, and hydraulic fracturing, EPA did
not find reasonably available monitoring data and estimated worker inhalation exposure using GSs and
ESDs with Monte Carlo modeling. The applied models are directly relevant to these conditions of use,
but the underlying distributions may not sufficiently capture variability across entire industry sectors.
For both measured and modeled data, the degree of certainty to which these data represent the true
distribution of exposure and the potential over- or underestimation of exposure is unknown.
Monitoring data and modeling approaches were not available to estimate occupational inhalation
exposures for ONUs. The ONU exposures are anticipated to be lower than worker exposures since
ONUs do not typically directly handle the chemical.
The PET byproduct and textile dyes conditions of use had the highest central tendency and high-end
worker inhalation exposure values, respectively. For PET byproduct, worker inhalation exposures were
estimated using OSHA monitoring data, which resulted in central tendency exposure of 4.7 mg/m3 and
high-end exposure of 47 mg/m3. For textile dyes, worker inhalation exposures were also estimated using
OSHA monitoring data, which resulted in central tendency exposure of 0.066 mg/m3 and high-end
exposure of 74 mg/m3.
The monitoring data sources and GSs and ESDs used to estimate occupational inhalation exposures all
had overall data quality determinations of either medium or high. The basis for determining overall data
Page 103 of 570
-------
quality is described in this section. In summary, each source is evaluated on multiple metrics based on
defined criteria. The individual metric ratings are used to obtain an overall study rating. All metrics have
an equal weight in determining the overall study rating. The resulting values are converted to an overall
data quality determination of "high," "medium," "low," "critically deficient," or "not rated/not
applicable." For more details on this process, see Section 5 of the 2021 Draft Systematic Review
Protocol
This section also includes information on the weight of scientific evidence conclusions for these
estimates, and a summary of the strengths, limitations, assumptions, and key sources of uncertainty for
these estimates.
3.1.2.2 Summary of Dermal Exposures Assessment
Table 3-1 presents the estimated dermal absorbed dose for workers in various OES. The dose estimates
assume one dermal exposure event (applied dose) per workday and that approximately 78 or 86 percent
of the applied dose is absorbed through the skin (depending on whether the OES is industrial or
commercial). The exposure estimates are provided for each OES, where the OES are "binned" based on
characteristics known to effect dermal exposure such as the maximum weight fraction of 1,4-dioxane
that could be present in that OES, open or closed system use of 1,4-dioxane, and large or small-scale
use. For a more detailed description of EPA's dermal assessment approach and each bin, see Appendix
F.3.
As shown in the Table 3-1, the calculated dermal absorbed dose for workers is lower in comparison to
those presented in the December 2020 Final Risk Evaluation for 1,4-Dioxane ( 320c). This is
due to the relatively lower concentrations of 1,4-dioxane found for the OES included in this supplement
than for those included in the 2020 RE. As noted previously, EPA did not assess dermal exposures to
ONUs as the Agency does not expect ONUs to have routine dermal exposures in the course of their
work. Depending on the OES, ONUs may have incidental dermal exposures due to surface
contamination. However, data (e.g., frequency and amount of liquid on the skin after contact) were not
identified to assess this exposure.
Table 3-1. Estimated Dermal Absorbed Dose (mg/day) for Workers in Various Conditions of Use
Weight
Fraction
(Max
Yilm,,)
No Gloves
(PF = 1)
Exposures Due to Glove Permeation/Chemical
Breakthrough (mg/day)
OES
Bin
Use Setting
Protective
Gloves
(PF = 5)
Protective
Gloves
(PF = 10)
Protective Gloves
(Industrial Uses
Only, PF = 20)
Textile dye
7
Industrial and
Commercial
4.7E-06
0.003 (CT)
0.009 (HE)
0.001 (CT)
0.002 (HE)
3.0E-4 (CT)
0.001 (HE)
1.5E-4 (CT)
4.5E-4 (HE)
Antifreeze
8
Commercial
8.6E-05
0.055 (CT)
0.165 (HE)
0.011 (CT)
0.033 (HE)
0.006 (CT)
0.017 (HE)
N/A
Surface
cleaner
9
Commercial
7.6E-05
0.049 (CT)
0.146 (HE)
0.010 (CT)
0.029 (HE)
0.005 (CT)
0.015 (HE)
N/A
Dish soap
10
Commercial
2.04E-04
0.131 (CT)
0.393 (HE)
0.026 (CT)
0.079 (HE)
0.013 (CT)
0.039 (HE)
N/A
Dishwasher
detergent
11
Commercial
5.8E-05
0.037 (CT)
0.111 (HE)
0.007 (CT)
0.022 (HE)
0.004 (CT)
0.011 (HE)
N/A
Page 104 of 570
-------
OES
Bin
Use Setting
Weight
Fraction
(Max
Y derm )
No Gloves
(PF = 1)
Exposures Due to Glove Permeation/Chemical
Breakthrough (mg/dav)
Protective
Gloves
(PF = 5)
Protective
Gloves
(PF = 10)
Protective Gloves
(Industrial Uses
Only, PF = 20)
Laundry
detergent
(industrial
and
institutional)
12
Industrial and
Commercial
1.3E-04
0.083 (CT)
0.248 (HE)
0.017 (CT)
0.050 (HE)
0.008 (CT)
0.025 (HE)
0.097 (CT)
0.290 (HE)
Paint and
floor lacquer
13
Industrial and
Commercial
3.0E-05
0.019 (CT)
0.058 (HE)
0.004 (CT)
0.012 (HE)
0.002 (CT)
0.006 (HE)
0.001 (CT)
0.003 (HE)
Polyethylene
terephthalate
(PET)
byproduct
14
Industrial
0.03
17.6 (CT)
52.8 (HE)
3.52 (CT)
10.6 (HE)
1.76 (CT)
5.28 (HE)
0.88 (CT)
2.64 (HE)
Ethoxylation
process
byproduct
15
Industrial
1.4E-03
0.827 (CT)
2.48 (HE)
0.165 (CT)
0.496 (HE)
0.083 (CT)
0.248 (HE)
0.041 (CT)
0.124 (HE)
Hydraulic
fracturing
16
Industrial and
Commercial
0.05
32.1 (CT)
96.2 (HE)
6.41 (CT)
19.2 (HE)
3.21 (CT)
9.62 (HE)
1.60 (CT)
4.81 (HE)
CT = central tendency; HE = high-end; PF = protection factor
3.1.2.3 Weight of Scientific Evidence Conclusions for Occupational Exposure
Information
Table 3-2 provides a summary of EPA's overall weight of scientific evidence conclusions for its
occupational exposure estimates for each of the assessed OES. These determinations are OES-specific.
For a description of overall confidence in all inhalation exposures, see Section 3.3.1.1. For an
explanation of EPA's judgement on the weight of scientific evidence conclusion, see Section 2.2.1.2.
Factors that increase and decrease the strength of the weight of scientific evidence are listed in
Table Apx C-5.
Due to a lack of data, EPA was not able to estimate ONU inhalation exposure from monitoring data or
models, so a qualitative assessment of potential ONU exposures was made. Similarly, EPA did not
assess dermal exposures to ONUs as EPA does not expect ONUs to have routine dermal exposures in
the course of their work. Depending on the COU, ONUs may have incidental dermal exposures due to
surface contamination. However, data (e.g., frequency and amount of liquid on the skin after contact)
were not identified to assess this exposure. Finally, due to the absence of dermal monitoring data, these
columns were omitted from Table 3-2.
Page 105 of 570
-------
Table 3-2. Summary of the Weight of Scientific Evidence for Occupational Exposure Estimates by PES
OES
Inhalation Exposure
Dermal Exposure
Monitoring
Monte Carlo Modeling
Weight of
Scientific Evidence
Modeling
Weight of Scientific
Evidence
Worker
# Data
Points
Data Quality
Rating"
Worker
Data
Quality
Rating"
Worker
Worker6
Worker
Textile dye
~
14
H
X
N/A
Moderate
~
Moderate
Antifreeze
X
N/A
N/A
~
H
Moderate
~
Moderate
Surface cleaner
~
49
H
X
N/A
Moderate to Robust
~
Moderate
Dish soap
X
N/A
N/A
~
N/A
Moderate
~
Moderate
Dishwasher detergent
X
N/A
N/A
~
N/A
Moderate
~
Moderate
Laundry detergent
(industrial and
institutional)
X
N/A
N/A
~
M
Moderate
~
Moderate
Paint and floor lacquer
~
17
H
X
N/A
Moderate
~
Moderate
PET byproduct
~
62
H
X
N/A
Moderate to Robust
~
Moderate
Ethoxylation process
byproduct
~
9
H
X
N/A
Moderate
~
Moderate
Hydraulic fracturing
X
N/A
N/A
~
M
Moderate to Robust
~
Moderate
"Data quality ratings of modeling approaches are based on the GS/ESD that was used in tandem with Monte Carlo modeling.
h Data quality ratings are not applicable for the dermal modeling approach because this modeling was conducted with an already-developed EPA model.
Page 106 of 570
-------
3.1.2.4 Strengths, Limitations, Assumptions, and Key Sources of Uncertainty for the
Occupational Exposure Assessment
3.1.2.4.1 Number of Workers
There are uncertainties surrounding the estimated number of workers potentially exposed to 1,4-dioxane.
First, BLS employment data for each industry/occupation combination are only available at the 3-, 4-, or
5-digit NAICS level, rather than at the full 6-digit NAICS level. This lack of specificity could result in
an overestimate of the number of exposed workers if some 6-digit NAICS are included in the less
granular BLS estimates but are not likely to use 1,4-dioxane for the assessed applications. EPA
addressed this issue by refining the OES estimates using total employment data from the U.S. Census'
SUSB. However, this approach assumes that the distribution of occupation types (SOC codes) in each 6-
digit NAICS is equal to the distribution of occupation types at the parent 5-digit NAICS level. If the
distribution of workers in occupations with 1,4-dioxane exposure differs from the overall distribution of
workers in each NAICS, then this approach will result in inaccuracy. The effects of this uncertainty on
the number of worker estimates are unknown, as the uncertainties may result in either over or
underestimation of the estimates depending on the actual distribution.
Second, EPA's determinations of industries (represented by NAICS codes) and occupations (represented
by SOC codes) that are associated with the OES assessed in this report are based on EPA's
understanding of how 1,4-dioxane is used in each industry. The designations of which industries and
occupations have potential exposures is a matter of professional judgement; therefore, the possibility
exists for the erroneous inclusion or exclusion of some industries or occupations. This may result in
inaccuracy but would be unlikely to systematically either overestimate or underestimate the count of
exposed workers.
3.1.2.4.2 Analysis of Inhalation Exposure Monitoring Data
The principal limitation of the monitoring data is the uncertainty in the representativeness of the data
due to some scenarios having limited exposure monitoring data in literature. Therefore the assessed
exposure levels may not be representative of worker exposures across all worker activities or the
industry as a whole. For example, monitoring data may not sufficiently capture activities that occur with
different frequency or duration than common production tasks. Additionally, monitoring data may only
be available for a limited number of sites. Differences in work practices and engineering controls across
sites can introduce variability and limit the representativeness of monitoring data. Age of the monitoring
data can also introduce uncertainty due to differences in workplace practices and equipment used at the
time the monitoring data were collected compared to those currently in use. Therefore, older data may
overestimate or underestimate exposures, depending on these differences. The effects of these
uncertainties on the occupational exposure assessment are unknown, as the uncertainties may result in
either overestimation or underestimation of exposures depending on the actual distribution of 1,4-
dioxane air concentrations and the variability of work practices among different sites.
In some scenarios where monitoring data were available, EPA did not find sufficient data to determine
complete statistical distributions. Ideally, EPA will present 50th and 95th percentiles for each exposed
population. In the absence of percentile data for monitoring, the mean or midpoint of the range may
serve as a substitute for the 50th percentile of the actual distributions. Similarly, the highest value of a
range may serve as a substitute for the 95th percentile of the actual distribution. However, these
substitutes are uncertain. The effects of these substitutes on the occupational exposure assessment are
unknown, as the substitutes may result in either overestimation or underestimation of exposures
depending on the actual distribution.
Page 107 of 570
-------
3.1.2.4.3 Modeled Inhalation Exposures
EPA addressed variability in inhalation models by identifying key model parameters to apply a
statistical distribution that mathematically defines the parameter's variability. EPA defined statistical
distributions for parameters using documented statistical variations where available. Where the
statistical variation was unknown, assumptions were made to estimate the parameter distribution using
available literature data, such as GSs and ESDs. However, there is uncertainty as to the
representativeness of the parameter distributions with respect to the modeled scenario because the data
are often not specific to sites that use 1,4-dioxane. In general, the effects of these uncertainties on the
exposure estimates are unknown, as the uncertainties may result in either overestimation or
underestimation on exposures depending on the actual distributions of each of the model input
parameters.
There is also uncertainty as to whether the model equations generate results that represent actual
workplace air concentrations. Some activity-based modeling does not account for exposures from other
activities. Another uncertainty is lack of consideration for engineering controls. The GS/ESDs assume
that all activities occur without any engineering controls or PPE, and in an open-system environment
where vapor and particulates freely escape and can be inhaled. Actual exposures may be less than
estimated depending on engineering control and PPE use.
A strength of the assessment is the variation of the model input parameters as opposed to using a single
static value. This parameter variation increases the likelihood of true occupational inhalation exposures
falling within the range of modeled estimates. An additional strength is that all data that EPA used to
inform the modeling parameter distributions have overall data quality determinations of either high or
medium from EPA's systematic review process.
3.1.2.4.4 Modeled Dermal Exposures
The Dermal Exposure to Volatile Liquids Model used to estimate dermal exposure to 1,4-dioxane in
occupational settings assumes a fixed fractional absorption of the applied dose; however, fractional
absorption may be dependent on skin loading conditions. The model also assumes a single exposure
event per day based on existing framework of the EPA/OPPT 2-Hand Dermal Exposure to Liquids
Model. The model does not address variability in exposure duration and frequency or uncertainty with
respect to the worker exposure activities and resulting exposed skin surface area which could result in
misestimation. Additionally, dermal exposures to 1,4-dioxane vapor that may penetrate clothing and the
potential for associated direct skin contact with clothing saturated with 1,4-dioxane vapor are not
included in quantifying exposures, which could potentially result in underestimates of exposures.
Although the extent of saturation of clothing with 1,4-dioxane vapors is unknown, it is expected to be
minimal given the low concentrations of 1,4-dioxane in formulations for the conditions of use in the
supplement.
A strength of the dermal assessment approach is the estimation of two different fractional absorption
values specific to industrial and commercial use settings as opposed to applying only one fractional
absorption value to both settings.
3.2 General Population Exposures
General population exposures occur when 1,4-dioxane is released into the environment and the media is
then a pathway for exposure. Figure 3-1 below provides a graphic representation of where and in which
media 1,4-dioxane may be found and the corresponding route of exposure.
Page 108 of 570
-------
1,4-Dioxane - General Population Exposures (Section 3.2):
Key Points
• EPA estimated oral and inhalation exposures to the concentrations of 1,4-dioxane in drinking
water and air estimated in Section 2.3 using equations and exposure factors described in
Appendix G.l and Appendix J.2.
• EPA estimated general population exposures to 1,4-dioxane in air and water with a particular
focus on populations that may be highly exposed
0 Fenceline communities. For exposures through air, EPA considered potential
exposures for communities within 10 km of a release site. For drinking water, EPA
considered potential exposures for communities relying on drinking water collected
downstream of release sites.
0 Lifestage. For drinking water, EPA evaluated lifestage-specific exposures for adults,
formula-fed infants, and children. For air exposures, the impacts of lifestage
differences were not able to be adequately quantified and so the air concentrations are
used for all lifestages.
0 High-end exposure estimates. EPA evaluated exposures based on high-end exposure
scenarios (e.g., air exposures include a range of modeled concentration predictions
[low-end, central tendency, and high-end]), although only high-end model predictions
of air concentrations are presented in this section).
Ambient Air
Inhalation
Landfills
(Industrial or
Muncipal)
BSfiBBBBBk. jSSjSSm
Wastewater
Drinking
Drinking
Water
Water
Recreation
Oral, Derma!
Figure 3-1. Potential Human Exposure Pathways to 1,4-Dioxane for the General Population"
" The diagram presents the media (white text boxes) and routes of exposure (italics for oral, inhalation, or dermal)
for the general population. Sources of drinking water from surface or water pipes is depicted with grey arrows.
Page 109 of 570
-------
3,2,1 Approach and Methodology
EPA used modeled concentrations of 1,4-dioxane in air, surface water, and groundwater estimated in
Section 2.3 to estimate acute and chronic general population exposures that could result from contact
with environmental media. These acute and chronic exposure estimates are used to evaluate cancer and
non-cancer risk described in Section 5. To estimate oral exposures to 1,4-dioxane in drinking water
(including groundwater used as drinking water), EPA used equations and exposure factors described in
Appendix G.l. To estimate inhalation exposures from 1,4-dioxane in air, EPA used equations and
exposure factors described in Appendix J.2. Longer exposure durations would result in greater
inhalation exposure. Individuals exposed through air over a full lifetime (78 years) could have exposures
approximately 2.36 times greater than those calculated for 33 years of exposure. Where possible,
available monitored data within these environmental media were used to provide context for modeled
results.
To estimate potential acute and chronic exposures through drinking water EPA calculated Acute Dose
Rates (ADR) and Average Daily Doses (ADD) for adults, formula-fed infants, and children. To estimate
lifetime exposures through drinking water, EPA calculated a Lifetime Average Daily Dose (LADD)
based on mean drinking water ingestion rates over 33 years11 of exposure starting from birth or 33 years
of exposure as an adult, averaged over a 78-year lifetime. Longer exposure durations or higher drinking
water ingestion rates would result in greater exposure. Individuals exposed through drinking water over
a full lifetime (78 years) could have exposure approximately 2.26 times greater than those calculated for
33 years of exposure. Lifetime cancer risk estimates based on 95th percentile drinking water ingestion
rates could result in 3-4 times higher exposures and risks than those based on mean ingestion rates,
depending on the age groups exposed (described in Appendix 5.2.5.41.1). Assumptions about drinking
water intake and body weight for each age group were based on information in the Exposure Factors
Handbook. EPA calculated ADs, ADDs, and LADDs based on the drinking water concentrations
estimated under a range of conditions in Section 2.3.1.3. Details of these calculations are presented in
Appendix I and 1,4-Dioxane Supplemental Information File: Drinking Water Exposure and Risk
Estimates for 1,4-Dioxane Release to Surface Water from Individual Facilities (U.S. EPA. 2024h).
To estimate potential acute and chronic exposures through air, EPA calculated ADs and ADCs based on
modeled air concentrations described in Section 2.3.3 To estimate potential lifetime exposures, EPA
calculated LADCs based on 33 years of exposure. Methods adequate to quantify the impact of lifestage
differences on 1,4-dioxane exposure are not available (see Section 4.3) and air concentration is used as
the exposure metric for all lifestages per EPA guidance (\ v << \ i lb). Specific equations,
inputs and assumptions are described in detail in Appendix IJ.2.
3,2.2 Drinking Water Exposure Assessment
EPA assessed general population drinking water exposures that could result from surface water or
groundwater used as drinking water. Exposures estimates presented below are based on surface water
concentrations modeled in Section 2.3.1 or groundwater concentrations modeled in Section 2.3.2.
Exposure estimates presented throughout this section focus on adults and formula-fed infants because
these are lifestages with the greatest drinking water intake relative to body weights and therefore the
greatest potential exposures.
11 Thirty-three years is the 95th percentile residential occupancy period (U.S. EPA Exposure Factors Handbook, Chapter 16,
Table 16-5).
Page 110 of 570
-------
3.2.2.1 Surface Water Exposure Assessment
EPA evaluated acute, chronic, and lifetime general population, exposures to 1,4-dioxane that may be
present in drinking water based on modeled surface water concentrations estimated in Section 2.3.1. For
drinking water exposures in this assessment, EPA focused on exposures in fenceline communities,
defined in this context as members of the general population who rely on drinking water from water
bodies receiving 1,4-dioxane releases from any industrial or DTD source.
Drinking water exposures were evaluated using a series of parallel analyses that provide information
about the individual contributions of specific COUs as well as information about aggregate exposures
that could result from multiple sources releasing to the same water body.
3.2.2.1.1 Exposures from Individual Facility Releases
To evaluate the individual contributions of releases associated with specific industrial and commercial
COUs to general population exposures, EPA calculated ADRs, ADDs, and LADDs based on modeled
water concentrations estimated in Section 2.3.1.3.1 (Table 3-3). A total of 125 release scenarios were
evaluated based on water concentrations estimated for annual releases that occur over a single day (a
peak exposure scenario), over 30 days of release, or over 250 to 365 days of release. Exposure estimates
are presented for both adults and formula-fed infants because these are lifestages with greatest drinking
water intake relative to body weights and therefore greatest exposures. ADRs based on a single day
release scenario range from 6.Ox 10~8 to 3,730 mg/kg for adults and 2.1 /10 1 to 1.3x 104 mg/kg for
infants. ADDs range from 1.1/10 " to 0.5 mg/kg/day for adults and 2,7/10 " to 1.3 mg/kg/day for
infants and are not influenced by the days of release. LADDS range from 4.2x 10~12 to 0.2 mg/kg/day for
adults exposed for 33 years and 3,5x10 13 to 1,6x 10 2 mg/kg/day for infants exposed for 1 year.
Complete exposure calculations are available in 1,4-Dioxane Supplemental Information File: Drinking
Water Exposure and Risk Estimates for 1,4-Dioxane Release to Surface Water from Individual Facilities
(U.S. EPA. 2024h\
Page 111 of 570
-------
Table 3-3. Adult and Infant Exposures Estimated from Facility-Specific Releases
OES
No. of
Releases
Modeled
Age Group'1
ADR (mg/kg-dav)
ADD (mg/kg-dav)
LADD (mg/kg-day)
Min
Exposure'
Mean
Exposure''
Max
Exposure1'
Min
Exposure'
Mean
Exposure''
Max
Exposure1'
Min
Exposure'
Mean
Exposure''
Max
Exposure1'
Disposal
25
Adult (21+
years)
6.03E-07
2.60E01
3.83E02
4.51E-10
1.44E-02
2.21E-01
1.91E—10
6.08E-03
9.36E-02
Infant (birth to
<1 year)
2.11E-06
9.11E01
1.34E03
1.15E-09
3.67E-02
5.65E-01
1.48E-11
4.71E-04
7.25E-03
Ethoxylation
byproduct
8
Adult (21+
years)
2.17E-07
1.04E02
8.31E02
9.06E-11
3.66E-02
2.93E-01
3.83E-11
1.55E-02
1.24E-01
Infant (birth to
<1 year)
7.61E-07
3.65E02
2.92E03
2.31E-10
9.36E-02
7.48E-01
2.97E-12
1.20E-03
9.59E-03
Functional
fluids (open-
system)
6
Adult (21+
years)
5.59E-04
6.33E-02
1.92E-01
1.83E-07
2.23E-05
6.66E-05
7.73E-08
9.44E-06
2.82E-05
Infant (birth to
<1 year)
1.96E-03
2.22E-01
6.75E-01
4.67E-07
5.70E-05
1.70E-04
5.99E-09
7.30E-07
2.18E-06
Import and
repackaging
12
Adult (21+
years)
4.35E-04
3.28E02
3.73E03
1.32E-07
3.05E-02
2.23E-01
5.59E-08
1.29E-02
9.43E-02
Infant (birth to
<1 year)
1.53E-03
1.15E03
1.31E04
3.38E-07
7.78E-02
5.69E-01
4.33E-09
9.97E-04
7.30E-03
Industrial uses
31
Adult (21+
years)
5.34E-07
2.05E01
1.87E02
1.97E-10
1.36E-02
1.55E-01
8.3 IE—11
5.77E-03
6.56E-02
Infant (birth to
<1 year)
1.87E-06
7.21E01
6.55E02
5.02E-10
3.48E-02
3.96E-01
6.44E-12
4.47E-04
5.08E-03
Manufacture
2
Adult (21+
years)
3.35E00
6.56E01
1.28E02
2.50E-03
4.91E-02
9.57E-02
1.06E-03
2.08E-02
4.05E-02
Infant (birth to
<1 year)
1.17E01
2.30E02
4.48E02
6.40E-03
1.25E-01
2.44E-01
8.20E-05
1.61E-03
3.13E-03
PET
manufacturing
19
Adult (21+
years)
1.11 E—04
4.32E01
6.67E02
3.86E-08
3.16E-02
5.00E-01
1.63E-08
1.34E-02
2.11 E—01
Infant (birth to
<1 year)
3.91E-04
1.52E02
2.34E03
9.86E-08
8.07E-02
1.28E00
1.26E-09
1.04E-03
1.64E-02
Printing inks
1
Adult (21+
years)
8.26E-02
8.26E-02
8.26E-02
6.18E-05
6.18E-05
6.18E-05
2.62E-05
2.62E-05
2.62E-05
Infant (birth to
<1 year)
2.90E-01
2.90E-01
2.90E-01
1.58E-04
1.58E-04
1.58E-04
2.02E-06
2.02E-06
2.02E-06
Remediation
16
Adult (21+
years)
6.04E-08
7.36E-02
7.19E-01
1.07E-11
4.56E-05
4.12E-04
4.51E-12
1.93E-05
1.74E-04
Infant (birth to
<1 year)
2.12E-07
2.58E-01
2.52E00
2.72E-11
1.17E-04
1.05E-03
3.49E-13
1.49E-06
1.35E-05
Page 112 of 570
-------
OES
No. of
Releases
Modeled
Age Group*
ADR (mg/kg-dav)
ADD (mg/kg-dav)
LADD (mg/kg-day)
Min
Exposure1'
Mean
Exposure''
Max
Exposure1'
Min
Exposure1'
Mean
Exposure''
Max
Exposure1'
Min
Exposure1'
Mean
Exposure''
Max
Exposure"
Overall
120
Adult (21+
years)
6.04E-08
5.84E01
3.73E03
1.07E-11
1.78E-02
5.00E-01
4.51E-12
7.55E-03
2.11E-01
Infant (birth to
<1 year)
2.12E-07
2.05E02
1.31E04
2.72E-11
4.56E-02
1.28E00
3.49E-13
5.84E-04
1.64E-02
Adult LADDs presented in this table were used to derive cancer risk estimates presented in Figure 5-1, Figure 5-2 and Figure 5-4.
ADRs presented here are calculated based on the assumption that all releases could occur on a single day of release (peak exposure scenario); ADDs and LADDs are based on
chronic exposure scenarios and are the same regardless of the number of days of release assumed. LADDs for adults are based on 33 years of exposure averaged over a 78-year
lifetime while LADDs for infant-specific exposures are based on 1 year of exposure averaged over a 78-year lifetime. LADDs for a full 78 years of exposure would be 2.26
times greater than those presented here. Similarly, LADDs based on 95th percentile drinking water ingestion rates would be approximately 3 -4 times greater, depending on the
age groups exposed.
"Adult refers to 21+years; infant refers to birth to <1 year.
4 These COUs are added since the 2020 RE was published.
c The minimum exposure for the identified days of release, within the identified OES, and for the identified age group.
dThe arithmetic mean exposure for the identified days of release, within the identified OES, and for the identified age group.
e The maximum exposure for the identified days of release, within the identified OES, and for the identified age group.
Page 113 of 570
-------
3.2.2.1.2 Exposures from Down-the-Drain Releases
To evaluate the potential contribution of DTD consumer and commercial releases, EPA calculated
ADRs, ADDs, and LADDs using modeled water concentrations estimated as described in Section
2.3.1.2.2. Water concentrations of 1,4-dioxane resulting from DTD releases depend on the population
size (an indicator of the number of people using products and contributing the releases) and the stream
flows of the receiving water bodies. Therefore, the adult LADDs presented in Table 3-4 are based on the
range of water concentrations estimated by Monte Carlo modeling of DTD release scenarios with
varying population size and stream flows. LADDs range from 1.7/10 10 to 5,1 / ] 0 4 mg/kg/day for
adults exposed for 33 years. Complete exposure calculations for adults and infants are available in 1,4-
Dioxane Supplemental Information File: Drinking Water Exposure and Risk Estimates for 1,4-Dioxane
Surface Water Concentrations Predicted with Probabilistic Modeling (U.S. EPA. 20240.
Table 3-4. Adult LADD Exposures (mg/kg/day) Estimated from 1,4-Dioxane DTD Consumer and
Commercial Releases
Population Contributing to DTD Releases
100
1,000
10,000
100,000
1,000,000
Stream Flow
(cfs)
100
5.1E-08
5.1E-07
5.1E-06
5.1E-05
5.1E-04
300
1.7E-08
1.7E-07
1.7E-06
1.7E-05
1.7E-04
1,000
5.1E-09
5.1E-08
5.1E-07
5.1E-06
5.1E-05
3,000
1.7E-09
1.7E-08
1.7E-07
1.7E-06
1.7E-05
10,000
5.1E-10
5.1E-09
5.1E-08
5.1E-07
5.1E-06
30,000
1.7E-10
1.7E-09
1.7E-08
1.7E-07
1.7E-06
The frequencies of each of these combinations of population size and flow rate are presented Table 2-11. Adult LADDs
presented in this table were used to derive the cancer risk estimates presented in Table 5-4.
LADDs for adults are based on 33 years of exposure averaged over a 78-year lifetime; LADDs for a full 78 years of
exposure would be 2.26 times greater than those presented here. Similarly, LADDs based on 95th percentile drinking
water ingestion rates would be approximately 3-4 times greater, depending on the age groups exposed.
3.2.2.1.3 Disposal of Hydraulic Fracturing Produced Waters
To evaluate the potential contribution of disposal of hydraulic fracturing produced waters to surface
water, EPA calculated ADRs, ADDs, and LADDs using the range of modeled water concentrations
estimated in Section 2.3.1.2.2. (Table 3-5). Water concentrations of 1,4-dioxane resulting from disposal
of hydraulic fracturing produced water vary substantially across sites. The estimated exposures
presented here are based on the range of water concentrations estimated by Monte Carlo modeling for a
range of site-specific factors. For this range of estimated surface water concentrations, Adult ADRs
range from 1.12xl0~14 to 6.32xl0~3 mg/kg and adult ADDs range from 3.07xl0~15 to 1.73><10~3
mg/kg/day. LADDs for adults exposed over 33 years over a 78-year lifetime range from 1.3/1 0 15 to
7.3xl0~4 mg/kg/day. Complete exposure calculations for adults and infants are available in 1,4-Dioxane
Supplemental Information File: Drinking Water Exposure and Risk Estimates for 1,4-Dioxane Surface
Water Concentrations Predicted with Probabilistic Modeling (U.S. EPA. 2024i).
Page 114 of 570
-------
Table 3-5. Adult ADR, ADD, and LADD Exposures Estimated from Disposal of Hydraulic
Fracturing Produced Waters to Surface Water
Monte Carlo
Distribution
Adult Acute Dose
Rate (mg/kg)
Adult Average Daily
Dose (mg/kg/day)
Adult Lifetime Average
Daily Dose (mg/kg/day)
Maximum
6.32E-03
1.73E-03
7.3E-04
99th percentile
3.04E-04
8.30E-05
3.5E-05
95th percentile
1.10E-04
3.00E-05
1.3E-05
Median
2.78E-06
7.59E-07
3.2E-07
5th percentile
1.36E-08
3.72E-09
1.6E-09
Minimum
1.12E-14
3.07E-15
1.3E-15
Adult LADDs presented in this table were used to derive cancer risk estimates presented in Table 5-5. LADDs for adults
are based on 33 years of exposure averaged over a 78-year lifetime; LADDs for a full 78 years of exposure would be 2.26
times greater than those presented here. Similarly, LADDs based on 95th percentile drinking water ingestion rates would
be approximately 3-4 times greater, depending on the age groups exposed.
3.2.2.1.4 Aggregate Exposure
Because multiple sources of 1,4-dioxane contribute to surface water and drinking water concentrations,
EPA also estimated aggregate general population exposures that could occur because of combined
contributions from DTD releases from consumer and commercial uses, upstream sources, and direct and
indirect industrial releases. EPA calculated ADRs, ADDs, and LADDs based on modeled water
concentrations estimated in Section 2.3.1.3.4 using probabilistic modeling of aggregate 1,4-dioxane
surface water concentrations that could occur downstream of industrial release sites for each COU.
LADDs estimated for adults exposed over 33 years over a 78-year lifetime range from 8.07x 1CT7 to
7.4x10 3 mg/kg/day based on median modeled water concentrations across COUs (Table 3-6). Complete
exposure calculations for adults and infants are available in 1,4-Dioxane Supplemental Information File:
Drinking Water Exposure and Risk Estimates for 1,4-Dioxane Surface Water Concentrations Predicted
with Probabilistic Modeling ( 20241).
Table 3-6. Adult LADD Exposures from Aggregate Concentrations Estimated Downstream of
Release Sites (Including DTD Releases and Direct and Indirect Industrial Releases)
COU
LADDs (mg/kg/day) Based on Modeled Aggregate Surface Water Concentrations
Estimated across the Monte Carlo Distribution
Min
5th
Percentile
25th
Percentile
Median
75th
Percentile
95th
Percentile
Max
Disposal
6.00E-09
6.82E-07
1.63E-06
3.93E-06
9.00E-06
9.64E-04
1.21E-01
Ethoxylation
byproduct
7.17E-09
3.72E-07
1.05E-06
1.93E-06
8.11E-06
1.98E-02
2.63E-01
Functional
fluids (open-
system)
5.62E-10
2.58E-07
4.99E-07
8.91E-07
7.20E-06
4.13E-05
6.22E-05
Import and
repackaging
1.82E-08
4.21E-07
2.10E-04
1.03E-03
4.53E-03
1.34E-02
1.18E00
Page 115 of 570
-------
cou
LADDs (mg/kg/day) Based on Modeled Aggregate Surface Water Concentrations
Estimated across the Monte Carlo Distribution
Min
5th
Percentile
25th
Percentile
Median
75th
Percentile
95th
Percentile
Max
Industrial Uses
4.61E-10
1.65E-07
3.90E-07
8.07E-07
4.66E-05
4.90E-03
5.91E-02
Manufacture
3.51E-04
1.06E-03
2.29E-03
7.40E-03
1.75E-02
4.04E-02
4.04E-02
PET
manufacturing
1.73E-08
4.54E-07
1.48E-06
1.43E-05
6.06E-04
2.53E-02
2.11E-01
Printing inks
4.80E-07
1.01E-06
1.27E-05
2.04E-05
2.38E-05
2.66E-05
2.71E-05
Remediation
2.48E-09
2.74E-07
6.29E-07
1.27E-06
3.10E-05
9.61E-05
2.29E-04
Adult LADDs presented in this table were used to derive cancer risk estimates presented in Figure 5-5. Percentiles reflect
concentrations estimated at various points in the Monte Carlo distribution. LADDs for adults are based on 33 years of
exposure averaged over a 78-year lifetime; LADDs for a full 78 years of exposure would be 2.26 times greater than those
presented here. Similarly, LADDs based on 95th percentile drinking water ingestion rates would be approximately 3-4
times greater, depending on the age groups exposed.
3.2.2.2 Groundwater Exposure Assessment
EPA evaluated general population exposures that could occur from disposals of 1,4-dioxane that
contaminate groundwater used as a primary source of drinking water. To estimate chronic exposures
through this drinking water pathway, EPA calculated ADDs and LADDs for adults and formula-fed
infants based on modeled groundwater concentrations of 1,4-dioxane estimated in Section 2.3.2. The
Agency did not evaluate acute exposures because methods used to estimate groundwater concentrations
provide an indication of potential concentrations occuring over many years, rather than peak
concentrations.
3.2.2.2.1 Disposal to Landfills
To evaluate general population exposure, EPA calculated ADDs and LADDs based on modeled
groundwater concentrations estimated in Section 2.3.2.3. Potential groundwater concentrations resulting
from disposal of 1,4-dioxane to municipal solid waste landfills vary across landfill loading rates and
concentrations of 1,4-dioxane in leachate. Estimated exposures presented here are therefore based on the
range of groundwater concentrations estimated under varying landfill conditions. Table 3-7 summarizes
LADD exposure estimates estimated for 33 years of exposure as an adult. Under the range of landfill
scenarios considered, adult LADDs range from 2.5><106 to 2,4/ 10 2 mg/kg/day. The highest LADDs
occur when leachate concentrations are above 100 mg/L and loading rates are above 10,000 lb. The
complete set of exposure estimates for adults and infants relying on groundwater as a primary drinking
water source are presented in 1,4-Dioxane Supplemental Information File: Drinking Water Exposure
and Risk Estimates for 1,4-Dioxane Land Releases to Landfills ( 324f).
Page 116 of 570
-------
Table 3-7. Adult LADD Exposures Estimated from Groundwater Contamination from Landfills
under Varying
^andfill Conditions
Loading Rate (lb)
Leachate
Concentration
(mg/L)
0.1
1
10
100
1,000
10,000
100,000
1,000,000
0.0001
2.5E-16
2.4E-15
3.0E-14
2.9E-13
2.7E-12
2.6E-11
2.5E-10
2.4E-09
0.001
2.5E-15
2.4E-14
3.0E-13
2.9E-12
2.7E-11
2.6E-10
2.5E-09
2.4E-08
0.01
2.5E-14
2.4E-13
3.0E-12
2.9E-11
2.7E-10
2.6E-09
2.5E-08
2.4E-07
0.1
2.5E-13
2.4E-12
3.0E-11
2.9E-10
2.7E-09
2.6E-08
2.5E-07
2.4E-06
1
2.5E-12
2.4E-11
3.0E-10
2.9E-09
2.7E-08
2.6E-07
2.5E-06
2.4E-05
10
2.5E-11
2.4E-10
3.0E-09
2.9E-08
2.7E-07
2.6E-06
2.5E-05
2.4E-04
100
2.5E-10
2.4E-09
3.0E-08
2.9E-07
2.7E-06
2.6E-05
2.5E-04
2.4E-03
1,000
2.5E-09
2.4E-08
3.0E-07
2.9E-06
2.7E-05
2.6E-04
2.5E-03
2.4E-02
10,000
2.5E-08
2.4E-07
3.0E-06
2.9E-05
2.7E-04
2.6E-03
2.5E-02
2.4E-01
Adult LADDs presented in this table were used to derive cancer risk estimates presented in Table 5-6. LADDs
for adults are based on 33 years of exposure averaged over a 78-year lifetime; LADDs for a full 78 years of
exposure would be 2.26 times greater than those presented here. Similarly, LADDs based on 95th percentile
drinking water ingestion rates would be approximately 3^ times greater, depending on the age groups exposed.
3.2.2.2.2 Disposal of Hydraulic Fracturing Produced Waters
To evaluate general population exposure resulting from disposal of hydraulic fracturing produced waters
to groundwater, EPA calculated ADDs and LADDs estimated in Section 2.3.2.4 (Table 3-8). Potential
groundwater concentrations resulting from disposal of hydraulic fracturing produced waters vary
substantially across sites. Estimated exposures presented here are based on the range of groundwater
concentrations estimated through Monte Carlo modeling. Under the range of hydraulic fracturing
scenarios considered, adult LADDs range from 4,9/ 10 9 to 2,1 /10 4 mg/kg/day. The complete set of
exposure estimates for adults and infants relying on groundwater as a primary drinking water source are
presented in 1,4-Dioxane Supplemental Information File: Drinking Water Exposure and Risk Estimates
for 1,4-Dioxane Land Releases to Surface Impoundments (U.S. EPA. 2024e).
Table 3-8. Estimated Exposures Resulting from Groundwater Contamination from Disposal of
Monte Carlo
Distribution
Modeled
Groundwater
Concentration (mg/L)
Adult ADD
(mg/kg/day)
Adult LADD
(mg/kg/day)
Infant ADD
(mg/kg/day)
Max
1.9E-02
2.1E-04
8.8E-05
5.3E-04
99th
1.5E-02
1.7E-04
7.1E-05
4.3E-04
95th
1.5E-02
1.7E-04
7.1E-05
4.3E-04
Mean
7.1E-04
7.9E-06
3.3E-06
2.0E-05
50th
1.2E-04
1.3E-06
5.6E-07
3.4E-06
5th
1.2E-04
1.3E-06
5.6E-07
3.4E-06
Min
4.4E-07
4.9E-09
2.1E-09
1.2E-08
Page 117 of 570
-------
Monte Carlo
Distribution
Modeled
Groundwater
Concentration (mg/L)
Adult ADD
(mg/kg/day)
Adult LADD
(mg/kg/day)
Infant ADD
(mg/kg/day)
Adult LADDs presentee
adults are based on 33 y
would be 2.26 times gre
ingestion rates would be
in this table were used to derive cancer risk estimates presented in Table 5-7. LADDs for
ears of exposure averaged over a 78-year lifetime; LADDs for a full 78 years of exposure
ater than those presented here. Similarly, LADDs based on 95th percentile drinking water
approximately 3-4 times greater, depending on the age groups exposed.
3,2.3 Air Exposure Assessment
EPA evaluated acute, chronic and lifetime general population, exposures to 1,4-dioxane in air. This
analysis focuses on potential fenceline community exposures that may occur within 10 km of release
sites.
3.2.3.1 Industrial COUs Reported to TRI
To evaluate general population exposures from industrial fugitive and stack emissions, EPA calculated
ACs, ADCs, and LADCs based on modeled air concentrations estimated in Section 2.3.3. The LADCs
presented in Table 3-9 are based maximum 95th percentile air concentrations estimated for the facilities
within each COU. LADCs within 10 km of release types considered here range from l.lxlO-11 to
6.9x10 3 ppm. These lifetime exposure estimates are based on 33 years of exposure over a 78-year
lifetime and are relevant to all lifestages. The complete set of inhalation exposure estimates for fenceline
communities are presented in 1,4-Dioxane Supplemental Information File: Air Exposures and Risk
Estimates for Single Year Analysis ( Q24e). EPA also considered how longer exposure
durations influence exposure. Individuals exposed over a full 78-year lifetime would have an exposure
2.36 times greater than those calculated for 33 years of exposure.
Page 118 of 570
-------
Table 3-9. Lifetime Average Dai y Concentrations Estimated within 10 km of 1,4-Dioxane Releases to Air
1,4-Dioxane OES
#
Facilities
LADCs Estimated within 5-10,000 m of Facilities with Greatest Exposures (ppm)
5 m
III in
30 m
60 m
Kill in
100 to
1,000 m
2,500 m
5,000 m
10,000 m
Disposal
15
1.8E-03
2.1E-03
7.6E-04
2.9E-04
1.3E-04
1.3E-05
8.0E-07
2.7E-07
8.8E-08
Dry film lubricant
8
6.8E-11
3.0E-09
2.2E-07
1.6E-06
2.7E-06
4.2E-07
1.2E-08
3.6E-09
1.6E-09
Ethoxylation
byproduct
6
2.8E-03
5.8E-03
3.1E-03
1.3E-03
6.9E-04
1.6E-04
9.3E-06
3.8E-06
1.5E-06
Film cement
1
5.3E-05
5.5E-05
1.9E-05
9.7E-06
5.3E-06
9.7E-07
5.8E-08
2.0E-08
6.4E-09
Functional fluids
(open-system)
2
5.4E-06
1.0E-05
4.4E-06
4.6E-06
7.7E-06
3.1E-06
2.9E-07
1.1E-07
3.7E-08
Import and
repackaging
1
1.1E-11
2.4E-10
2.3E-08
1.8E-07
3.7E-07
1.4E-07
2.8E-08
1.7E-08
9.4E-09
Industrial uses
12
1.8E-03
2.0E-03
6.5E-04
2.4E-04
1.2E-04
3.0E-05
3.7E-06
1.4E-06
4.8E-07
Laboratory chemical
1
8.7E-04
9.1E-04
3.1E-04
1.6E-04
8.7E-05
1.6E-05
9.6E-07
3.2E-07
1.1E-07
Manufacturing
1
3.7E-03
6.9E-03
3.3E-03
1.4E-03
6.7E-04
6.0E-05
3.5E-06
1.1E-06
3.4E-07
PET manufacturing
13
3.4E-03
4.0E-03
1.5E-03
5.9E-04
2.7E-04
4.5E-05
8.8E-06
5.3E-06
2.8E-06
Spray foam
application
1
3.3E-07
3.6E-07
1.2E-07
6.4E-08
3.6E-08
6.6E-09
7.3E-10
2.7E-10
1.0E-10
LADCs are based on 33 years exposure duration and the maximum 95th percentile air concentration predictions for the facility in each COU with the
greatest exposures. Adult LADCs presented in this table were used to derive the cancer risk estimates presented in Table 5-8. LADCs for individuals
exposed for a full 78 years would be 2.36 times greater than values presented here.
Page 119 of 570
-------
3.2.3.2 Hydraulic Fracturing
To evaluate general population exposures to fugitive emissions from hydraulic fracturing operations,
EPA calculated ACs, ADCs, and LADCs based on modeled air concentrations estimated in Section
2.3.3.2.4 under a range of different release scenarios and topographical conditions (Table 3-10). LADCs
within 1,000 m of hydraulic fracturing operations range from 8.7/10 4 to 5.2 ppm. These lifetime
exposure estimates are based on 33 years of exposure over a 78-year lifetime and are relevant to all
lifestages. The complete set of inhalation exposure estimates from fugitive emissions of hydraulic
fracturing operations are presented in 1,4-Dioxane Supplemental Information File: Air Exposure and
Risk Estimates for 1,4-Dioxane Emissions from Hydraulic Fracturing Operations (U.S. EPA. 2024b).
Page 120 of 570
-------
Table 3-10. Exposures from Fugitive Emissions Estimated within 1,000 m of Hydraulic Fracturing Operations"
Air Concentrations for 95th Percentile Modeled Releases
Air Concentrations for 50th Percentile Modeled Releases
Fugitive
(ppm)
(ppm)
Emissions
Exposure
High-End Modeled Air
Central Tendency (Mean)
High-End Modeled Air
Central Tendency (Mean)
Release
Duration
Concentrations
Modeled Air Concentrations
Concentrations
Modeled Air Concentrations
Scenario
100 m
1,000 m
100 to
1,000 m
100 m
1,000 m
100 to
1,000 m
100 m
1,000 m
100 to
1,000 m
100 m
1,000 m
100 to
1,000 m
AC
5.3E-02
2.4E-03
6.2E-03
4.2E-02
1.8E-03
4.8E-03
3.1E-03
1.4E-04
3.6E-04
2.4E-03
1.0E-04
2.7E-04
South
(Coastal)-
ADC
1.1E-02
4.8E-04
1.2E-03
8.3E-03
3.6E-04
9.4E-04
6.0E-04
2.7E-05
7.0E-05
4.7E-04
2.0E-05
5.4E-05
Rural-24
LADC
4.5E-03
2.0E-04
5.2E-04
3.5E-03
1.5E-04
4.0E-04
2.6E-04
1.2E-05
3.0E-05
2.0E-04
8.6E-06
2.3E-05
AC
4.1E-02
2.4E-03
5.5E-03
3.1E-02
1.5E-03
3.7E-03
2.3E-03
1.3E-04
3.1E-04
1.8E-03
8.3E-05
2.1E-04
West North
Central-
ADC
8.0E-03
4.6E-04
1.1E-03
6.1E-03
2.9E-04
7.3E-04
4.6E-04
2.7E-05
6.2E-05
3.5E-04
1.6E-05
4.2E-05
Rural-24
LADC
3.4E-03
2.0E-04
4.6E-04
2.6E-03
1.2E-04
3.1E-04
1.9E-04
1.1E-05
2.6E-05
1.5E-04
6.9E-06
1.8E-05
AC
2.6E-02
5.8E-04
1.8E-03
2.3E-02
5.0E-04
1.6E-03
1.5E-03
3.3E-05
1.0E-04
1.3E-03
2.9E-05
9.0E-05
South
(Coastal)-
ADC
5.1E-03
1.1E-04
3.5E-04
4.5E-03
9.9E-05
3.1E-04
2.9E-04
6.5E-06
2.0E-05
2.6E-04
5.7E-06
1.8E-05
Urban-24
LADC
2.2E-03
4.8E-05
1.5E-04
1.9E-03
4.2E-05
1.3E-04
1.2E-04
2.8E-06
8.5E-06
1.1E-04
2.4E-06
7.5E-06
AC
2.4E-02
6.2E-04
1.9E-03
1.9E-02
4.6E-04
1.4E-03
1.4E-03
3.6E-05
1.1E-04
1.1E-03
2.6E-05
8.1E-05
West North
Central-
ADC
4.8E-03
1.2E-04
3.7E-04
3.8E-03
9.0E-05
2.8E-04
2.7E-04
7.0E-06
2.1E-05
2.2E-04
5.2E-06
1.6E-05
Urban-24
LADC
2.0E-03
5.2E-05
1.6E-04
1.6E-03
3.8E-05
1.2E-04
1.2E-04
3.0E-06
8.9E-06
9.3E-05
2.2E-06
6.7E-06
AC
9.6E-03
8.8E-05
3.8E-04
8.3E-03
6.8E-05
3.2E-04
5.5E-04
5.1E-06
2.2E-05
4.8E-04
3.9E-06
1.8E-05
South
(Coastal)-
ADC
1.9E-03
1.7E-05
7.6E-05
1.6E-03
1.3E-05
6.2E-05
1.1E-04
1.0E-06
4.3E-06
9.4E-05
7.6E-07
3.6E-06
Rural-8
LADC
8.0E-04
7.4E-06
3.2E-05
7.0E-04
5.6E-06
2.6E-05
4.6E-05
4.2E-07
1.8E-06
4.0E-05
3.2E-07
1.5E-06
AC
2.0E-02
7.7E-04
1.9E-03
1.1E-02
2.1E-04
6.6E-04
6.1E-04
1.2E-05
3.8E-05
6.1E-04
1.2E-05
3.8E-05
West North
Central-
ADC
4.0E-03
1.5E-04
3.7E-04
2.1E-03
4.2E-05
1.3E-04
1.2E-04
2.4E-06
7.4E-06
1.2E-04
2.4E-06
7.4E-06
Rural-8
LADC
1.7E-03
6.4E-05
1.5E-04
8.9E-04
1.8E-05
5.5E-05
5.1E-05
1.0E-06
3.1E-06
5.1E-05
1.0E-06
3.1E-06
Page 121 of 570
-------
Fugitive
Emissions
Release
Scenario
Exposure
Duration
Air Concentrations for 95th Percentile Modeled Releases
(ppm)
Air Concentrations for 50th Percentile Modeled Releases
(ppm)
High-End Modeled Air
Concentrations
Central Tendency (Mean)
Modeled Air Concentrations
High-End Modeled Air
Concentrations
Central Tendency (Mean)
Modeled Air Concentrations
100 m
1,000 m
100 to
1,000 m
100 m
1,000 m
100 to
1,000 m
100 m
1,000 m
100 to
1,000 m
100 m
1,000 m
100 to
1,000 m
South
(Coastal)-
Urban-8
AC
8.7E-03
6.8E-05
3.2E-04
8.0E-03
6.0E-05
2.9E-04
4.6E-04
3.4E-06
1.7E-05
4.6E-04
3.4E-06
1.7E-05
ADC
1.7E-03
1.3E-05
6.4E-05
1.6E-03
1.2E-05
5.8E-05
9.0E-05
6.8E-07
3.3E-06
9.0E-05
6.8E-07
3.3E-06
LADC
7.3E-04
5.6E-06
2.7E-05
6.7E-04
5.0E-06
2.4E-05
3.8E-05
2.9E-07
1.4E-06
3.8E-05
2.9E-07
1.4E-06
West North
Central-
Urban-8
AC
1.5E-02
2.9E-04
9.4E-04
9.2E-03
1.2E-04
4.5E-04
5.3E-04
6.8E-06
2.6E-05
5.3E-04
6.8E-06
2.6E-05
ADC
2.9E-03
5.8E-05
1.8E-04
1.8E-03
2.4E-05
8.9E-05
1.0E-04
1.4E-06
5.1E-06
1.0E-04
1.4E-06
5.1E-06
LADC
1.2E-03
2.4E-05
7.8E-05
7.7E-04
1.0E-05
3.8E-05
4.4E-05
5.7E-07
2.2E-06
4.4E-05
5.7E-07
2.2E-06
" Lifetime Average Daily Concentrations (LADCs) presented in this table are based on 33 years exposure duration and correspond to the cancer risk estimates
presented in Table 5-7. LADCs for individuals exposed for a full 78 years would be 2.36 times greater than values presented here.
AC = Acute Concentration; ADC = Average Daily Concentration; LADC = Lifetime Average Daily Concentration
Page 122 of 570
-------
3.2.3.3 Industrial and Institutional Laundry Facilities
To evaluate exposures to emissions from industrial and institutional laundry facilities, EPA calculated
ACs, ADCs, and LADCs based on vapor and particulate air concentrations estimated in Section
2.3.3.2.4. High-end and central tendency air exposures estimated under the more conservative exposure
scenario evaluated (rural south coastal topography, assuming 24 hours of releases each day) are
presented for each type of laundry in Table 3-11. LADCs estimated within 1,000 m of laundry facilities
operations range from 8,7/10 4 to 2,4/ 10 6 ppm. These lifetime exposure estimates are based on 33
years of exposure over a 78-year lifetime and are relevant to all lifestages. The complete set of
inhalation exposure estimates from fugitive emissions of commercial laundry facilities are presented in
1,4-Dioxane Supplemental Information File: Air Exposures and Risk Estimates for Industrial Laundry
( 2024c).
Page 123 of 570
-------
Table 3-11. Exposures from Fugitive Emissions Estimated near Industrial and Institutional
Laundry Facilities"
Facility
Type
Detergent
and
Emission
Type
Exposure
Duration
Modeled Air Concentrations for Maximum Release Estimates (ppm)
High-End
Central Tendency (Mean)
100 m
1,000 m
100 to
1,000 m
100 m
1,000 m
100 to 1,000 m
Industrial
laundry
Liquid-
Vapor
AC
6.9E-06
3.3E-07
8.2E-07
5.0E-06
2.1E-07
5.6E-07
ADC
5.4E-06
2.5E-07
6.3E-07
4.9E-06
2.1E-07
5.6E-07
LADC
2.3E-06
1.1E-07
2.7E-07
2.1E-06
9.0E-08
2.4E-07
Powder-
Vapor
AC
6.9E-06
3.3E-07
8.2E-07
4.9E-06
2.1E-07
5.6E-07
ADC
5.4E-06
2.5E-07
6.2E-07
4.9E-06
2.1E-07
5.6E-07
LADC
2.3E-06
1.1E-07
2.6E-07
2.1E-06
8.9E-08
2.3E-07
Powder-
PM10
AC
7.2E-06
1.6E-07
5.9E-07
5.1E-06
1.2E-07
4.2E-07
ADC
5.6E-06
1.3E-07
4.7E-07
5.0E-06
1.2E-07
4.2E-07
LADC
2.4E-06
5.5E-08
2.0E-07
2.1E-06
4.9E-08
1.8E-07
Powder-
PM2.5
AC
6.9E-06
3.1E-07
8.0E-07
4.9E-06
2.0E-07
5.5E-07
ADC
5.4E-06
2.4E-07
6.1E-07
4.9E-06
2.0E-07
5.4E-07
LADC
2.3E-06
1.0E-07
2.6E-07
2.1E-06
8.5E-08
2.3E-07
Institutional
laundry
Liquid-
Vapor
AC
3.6E-06
1.6E-07
4.1E-07
3.1E-06
1.3E-07
3.5E-07
ADC
3.4E-06
1.6E-07
4.0E-07
3.1E-06
1.3E-07
3.5E-07
LADC
1.4E-06
6.7E-08
1.7E-07
1.3E-06
5.7E-08
1.5E-07
Powder-
vapor
AC
1.1E-07
4.8E-09
1.2E-08
9.2E-08
4.0E-09
1.0E-08
ADC
1.0E-07
4.7E-09
1.2E-08
9.2E-08
3.9E-09
1.0E-08
LADC
4.2E-08
2.0E-09
4.9E-09
3.9E-08
1.7E-09
4.4E-09
Powder-
PM10
AC
1.1E-07
2.5E-09
8.9E-09
9.4E-08
2.2E-09
7.9E-09
ADC
1.0E-07
2.4E-09
8.7E-09
9.4E-08
2.2E-09
7.9E-09
LADC
4.4E-08
1.0E-09
3.7E-09
4.0E-08
9.2E-10
3.3E-09
Powder-
PM2.5
AC
1.1E-07
4.6E-09
1.2E-08
9.2E-08
3.8E-09
1.0E-08
ADC
1.0E-07
4.5E-09
1.1E-08
9.2E-08
3.8E-09
1.0E-08
LADC
4.2E-08
1.9E-09
4.8E-09
3.9E-08
1.6E-09
4.3E-09
" LADCs presented in this table are based on 33 years exposure duration and correspond to the cancer risk estimates
presented in Table 5-10. LADCs for individuals exposed for a full 78 years would be 2.36 times greater than values
presented here.
AC = Acute Concentration; ADC = Average Daily Concentration; LADC = Lifetime Average Daily Concentration
3.3 Weight of Scientific Evidence Conclusions
As described in the 2021 Draft Systematic Review Protocol ( ), the weight of scientific
evidence supporting exposure assessments is evaluated based on the availability and strength of
exposure scenarios and exposure factors, measured and monitored data, estimation methodology and
model input data, and, if appropriate, comparisons of estimated and measured exposures. The strength of
each of these evidence streams can be ranked as either robust, moderate, slight, or indeterminate. For
Page 124 of 570
-------
each component of this exposure assessment, EPA evaluated the weight of scientific evidence for
individual evidence streams and then used that information to evaluate the overall weight of evidence
supporting each set of exposure estimates. General considerations for evaluating the strength of evidence
for each evidence stream are summarized in TableApx C-5. Specific examples of how these
considerations are applied to overall weight of evidence conclusions are provided in Table Apx C-6.
Overall confidence descriptions of high, medium, or low are assigned to the exposure assessment based
on the strength of the underlying scientific evidence. When the assessment is supported by robust
evidence, overall confidence in the exposure assessment is high; when supported by moderate evidence,
overall confidence is medium; when supported by slight evidence, overall confidence is low.
3.3.1 Occupational Exposures
The weight of scientific evidence for occupational exposure estimates is determined by several different
evidence streams, including the following:
• Evidence supporting the exposure scenarios (Section 3.1.1 and Appendix F.4);
• The quality and representativeness of available monitoring data (Appendix F.4);
• Evidence supporting modeling approaches (Section 3.1.1 and Appendix F.4); and
• Evidence supporting model input data (Appendix F.4).
3.3.1.1 Inhalation Exposure
Occupational inhalation exposure estimates are supported by moderate to robust evidence (see Appendix
F.6).
• Exposure Scenarios and Exposure Factors. The exposure scenarios and exposure factors
underlying the inhalation assessment are supported by moderate to robust evidence.
Occupational inhalation exposure scenarios and exposure factors, including duration of exposure,
body weight, and breathing rate, were informed by sources of data with medium to high data
quality ratings, increasing the strength of evidence. For most OESs/COUs, EPA used
information directly relevant to the evaluated exposure scenarios; however, for some
OESs/COUs, EPA used information from surrogate scenarios, decreasing the strength of
evidence for those scenarios. Additionally, there is uncertainty in the extent to which the entire
population of workers within an OES/COU are represented by the available data.
• Measured and Monitored Data. Measured/monitored data are supported by moderate to robust
evidence. EPA used sources of data such as OSHA and NIOSH, which have medium to high data
quality ratings, increasing the strength of the evidence. For the OESs/COUs with available
monitoring data, the data was directly applicable to the assessed exposure scenario, as opposed to
from a surrogate exposure scenario. However, the available monitoring data were limited to a
single source for each OES/COU and often consisted of a small or dated dataset. Additionally,
these data often only included one or a limited number of sites at which the data were measured,
decreasing the strength of evidence for those OESs/COUs.
• Modeling Methodologies. The modeling methodologies are supported by moderate to robust
evidence. Modeling was implemented to assess occupational inhalation exposures for three of
the OESs/COUs, using methodologies from GS/ESD that are generally well described. The
modeling incorporates Monte Carlo simulation to allow for variation in the model input data,
which increases the representativeness of the approach towards the true population of potentially
exposed workers and increases the strength of the evidence. However, EPA was unable to
develop distributions for all input parameters, increasing the uncertainty in the parameterization
and applicability.
Page 125 of 570
-------
• Model Input Data. Model input data are supported by moderate to robust evidence. For some
model input data, EPA used 1,4-dioxane-specific data from sources such as process information,
product concentration information, and FracFocus 3.0. For other model input parameters, generic
data from the GS/ESD used for the modeling methodology was used due to lack of 1,4-dioxane
data.
• Comparison of Modeled and Monitored Data. The comparison of modeled and measured
occupational inhalation exposures is not rated because no comparisons between modeled and
measured exposures were made.
Overall Confidence in Occupational Inhalation Exposure Estimates
The overall confidence in the occupational inhalation exposure estimates (Section 3.3.1.1) ranges from
low to high, depending on the OES/COU. Measured/monitored data are supported by moderate to robust
evidence. Additionally, the modeling methodologies and underlying model input data is supported by
moderate to robust evidence. However, there is uncertainty in the representativeness of the assessed
exposure scenarios towards all potential exposures for the given OES/COU, limitations in the amount
and age of monitoring data, and limitations in the modeling approaches towards 1,4-dioxane-specific use
within the OES/COU. Therefore, while the underlying data and methods used to estimate occupational
inhalation exposures is supported by moderate to robust evidence, the overall confidence of these
estimates is low to high depending on the OES/COU. OES/COU-specific discussions of the available
inhalation exposure data and overall confidence are presented in Appendix F.6.
3.3.1.2 Dermal Exposure
Occupational dermal exposure estimates are supported by slight to robust evidence (see Appendix F.3).
• Exposure Scenarios and Exposure Factors. The exposure scenarios and exposure factors
underlying the dermal assessment are supported by moderate to robust evidence. Dermal
exposure scenarios were informed by process information and GS/ESD with medium to high
data quality ratings, increasing the strength of evidence. Exposure factors, including amount of
material on skin, surface area of skin exposed, and absorption of 1,4-dioxane through the skin,
were informed by literature sources, the ChemSTEER User Guide (x v << \ a) for
standard exposure parameters, and a European model, which have medium to high data quality
ratings. EPA used information directly relevant to the evaluated exposure scenarios; however,
there is uncertainty in the extent to which the entire population of workers within an OES/COU
are represented by the available data.
• Measured and Monitored Data. No measured/monitored dermal exposure data were used in
the occupational dermal exposure assessment. EPA did use measured data on 1,4-dioxane
concentrations in various products from process information and other literature sources, which
have medium to high data quality ratings, depending on the data source.
• Modeling Methodologies. The modeling methodologies are supported by moderate evidence.
EPA used the EPA Dermal Exposure to Volatile Liquids Model to calculate the dermal retained
dose for each OES/COU. This model modifies the EPA/OPPT 2-Hand Dermal Exposure to
Liquids Model by incorporating a "fraction absorbed (fabs)" parameter to account for the
evaporation of volatile chemicals. Additionally, the model incorporates a glove "protection
factor" to inform risk management decisions. These modifications improve the modeling
methodology and allow EPA to differentiate dermal exposures between commercial and
industrial settings by varying the absorption and dermal protection factors. However, the
modeling approach is still limited by the low variability for different worker activities/exposure
scenarios.
Page 126 of 570
-------
• Model Input Data. Model input data are supported by slight to moderate evidence. As discussed
above, model parameters were informed by sources with medium to high data quality ratings.
However, a limitation is that some of the model input data was generic and not specific to 1,4-
dioxane OES/COU.
• Comparison of Modeled and Monitored Data. All occupational dermal exposures were
modeled, and no measured dermal exposures were used in this assessment, therefore there is no
comparison.
Overall Confidence in Occupational Dermal Exposure Estimates
The overall confidence in the occupational dermal exposure estimates (Section 3.1.2.2) is medium for all
OES/COU because the same modeling approach was used for all OES/COU. The modeling
methodology is supported by moderate evidence, with model input parameters from literature sources, a
European model, standard defaults from the ChemSTEER User Guide ( a), and 1,4-
dioxane product concentration data from process information. These sources range from slight to robust,
depending on factors such as age and applicability to OES/COU. The modeling is limited by the use of
standard input parameters that are not specific to 1,4-dioxane and a lack of variability in dermal
exposure for different worker activities. Therefore, EPA's overall confidence in the occupational dermal
exposure estimates is medium.
3.3.2 Drinking Water
3.3.2.1 Drinking Water Exposure Estimates Based on Surface Water Concentrations
The weight of evidence for drinking water exposure estimates is determined by several different
evidence streams, including the following:
• Evidence supporting the general population exposure scenarios (Section 3.2.1);
• The quality and representativeness of available surface water and drinking water monitoring
data (Section 2.3.1.1);
• Evidence supporting modeling approaches (Section 2.3.1.3 and Appendix G.2);
• Evidence supporting release data used as model input data (Section 2.2 and Appendix E.3); and
• Consistency between modeled and monitored water concentrations (Section 2.3.1.4).
As described in Section 2.3.1, multiple approaches were used to predict surface water concentrations
resulting from several sources. These included the evaluation of facility-specific releases, down the
drain releases to surface water, hydraulic fracturing releases and aggregation of surface water releases.
The associated strengths, limitations and confidence in these estimated environmental concentrations are
described in Section 2.3.1.4. The general population drinking water exposure scenarios and exposure
factors used to estimate exposures that could result from estimated water concentrations are described in
Section 3.2.
Drinking water exposure estimates based on modeled surface water concentrations are supported by
overall moderate to robust evidence, with the strength of the evidence varying across analysis
approaches and COUs/OESs.
• Exposure Scenarios and Exposure Factors. The exposure scenarios and exposure factors
underlying all drinking water exposure estimates are supported by moderate to robust evidence.
Exposure factors for drinking water are based on robust data on drinking water intake rates and
body weights as derived from exposure factors from the EPA's Exposure Factors Handbook
( ). For chronic drinking water exposure scenarios, mean water ingestion values
were applied, where 95th percentile ingestion values could result in as much as 3-4 times higher
Page 127 of 570
-------
exposures, depending on the age groups considered. The drinking water exposure scenarios
generally rely on the assumption that little or no dilution occurs prior to drinking water intakes.
That assumption may not be representative of exposures at all locations. Although there are
locations where this assumption is expected to the accurate, the extent of downstream dilution
that occurs prior to drinking water intakes is highly variable across locations. The proximity of
facility releases to actual drinking water intakes is evaluated in Section 2.3.1.2.4 and Appendix
G.2.4. Uncertainties related to downstream dilution decrease the overall strength of evidence for
these exposure scenarios. However, EPA has performed several analyses that calculate exposures
and risks under alternate assumptions about downstream dilution and illustrate the quantitative
impact of those assumptions (see Section 5.2.2.1.2), increasing the overall strength of evidence.
Drinking water exposure scenarios also rely on the data-driven assumption that 1,4-dioxane is
not removed through treatment. Moderate to robust data provide support for this assumption
under many treatment scenarios. These assumptions may over-estimate exposure for some
locations, but provide an overall distribution that is generally expected to be representative of
exposure scenarios.
• Measured and Monitored Data. The measured/monitored data are supported by moderate
evidence. The high number of monitoring data points for surface water and drinking water from
high quality sources in multiple locations over multiple years increases the strength of the
evidence from monitoring data. Monitoring data confirm that 1,4-dioxane is present in some
surface water and drinking water in some locations. However, evidence from monitoring data
may not be representative of all sites where 1,4-dioxane is released to surface water from TSCA
sources, decreasing the strength of evidence from monitoring data. The lack of temporal and/or
spatial alignment between most monitoring data and reported release locations makes direct
comparison challenging for most locations. However, a limited number of sites with monitoring
data are co-located with sites where 1,4-dioxane releases are reported, supporting comparisons of
monitoring and modeled estimates that increase the overall strength of the evidence. In addition,
as described in Section 2.3.1.4, monitoring data for surface water directly downstream from
releases show concentrations multiple orders of magnitude greater than typical ambient surface
water concentrations, aligning with patterns of modeled results.
• Modeling Methodologies. The modeling methodologies are supported by moderate to robust
evidence.
o The methodology for deriving exposure estimates for facility releases is moderate and is
applicable to the populations included in the exposure scenarios. This approach makes
some conservative assumptions about flow rates and release frequency and amount.
Additionally, the modeling does not take into account downstream fate or transport, but
the physical chemical properties of 1,4-dioxane are expected to moderate the impact
these influences could have on the modeled instream concentrations. The model is
designed to estimate possible higher end water concentrations expected at specific
locations.
o The probabilistic methodology used for deriving exposure estimates for DTD releases,
hydraulic fracturing releases, and aggregate releases from all sources is robust. This
approach incorporates the full distribution of facility releases over multiple years and
corresponding instream flow rate data rather than relying on the most conservative model
inputs. It is designed to provide a nationally representative distribution of estimated water
concentrations under varying conditions.
• Model Input Data. Model input data are supported by slight to robust evidence, with the
strength of the evidence varying across individual COUs/OESs. The strength of evidence
Page 128 of 570
-------
supporting modeled water concentrations relies heavily on the quality of the facility or OES-
specific release data used as inputs for the model, including both the amount of release, location
of the release, and the corresponding flow in the receiving water body. A summary of sources of
flow and release data for facility release modeling is presented in Table 2-7. A more detailed
OES-specific discussion of the confidence in sources of release information is presented in
Appendix E.3.4.
o For overall distributions of industrial releases across sites, model input data are supported
by robust evidence. As illustrated in Section 5.2.2.1.2, EPA estimated exposures and risks
across the full distribution of facility releases both for the whole dataset and for a subset
of facilities with high quality reporting information. Comparison of these distributions
demonstrates that inclusion of locations relying on more limited release information had
limited impact on the overall distributions of exposures.
o For COUs/OESs that rely primarily on release data reported to TRI via Form R, or
reported to ICIS-NPDES via DMR, site-specific release estimates are supported by
moderate to robust evidence. As described in Appendix E.3.1, these release estimates are
based on release amounts reported by facilities. Most COUs/OESs are included in this
group.
o For COUs/OESs that rely primarily on release data reported to TRI via Form A, site-
specific release estimates are supported by moderate evidence. As described in Appendix
E.3.1, Form A simply indicates that releases are below the reporting thresholds and
specific release estimates require assumptions about amounts, locations, and media of
release. The Import and repackaging OES releases used in this analysis are entirely based
on Form A reporting of releases, and just under half of the Industrial Uses OES releases
were reported via Form A.
o For COUs/OESs that rely primarily on other sources of release information or generic
scenarios, site-specific release estimates are supported by slight to moderate evidence.
For these scenarios, EPA estimated daily wastewater discharges by using various
modeling approaches, including the use of surrogate TRI and DMR data and modeling
using data from literature, GSs, and ESDs.
¦ For DTD sources, release information is supported by slight to moderate
evidence. Although confidence in the individual contribution from some specific
COUs (ie specific consumer or commercial product categories) is lower,
confidence in estimates of overall DTD releases is moderate. The presented model
is intended to inform the total contribution of DTD releases to overall aggregate
instream concentration as well as providing evidence of individual COUs that
may be most influential. Presented results should be taken in relation to one
another qualitatively rather than discrete quantitative values. Distributions of
DTD releases of consumer and commercial products were estimated for each
COU on a per capita basis using the SHEDS-HT model. There is slight to
moderate evidence that the proportions of populations applied for the commercial
users. The selection of high-end consumer loading rates to represent commercial
uses, may not reflect all communities and commercial use patterns, particularly
for the paint and dishwashing COUs.
¦ For hydraulic fracturing releases, release information is supported by moderate
evidence. Releases were estimated using Monte Carlo modeling with information
from the Revised ESD on Hydraulic Fracturing and FracFocus 3.0.
Page 129 of 570
-------
• Comparison of Modeled and Monitored Data. Comparisons of estimated and measured
exposures provide moderate evidence. Because most of the available monitoring data are not
collected in locations with known releases temporally or spatially, it is difficult to make direct
comparisons in most locations. However, in case study locations where monitoring data re
located near release sites, comparisons demonstrate that there is general consistency between
measured and/or reported and modeled estimates (Section 2.3.1.4), increasing the overall
strength of the evidence. Monitoring data confirm that 1,4-dioxane is present in some surface
water and drinking water. Uncertainty as to whether trends observed in case study locations are
representative of all of the sites decreases overall confidence in these comparisons.
Overall Confidence in Exposure Estimates
Overall confidence in drinking water exposure estimates for surface water concentrations modeled from
facility releases (Section 3.2.2.1.1) is high across the overall distribution, particularly when limited to
sites with high quality sources of release data. For individual facilities and COUs, overall confidence in
exposure estimates varies depending on the confidence in source-specific release data. The modeling
methodology used for this analysis is supported by moderate evidence. This approach makes some
conservative assumptions about flow rates and release frequency and amount. It is designed to estimate
water concentrations expected at specific locations. Available monitoring data confirm that 1,4-dioxane
is present in some surface water and drinking water, though most of the available data were not collected
near release sites are therefore not directly comparable. The overall level of confidence in OES/COU-
specific exposure estimates depends on the source of OES/COU-specific release data described in
Appendix E.3:
• Overall confidence in drinking water exposure estimates is medium to high for OESs/COUs that
rely primarily on site-specific release data reported to DMR or to TRI via Form R.
• Overall confidence in site-specific drinking water exposure estimates is medium for OESs/COUs
for which site-specific release estimates are based on reporting to TRI via Form A.
• Overall confidence in drinking water exposure estimates is low to medium for OESs/COUs for
which site-specific release estimates are based on surrogate or modeled information.
Overall confidence in drinking water exposure estimates for DTD releases under varying conditions
(Section 3.2.2.1.2), is medium. The modeling methodology used for this analysis is supported by robust
evidence. This analysis is designed to provide a nationally representative distribution of estimated water
concentrations under varying conditions. This analysis defines the conditions under which exposures are
higher, but is not designed to predict the specific levels of exposure resulting from DTD releases at
specific locations with precision. Exposure estimates rely on estimated distributions of DTD releases of
specific consumer and commercial products categories associated with each COU. Distributions of DTD
releases of consumer and commercial products were estimated for each COU on a per capita basis using
the SHEDS-HT model. Although confidence in the individual contribution from some specific COUs is
lower, confidence in estimates of overall DTD releases is moderate.
Overall confidence in drinking water exposure estimates for hydraulic fracturing releases (Section
3.2.2.1.3) is medium. The modeling methodology used for this analysis is supported by robust evidence
and is designed to provide a nationally representative distribution of estimated water concentrations
under varying conditions. Releases used as inputs in the model were estimated using Monte Carlo
modeling that captures variability across sites. However, the modeled exposure estimates are not directly
tied to specific releases at known locations, decreasing the strength of the evidence related to the
representativeness of the exposure estimates for actual exposures.
Overall confidence in drinking water exposure estimates for aggregate surface water concentrations
predicted by probabilistic modeling (Section 3.2.2.1.4) is high across the overall distribution. For
Page 130 of 570
-------
individual facilities and COUs, overall confidence in exposure estimates varies depending on the
confidence in source-specific release data. The modeling methodology used for this analysis is
supported by robust evidence and is designed to provide a nationally representative distribution of
estimated water concentrations under varying conditions. The estimated drinking water concentrations
modeled in this analysis incorporate contributions from direct and indirect industrial releases, DTD
releases, and other upstream sources. Available monitoring data confirm that 1,4-dioxane is present in
some surface water and drinking water, though most of the available data were not collected near release
sites and are therefore not directly comparable. The overall level of confidence in resulting exposure
estimates depends on the source of OES/COU-specific release data described in Appendix E.3:
• Overall confidence in drinking water exposure estimates is medium to high for OESs/COUs that
rely primarily on release data reported to DMR or to TRI via Form R. Most COUs/OESs are
included in this group.
• Overall confidence in drinking water exposure estimates is medium for OESs/COUs for which
release estimates are based on reporting to TRI via Form A. The Import and repackaging OES
releases used in this analysis are entirely based on Form A reporting of releases, and just under
half of the Industrial Uses OES releases were reported via Form A.
• Overall confidence in drinking water exposure estimates is low to medium for OESs/COUs for
which release estimates are based on surrogate or modeled information.
3.3.2.2 Drinking Water Exposure Estimates Based on Groundwater Concentrations
The weight of evidence for exposure estimates presented in this section is determined by several
different evidence streams, including the following:
• Evidence supporting the exposure scenarios (Section 3.2.1);
• The quality and representativeness of available groundwater monitoring data (Section 2.3.2.1);
• Evidence supporting modeling approaches and input data (Sections 2.3.2.3.1 and 2.3.2.4.1);
• Evidence supporting release data used as model input data (Section 2.2 and Appendix E.4); and
• Agreement between modeled and monitored water concentrations.
3.3.2.2.1 Groundwater Concentrations Resulting from Disposal to Landfill
Drinking water exposure estimates based on groundwater concentrations modeled for landfill disposal
scenarios are supported by overall slight to moderate evidence.
• Exposure Scenarios and Exposure Factors. The exposure scenarios and exposure factors
underlying these drinking water exposure estimates are supported by slight to moderate
evidence. Exposure factors for drinking water are based on robust data on drinking water intakes,
body weight, and other standard exposure factors from the EPA's Exposure Factors Handbook
( ). However, the drinking water exposure scenario relies on the assumption that
the groundwater concentrations estimated with the DRAS model may occur in locations where
groundwater is used as a primary drinking water source. Although there is uncertainty around
this assumption, this analysis is intended to capture a scenario where the greatest exposures are
likely to occur.
• Measured and Monitored Data. Measured/monitored data are supported by moderate evidence.
Monitoring data were available to sufficiently cover most or all of the population groups
included within the exposure scenarios but there are a limited number of studies to corroborate
findings. Since little data is readily available on the concentration of 1,4-dioxane near or around
landfills in groundwater, some caution is required when interpreting monitoring data as it may
not be fully representative of conditions around all landfills.
Page 131 of 570
-------
• Modeling Methodology. The modeling methodology is supported by robust evidence. The
DRAS methodology for deriving the estimate is well described. The underlying computational or
scientific basis is robust and has an empirical basis considering chemical specific properties.
• Model Input Data. The release data relied on as a model input is supported by slight evidence.
Model inputs for the DRAS model include chemical properties of 1,4-dioxane that are well-
defined and reviewed and therefore supported by robust evidence. However, model inputs for
leachate concentrations and loading rates are more uncertain. EPA does not have reasonably
available information on actual concentrations of 1,4-dioxane in leachate for most landfills and
therefore selected landfill leachate concentrations are based on potential for risk to human health.
Loading rates are based on the range reported in TRI for RCRA subtitle C landfills and therefore
may not be representative of nonhazardous landfills evaluated in this analysis. These
uncertainties around landfill leachate concentrations and loading rates decrease the strength of
the evidence for model input data.
• Comparison of Modeled and Monitored Data. Comparison of estimated and measured
exposures provides moderate evidence because monitoring data confirm the presence of 1,4-
dioxane in groundwater in some locations and modeled estimates and measured exposure values
are comparable, however differences in methodology, collection, or context make it difficult to
arrive at full agreement.
Overall Confidence in Exposure Estimates
Overall confidence in drinking water exposure estimates resulting from disposal to landfills (Section
3.2.2.2.1) is low to medium. The modeling methodology is robust. However, the release information
relied on as model input data is supported by slight to moderate evidence, decreasing overall confidence.
In addition, this drinking water exposure scenario relies on the assumption that the groundwater
concentrations estimated with the DRAS model may occur in locations where groundwater is used as a
primary drinking water source. Although the substantial uncertainty around the extent to which these
exposures occur decreases overall confidence in the exposure scenario, this scenario represents a PESS
exposure.
3.3.2.2.2 Groundwater Concentrations Resulting from Disposal of Hydraulic
Fracturing Waste
Drinking water exposure estimates based on modeled groundwater concentrations estimated under a
range of hydraulic fracturing waste disposal scenarios are supported by slight to moderate evidence.
• Exposure Scenarios and Exposure Factors. The exposure scenario factors underlying these
exposure estimates are supported by slight to moderate evidence. Exposure factors for drinking
water are based on robust data on drinking water intakes, body weight, and other standard
exposure factors from EPA's Exposure Factors Handbook. However, the drinking water
exposure scenario relies on the assumption that the estimated groundwater concentrations may
occur in locations where groundwater is used as a primary drinking water source.
• Measured and Monitored Data. The measured/monitored data are supported by indeterminate
evidence. Available groundwater monitoring data are not located near hydraulic fracturing
operations and do not provide information about the potential for hydraulic fracturing operations
to contribute to groundwater contamination.
• Modeling Methodologies. The modeling methodology and input data are supported by robust
evidence. The methodology for deriving the estimate is well described, the underlying
computational or scientific basis is robust, and has an empirical basis considering chemical
specific properties.
Page 132 of 570
-------
• Model Input Data. Hydraulic fracturing releases are supported by moderate evidence. As
described in Appendix E.4.4, releases were estimated using Monte Carlo modeling with
information from the Revised ESD on Hydraulic Fracturing and FracFocus 3.0. DRAS modeling
was based on very limited data on concentrations of 1,4-dioxane in produced water as reported in
the literature. Reliance on limited data and uncertainty around the representativeness of that data
decrease the strength of the evidence for model input data.
• Comparison of Modeled and Monitored Data. The comparison of estimated and measured
exposures is not rated because no comparisons between estimated and measured exposures were
made.
Overall Confidence in Exposure Estimates
Overall confidence in drinking water exposure estimates resulting from disposal of hydraulic fracturing
waste (Section 3.2.2.2.2) is low to medium. The modeling methodology is robust and the release
information relied on as model input data is supported by moderate evidence. However, no monitoring
data are available to confirm detection of 1,4-dioxane in groundwater near hydraulic fracturing
operations. This drinking water exposure scenario relies on the assumption that the estimated
groundwater concentrations may occur in locations where groundwater is used as a primary drinking
water source. Although the substantial uncertainty around the extent to which these exposures occur
decreases overall confidence in the exposure scenario, this scenario represents a PESS exposure.
3,3.3 Air
The weight of scientific evidence for exposure estimates presented in this section is determined by
several different evidence streams, including the following:
• Evidence supporting the exposure scenarios (Section 3.2.1);
• The quality and representativeness of available groundwater monitoring data (Section 2.3.3.1);
• Evidence supporting modeling approaches and input data (Section 2.3.3.2);
• Evidence supporting release data used as model input data (Section 2.3.3.2); and
• Consistency between modeled and monitored water concentrations.
As described in Section 2.3.3, 1,4-dioxane concentrations in air were estimated for areas around
industrial COUs reported to TRI, hydraulic fracturing operations, and institutional and industrial laundry
facilities. The associated strengths and limitations of these estimated environmental concentrations are
described in Section 2.3.3.3. The general population air exposure scenarios and exposure factors used to
estimate exposures are described in Section 3.2.3.
3.3.3.1 Modeled Air Concentrations for Industrial COUs Reported to TRI
Inhalation exposure estimates resulting from 1,4-dioxane releases for industrial COUs reported to TRI
are supported by overall moderate evidence.
• Exposure Scenarios and Exposure Factors. Exposure scenarios underlying these exposure
estimates are supported by moderate evidence. The exposure factors used to build the exposure
scenarios are directly relevant to general population exposures for communities living in close
proximity to releasing facilities. While the long-term exposure scenarios are most directly
relevant for individuals who reside in fenceline communities for many years, these scenarios are
expected to be within the range of normal habits and exposure patterns expected in the general
population. However, there is uncertainty around the extent to which people actually live and
work around the specific facilities where exposures are highest, decreasing the overall strength of
evidence for these exposure scenarios, particularly at the distances nearest to facilities.
Page 133 of 570
-------
• Measured and Monitored Data. The measured/monitored data are supported by indeterminate
evidence. No measured or monitored data were available.
• Modeling Methodologies. The modeling methodology used to estimate exposure concentrations
via the ambient air pathway is supported by robust evidence. Air concentrations were estimated
using AERMOD and IIOAC. AERMOD is EPA's regulatory model and has been thoroughly
peer reviewed; therefore, the general confidence in results from the model is high but reliant on
the integrity and quality of the inputs used and interpretation of the results. Confidence in
modeled air concentrations resulting from stack releases is lower at distances less than 100 m of
release sites, but confidence in modeled concentrations for fugitive emissions is higher near
release sites. Although this is a source of uncertainty, air concentrations from fugitive emissions
tend to peak within 10 m of release sites while stack releases were found to peak around 100 m,
indicating that air concentrations modeled at distances less than 100 m of release sites are
generally driven by fugitive emissions. IIOAC is an Excel-based model with results based on
pre-run AERMOD exposure scenarios under a variety of environmental and release conditions.
There is a moderate to high confidence in air concentrations estimated using IIOAC because,
although IIOAC results are based on pre-run AERMOD exposure scenarios (high confidence),
some key sources of uncertainty identified in Section 2.3.3.3 (like limited set of distances
evaluated (100, 100 to 1,000, and 1,000 m) and assumptions made about meteorological
conditions necessary to provide a more conservative exposure estimate) can lead to a slightly
lower confidence (moderate).
• Model Input Data. Model input data on air releases are supported by slight to robust evidence,
with the strength of the evidence varying across COUs/OESs. A more detailed OES-specific
discussion of the confidence in sources of release information is presented in Appendix E.5.4.
o For COUs/OESs that rely primarily on release data reported to TRI via Form R, site-
specific release estimates are supported by moderate to robust evidence. As described in
Appendix E.5.4, these release estimates are based on specific release amounts and other
source-specific information reported by facilities as a regulatory requirement.
o For COUs/OESs that rely primarily on release data reported to TRI via Form A, site-
specific release estimates are supported by moderate evidence. As described in Appendix
E.5.4, Form A simply indicates that releases are below the reporting thresholds and
specific release estimates require assumptions about exact amounts and locations of
releases.
o For COUs/OESs that rely primarily on other sources of release information or generic
scenarios, release estimates are supported by evidence ranging from slight to moderate
evidence. For these scenarios, EPA estimated daily and annual air releases using various
modeling approaches, including the use of surrogate TRI data and modeling using data
from literature, GSs, and ESDs.
• Comparison of Modeled and Monitored Data. Comparison of estimated and measured
exposures provide indeterminate evidence. No measured or monitored data were available for
comparison.
Overall Confidence in Exposure Estimates
Overall confidence in inhalation exposure estimates resulting for air concentrations modeled based on
industrial releases (Section 3.2.3.1) varies across COUs. The AERMOD modeling methodology used for
this analysis is robust and considers contributions from both stack and fugitive emissions. The exposure
scenarios considered are most relevant to long-term residents in fenceline communities. There is
uncertainty around the extent to which people live and work in the specific locations where exposures
Page 134 of 570
-------
are highest, decreasing confidence in the exposure scenarios, particularly at distances nearest to
facilities. Overall confidence varies due to variable levels of confidence in underlying release
information used to the support the analysis:
• Overall confidence in site-specific inhalation exposure estimates is medium to high for
OESs/COUs that rely primarily on release data reported to TRI via Form R.
• Overall confidence in site-specific inhalation exposure estimates is medium for OESs/COUs for
which release estimates are based on data reported to TRI via Form A.
• Overall confidence in inhalation exposure estimates is low to medium for OESs/COUs for which
release estimates are based on surrogate or modeled information.
3.3.3.2 Air Concentrations Modeled near Hydraulic Fracturing Operations and
Industrial/Institutional Laundries
Inhalation exposure estimates resulting from 1,4-dioxane released to air from hydraulic fracturing
operations and industrial/institutional laundries are supported by overall moderate evidence.
• Exposure Scenarios and Exposure Factors. Exposure scenarios underlying these exposure
estimates are supported by moderate evidence. The factors used to build the exposure scenarios
are directly relevant to general population exposures for communities living in close proximity to
releasing facilities. While the long-term exposure scenarios are most directly relevant for
individuals who reside in fenceline communities for many years, these scenarios are expected to
be within the range of normal habits and exposure patterns expected in the general population.
However, there is some uncertainty around the extent to which people actually live and work
around the specific locations where exposures are highest, decreasing the overall strength of
evidence for these exposure scenarios.
• Measured and Monitored Data. The measured/monitored data are supported by indeterminate
evidence. No measured or monitored data were available.
• Modeling Methodologies. The modeling methodology used to estimate exposure concentrations
via the ambient air pathway is supported by robust evidence. Air concentrations were estimated
using IIOAC. IIOAC is an Excel-based model with results based on pre-run AERMOD exposure
scenarios under a variety of environmental and release conditions. There is a moderate to high
confidence in air concentrations estimated using IIOAC because, although IIOAC results are
based on pre-run AERMOD exposure scenarios (high confidence), some key sources of
uncertainty identified in Section 2.3.3.3 (like limited set of distances evaluated (100, 100 to
1,000, and 1,000 m) and assumptions made about meteorological conditions necessary to provide
a more conservative exposure estimate) can lead to a slightly lower confidence (moderate).
• Model Input Data. Input data used for modeling exposures from hydraulic fracturing operations
and industrial/institutional laundries are supported by moderate evidence. As described in
Appendix E.5.4, these modeled exposure estimates are based on alternative release estimates and
scenario conditions found in the literature and derived with Monte Carlo models of release
estimate, some of which have been peer reviewed, others which may not be peer reviewed. Since
the modeled exposures are based on alternative release estimates, which in turn are based on
modeled data and outputs, there is a lower overall confidence in the modeled exposures from
such input data. Additionally, exposure estimates using this input data requires certain
assumptions which can lead to a lower overall confidence in the estimated exposure
concentrations.
Page 135 of 570
-------
• Comparison of Modeled and Monitored Data. Comparison of estimated and measured
exposures provide indeterminate evidence. No measured or monitored data were available for
comparison.
Overall Confidence in Exposure Estimates
Overall confidence in inhalation exposure estimates resulting for air concentrations modeled based on
releases from hydraulic fracturing operations (Section 3.2.3.2) is medium. The modeling methodologies
used to estimate air concentrations are robust. The distribution of air releases used as model input data
were estimated using Monte Carlo modeling and rely on assumptions. No air monitoring data were
available to confirm detection of 1,4-dioxane in air near hydraulic fracturing operations. There is
uncertainty around the extent to which people live and work in the specific locations where exposures
are highest, decreasing confidence in the exposure scenarios.
Overall confidence in inhalation exposure estimates resulting from air concentrations modeled based on
releases from industrial and institutional laundries ion 0) is medium. The modeling methodologies are
robust. The distribution of air releases used as model input data were estimated using Monte Carlo
modeling and rely on assumptions. No air monitoring data were available to determine whether 1,4-
dioxane is detected near industrial and institutional laundry facilities.
Page 136 of 570
-------
4 HUMAN HEALTH HAZARD
1,4-Dioxane - Human Health Hazards (Section 4):
Key Points
EPA previously evaluated reasonably available information for human health hazards and
identified hazard endpoints for non-cancer effects and cancer effects following acute and chronic
exposures. This section describes adjustments made to previously published hazard values to
align with the exposure scenarios evaluated in this supplemental evaluation.
• EPA considered the potential for increased susceptibility across PESS factors throughout
the hazard assessment and dose-response analysis. PESS categories identified in the
assessment include lifestage, genetics, and preexisting disease.
• The primary acute/short-term, non-cancer endpoint for 1,4-dioxane is liver toxicity
following inhalation exposure.
• The primary chronic, non-cancer endpoints for 1,4-dioxane are liver toxicity and systemic
effects on the olfactory epithelium.
• Inhalation cancer endpoint for 1,4-dioxane is based on combined tumor risk at multiple
sites.
• Oral and dermal cancer endpoints for 1,4-dioxane are based on liver tumors following
oral exposures.
4.1 Summary of Hazard Endpoints Previously Identified in the 2020 Risk
Evaluation
This supplement relies on the Hazard Identification and Dose-Response Assessment that was previously
described in the 2020 RE. All hazard values used to calculate risks for 1,4-dioxane in this supplement
were derived from the previously peer-reviewed PODs published in the 2020 RE and amended in the
2023 correction memo.12
Hazard values used in the 2020 RE include human equivalent concentrations (HECs) and human
equivalent doses (HEDs) for non-cancer endpoints. Additionally, an inhalation unit risk (IUR) and
cancer slope factor (CSF) for lifetime cancer risk were derived for both occupational and consumer
scenarios for COUs where it was applicable. The hazard values published in the 2020 RE and used as
the basis for hazard values in this supplement were developed with consideration for potentially
susceptible subpopulations. Several potential sources of susceptibility were discussed qualitatively
including lifestage, genetic variability, liver disease, and other chronic diseases that may influence
metabolism or target organ susceptibility. EPA applied a 10x uncertainty factor to non-cancer hazard
values to account for these sources of human variability.
4.2 Summary of Adjustments to Previously Established Hazard Values
For many of the exposure scenarios evaluated in this supplement, the previously established peer-
reviewed hazard values were applied without modification. For example, risks from occupational
12 In June 2023, EPA posted a correction of dermal hazard values to the docket. Correction of Dermal Acute and Chronic
Non-cancer Hazard Values Used to Evaluate Risks from Occupational Exposures in the Final Risk Evaluation for 1,4-
Dioxane is available at https://www.regulations.gov/document/EPA-HO-OPPT-2016-0723-0Q99.
Page 137 of 570
-------
exposures to products containing 1,4-dioxane as a byproduct can be evaluated using the acute, chronic,
and cancer hazard values previously developed for OESs.
Some of the exposure scenarios included in this supplement require duration adjustments to the
previously established PODs. For example, to evaluate risks from ambient air exposures for fenceline
communities, EPA assumes continuous exposure to air for 24 hours/day, 7 days/week. As described in
more detail below, EPA adjusted the previously established HEC and IUR values (originally developed
for 8 hours/day, 5 days/week exposures) to identify hazard values appropriate for continuous exposure
scenarios.
In addition, acute and chronic non-cancer oral and dermal HEDs extrapolated from occupational HECs
were corrected to apply consistent breathing rates assumptions.
The full set of hazard values used to evaluate risk from the exposure scenarios in this supplement are
presented in Table 4-1.
Page 138 of 570
-------
Table 4-1. Hazard Values Used for 1,4-Dioxane in this Supplement
Scenario
(Population)
Endpoints
Inhalation
HEC/IUR
Dermal
HED/CSF
Oral HED/CSF
Total
Uncertainty
Factors
Reference(s)
Acute/short-term non-
cancer
(general population)
Systemic liver
toxicity
26.2 ppm
(94.5 mg/m3)
24 hours
17.4 mg/kg-d
(extrapolated from HEC)
17.4 mg/kg-d
(extrapolated from HEC)
300
(Putz et al.. 1979)
(Mattie et al.. 2012)
Acute
non-cancer
(occupational)
Systemic liver
toxicity
78.7 ppm
(284 mg/m3)
8 hours
17.4 mg/kg-d
(extrapolated from HEC)'1
17.4 mg/kg-d
(extrapolated from
HEC)17
300
(Mattie et al.. 2012)
Chronic
non-cancer
(general population)
Olfactory epithelium
effects attributed to
systemic delivery
(inhalation)l1; liver
toxicity (oral)
0.846 ppm
(3 mg/m3)
24 hours, 7
days/week
0.56 mg/kg-d
(extrapolated from HEC)
2.6 mg/kg-d
30
(Kano et al.. 2009;
Kasai et al.. 2009)
Chronic
non-cancer
(occupational)
Olfactory epithelium
effects attributed to
systemic delivery
(inhalation)a; liver
toxicity (oral)
3.6 ppm
(12.8 mg/m3)
8 hours, 5
days/week
0.56 mg/kg-d
(extrapolated from HEC)''
2.6 mg/kg-d
30
(Kano et al.. 2009;
Kasai et al.. 2009)
Cancer
(general population)
Inhalation cancer
risk based on
combined tumor risk
at multiple sites;
oral/dermal cancer
risk based on liver
tumors
IUR:
1.6E-02 per ppm
4.3E-06
(Hg/m3) 1
24 hours, 365
days/ year
CSF:
1.2E-01 (mg/kg-d)1
(extrapolated from oral
CSF)
CSF:
1.2E-01 (mg/kg-d)1
(Kano et al.. 2009;
Kasai et al.. 2009;
NTP. 1986)
Cancer
(occupational)
Inhalation cancer
risk based on
combined tumor risk
at multiple sites;
oral/dermal cancer
risk based on liver
tumors
IUR:
3.7E-03 per ppm
1.0E-06
(Hg/m3) 1
8 hours, 5
days/week
CSF: 1.2E-01 (mg/kg-d)1
(extrapolated from oral
CSF)
CSF:
1.2E-01 (mg/kg-d)1
(Kano et al.. 2009;
Kasai et al.. 2009)
"Due to the uniform distribution of lesions (rather than a distribution consistent with airflow), EPA concluded that effects in the olfactory epithelium may be due to
systemic delivery rather than portal of entry effects due to the (see discussion in the 2020 RE on p 183, p. 188).
b Occupational HEDs extrapolated from occupational HECs were corrected as described in the correction memo.
HEC = Human Equivalent Concentration; HED = Human Equivalent Dose; CSF = Cancer Slope Factor; IUR = Inhalation Unit Risk
Page 139 of 570
-------
4,2,1 Derivation of Acute/Short-Term Hazard Values
4.2.1.1 Inhalation HEC
The acute/short-term HECs are based on the lowest-observed-adverse-effect concentration (LOAEC) for
systemic liver toxicity observed in a short-term inhalation toxicity study in rats exposed 6 hours/day for
5 days/week. In the 2020 RE, EPA derived an HEC for 8-hour occupational exposures by applying a
duration adjustment for an 8-hour exposure and a dosimetric adjustment factor of 1 (the default value
when the calculated ratio of animal to human blood:air partition coefficients is greater than 1 (U.S. EPA.
1994b). The occupational HEC derived in the 2020 RE is based on default breathing rate assumptions
and did not use adjustments for occupational breathing rates. For this supplement, EPA also derived an
HEC for continuous general population exposures by applying a 24-hour duration adjustment to the
original HEC.
4.2.1.2 Oral and Dermal HEDs
In the absence of acute oral or dermal toxicity studies, the acute/short-term HED was derived from the
acute HEC using route-to-route extrapolation. An acute HED for the general population was derived
from the duration-adjusted 24-hour HEC using the following equation:
dermal or oral HED (mg/kg-d) = PODhec (mg/m3) x inhalation volume x 100% inhalation
absorption ^ body weight
where the inhalation volume for the general population is 14.7 m3/day and body weight is 80 kg, based
on EPA's Exposure Factors Handbook (\ c< < i1 \ JO I I). Inhalation absorption was estimated based on
experimental data from inhalation exposures in humans (Young et at.. 1977: Young et at.. 1976) that
indicated that 1,4-dioxane is readily absorbed; however, the available studies did not measure the
parameters needed to generate a quantitative estimate of the fraction absorbed. Given this qualitative
indication of rapid systemic uptake and the absence of quantitative inhalation absorption data, 100
percent inhalation absorption is assumed.
In the 2020 RE, an occupational acute HED was derived from the occupational HEC using the same
equation but with an inhalation volume for workers based on higher breathing rates. As described in the
2023 correction memo,13 that derivation was incorrect. Because the occupational HEC was derived
based on a normal general population breathing rate, the HED derivation should apply the same
breathing rate assumptions. This supplement for 1,4-dioxane uses the revised acute occupational HED,
which is equal to the general population HED.
4.2,2 Derivation of Chronic Hazard Values
4.2.2.1 Inhalation HEC
The chronic HECs are based on BMCLio {i.e., the lower confidence limit of the benchmark
concentrations associated with a benchmark response of 10%) for effects in the olfactory epithelium
following inhalation exposures to rats for 6 hours/day, 5 days/week for 2 years. In the 2020 RE, EPA
derived an HEC for chronic worker exposures by applying a duration adjustment for 8 hours/day and a
dosimetric adjustment factor of 1 (the default value using the RGDR approach for systemic effects when
the calculated ratio of animal to human blood:air partition coefficients is greater than 1 (U.S. EPA.
1994b)). The occupational HEC derived in the 2020 RE used default breathing rate assumptions and did
13 Available at https://www.regulations.gov/document/EPA-HO-QPPT-2016-0723-0099.
Page 140 of 570
-------
not adjust for occupational breathing rates. For this supplement, EPA also derived an HEC for
continuous general population exposures by applying an alternate duration adjustment for 24 hours/day,
7 days/week.
4.2.2.2 Oral HEDs
A chronic oral HED was calculated based on a benchmark dose level (BMDL) for liver toxicity
observed following chronic drinking water exposures to male rats and a nearly identical NOAEL value
for liver toxicity in male rats in a similar chronic toxicity study. In the 2020 RE, EPA derived an HED
by multiplying the nearly identical rodent BMDL and NOAEL values by (BWa/BWh)0'25, where BWa is
the bioassay-specific rodent body weight, and BWh is the default human body weight of 70 kg. Because
the chronic HED is based on a daily dose rate (as opposed to an intermittent exposure concentration), it
is equally applicable to both occupational and general population exposures and no additional
conversion is required.
4.2.2.3 Dermal HEDs
In the absence of chronic dermal toxicity studies, chronic dermal HEDs were derived from both the
chronic HEC and from the oral HED using route-to-route extrapolation. In the 2020 RE, the dermal
HED used for occupational risk calculations was extrapolated from the chronic worker HEC. For this
supplement, EPA also derived an HED from the HEC for continuous general population exposure. The
duration-adjusted chronic HEC for general populations was converted to a chronic HED for the general
population using the following equation:
dermal HED (mg/kg-d) = inhalation BMDLhec (mg/m3) x inhalation volume x 100% inhalation
absorption ^ body weight
where the inhalation volume for the general population is 14.7 m3/day ( ) for a 24-hour
general population exposure and the body weight is 80 kg. As described above for the acute hazard
values, EPA assumed 100 percent inhalation absorption. In the 2020 RE, an occupational HED was
derived from the occupational HEC using the same basic equation but with an inhalation volume for
workers based on higher breathing rates. The difference in the HEDs derived from occupational and
general population HECs reflect differences in breathing rate assumptions for the two populations.
In the 2020 RE, an occupational chronic dermal HED was derived from the occupational HEC using the
same equation but with an inhalation volume for workers based on higher breathing rates. As described
in the correction memo, that derivation was incorrect. Because the occupational HEC was derived based
on a normal general population breathing rate, the HED derivation should apply the same breathing rate
assumptions. This assessment uses the revised occupational chronic dermal HED, which is equal to the
general population HED.
4.2.3 Derivation of Cancer Hazard Values
For cancer, the inhalation unit risk (IUR) value was derived using the MS-Combo model to evaluate the
combined cancer risk for multiple tumor sites observed in male rats following inhalation exposure for 6
hours/day, 5 days/week for 2 years. Tumor types included in the MS-Combo model include nasal cavity
squamous cell carcinoma, Zymbal gland adenoma, hepatocellular adenoma or carcinoma, renal cell
carcinoma, peritoneal mesothelioma, mammary gland fibroadenoma, and subcutis fibroma. In the 2020
RE, EPA derived an IUR for chronic worker exposures by applying a dosimetric adjustment factor of 1
and a duration adjustment for 8 hours/day. The occupational IUR derived in the 2020 RE applied default
breathing rate assumptions and did not use adjustments for occupational breathing rates. The
occupational IUR was rounded to 1 x 10~6 (|ig/m3) 1 for application in risk calculations. For this
Page 141 of 570
-------
supplement, EPA used that rounded occupational IUR to derive an IUR for continuous general
population exposures to 1,4-dioxane by applying a duration adjustment for 24 hours/day, 7 days/week.
The oral and dermal cancer slope factor was derived using the Multistage Weibull Model for the liver
tumors in female mice that had been exposed continuously via drinking water. In the 2020 RE, EPA
calculated an HED for each tumor type by multiplying rodent doses by (BWa/BWh)0,25, where BWa is
the bioassay-specific rodent body weight and BWh is the default human body weight of 70 kg. The CSF
was then calculated by dividing the benchmark response rate (0.5) by the HED. This CSF was applied to
both occupational and consumer/general population scenarios using scenario-specific risk benchmarks
and lifetime exposure estimates.
4.3 Strengths, Limitations, Assumptions, and Key Sources of Uncertainty
in the Hazard and Dose-Response Analysis
All assumptions or uncertainties inherent to the human health hazard assessment and dose-response
analysis that were peer-reviewed in the 2020 RE are still applicable for this supplement. As described in
the 2020 RE, EPA has medium confidence in the acute non-cancer PODs and high confidence in the
chronic non-cancer PODs for oral, dermal, and inhalation exposures. EPA has high confidence in the
cancer inhalation unit risk and medium to high confidence in the oral and dermal cancer slope factor.
These conclusions are based on the fact that there is a robust set of high quality chronic and sub-chronic
inhalation and oral exposure studies in rats and mice. The available evidence demonstrates consistent
systemic toxicity and tumor formation in rats exposed via inhalation and in both rats and mice exposed
via drinking water. Key sources of uncertainty include limited data on some sensitive reproductive and
developmental endpoints, reliance on route-to-route extrapolation, uncertainty around the mode of
action for 1,4-dioxane carcinogenicity, and the potential for subpopulations or lifestages with increased
biological susceptibility to 1,4-dioxane. Available methods indicate potential higher inhaled doses in
young children than adults, consistent with 1,4-dioxane specific studies integrating lifestage differences
in ventilation, anatomy and metabolism via CYP2E1 ( ). The preferred method to
quantify these lifestage differences is a 1,4-dioxane specific PBPK model; however, the available PBPK
models for 1,4-dioxane are not adequate and there are not generally accepted default methods not
specific to 1,4-dioxane. Therefore, the air concentration is used as the exposure metric for all lifestages
and the 10x uncertainty factor accounts for these lifestage differences per EPA guidance (
2012. 1994bY
Page 142 of 570
-------
5 HUMAN HEALTH RISK CHARACTERIZATION
1,4-Dioxane - Human Health Risk Characterization (Section 5):
Key Points
EPA estimated cancer and non-cancer risks for each exposure pathway for a range of central tendency and
high-end exposure scenarios. Overall confidence in risk estimates varies across exposure pathways and
COUs, depending on the data and assumptions used to derive exposure and risk estimates. Differences in
estimates between central tendency and high-end exposure scenarios may reflect both variability across the
population and uncertainty in the exposure assessment.
• Cancer and non-cancer risks were evaluated for occupational inhalation and dermal exposures to
1,4-dioxane present as a byproduct.
o Cancer risk estimates for inhalation exposure range from 4.8/ 10 " to 1.9x10 4 for central
tendency exposures and 4.8/ 10 10 to 7.4/ 10 1 for high-end exposures,
o Cancer risk estimates for dermal exposure range from 8.1 x 1CT7 to 7.3 x 10 ^ for central
tendency exposures and from 5.0x10-6 to 2.8xl0~2 for high-end exposures.
• Cancer and non-cancer risks were evaluated for drinking water exposures resulting from releases to
surface water, including facility releases, down-the-drain releases, hydraulic fracturing releases, and
aggregate releases from multiple sources.
o Risk from individual facilities vary substantially within and across COUs, with cancer risk
estimates ranging from 5.4x10 ^" to 0.025.
o Cancer risk estimates from modeled down-the-drain releases are highest in locations where
large populations are contributing to these releases and where they are ultimately discharged
to streams with low flow,
o Cancer risk estimates from modeled hydraulic fracturing waste releases to surface water are
3.9x10-8 for median modeled releases and 1.5xl0~6 for 95th percentile modeled releases,
o Probabilistic modeling provides a distribution of risk estimates reflecting a range of
drinking water scenarios that account for aggregate sources of 1,4-dioxane in water.
• Cancer risks were evaluated for drinking water exposures resulting from releases to land with
potential to reach groundwater.
o Risk estimates from landfill leachate are highest under disposal scenarios resulting in higher
1,4-dioxane concentrations in leachate and higher landfill loading rates,
o Cancer risk estimates for drinking water exposures resulting from hydraulic fracturing waste
released to land/groundwater range from 4.Ox 10 7 for median modeled releases to 8.6x 10 6
for 95th percentile modeled releases.
• Cancer and non-cancer risks were evaluated for general population exposure to 1,4-dioxane in air.
o Cancer risk estimates for industrial air releases reported to TRI were generally highest
within 1,000 m of the facilities and lower at greater distances,
o Cancer risk estimates within 1,000 m of hydraulic fracturing operations range from 0.2x 10~8
to 7.1 x 10~5 for a range of model scenarios across a range of high-end and central tendency
release scenarios.
o Cancer risk estimates within 1,000 m of industrial and institutional laundries range from
1.5 xlO-11 to 3.8xl0~8 across a range of high-end and central tendency air concentrations
modeled for maximum release scenarios.
Page 143 of 570
-------
5.1 Risk Characterization Approach
The exposure scenarios, populations of interest, and toxicological endpoints used for evaluating risks
from acute and chronic exposures are summarized below in Table 5-1. To estimate risks from
occupational and general population exposure scenarios evaluated in this supplement, EPA used the
same methods described in the 2020 RE, as summarized below.
Table 5-1. Use Scenarios, Populations of Interest, and Toxicological Endpoints Used for Acute and
Chronic Exposures
Populations
of Interest
and Exposure
Seenarios
Workers"
Acute - Adolescent (>16 vears old) and adult workers exposed to 1.4-dioxane for a sinele 8-hour exposure
Chronic - Adolescent (>16 vears old) and adult workers exposed to 1.4-dioxane for the entire 8-hour
workday for 260 days per year for 40 working years
General Population Drinking Water Exposures b
Acute - Adults, children, and formula-fed infants exoosed to 1.4-dioxane throueh drinkine water over a 24-
hour period
Chronic - Adults, children, and formula-fed infants exoosed to 1.4-dioxane throueh drinkine water for 33 or
78 years d
General Population Ambient Air Exposurec
Acute - Peoole exoosed to 1.4-dioxane throueh ambient air over a 24-hour ocriod
Chronic - People exposed to 1.4-dioxane through ambient air continuouslv for 33 or 78 vears d
Health
Effects,
Hazard
Values and
Benchmarks
Non-cancer Acute/Short-term Hazard Values
Sensitive acute/short-term health effect: liver toxicity
Acute Uncertainty Factors (Benchmark MOE) = 300 (UFA = 3; UFH = 10; UFL = 10)
• 8-hour HEC (occupational exposure) = 78.7 DDin
• 24-hour HEC (continuous eeneral DODulation exposure) = 26.2 DDin
• Acute Oral and Dermal HED (occupational and general population exposure) = 17.4 mg/kg
Non-cancer Chronic Hazard Values
Sensitive chronic health effects:
• Liver toxicity (oral)
• Effects on the olfactory epithelium due to systemic exposures (inhalation and dermal)
Chronic Uncertainty Factors (Benchmark MOE) = 30 (UFa = 3; UFh = 10)
• HEC (8-hour occupational exposure) = 3.6 ppm
• HEC (continuous exposure general population exposure) = 0.846 ppm
• Oral HED (for both occupational and general population scenarios) = 2.6 mg/kg/day
• Dermal HED (extrapolated from HECs for both occupational and general population scenarios) =
0.56 mg/kg/day
Cancer Hazard Values
Inhalation cancer hazard for 1,4-dioxane is based on combined tumor hazard at multiple sites
• IUR (occupational) = 3.7E-03 per ppm
• IUR (continuous) = 1.6E-02 per ppm
Oral and dermal cancer hazards for 1,4-dioxane are based on liver tumors following oral exposures
• Oral/dermal slope factor = 1.2E-01 (mg/kg/day)-1
MOE = margin of exposure; UFA = Interspecies uncertainty factor for animal-to-human extrapolation; UFH = Intraspecies
uncertainty factor for human variability; UFL = LOAEC-to-NOAEC uncertainty factor for reliance on a LOAEC as the POD
" Adult workers (>16 years old) include both female and male workers. Risks to ONUs were not calculated separately because
exposure data were not available for ONUs for the OESs being evaluated. Risks to ONUs are assumed to be equal to or less
than risks to workers who handle materials containing 1,4-dioxane as part of their job.
b These scenarios are used to evaluate potential risks from 1,4-dioxane in surface water, drinking water sources and
groundwater that may be used as drinking water.
c Inhalation exposures are described in terms of air concentrations and do not include lifestage-specific adjustments; risk
estimates based on air concentrations are intended to address risks to all lifestages (see Section 4.3).
d33 vears is the 95th percentile residential occupancy period inEPA's Exposure Factors Handbook (U.S. EPA, 2011),
Chapter 16, Table 16-5; 78 years is equal to the duration of a full lifetime used in these analyses.
Page 144 of 570
-------
.1.1 Estimation of Non-cancer Risks
EPA used a margin of exposure (MOE) approach to identify potential non-cancer risks. The MOE is the
ratio of the non-cancer POD divided by a human exposure dose. Acute and chronic MOEs for non-
cancer inhalation and dermal risks were calculated using the following equation:
Non — cancer Hazard value (POD)
MOEacute or chronic ~
Human Exposure
Where:
MOE = Margin of exposure (unitless)
Hazard value (POD) = HEC (ppm) or HED (mg/kg-d)
Human Exposure = Exposure estimate (in ppm or mg/kg-d)
MOE risk estimates may be interpreted in relation to benchmark MOEs. Benchmark MOEs are typically
the total UF for each non-cancer POD. The MOE estimate is interpreted as indicating a human health
risk if the MOE estimate is less than the benchmark MOE (i.e., the total UF). On the other hand, if the
MOE estimate is equal to or exceeded the benchmark MOE, risk is not indicated. Typically, the larger
the MOE, the more unlikely it is that a non-cancer adverse effect occurs relative to the benchmark.
When determining whether a chemical substance presents unreasonable risk to human health or the
environment, calculated risk estimates are not "bright-line" indicators of unreasonable risk, and EPA has
discretion to consider other risk-related factors apart from risks identified in risk characterization.
5.1.2 Estimation of Cancer Risks
Extra cancer risks for repeated exposures to a chemical were estimated using the following equations:
Inhalation Cancer Risk = Human Exposure x IUR
or
Dermal/Oral Cancer Risk = Human Exposure x CSF
Where:
Risk = Extra cancer risk (unitless)
Human exposure = Exposure estimate (LADC in ppm)
IUR = Inhalation unit risk
CSF = Cancer slope factor
Estimates of extra cancer risks are interpreted as the incremental probability of an individual developing
cancer over a lifetime following exposure (i.e., incremental, or extra individual lifetime cancer risk).
5.2 Human Health Risk Characterization
5.2,1 Summary of Risk Estimates for Occupational Exposures
EPA estimated cancer and non-cancer risks for workers exposed to 1,4-dioxane based on the
occupational exposure estimates that were described in Section 3.1. Risks to ONUs were not calculated
separately because exposure data were not available for ONUs for the OESs being evaluated. Risks to
ONUs are assumed to be equal to or less than risks to workers who handle materials containing 1,4-
dioxane as part of their job.
Page 145 of 570
-------
Below are summaries of the cancer risk estimates for the inhalation and dermal exposures as well as key
sources of uncertainty for all occupational exposure scenarios assessed in this supplement. These risk
estimates are based on exposures to workers in the absence of PPE such as gloves or respirators. Section
3.1.2.4 contains an overall discussion on strengths, limitations, assumptions, and key sources of
uncertainty for the occupational exposure assessment. Additionally, Appendix F contains a
comprehensive weight of scientific evidence summary table which presents an OES-by-OES discussion
of the key factors that contributed to each weight of scientific evidence conclusion. Results for the risk
calculations and occupational OES/COUs from the current analysis as well as those previously
presented in the 2020 RE are available in 1,4-Dioxane Supplemental Information File: Occupational
Exposure and Risk Estimates ( Mu).
Risk estimates vary across OES/COUs. Because cancer risk is the primary risk driver in most exposure
scenarios, this summary of results focuses on cancer risk estimates. For 7 of the 10 COU subcategories
evaluated, high-end cancer risk estimates were above 1 in 10,000. For many of those COUs, acute
and/or chronic non-cancer risk estimates were below the corresponding benchmark MOEs, indicating
that non-cancer risks may also be a concern. Cancer risk estimates for inhalation exposure range from
4.8xl0~u to 1.9xl0~4 for central tendency exposures and 4.8xlO~10 to 7.4><10~3 for high-end exposures.
Cancer risk estimates for dermal exposure range from 8.1 x 10~7 to 7.3x10 3 for central tendency
exposures and from 5.Ox 10~6 to 2.8x ] 0 2 for high-end exposures. Risks are highest for PET
manufacture, hydraulic fracturing operations, ethoxylation processes, and textile dyes. For these OESs,
cancer risk estimates were greater than 1 in 10,000 for both central tendency and high-end exposures.
For these OESs, the key uncertainties include limited exposure monitoring data, age of data,
representativeness of key modeling parameters, and the extent to which the data collected under past
practices and operations are representative of modern practice and operations.
Overall confidence in risk estimates for occupational inhalation exposures ranges from low to high,
depending on the confidence in exposure assessment for each OES/COU. As described in Section 4.3,
overall confidence in the cancer inhalation unit risk underlying these risk estimates is high. As described
in Section 3.3.1.1, the measured and monitored inhalation exposure data are supported by moderate to
robust evidence. Additionally, the exposure modeling methodologies and underlying model input data is
supported by moderate to robust evidence. However, there is uncertainty in the representativeness of the
assessed exposure scenarios towards all potential exposures for the given OES/COU, limitations in the
amount and age of monitoring data, and limitations in the modeling approaches towards 1,4-dioxane-
specific use within the OES/COU. Therefore, while the underlying data and methods used to estimate
occupational inhalation risk is supported by moderate to robust evidence, the overall confidence of these
estimates ranges from low to high depending on the OES/COU. Key exposure considerations along with
the corresponding risk estimates are below.
• Industrial/Commercial Use of Textile Dye. Risk estimates were derived using personal
breathing zone and area monitoring data collected from 1991 to 2010 at four facilities linked to
the use of textile dyes. Cancer risk estimates for inhalation exposure range from 1,9x 10~4 for
central tendency exposures to 7.4x10 3 for high-end exposures. However, there is uncertainty in
the risk estimates. The monitoring data used in this analysis are limited {i.e., 14 samples from
four sites). It also is not known how manufacturing processes and workplace conditions have
changed since the 1990s, when approximately half of the data was collected. For instance, EPA
does not have information available about the actual activities of the sampled workers and the
Page 146 of 570
-------
representativeness of the facility engineering controls to the modern practice. EPA identified this
uncertainty and solicited public comment but did not receive further information on this COU.
Industrial/Commercial Use of Antifreeze. Risk estimates were derived from occupational
exposures modeled using Monte Carlo simulations for the worker activity of container
unloading. Cancer risk estimates for inhalation exposure range from 4.8x 1CT11 for central
tendency exposures to 4,8/10 10 for high-end exposures. However, there is uncertainty in the
risk estimates. Specifically, there is uncertainty as to the representativeness of some of the model
input data and, therefore, subsequent calculated exposures to the actual distribution of antifreeze
occupational exposures. This is due to limitations of using generic industry values identified for
the automotive industry. Also contributing to the uncertainty is that EPA used use rates from the
consumer exposure model for commercial/industrial use in the Monte Carlo modeling.
Industrial/Commercial Use of Surface Cleaner. Risk estimates were derived using 49 personal
breathing zone samples taken in 2019 during the use of surface cleaners in domestic kitchens and
bathrooms. Cancer risk estimates for inhalation exposure range from 2.8x 10~7 for central
tendency exposures to 3,7/10 6 for high-end exposures. However, there is uncertainty in the risk
estimates. Specifically, the monitoring data summary did not provide discrete monitoring points
and only provided summary statistics such as the geometric mean and maximum. Therefore,
EPA could not calculate the 50th and 95th percentile exposures. Also, it is uncertain the extent to
which the cleaning activities captured in this study reflect all occupational surface cleaning
scenarios, as they were measured in a consumer setting.
Industrial/Commercial Use of Dish Soap and Dishwasher Detergent. Risk estimates were
derived from occupational exposures modeled using Monte Carlo simulations for the worker
activities of container unloading and cleaning dishes. Cancer risk estimates for inhalation
exposure range from 4.4x 10~7 for central tendency exposures to 5.1 / 1 0 6 for high-end
exposures. However, there is uncertainty in these risk estimates. Due to a lack of data specific to
1,4-dioxane for this use, EPA used industry-specific data from a public comment along with
standard default values from sources like the ChemSTEER User Guide for the model input
parameters. In addition, the use rate of dish soaps in the model is based on values from the
Consumer Exposure Model which were adjusted for commercial use. This approach adds
uncertainty to the assessment.
Industrial/Commercial Use of Laundry Detergent. Risk estimates were derived from
occupational exposures modeled using Monte Carlo simulations for the worker activity of
unloading detergent into machines, container cleaning, and laundry operations. For industrial
laundries, cancer risk estimates for vapor inhalation exposure range from 3,3/10 7 for central
tendency exposures to 1.0/10 5 for high-end exposures. For institutional laundries, cancer risk
estimates for vapor inhalation exposure range from 2.5x 10~7 for central tendency exposures to
7,9x 10 6 for high-end exposures. In both cases, cancer risk estimates for total particulates
inhalation range from 2.2x 10~8 for central tendency exposures to 7.Ox 10 7 for high-end
exposures. Cancer risk estimates for respirable particulates inhalation range from 5.5x 10~9 for
central tendency exposures to 2,Ox 10 7 for high-end exposures. However, there is uncertainty in
the risk estimates. Specifically, there is uncertainty as to the representativeness of some of the
model inputs and, therefore, subsequent calculated exposures to the actual distribution of laundry
detergent occupational exposures. This is due to limitations of using generic industry values
identified for institutional and industrial laundries.
Industrial/Commercial Use of Paint and Floor Lacquer. Risk estimates were derived using 17
personal breathing zone samples collected by NIOSH in 1987 at a military vehicle painting site.
Cancer risk estimates for inhalation exposure range from 8.Ox 10~5 for central tendency exposures
Page 147 of 570
-------
to 5.9x 1CT4 for high-end exposures. However, there is uncertainty in the risk estimates. The
monitoring data used in this analysis are limited {i.e., 17 samples taken at one site). It also is not
known how processes and workplace conditions have changed since 1987. For instance, EPA
does not have information available about the actual activities of the sampled workers and the
representativeness of the facility engineering controls to modern practice. EPA identified this
uncertainty and solicited public comment, but did not receive further information on this COU.
• Polyethylene Terephthalate (PET) Manufacturing. Risk estimates were derived using
personal breathing zone monitoring data collected from 1985 to 1994 at five facilities linked to
PET manufacturing, and personal breathing zone and area monitoring data from two public
comments (collected 1998-2023, and 2019). Cancer risk estimates for inhalation exposure range
from 2,8/10 4 for central tendency exposures to 2,9/ 10 3 for high-end exposures. However,
there is uncertainty in the risk estimates with respect to the 1994 data since it is unknown how
manufacturing processes and workplace conditions have changed. In addition, there is
uncertainty in the representativeness of the monitoring data for all sites and worker activities in
this OES.
• Ethoxylation Processes. Risk estimates were derived using eight personal breathing zone data
points from a public comment for the worker activities of unloading and laboratory activities. In
addition, one composite 8-hour time-weighted average personal breathing zone sample was
collected from one worker in 2000 at a soap and detergent manufacturing facility. Cancer risk
estimates for inhalation exposure range from 2.1 x 10~4 for central tendency exposures to 5,4/ 10 4
for high-end exposures. However, there is uncertainty in the risk estimates. There is uncertainty
as to the worker activities covered by this monitoring data and whether all foreseeable activities,
corresponding exposures, and workplace operations are represented.
• Hydraulic Fracturing Operations. Risk estimates were derived from occupational exposures
modeled using Monte Carlo simulations for the worker activities of container unloading,
container cleaning, and equipment cleaning. Cancer risk estimates for inhalation exposure range
from 2,2/ 10 6 for central tendency exposures to 2,5/10 4 for high-end exposures. However,
there is uncertainty in the risk estimates. Specifically, there is uncertainty as to the
representativeness of some of the model input data and, therefore, the subsequent calculated
exposures to the actual distribution of hydraulic fracturing occupational exposures. This is due to
limitations of using generic industry values identified for the hydraulic fracturing industry as
well as self-reported values from FracFocus as model parameters. FracFocus data may not fully
represent operations across multiple sites throughout the United States as only certain sites
volunteered to submit data.
Overall confidence in risk estimates for occupational dermal exposures is medium for all OES/COUs
because the same modeling approach was used for all OES/COUs. As described in Section 4.3 overall
confidence in the oral and dermal cancer slope factor underlying these risk estimates is medium to high.
As described in Section 3.3.1.2, the dermal exposure modeling methodology is supported by moderate
evidence, with model input parameters from literature sources, a European model, standard defaults
from the ChemSTEER User Guide ( 1015a). and 1,4-dioxane product concentration data from
process information. These sources range from slight to robust, depending on factors such as age and
applicability to OES/COU. The modeling is limited by the use of standard input parameters that are not
specific to 1,4-dioxane and a lack of variability in dermal exposure for different worker activities.
Differences in the dermal exposure modeling across COUs are driven primarily by COU-specific weight
fractions of 1,4 dioxane and the independent assessment of evaporative impacts in commercial and
industrial settings. Therefore, EPA's overall confidence in the occupational dermal risk estimates is
medium.
Page 148 of 570
-------
5.2,2 Summary of Risk Estimates for the General Population
5.2.2.1 Drinking Water - Surface Water Pathway
Risks from drinking water exposures were evaluated using a series of analyses that provide information
about the specific contributions of releases associated with individual OESs as well as information about
aggregate exposures and risks that could result from multiple sources releasing to the same water body.
Because most reasonably available surface water and drinking water monitoring data are not co-located
with 1,4-dioxane release sites, this analysis relies primarily on drinking water concentrations modeled
based on reasonably available release information. Risks predicted based on reasonably available
monitoring data are presented in 5.2.2.1.1.
EPA estimated cancer and non-cancer risks for adults, children, and formula-fed infants exposed to 1,4-
dioxane in drinking water. All risk estimates presented in this summary focus on the scenario with the
greatest potential exposure and risk. Because adult drinking water exposures relative to body weight are
greater than exposures relative to body weight averaged over the course of childhood (as illustrated in
1,4-Dioxane Supplemental Information File: Drinking Water Exposure and Risk Estimates for 1,4-
Dioxane Release to Surface Water from Individual Facilities ( 2024h)) the 33 - year drinking
water exposure scenario with the greatest lifetime cancer risk is 33 years of drinking water exposure as
an adult. Similarly, because cancer risk is the primary risk driver in most exposure scenarios, this
summary of results focuses on cancer risk estimates. More comprehensive sets of risk estimates for non-
cancer effects and other exposure scenarios are presented in the supplemental files referenced
throughout this section.
While most cancer risk estimates summarized in this section are based on exposures resulting from 33-
year exposure durations and mean drinking water ingestion rates, longer exposure durations or higher
drinking water ingestion rates would result in greater exposure and risk. Individuals exposed over a full
lifetime (78 years) could have exposure and risk approximately 2.3 times greater than those calculated
for 33 years of exposure. As some people may live in a community near releases for longer durations,
EPA agrees with the SACC recommendation to utilize a full lifetime of exposure for assessing lifetime
cancer risks for fenceline communities. Lifetime cancer risk estimates based on 95th percentile drinking
water ingestion rates could result in 3 to 4 times higher exposures and risks than those based on mean
ingestion rates, depending on the age groups exposed (described in Appendix I). Although consideration
of alternate exposure factors such as lifetime and ingestion rates result in increased risks of less than an
order of magnitude, where the original estimates are close to the applicable benchmark, this could result
in changes to overall risk conclusions.
Drinking water exposure and risk estimates are highly dependent on the amount of 1,4-dioxane released
and the flow of the receiving water body. Both of these factors vary substantially across facilities within
each COU/OES, making release amount and flow much more important predictors of risk than a
facility's identified COU/OES. Exposure and risk estimates are also influenced by whether there is a
drinking water intake downstream of a release and the degree of dilution that occurs between the point
of release and the drinking water intake. Many of the risk estimates presented in the sections that follow
(for facility-specific releases, DTD, hydraulic fracturing, and aggregate modeling) assume that no
additional downstream dilution occurs prior to reaching drinking water intakes. This represents an upper
end estimate of exposure and risk based on the available data and the potential for intakes to be directly
downstream of a releasing facility. EPA conducted further analysis of the facility-specific releases to
consider the potential impact of downstream dilution on actual concentrations at drinking water intakes
and resulting risk estimates. Even when accounting for dilution between known releases and identified
drinking water intake locations, water concentrations estimated at drinking water intakes, instances of
Page 149 of 570
-------
cancer risks greater than 1 in 1 million for some public water systems are identified. Proximity of
releases to drinking water intakes and dilution are further discussed in Section 2.3.1.2.4/Appendix G.2.4
and Section 5.2.2.1.2.
1,4-Dioxane is not readily removed through typical wastewater or drinking water treatment processes.
Therefore, the drinking water risk estimates presented below are derived based on the assumptions that
drinking water intakes are located near 1,4-dioxane release sites and that no 1,4-dioxane is removed by
POTWs or through drinking water treatment. Use of source water estimated concentrations of 1,4-
dioxane to calculate cancer risk estimates is considered protective of all systems. These assumptions are
further discussed in Section 2.3.1.1/Appendix G.1.2).
5.2.2.1.1 Risks from Exposure to Drinking Water Concentrations Indicated in
Finished Drinking Water Monitoring Data
EPA evaluated risks for 1,4-dioxane concentrations reported in the reasonably available finished
(treated) drinking water monitoring data. Monitoring data included in this analysis were from
generalized, broad monitoring strategies, rather than targeted efforts to assess areas of known
contamination. As previously illustrated in Figure 2-10, 1,4-dioxane was below limits of detection for 89
percent of finished drinking water samples included in UCMR3 and state databases. Table 5-2
summarizes the distribution of lifetime cancer risk estimates from 1,4-dioxane concentrations detected
in finished drinking water reported in these databases (described in Section 2.3.1.1). This drinking water
monitoring data provides evidence that 1,4-dioxane is present in some finished drinking water and may
contribute to cancer risks in locations at the high-end of monitored drinking water concentrations.
Monitoring data may not include the full range of 1,4-dioxane concentrations that result from industrial
releases. As discussed in Section 2.3.1.1, available drinking water monitoring data do not necessarily
capture locations that are most impacted by releases temporally or spatially and they often reflect
concentrations at a single point in time rather than average concentrations. However, as described in
Appendix G.2.3.2, in locations where monitoring data are available near release sites, comparisons
demonstrate strong consistency between modeled concentrations and monitoring data. EPA's evaluation
of drinking water risks therefore primarily relied on modeled estimates of 1,4-dioxane concentrations
that occur near release sites.
Page 150 of 570
-------
Table 5-2. Lifetime Cancer Risk Estimates for 1,4-Dioxane Concentrations Detected in Finished
Drinking Water
Percentile Drinking Water Monitoring Data
Min
5%
25%
Median
75%
90%
95%
Max
Water conc.
(Mg/L)
2.00E-03
3.50E-02
3.50E-02
3.50E-02
3.50E-02
7.93E-02
0.16
13.3
Lifetime
Cancer Risk
4.02E-11
1.95E-08
1.95E-08
1.95E-08
8.37E-08
2.79E-07
3.46E-07
7.42E-06
Lifetime cancer risk estimates are based on mean drinking water ingestion rates over 33 years of oral exposure
through drinking water as an adult. Lifetime cancer risk estimates for a full 78 years of exposure would be 2.26 times
greater than the risk estimates presented here. Similarly, lifetime cancer risk estimates based on 95th percentile
drinking water ingestion rates would be approximately 3^ times greater, depending on the age groups exposed.
Percentiles reflect concentrations across the distribution of available drinking water monitoring data (this distribution
includes non-detects as half the detection limit).
5.2.2.1.2 Risks from Exposures to Water Concentrations Modeled from Industrial
Releases
To estimate the contribution of industrial releases to general population risks from drinking water, EPA
calculated cancer and non-cancer risk estimates based on modeled surface water concentrations in
receiving water bodies described in Section 2.3.1.3.1 and the resulting drinking water exposures
calculated as described in Section 3.2.2. Because there is substantial variation and uncertainty around
the extent of dilution that may occur in the receiving water body between the point of release and the
locations of drinking water intakes, EPA calculated cancer risk estimates under a range of reasonable
downstream dilution assumptions.
Figure 5-1 shows the distribution of cancer risk estimates for industrial releases reported to TRI and
DMR, assuming that concentrations at drinking water intakes are the same as concentrations that occur
at the point of release after initial mixing in the receiving water body. Based on available data, this is a
plausible scenario in some locations. Lifetime cancer risk estimates are based on median drinking water
ingestion rates over 33 years of exposure as an adult and range from 5.41 x 1CT13 to 2,54/ 10 2, The
median cancer risk estimate for these modeled concentrations is 2.32x 1CT6 and the 95th percentile risk
estimate is 4.92x 1CT3. Lifetime cancer risk estimates for a full 78 years of exposure would be 2.26 times
greater than the risk estimates based on 33 years. Similarly, lifetime cancer risk estimates based on 95th
percentile drinking water ingestion rates would be approximately 3-4 times greater, depending on the
age groups exposed. Acute and chronic non-cancer risk estimates for some facilities (data not shown)
also indicate potential for non-cancer risk relative to benchmark MOEs. This analysis represents an
upper bound drinking water exposure scenario in which intakes are located near the point of release or in
which minimal additional dilution occurs downstream. Complete cancer and non-cancer risk estimates
for facility and OES-specific releases are presented in 1,4-Dioxane Supplemental Information File:
Drinking Water Exposure and Risk Estimates for 1,4-Dioxane Release to Surface Water from Individual
Facilities (U.S. EPA. 2024hY
Page 151 of 570
-------
20-
16-
0 12-
CH CO
M— y
° «
C
0
=3
cr
a>
4-
0-
10"
10"
10"
10"
10"
10"° 10"' 10"° 10"'
Adult Lifetime Cancer Risk
10"'
10"
10"
10"1
Figure 5-1. Distribution of Adult Lifetime Cancer Risk across all
Facilities, Assuming No Additional Dilution Occurs between the Point of
Release and the Location of Drinking Water Intakes
There is substantial variation in cancer risk estimates both within and across OESs. The large ranges of
modeled water concentrations and corresponding risk estimates reflect the large differences in the
amount of 1,4-dioxane released from facilities, the magnitude of flow within the receiving water body or
both.
For facilities where specific release amounts or locations are not reported, release amounts and flow
rates are based on conservative assumptions that may result in high risk estimates. There is uncertainty
around risk estimates for those facilities with limited release information, but facility-specific
information on release amounts and locations was available for most facilities. Therefore, while facility-
specific risk estimates based on facilities with limited information should be interpreted with caution,
most estimates are informed by moderate to robust modeling approaches and input data. To determine
the extent to which inclusion of facilities with limited release information influences the overall
distribution, EPA repeated this risk estimate analysis presented in Figure 5-1 using only facilities for
which high quality release data are available (Figure 5-2). Specifically, this additional analysis is limited
to facilities for which the annual release amount was sourced from either TRI Form R or DMR, and the
receiving water body reach code was identified in the facility's NPDES permit. Out of the 120 total
direct and indirect releases evaluated in this section, 80 met these strict data criteria. The resulting
distribution of risk estimates are similar to the results of the analysis including all facilities, ranging
from 5.41 xl0~13 to 2.54xl0~2, with a median of 8.51xl0~7 and 95th percentile of 4.92><10~3.
Page 152 of 570
-------
12-
LO
O)
0_
(f)
Cl)
(/)
CD
_0
0 .E
a: cq
o w
c
0
D
cr
0
4-
0-
10"
10"
10
10"
10"° 10"' 10° 10"
Adult Lifetime Cancer Risk
10 '
10"
10
10
Figure 5-2. Distribution of Adult Lifetime Cancer Risk across Facilities with
High Quality Release Data, Assuming No Additional Dilution Occurs between
the Point of Release and the Location of Drinking Water Intakes
The risk estimates summarized in Figure 5-1 and Figure 5-2 rely on the assumption that concentrations
at drinking water intakes are the same as concentrations estimated near the point of release. To evaluate
the validity of that assumption, EPA considered the proximity of release sites to downstream drinking
water intake locations for community and non-community non-transient PWSs. As shown in Table 5-3,
of the 69 facilities with cancer risk greater than 1 x 1CT6, 22 (32%) have a downstream drinking water
intake within 250 km and 4 of those have a drinking water intake within 10 km. A detailed description
of this analysis is provided in Appendix G.2.4.
Table 5-3. Proximity of Nearest Downstream Drinking Water Intakes to Facilities Resulting in
Cancer Risk Greater than lxlO-6
Total Facilities
Evaluated
Facilities with Cancer
Risk > 1E-06
DWI within
250 km
DWI within
100 km
DWI within
50 km
DWI within
25 km
DWI within
10 km
120
69
22
17
11
7
4
The portion of 1,4-dioxane that remains after the additional dilution that occurs as it travels downstream
is highly variable based on site-specific characteristics, ranging from less than 1 percent to nearly 100
percent of the original concentrations (Figure 5-3). The site-specific factors that influence this additional
downstream dilution may not be fully captured in a national-scale assessment. Based on available site-
specific information for each facility, the mean modeled dilution predicted at downstream drinking
water intakes is diluted to 1 percent of original concentrations estimated in receiving water bodies near
the point of release.
Page 153 of 570
-------
I ^<9
LO LJ ^
CM ^ 0 40
•E | g
(/J o .<Ł
d) (0 fV
1 0
3 CO
o Ł CO 20
i_ c50
CD 0 i_
-Q Ł= 0
E <2 TO
Q
o-
1 1 1 1 1—
0 25 50 75 100
Diluted Concentration at Intake as a Percent
of the Concentration at the Point of Release
Figure 5-3. Distribution of Dilution of 1,4-Dioxane Concentrations at Downstream
Drinking Water Intakes
Figure 5-4 shows the distribution of cancer risk estimates for industrial releases, assuming that
concentrations at drinking water intake locations are diluted to 1 percent of the original 1,4-dioxane
concentrations in surface water estimated at the point of release. Lifetime cancer risk estimates for these
modeled concentrations range from 5.41 x 1CT15 to 2.54x 1CT4 The median cancer risk estimate is
8.51>
-------
intakes located downstream of releasing facilities, based on location-specific estimates of dilution. For
the 22 facilities with cancer risk greater than 1 in 1 million and drinking water intakes located within
250 km downstream, EPA identified downstream intakes associated with 73 distinct PWSs.
Even when accounting for site-specific influences on dilution, EPA modeled concentrations that would
result in adult lifetime cancer risk in excess of 1 in 1 million at intakes for 20 of the PWSs identified
through this assessment, serving a combined population of 2,124,000 people. Adult lifetime cancer risk
estimates were greater than 1 in 100,000 for 5 of these public water systems, serving a combined
population of 834,000 people. This analysis also identified locations with multiple releasing facilities
upstream of the same drinking water intake; however, in all such cases the aggregated adult lifetime
cancer risk calculated at the intake for the aggregated diluted concentration was less than 1 in a million.
A detailed description of this analysis is provided in Appendix G.2.4. Overall confidence in these
dilution-adjusted risk estimates is high for drinking water intakes located at or near the point of release,
but confidence decreases substantially with increasing distance downstream. This analysis does not
provide a comprehensive survey of modeled 1,4-dioxane concentrations at all drinking water intakes.
There may be additional drinking water intakes downstream of facilities releasing 1,4-dioxane that are
not accounted for in the intake database used in this analysis.
Overall, these analyses indicate that in many locations, downstream dilution may be expected to
substantially reduce 1,4-dioxane concentrations at the point of drinking water intakes. However, even
when accounting for dilution, upstream industrial releases reported to TRI or DMR contribute to cancer
risk estimates greater than 1 in a million or 1 in 100,000 at known drinking water intake locations.
The set of distributions presented in Figure 5-1, Figure 5-2, and Figure 5-4 indicate that high risks can
occur in specific locations downstream of release sites due to factors such as the size of the releasing
event(s), stream flow volume, proximity of the release site to drinking water intake, and limited drinking
water treatment removal from typical treatment methods.
Overall confidence in the overall distribution of risk estimates for drinking water exposures resulting
from facility releases is medium to high. Overall confidence in site-specific risk estimates for individual
facility releases varies both within and across OES, depending on the confidence in the source-specific
release data. As described in Section 4.3 overall confidence in the oral and dermal cancer slope factor
underlying these risk estimates is medium to high. As described in Section 3.3.2.1, the overall exposure
modeling methodology used for this analysis is supported by moderate evidence. It is designed to
estimate water concentrations expected at specific locations. Exposure estimates for this scenario are
based on some conservative assumptions about flow rates and release frequency and amount. For most
COUs, this analysis is limited to facilities that report via TRI and/or DMR. Other sources releasing
smaller amounts of 1,4-dioxane are not directly captured. Available monitoring data confirm that 1,4-
dioxane is present in some surface water and drinking water, though most of the available data were not
collected near release sites are therefore not directly comparable.
The overall level of confidence in facility-specific release estimates and resulting risk estimates depends
on the source of the release data described in Appendix E.3:
• Overall confidence in drinking water exposure estimates is medium to high for OESs/COUs that
rely primarily on release data reported to DMR or to TRI via Form R. Most COUs/OESs are
included in this group.
• Overall confidence in drinking water exposure estimates is medium for OESs/COUs for which
release estimates are based on reporting to TRI via Form A. The Import and repackaging OES
Page 155 of 570
-------
releases used in this analysis are entirely based on Form A reporting of releases, and just under
half of the Industrial Uses OES releases were reported via Form A.
Although confidence in facility-specific risk estimates varies, estimates for most facilities are informed
by moderate to robust modeling approaches and input data. Furthermore, the overall distribution is not
meaningfully altered by exclusion of facility-specific data based on more limited release information (as
illustrated in Figure 5-1 and Figure 5-2). There is some uncertainty around the proximity of releases to
drinking water intake locations and the extent to which 1,4-dioxane is further diluted prior to reaching
intake locations. EPA therefore estimated distributions of cancer risk estimates under a range of
assumptions about downstream dilution, reflecting the range of plausible drinking water intake
scenarios, as indicated by available site-specific information.
5.2.2.1.3 Risks from Exposures to Water Concentrations Modeled from DTD
Releases (from POTWs), Assuming No Downstream Dilution
EPA evaluated the potential contribution of DTD releases of consumer and commercial products to
drinking water exposure and risk. Surface water concentrations at the point of DTD releases via POTWs
are primarily determined by the size of the population contributing to DTD releases and the flow rates of
receiving water bodies. Risk estimates presented in this section are not tied to known releases at specific
locations. Rather, this analysis defines the conditions under which DTD releases would result in varying
levels of risk. Further information on the specific COUs contributing to DTD releases and the
contributions of each are presented in Appendix G.2.3.4 and Table Apx G-4.
Cancer risk estimates shown in Table 5-4 were calculated based on drinking water exposure estimates
presented in Section 3.2.2.1.2, which correspond to surface water concentrations estimated by
probabilistic modeling of DTD releases under varying population sizes and stream flows. The resulting
risk estimates indicate that risk is highest in locations where large populations are contributing to DTD
releases and those releases are ultimately discharged to streams with low flow. Cancer risk estimates
greater than 1 in a million were seen in combinations of population size and receiving waterbody flow
rates that can be found across the country. Areas with drier climates may be more likely to have
intermittent streams and generally have greater likelihood for elevated environmental concentrations of
1,4-dioxane resulting from DTD loading via POTWs. However, review of a limited dataset of POTW
data demonstrated that the conditions in Table 5-4 resulting in higher levels of risk do occur on a site-
specific basis throughout the country, regardless of climate.
Page 156 of 570
-------
Table 5-4. Lifetime Cancer Risk" Estimates from DTD Releases Alone (at the Point of Release)
under a Range of Population and Flow Rate Scenarios
Population Contributing to Down-the-Drain Releases
100
1,000
10,000
100,000
1,000,000
Annual
Average
Stream Flow
(cfs)
100
6.11E-09
6.11E-08
6.11E-07
6.1 IE 06
6.11E-05
300
2.04E-09
2.04E-08
2.04E-07
2.04E-06
2.04E-05
1,000
6.11E-10
6.11E-09
6.11E-08
6.11E-07
6.11E-06
3,000
2.04E-10
2.04E-09
2.04E-08
2.04E-07
2.04E-06
10,000
6.11 E— 11
6.11E-10
6.11E-09
6.11E-08
6.11E-07
30,000
2.04E-11
2.04E-10
2.04E-09
2.04E-08
2.04E-07
"Lifetime cancer risk estimates are based on mean drinking water ingestion rates over 33 years of oral exposure
through drinking water as an adult. Lifetime cancer risk estimates for a full 78 years of exposure would be 2.26
times greater than the risk estimates presented here. Similarly, lifetime cancer risk estimates based on 95th
percentile drinking water ingestion rates would be approximately 3-4 times greater, depending on the age
groups exposed.
The frequencies of each of these combinations of population size and flow rate are presented in Table 2-11.
As described in Section 2.3.1.3.2, EPA considered the frequency of the varying combinations of
population sizes and flow rates. For communities with single POTWs treating wastewater, most fell into
the range of 100 to 10,000 people, with the annual average flow of the receiving water body less than
300 cfs (Table 2-11). Cancer risk estimates for communities in this range of population sizes are as low
as 2.04x 10~8 at flows of 300 cfs and increase at lower flows. For example, cancer risk estimates for 33
years of exposure resulting from releases from a population size of 10,000 could be as high as 2,04/ 10 6
at a flow of 30 cfs. Acute and chronic non-cancer risk estimates for these scenarios do not indicate non-
cancer risk relative to benchmark MOEs. For reference, stream flows of 100 cfs might be considered a
small river, while anything less than 100 cfs would be considered a stream or creek. Complete cancer
and non-cancer risk estimates for the range of water concentrations from DTD releases estimated under
varying conditions using probabilistic modeling are presented in 1,4-Dioxane Supplemental Information
File: Drinking Water Exposure and Risk Estimates for 1,4-Dioxane Surface Water Concentrations
Predicted with Probabilistic Modeling ( |i).
Lifetime cancer risk estimates in Table 5-4 are based on mean drinking water ingestion rates over 33
years of oral exposure through drinking water as an adult. In response to SACC recommendations, EPA
considered the impacts of a full lifetime of exposure to 1,4-dioxane in drinking water. Lifetime cancer
risk estimates for a full 78 years of exposure are 2.26 times greater than the risk estimates presented in
the table, resulting in risk estimates as high as 1.4 /10 4 for the combinations of population size and
stream flow considered. As some people may live in a community near releases for longer durations,
EPA agrees with the SACC recommendation to utilize a full lifetime of exposure for assessing lifetime
cancer risks for fenceline communities. Similarly, lifetime cancer risk estimates based on 95th percentile
drinking water ingestion rates would be approximately 3 to 4 times greater, depending on the age groups
exposed (95th percentile ingestion rates averaged across all ages are 3.7 times greater than mean
ingestion rates), resulting in risk estimates as high as 2.3 x 1CT4 for the combinations of population size
and stream flow considered. Although consideration of alternate exposure factors such as lifetime and
ingestion rates result in increased risks of less than an order of magnitude, where the original estimates
are close to the applicable benchmark, this could result in changes to overall risk conclusions.
Page 157 of 570
-------
Overall confidence in risk estimates for drinking water exposures resulting from DTD releases under
varying population and stream flow conditions is medium. As described in Section 4.3 overall
confidence in the oral and dermal cancer slope factor underlying these risk estimates is medium to high.
As described in Section 3.3.2.1, the exposure modeling methodology used for this analysis is supported
by robust evidence and is designed to provide a nationally representative distribution of estimated water
concentrations under varying conditions. Exposure estimates rely on estimated distributions of DTD
releases of consumer and commercial products for each COU. Distributions of DTD releases of
consumer and commercial products were estimated for each COU on a per capita basis using the
SHEDS-HT model. Because this analysis is not tied to specific sites, there is uncertainty around the
proximity of releases to drinking water intake locations and the extent to which 1,4-dioxane is further
diluted prior to reaching intake locations. For this analysis, EPA assumed that no additional dilution
occurs prior to reaching drinking water intakes. Although confidence in the individual contribution from
some specific COUs is lower, confidence in estimates of overall DTD releases is moderate.
5.2.2.1.4 Risks from Exposure to Drinking Water Concentrations Modeled from
Disposal of Hydraulic Fracturing Produced Waters to Surface Water,
Assuming No Downstream Dilution
EPA evaluated the potential contribution of the disposal of hydraulic fracturing produced waters to
surface water by aggregating exposures and risks. The range of water concentrations that may result
from releases of hydraulic fracturing waste to surface water were estimated using probabilistic
modeling. Risk estimates presented in this section are not tied to known releases at specific locations.
Rather, this analysis defines the conditions under which releases from hydraulic fracturing would result
in varying levels of risk. These risk estimates are based on the assumption that 1,4-dioxane is not
removed by POTWs or through drinking water treatment.
Cancer risk estimates across the full distribution of modeled releases are presented in Table 5-5. Cancer
risk estimates based on median drinking water ingestion rates over 33 years of exposure are 3,85/10 8
for median modeled releases and 1,52x 1CT6 for 95th percentile modeled releases. Lifetime cancer risk
estimates for a full 78 years of exposure would be 2.26 times greater than the risk estimates based on 33
years. However, it is unlikely that there will be exposures that result in the 95th percentile lifetime
cancer risks, whether based on 33 years or a full lifetime. While hydraulic fracturing produced water
continues to be returned throughout the life of the well, the percentage of produced water drops off after
the first few weeks or months and is replaced by produced oil or gas and it is not known how much and
for how long these wells will ultimately produce ( a). Similarly, lifetime cancer risk
estimates based on 95th percentile drinking water ingestion rates would be approximately 3 to 4 times
greater, depending on the age groups exposed. Acute and chronic non-cancer risk estimates based on
95th percentile modeled releases do not indicate risk relative to benchmark MOEs. The maximum water
concentration estimated by the model reflects a scenario in which waste is released to a stream with very
low flow. EPA does not have site-specific information to indicate that such a scenario combining a high
release with a low flow actually occurs. Complete risk estimates for the range of water concentrations
from DTD releases estimated under varying conditions using probabilistic modeling are presented in
1,4-Dioxane Supplemental Information File: Drinking Water Exposure and Risk Estimates for 1,4-
Dioxane Surface Water Concentrations Predicted with Probabilistic Modeling ( 2024i).
Overall confidence in risk estimates for drinking water exposures resulting from hydraulic fracturing
releases is medium. As described in Section 4.3 overall confidence in the oral and dermal cancer slope
factor underlying these risk estimates is medium to high. As described in Section 3.3.2.1, the exposure
modeling methodology used for this analysis is supported by robust evidence and is designed to provide
a nationally representative distribution of estimated water concentrations under varying conditions.
Page 158 of 570
-------
Releases used as inputs in the model were estimated using Monte Carlo modeling that captures
variability across sites. However, the modeled exposure estimates are not directly tied to specific
releases at known locations, decreasing the strength of the evidence related to the representativeness of
the exposure estimates for actual exposures. There is uncertainty around the years of exposure that may
be expected to result from hydraulic fracturing given the limited lifetime of a typical hydraulic
fracturing well. There is also some uncertainty around the proximity of releases to drinking water intake
locations and the extent to which 1,4-dioxane is further diluted prior to reaching intake locations. For
this analysis, EPA assumed that no additional dilution occurs prior to reaching drinking water intakes.
Table 5-5. Lifetime Cancer Risks Estimated from Hydraulic
Fracturing Produced Waters Disposed to Surface Water
under a Range of Scenarios
Monte Carlo Distribution
Adult Lifetime Cancer Risk
Maximum
8.76E-05
99th Percentile
4.21E-06
95th Percentile
1.52E-06
Median
3.85E-08
5th Percentile
1.89E-10
Minimum
1.56E-16
Lifetime cancer risk estimates are based on mean drinking water
ingestion rates over 33 years of oral exposure through drinking
water as an adult. There is uncertainty around the years of exposure
that may be expected to result from hydraulic fracturing given the
limited lifetime of atypical hydraulic fracturing well. LADDs used
to calculate these cancer risk estimates are presented in Table 3-5.
5.2.2.1.5 Aggregate Risks from Drinking Water Exposures Modeled from Multiple
Sources Releasing to Surface Water, Assuming No Downstream Dilution
Multiple sources may contribute to 1,4-dioxane concentrations in drinking water sourced from surface
water in a single location. EPA therefore estimated aggregate general population exposures and risks
that could occur as a result of combined contributions from multiple sources. As described in Section
2.3.1.3.4, EPA used probabilistic modeling to predict aggregate surface water concentrations that could
occur when accounting for DTD releases, indirect releases, and other upstream sources. EPA estimated
cancer and non-cancer risks for the drinking water exposure estimates in Section 3.2.2, which
correspond to the modeled aggregate surface water concentrations described in Section 2.3.1.3.4 and
assume that no 1,4-dioxane is removed through treatment. This analysis also assumes that
concentrations at drinking water intakes are not further diluted from the concentrations modeled near the
point of release. There is wide variation in both cancer and non-cancer risk within and across
OESs/COUs when taking into account aggregate contributions from other sources. This variation is
illustrated in the cancer risk estimates shown in the distributions of cancer risk estimates for exposures
modeled for each OES/COU in Figure 5-5. The large ranges of risk estimates for some OESs/COUs
reflect substantial variation in releases and characteristics of receiving water bodies across the set
facilities associated with those OESs. High-end cancer risk estimates in this analysis are very similar to
high-end risk estimates for individual facility releases alone, indicating that high-end estimates are
driven primarily by high-end industrial releases. Complete cancer and non-cancer risk estimates for the
range of aggregate water concentrations estimated for each COU using probabilistic modeling are
Page 159 of 570
-------
presented in 1,4-Dioxane Supplemental Information File: Drinking Water Exposure and Risk Estimates
for 1,4-Dioxane Surface Water Concentrations Predicted with Probabilistic Modeling (U.S. EPA.
20241)-
Overall confidence in distributions of risk estimates for drinking water exposures resulting from for
aggregate surface water concentrations predicted by probabilistic modeling varies across OES/COU.
Although confidence is not uniform for all facilities within an OES, overall confidence ratings for each
OES are intended to communicate how the factors that contribute to confidence and uncertainty vary
across COUs. As described in Section 4.3 overall confidence in the oral and dermal cancer slope factor
underlying these risk estimates is medium to high. As described in Section 3.3.2.1, the exposure
modeling methodology used for this analysis is supported by robust evidence and is designed to provide
a nationally representative distribution of estimated water concentrations under varying conditions. For
most COUs, this analysis is limited to facilities that report releases via TRI and/or DMR. Other sources
releasing smaller amounts of 1,4-dioxane are not directly captured, though the distribution of surface
water monitoring data used to represent background concentrations in the model is intended to capture
these other upstream sources. Available monitoring data confirm that 1,4-dioxane is present in some
surface water and drinking water, though most of the available data were not collected near release sites
and are therefore not directly comparable. In release locations where monitoring data are available, case
studies demonstrate strong consistency between modeled estimates and measured surface water
concentrations.
There is some uncertainty around the proximity of releases to drinking water intake locations and the
extent to which 1,4-dioxane is further diluted prior to reaching intake locations. For this analysis, EPA
assumed that no additional dilution occurs prior to reaching drinking water intakes. The characterization
of downstream dilution presented in 5.2.2.1.2 for individual facility releases illustrates the extent to
which downstream dilution may impact overall risk estimates.
The overall level of confidence in resulting exposure estimates depends on the source of OES/COU-
specific release data described in Appendix E.3:
• Overall confidence in drinking water exposure estimates is medium to high for OESs/COUs that
rely primarily on release data reported to DMR or to TRI via Form R. Most COUs/OESs are
included in this group.
• Overall confidence in drinking water exposure estimates is medium for OESs/COUs for which
release estimates are based on reporting to TRI via Form A. The Import and repackaging OES
releases used in this analysis are entirely based on Form A reporting of releases, and just under
half of the Industrial uses OES releases were reported via Form A.
• Overall confidence in drinking water exposure estimates is low to medium for OESs/COUs for
which release estimates are based on surrogate or modeled information.
Page 160 of 570
-------
Disposal
c
TO
T3
a
1-r-u-JTL
Lf)
CL
Import and Repackaging
_dn tH
LO
CD
Q_
PET Manufacturing
~u
0)
~Ln-r~l~l~rhr-r-r->--.
LO
cd
Q_
Ethoxylation byproduct
c
gj
a>
-=d
IT)
CD
CL
Industrial Uses
LO
CD
CL
Printing Inks
Functional Fluids (Open-System)
c
CO
TJ
CD
J]
LO
CD
0.
~dJlb
Manufacture
LO
CD
CL
Remediation
-=~Ł1
LO
CD
CL
Imlh.
10~J 10"^ 10"1 10"1U 10"3 10~B 10"' 10~b 10"* 10"4 10"J 10"* 10"1 10"1U 10"
Modeled Adult Lifetime Cancer Risk from Total 1,4-Dioxane (Releases + Background)
Figure 5-5. Histograms of Lifetime Cancer Risk Estimates for Aggregate Water Concentrations Estimated Downstream of COUs
with Vertical Lines Showing the Median and 95th Percentile (P95) Values
Page 161 of 570
-------
5.2.2.1.6 Integrated Summary of Drinking Water Risk Estimates across Multiple
Lines of Evidence for Surface Water
Risks from drinking water exposures were evaluated using a series of analyses that provided information
about the specific contributions of releases associated with individual sources as well as aggregate
exposures and risks to the general population. This analysis finds cancer risk estimates greater than 1 in
1 million from drinking water exposures informed by both monitoring data and modeled surface water
concentrations. Modeled concentrations result in cancer risk estimates greater than 1 in 1 million across
a range of individual sources and aggregate sources utilizing plausible drinking water exposure
scenarios.
Monitoring data demonstrates that 1,4-dioxane is present in some source water and finished drinking
water samples. Measured concentrations in finished drinking water samples resulted in cancer risk
estimates greater than 1 in 1 million at the high-end of the distribution of monitoring samples. Most
drinking water treatment systems are not expected to remove 1,4-dioxane from water, suggesting that
concentrations detected in source water can also be an indication of concentrations in drinking water.
Available monitoring data provided information about general population exposures but did not capture
high concentrations occuring in specific locations or at specific times from direct and indirect releases
into water bodies. Therefore, EPA relied on estimated concentrations modeled for a range of specific
release scenarios, including direct and indirect industrial releases, DTD releases, disposal of hydraulic
fracturing waste, and aggregate concentrations resulting from varying combinations of multiple sources
to characterize risks from the water pathway. EPA evaluated the performance of the models used to
estimate water concentrations with monitoring data from site-specific locations serving as cases studies.
These case study comparisons demonstrated strong consistency between modeled concentrations and
monitoring data, thereby increasing confidence in risk estimates based on modeled concentrations.
Across all modeled scenarios, 1,4-dioxane concentrations in water are primarily determined by the
amount of release from varying sources and the flow of the receiving water body. These two factors are
highly location and source-specific, resulting in very wide ranges of modeled water concentrations and
risk estimates for each set of analyses presented in the previous section above.
Risk estimates based on 1,4-dioxane concentrations modeled in the receiving water bodies at the point of
release show potential for risk greater than 1 in 1 million or 1 in 100,000 from each of the sources
assessed.
As described in Section 5.2.2.1.2, dilution that occurs between the point of release and drinking water
intake locations may be expected to reduce 1,4-dioxane concentrations in some locations. However,
even when accounting for dilution, upstream releases contribute to cancer risk estimates greater than 1 in
a million or 1 in 100,000 at some drinking water intake locations. EPA evaluated risks based on
modeled water concentrations for a sample of drinking water intake locations downstream of releases
where risk was greater than 1 in 1 million. After accounting for additional dilution, cancer risk estimates
remained greater than 1 in 1 million for 27 percent of the public water systems evaluated, serving a
combined population of over 2 million people.
The potential relative contribution from different sources varies under different conditions and is likely
to be site-specific. For example, high-end risk estimates in the aggregate model (presented in Section
5.2.2.1.5) are very similar to high-end risk estimates for facility releases alone (presented in Section
5.2.2.1.2), suggesting that in cases where industrial releases are high, those releases will be the dominant
Page 162 of 570
-------
source of 1,4-dioxane in water. On the other hand, under some conditions (presented in Section
5.2.2.1.3), plausible DTD release scenarios may present risk greater than 1 in 1 million in the absence of
industrial releases. Taken together, the analyses presented throughout this section demonstrate that each
of these sources may contribute to drinking water risks under some conditions. These analyses define the
conditions under which different levels of risk may occur.
5.2.2.2 Drinking Water - Groundwater and Disposal Pathways
EPA estimated risks from general population exposures that could occur if groundwater containing 1,4-
dioxane is used as a source of drinking water. Risk estimates presented in this section are not tied to
known releases at specific locations. Rather, this analysis defines the conditions under which 1,4-
dioxane disposal to landfills or from hydraulic fracturing operations could result in varying levels of
risk.
Cancer and non-cancer risk estimates were calculated based on modeled groundwater concentrations
described in Section 2.3.1.4 and corresponding drinking water exposures estimates described in Section
3.2.2.2. All risk estimates presented in this summary focus on the scenario with the greatest potential
exposure and risk. Because adult drinking water exposures relative to body weight are greater than
exposures relative to body weight that occur over the course of childhood (as illustrated in 1,4-Dioxane
Supplemental Information File: Drinking Water Exposure and Risk Estimates for 1,4-Dioxane Release
to Surface Water from Individual Facilities (U.S. EPA. 2024hV). the scenario with the greatest lifetime
cancer risk is 33 years of exposure as an adult. Lifetime cancer risk estimates presented in this section
are based on median drinking water ingestion rates over 33 years of exposure as an adult. Lifetime
cancer risk estimates for a full 78 years of exposure would be 2.26 times greater than the risk estimates
based on 33 years. Similarly, lifetime cancer risk estimates based on 95th percentile drinking water
ingestion rates would be approximately 3 to 4 times greater, depending on the age groups exposed.
For potential groundwater concentrations resulting from landfill leachate, EPA estimated cancer and
non-cancer risks for adults and formula-fed infants at concentrations estimated under varying
hypothetical combinations of leachate concentrations and loading rates. As shown in Table 5-6, lifetime
cancer risk estimates increase under scenarios with higher leachate concentrations and loading rates.
Chronic non-cancer risk estimates (not shown) indicate risk relative to the benchmark MOE only at the
highest leachate concentrations and loading rates. These concentrations and loading rates represent a
scenario where 1,4-dioxane is either delisted and released to a municipal solid waste landfill or when
trace concentrations present in consumer and commercial products are disposed to those same landfills.
Though the higher concentrations of 1,4-dioxane in leachate and higher loading rates are less likely, they
may represent a high-end PESS exposure. Complete results for cancer and non-cancer risk are available
in 1,4-Dioxane Supplemental Information File: Drinking Water Exposure and Risk Estimates for 1,4-
Dioxane Land Releases to Landfills ( 2024f).
Overall confidence in risk estimates for drinking water exposures resulting from disposal to landfills is
low to medium. As described in Section 4.3 overall confidence in the oral and dermal cancer slope
factor underlying these risk estimates is medium to high. As described in Section 3.3.2.2.1 the modeling
methodology is robust. However, the release information relied on as model input data is supported by
slight to moderate evidence, decreasing overall confidence. In addition, this drinking water exposure
scenario relies on the assumption that the estimated groundwater concentrations may occur in locations
where groundwater is used as a primary drinking water source. Although the substantial uncertainty
around the extent to which these exposures occur decreases overall confidence in the exposure scenario,
this scenario represents a PESS exposure.
Page 163 of 570
-------
Table 5-6. Lifetime Cancer Risks" Estimated for Modeled Groundwater Concentrations Estimated
under Varying Landfill Conditions
Leachate
Concentration
(mg/L)
Loading Rate (lb)
0.1
1
10
100
1,000
10,000
100,000
1,000,000
0.0001
3.1E-17
2.9E-16
3.6E-15
3.4E-14
3.3E-13
3.1E-12
3.0E-11
2.9E-10
0.001
3.1E-16
2.9E-15
3.6E-14
3.4E-13
3.3E-12
3.1E-11
3.0E-10
2.9E-09
0.01
3.1E-15
2.9E-14
3.6E-13
3.4E-12
3.3E-11
3.1E-10
3.0E-09
2.9E-08
0.1
3.1E-14
2.9E-13
3.6E-12
3.4E-11
3.3E-10
3.1E-09
3.0E-08
2.9E-07
1
3.1E-13
2.9E-12
3.6E-11
3.4E-10
3.3E-09
3.1E-08
3.0E-07
2.9E-06
10
3.1E-12
2.9E-11
3.6E-10
3.4E-09
3.3E-08
3.1E-07
3.0E-06
2.9E-05
100
3.1E-11
2.9E-10
3.6E-09
3.4E-08
3.3E-07
3.1E-06
3.0E-05
2.9E-04
1,000
3.1E-10
2.9E-09
3.6E-08
3.4E-07
3.3E-06
3.1E-05
3.0E-04
2.9E-03
10,000
3.1E-09
2.9E-08
3.6E-07
3.4E-06
3.3E-05
3.1E-04
3.0E-03
2.9E-02
a Lifetime cancer risk estimates based on mean drinking water ingestion rates over 33 years of oral exposure through
drinking water as an adult. Lifetime cancer risk estimates for a full 78 years of exposure would be 2.26 times greater than
the risk estimates presented here. Similarly, lifetime cancer risk estimates based on 95th percentile drinking water
ingestion rates would be approximately 3-4 times greater, depending on the age groups exposed.
For groundwater concentrations resulting from disposal of hydraulic fracturing produced water, EPA
calculated cancer and non-cancer risks for adults and formula-fed infants. Cancer risk estimates across
the full distribution of modeled releases are presented in Table 5-7. Cancer risk estimates are 4.0E-07 for
median modeled releases and 8.6xlCT6 for 95th percentile modeled releases. Chronic non-cancer risk
estimates are above the corresponding benchmark MOE for all modeled groundwater concentrations,
indicating lower non-cancer risk from non-cancer effects. Complete cancer and noncancer risk
calculations are available in 1,4-Dioxane Supplemental Information File: Drinking Water Exposure and
Risk Estimates for 1,4-Dioxane Land Releases to Surface Impoundments ( 24g).
The risk estimates presented here are based on groundwater concentrations modeled using the original
release assessment published in the draft supplement. Although EPA revised the release assessment for
hydraulic fracturing based on SACC recommendations, the shift in release estimates is not sufficient to
result in changes to overall risk conclusions. Therefore, the Agency did not revise subsequent modeling
or exposure and risk estimates for releases from hydraulic fracturing operations. Because the revised
release assessment resulted in lower release values at the high-end, the risk estimates presented here
based on the original release assessment may overestimate risk at the high-end.
Overall confidence in risk estimates for drinking waters resulting from disposal of hydraulic fracturing
waste is low to medium. As described in Section 4.3 overall confidence in the oral and dermal cancer
slope factor underlying these risk estimates is medium to high. As described in Section 3.3.2.2.2, the
modeling methodology is robust and the release information relied on as model input data is supported
by moderate evidence. However, no monitoring data are available to confirm detection of 1,4-dioxane in
groundwater near hydraulic fracturing operations. This drinking water exposure scenario relies on the
assumption that the estimated groundwater concentrations may occur in locations where groundwater is
used as a primary drinking water source. There is uncertainty around the years of exposure that may be
Page 164 of 570
-------
expected to result from hydraulic fracturing given the limited lifetime of a typical hydraulic fracturing
well. Although the substantial uncertainty around the extent to which these exposures occur decreases
overall confidence in the exposure scenario, this scenario represents a PESS exposure.
Table 5-7. Lifetime Cancer Risks" Estimated for Modeled Groundwater Concentrations Resulting
from Disposal of Hydraulic Fracturing Produced Water
Monte Carlo
Distribution
Modeled Groundwater
Concentration (mg/L)
Adult LADD
(mg/kg/day)
Adult Cancer Risk
Estimate
Max
1.9E-02
8.8E-05
1.1E-05
99th
1.5E-02
7.1E-05
8.6E-06
95th
1.5E-02
7.1E-05
8.6E-06
Mean
7.1E-04
3.3E-06
4.0E-07
50th
1.2E-04
5.6E-07
6.8E-08
5th
1.2E-04
5.6E-07
6.8E-08
Min
4.4E-07
2.1E-09
2.5E-10
a Lifetime cancer risks based on mean drinking water ingestion rates over 33 years of oral exposure through drinking
water as an adult. Lifetime cancer risk estimates for a full 78 years of exposure would be 2.26 times greater than the
risk estimates presented here. There is uncertainty around the years of exposure that may be expected to result from
hydraulic fracturing given the limited lifetime of a typical hydraulic fracturing well. Similarly, lifetime cancer risk
estimates based on 95th percentile drinking water ingestion rates would be approximately 3-4 times greater,
depending on the age groups.
5.2.2.3 Air Pathway
EPA estimated risks from general population exposures to 1,4-dioxane released to air, with a focus on
exposures in fenceline communities. Risks were evaluated for air releases from industrial COUs,
hydraulic fracturing operations, and industrial and institutional laundry facilities based on exposure
estimates in Section 3.2.3.
5.2.2.3.1 Industrial COUs Reported to TRI
EPA estimated risks from general population exposures that could occur in communities neighboring
industrial releases associated with stack and fugitive emissions. Cancer and non-cancer risk estimates for
general population exposures within 10,000 m of industrial releases were calculated for the 10th, 50th,
and 95th percentiles of modeled exposure concentrations estimated in Section 3.2.3.1. Table 5-8
summarizes the cancer risk estimates based on 33 years exposure duration and for 95th percentile
exposure concentrations within 1,000 m of the facilities with the greatest risk in each OES/COU,
ranging from 1,05/10 10 to 1.1 x ] o 4, Cancer risk estimates based on 33 years exposure duration and for
50th percentile modeled exposure concentrations within 1,000 m of the highest risk facilities range from
2.5xl0~u to 8.3xl0~5 (data not shown).
Lifetime cancer risk estimates in Table 5-8 are based on 33 years of continuous inhalation exposure
averaged over a 78-year lifetime. EPA agrees with the SACC recommendation for EPA to utilize a full
lifetime of exposure for fenceline communities. Lifetime cancer risk estimates for a full 78 years of
continuous inhalation exposure would be 2.36 times greater than the risk estimates presented here,
resulting in risk estimates as high as 2.6xl0~4 (for manufacturing, within 10m of facilities). Risk
estimates were generally highest within 1,000 m of the facilities and lower at greater distances. As
discussed in Section 2.3.3.3, exposure estimates very near facilities (5-10 m) may be impacted by
assumptions made for modeling around an area source (10^10 area source places people at 5 m on top of
Page 165 of 570
-------
the release point). This, in combination with other factors like meteorological data, release heights, and
plume characteristics can result in lower or higher exposures at 5 m than just off the release point at 10
m. Air concentrations from fugitive emissions tend to peak within 10 m of release sites while
contributions from stack releases generally peak around 100 m, meaning that risks nearest to release
sites are often driven by fugitive releases. Acute and chronic non-cancer risk estimates (not shown) do
not indicate risk relative to benchmark MOEs for any of the estimated exposure concentrations at any
facilities evaluated. Complete cancer and non-cancer risk results for air concentrations modeled from
stack, fugitive and combined air emissions are provided in 1,4-Dioxane Supplemental Information File:
Air Exposures and Risk Estimates for Single Year Analysis ( )24e).
Air exposure and risk estimates are dependent on release amounts, stack heights, contributions from
stack releases and fugitive emissions, topography, and meteorological conditions. These factors vary
substantially across facilities within each OES/COU, making release amount, stack height, and
meteorological conditions more important predictors of risk than a facility's identified OES/COU.
Overall confidence in site-specific risk estimates for inhalation exposure resulting from industrial
releases varies across OES/COUs. As described in Section 4.3, overall confidence in the cancer
inhalation unit risk underlying these risk estimates is high. As described in Section 3.3.3.1, the
AERMOD modeling methodology used for this analysis is robust and accounts for both stack and
fugitive emissions. The exposure scenarios considered are most relevant to long-term residents in
fenceline communities. There is some uncertainty around the extent to which people actually live and
work around the specific facilities where risks are highest, decreasing overall confidence in the exposure
scenario, particularly at distances nearest release sites. Overall confidence varies due to variable levels
of confidence in underlying release information used to estimate exposures. An OES-specific discussion
of the confidence in sources of release information is presented in Appendix E.5E.5.4, but in general
terms
• Overall confidence in risk estimates is medium to high for OESs/COUs that rely primarily on
release data reported to TRI via Form R.
• Overall confidence in risk estimates is medium for OESs/COUs for which release estimates are
based on data reported to TRI via Form A.
• Overall confidence in risk estimates is low to medium for OESs/COUs for which release
estimates are based on surrogate or modeled information.
Page 166 of 570
-------
Table 5-8. Inhalation Lifetime Cancer Risks" within 10 km of Industrial Air Releases Based on 95th Percentile Modeled Exposure
Concentrations
OES
Corresponding COUs
# Facilities
Distance from Facility with Greatest Risk (m)6
Overall
Confidence
Life Cvelc
Stage
Category
Subcategory
Total
Risk
>lE-06
5
10
30
60
100
100-1,000
Disposal
Disposal
Disposal
Hazardous waste
incinerator
Off-site waste
transfer
Underground
injection
Hazardous landfill
15
5
2.88E-05
3.42E-05
1.22E-05
4.67E-06
2.13E-06
2.00E-07
Medium to
High
Dry film
lubricant
Industrial use,
commercial use
Other uses
Dry film lubricant
8
0
1.09E-12
4.83E-11
3.46E-09
2.62E-08
4.26E-08
6.72E-09
Low to
Medium
Ethoxylation
byproduct
Processing
Byproduct
Byproduct produced
during the
ethoxylation
process to make
ethoxylated
ingredients for
personal care
products
6
3
4.42E-05
9.21E-05
4.96E-05
2.09E-05
1.11E-05
2.58E-06
Medium to
High
Film cement
Industrial use,
commercial use
Adhesives
and sealants
Film cement
1
0
8.46E-07
8.86E-07
2.99E-07
1.54E-07
8.46E-08
1.55E-08
Low to
Medium
Functional
fluids (open-
system)
Industrial use
Functional
fluids (open
and closed
systems)
Polyalkylene glycol
lubricant
Synthetic
metalworking fluid
Cutting and tapping
fluid
2
0
8.67E-08
1.60E-07
6.98E-08
7.31E-08
1.23E-07
5.02E-08
Medium to
High
Import and
repackaging
Manufacturing
Import
Import
Repackaging
1
0
1.82E-13
3.78E-12
3.74E-10
2.82E-09
5.89E-09
2.18E-09
Medium to
High
Page 167 of 570
-------
Corresponding
COLs
# Facilities
Distance from Facility with Greatest Risk (m)6
Overall
OES
Life Cvelc
Stage
Category
Subcategory
Total
Risk
> 1E—06
5
10
30
60
100
100-1,000
Confidence
Processing
Processing as
a reactant
Polymerization
Catalyst
Processing
Non-
incorporative
Basic organic
chemical
manufacturing
(process solvent)
Industrial
Uses
Industrial use
Intermediate
use
Plasticizer
intermediate
Catalysts and
reagents for
anhydrous acid
reactions,
brominations, and
sulfonations
12
6
2.84E-05
3.24E-05
1.04E-05
3.84E-06
1.89E-06
4.85E-07
Medium to
High
Laboratory
Industrial use,
Laboratory
Chemical reagent
1
1
1.40E-05
1.46E-05
4.91E-06
2.54E-06
1.40E-06
2.54E-07
Low to
Chemical
commercial use
chemicals
Reference material
Medium
Use
Spectroscopic and
photometric
measurement
Liquid scintillation
counting medium
Stable reaction
medium
Cryoscopic solvent
for molecular mass
determinations
Preparation of
histological
sections for
microscopic
examination
Page 168 of 570
-------
OES
Corresponding COLs
# Facilities
Distance from Facility with Greatest Risk (m)6
Overall
Confidence
Life Cvelc
Stage
Category
Subcategory
Total
Risk
> 1E—06
5
10
30
60
100
100-1,000
Manufactur-
ing
Manufacturing
Domestic
manufacture
Domestic
manufacture
1
1
5.91E-05
1.10E-04
5.20E-05
2.18E-05
1.08E-05
9.62E-07
Medium to
High
PET
Manufactur-
ing
Processing
Byproduct
Byproduct produced
during the
production of
polyethlene
terephtalate
13
10
5.42E-05
6.48E-05
2.37E-05
9.47E-06
4.35E-06
7.25E-07
Medium to
High
Spray foam
application
Industrial use,
commercial use
Other uses
Spray polyurethane
foam
1
0
5.28E-09
5.68E-09
1.94E-09
1.02E-09
5.79E-10
1.05E-10
Low to
Medium
"Lifetime cancer risks based on 33 years of continuous inhalation exposure averaged over a 78-year lifetime. Lifetime cancer risks for a full 78 years of continuous
inhalation exposure would be 2.36 times greater than the risk estimates presented here.
h Cancer risks were also calculated at 2,500, 5,000 and 10,000 m from all facilities.
Page 169 of 570
-------
Land Use Analysis
For locations where lifetime cancer risk is greater than 1 x 1CT6, EPA evaluated land use patterns to
determine whether fenceline community exposures may be reasonably anticipated. Detailed results of
this analysis are described in Appendix J.3 and are consistent with the methods described in the 2022
Draft TSCA Screening Level Approach for Assessing Ambient Air and Water Exposures to Fenceline
Communities Version 1.0 (U.S. EPA. 2022d). In short, EPA determined whether residential,
industrial/commercial businesses, or other public spaces are present within those radial distances where
cancer risk estimates for 95th percentile modeled air concentrations are greater than 1 x 1CT6 for each
facility. In all cases, risks greater than 1 / 10 6 were within 1,000 m or less of releasing facilities. This
analysis was limited to facilities that could be mapped to a GIS location. Based on this characterization
of land use patterns, fenceline community exposures have the potential to occur at 50 percent of
facilities (11 of 22 GIS-mapped facilities) where cancer risk is greater than 1 x 10~6 based on modeled
fenceline air concentrations.
Aggregate Risk
EPA also evaluated potential risks from aggregate exposures from multiple neighboring facilities using a
conservative screening methodology. EPA identified five groups of two to four facilities reporting 1,4-
dioxane releases in proximity to each other (i.e., within 10 km). Aggregating risks estimated for these
groups of facilities were generally dominated by the facility with the greatest risk. This aggregate
analysis did not identify locations with cancer risk greater than 1 x 10~6 that did not already have cancer
risk above that level from an individual facility. Details of the methods and results of this aggregate
analysis are described in Appendix J.4.
5.2.2.3.2 Hydraulic Fracturing
Cancer and non-cancer risk estimates for potential general population exposures within 1,000 m of
hydraulic fracturing operations were calculated for a range of air concentrations modeled across the
distribution of release estimates, as described in Section 3.3.3.2. Table 5-9 presents lifetime cancer risk
estimates for exposure to high-end and central tendency air concentrations modeled for both high-end
(95th percentile) and central tendency (50th percentile) modeled releases for a range of topographical
and meteorological scenarios. Lifetime cancer risk estimates for distances within 1,000 m of hydraulic
fracturing operations range from 3.9x 10~7 to 7.1 x 10 5 for high-end release estimates and 2,2/ 10 8 to
4.1xio-6 for central tendency release estimates across a range of model scenarios. Acute and chronic
non-cancer risk estimates (not shown) do not indicate risk relative to benchmark MOEs for any exposure
concentrations estimated for hydraulic fracturing operations. Complete results are provided in 1,4-
Dioxane Supplemental Information File: Air Exposure and Risk Estimates for 1,4-Dioxane Emissions
from Hydraulic Fracturing Operations (U.S. EPA. 2024b).
The risk estimates presented here are based on air concentrations modeled using the original release
assessment published in the draft supplement. Although EPA revised the release assessment for
hydraulic fracturing based on SACC recommendations, the shift in release estimates is not sufficient to
result in changes to overall risk conclusions. Therefore, EPA did not revise subsequent air modeling or
exposure and risk estimates for air releases from hydraulic fracturing operations. Because the revised
release assessment resulted in lower release values, the risk estimates presented here based on the
original release assessment may overestimate risk.
Overall confidence in risk estimates for inhalation exposures resulting for air concentrations modeled
based on releases from hydraulic fracturing operations is medium. As described in Section 4.3, overall
confidence in the cancer inhalation unit risk underlying these risk estimates is high. As described in
Section 3.3.3.2 the modeling methodologies used to estimate air concentrations are robust. The
Page 170 of 570
-------
distribution of air releases used as model input data were estimated using Monte Carlo modeling and
rely on assumptions. No air monitoring data were available to confirm detection of 1,4-dioxane is air
near hydraulic fracturing operations. Because the air concentrations underlying this analysis are based
on releases estimated using probabilistic modeling, they are not tied to specific locations that can be
evaluated for land use patterns. There is therefore substantial uncertainty around the extent to which
people actually live and work around the specific locations where risks are highest, decreasing overall
confidence in the exposure scenario. There is also uncertainty around the years of exposure that may be
expected to result from hydraulic fracturing given the limited lifetime of a typical hydraulic fracturing
well.
Page 171 of 570
-------
Table 5-9. Lifetime Cancer Risk Estimates for Fugitive Emissions from Hydraulic Fracturing'
Ul I)
Fugitive
Emissions Release
Scenario
Cancer Risk Estimates for 95th Percentile Modeled Releases
Cancer Risk Estimates for 50th Percentile Modeled Releases
High-End Modeled Air
Concentrations
100 m
1,000 m
100 to
1,000 m
Central Tendency Modeled Air
Concentrations
High-End Modeled Air
Concentrations
100 m
1,000 m
100 to
1,000 m
100 m
1,000 m
100 to
1,000 m
Central Tendency Modeled
Air Concentrations
100 m
1,000 m
100 to
1,000 m
South (Coastal)-
Rural-24
7.1E-05
3.2E-06
8.3E-06
5.6E-05
2.4E-06
6.4E-06
4.1E-06
1.8E-07
4.7E-07
3.2E-06
1.4E-07
3.6E-07
West North
Central-Rural-24
5.4E-05
3.1E-06
7.3E-06
4.1E-05
1.9E-06
4.9E-06
3.1E-06
1.8E-07
4.2E-07
2.4E-06
1.1E-07
2.8E-07
South (Coastal)-
Urban-24
3.4E-05
7.7E-07
2.4E-06
3.0E-05
6.7E-07
2.1E-06
2.0E-06
4.4E-08
1.4E-07
1.7E-06
3.8E-08
1.2E-07
West North
Central-Urban-24
3.2E-05
8.3E-07
2.5E-06
2.6E-05
6.1E-07
1.9E-06
1.8E-06
4.8E-08
1.4E-07
1.5E-06
3.5E-08
1.1E-07
South (Coastal)-
Rural-8
1.3E-05
1.2E-07
5.1E-07
1.1E-05
9.0E-08
4.2E-07
7.3E-07
6.8E-09
2.9E-08
6.4E-07
5.2E-09
2.4E-08
West North
Central-Rural-8
2.7E-05
1.0E-06
2.5E-06
1.4E-05
2.9E-07
8.8E-07
1.5E-06
5.9E-08
1.4E-07
8.2E-07
1.6E-08
5.0E-08
South (Coastal)-
Urban-8
1.2E-05
9.0E-08
4.3E-07
1.1E-05
8.0E-08
3.9E-07
6.7E-07
5.2E-09
2.5E-08
6.1E-07
4.6E-09
2.2E-08
West North
Central-Urban-8
1.9E-05
3.9E-07
1.2E-06
1.2E-05
1.6E-07
6.0E-07
1.1E-06
2.2E-08
7.1E-08
7.0E-07
9.1E-09
3.4E-08
" Lifetime cancer risks based on 33 years of continuous inhalation exposure averaged over a 78-year lifetime. Lifetime cancer risks for a full 78 years of
continuous inhalation exposure would be 2.36 times greater than the risk estimates presented here.
b Cancer risk estimates shown here are based on modeled releases and air concentrations estimated for 72 days of release.
Page 172 of 570
-------
5.2.2.3.3 Industrial and Institutional Laundry Facilities
Cancer and non-cancer risk estimates for potential general population exposures within 1,000 m of
industrial and institutional laundry facilities were calculated for a range of air concentrations modeled
for a range of releases, as described in Section 3.2.3.3. Table 5-10 presents lifetime cancer risk estimates
for exposures estimated from both high-end and central tendency air concentrations modeled based on
the maximum release scenario for each type of laundry under the most conservative exposure scenario
evaluated (rural south coastal topography, assuming 24 hours of releases each day). Lifetime cancer risk
estimates for distances within 1,000 m of laundry facilities range from 1.5/10 " to 3,8/10 8 across a
range of high-end and central tendency exposure scenarios. Acute and chronic non-cancer risk estimates
(not shown) do not indicate risk for any estimated exposure concentrations for laundries relative to the
benchmark MOEs. Complete results are provided in 1,4-Dioxane Supplemental Information File: Air
Exposures and Risk Estimates for Industrial Laundry ( 2024c).
The risk estimates presented here are based on air concentrations modeled using the original release
assessment published in the draft supplement. Although EPA revised the release assessment for
industrial and institutional laundries based on SACC recommendations, the shift in release estimates is
not expected to be sufficient to result in changes to overall risk conclusions. Therefore, the Agency did
not revise subsequent air modeling or exposure and risk estimates for air releases from industrial and
institutional laundries. Because the revised release assessment resulted in values roughly an order of
magnitude higher, the risk estimates presented here based on the original release assessment may
underestimate risk.
Overall confidence in risk estimates from inhalation exposures resulting from industrial and institutional
laundries is medium. As described in Section 4.3, overall confidence in the cancer inhalation unit risk
underlying these risk estimates is high. As described in Section 3.3.3.2, the modeling methodologies are
robust. The distribution of air releases used as model input data were estimated using Monte Carlo
modeling and rely on assumptions. No air monitoring data were available to determine whether 1,4-
dioxane is detected near industrial and institutional laundry facilities. Because the air concentrations
underlying this analysis are based on probabilistic modeling, they are not tied to specific locations that
can be evaluated for land use patterns. There is therefore substantial uncertainty around the extent to
which people actually live and work around the specific locations where risks are highest, decreasing
overall confidence in the exposure scenario.
Table 5-10. Lifetime Cancer Risk Estimates for Fugitive Emissions from Industrial and
Institutional Laundry Facilities"
Facility Type
Detergent and
Emissions Type
Cancer Risk Estimates for Maximum Modeled Releases
High-End Modeled Air
Concentrations
Central Tendency Modeled Air
Concentrations
100 m
1,000 m
100 to
1,000 m
100 m
1,000 m
100 to
1,000 m
Industrial
Laundry
Liquid - vapor
3.7E-08
1.7E-09
4.3E-09
3.3E-08
1.4E-09
3.8E-09
Powder - vapor
3.6E-08
1.7E-09
4.2E-09
3.3E-08
1.4E-09
3.8E-09
Powder - PM10
3.8E-08
8.8E-10
3.2E-09
3.4E-08
7.9E-10
2.9E-09
Powder - PM2.5
3.6E-08
1.6E-09
4.1E-09
3.3E-08
1.4E-09
3.7E-09
Page 173 of 570
-------
Facility Type
Detergent and
Emissions Type
Cancer Risk Estimates for Maximum Modeled Releases
High-End Modeled Air
Concentrations
Central Tendency Modeled Air
Concentrations
100 m
1,000 m
100 to
1,000 m
100 m
1,000 m
100 to
1,000 m
Institutional
Laundry
Liquid - vapor
2.3E-08
1.1E-09
2.7E-09
2.1E-08
9.0E-10
2.4E-09
Powder - vapor
6.8E-10
3.2E-11
7.9E-11
6.2E-10
2.7E-11
7.0E-11
Powder - PM10
7.1E-10
1.6E-11
5.9E-11
6.4E-10
1.5E-11
5.3E-11
Powder - PM2.5
6.8E-10
3.0E-11
7.7E-11
6.2E-10
2.6E-11
6.9E-11
a Lifetime cancer risks based on 33 years of continuous inhalation exposure averaged over a 78-year lifetime.
Lifetime cancer risks for a full 78 years of continuous inhalation exposure would be 2.36 times greater than the risk
estimates presented here.
5,2,3 Potentially Exposed or Susceptible Subpopulations
EPA considered PESS throughout the exposure assessment presented in this supplement and throughout
the hazard identification and dose-response analysis described in the 2020 RE. Table 5-11 summarizes
how PESS were incorporated into the supplement through consideration of increased exposures and/or
increased biological susceptibility. The table also summarizes the remaining sources of uncertainty
related to consideration of PESS.
Table 5-11. Summary of PESS Considerations Incorporated throughout the Analysis and
Remaining Sources of Uncertainty
PESS
Categories
Potential Exposures Identified in Specific
Subpopulations and Incorporated into
Exposure Assessment
Potential Sources of Biological
Susceptibility Identified and
Incorporated into Hazard Assessment
Lifestage
General population drinking water exposure
scenarios include lifestage-specific exposure
factors for adults, children, and formula-fed
infants (Section 5.2.2.1); Inhalation exposures
are based on air concentrations and are therefore
consistent across lifestages (Section 5.2.2.3).
Based on pchem properties and a lack of studies
evaluating potential for accumulation in milk,
EPA did not quantitatively evaluate the milk
pathway and this is a source of uncertainty.
EPA qualitatively described the potential
for biological susceptibility due to lifestage
differences and developmental toxicity but
did not identify quantitative evidence of
lifestage-specific susceptibilities to 1,4-
dioxane; A 10/ UF was applied for human
variability. The magnitude of potential
lifestage differences in metabolism and
toxicity are not well quantified and are a
remaining source of uncertainty.
Pre-existing
Disease
EPA did not identify health conditions that may
influence exposure. The potential for pre-
existing disease to influence exposure (due to
altered metabolism, behaviors, or treatments
related to the condition) is a source of
uncertainty.
EPA qualitatively described the potential
for pre-existing health conditions, such as
liver disease, to increase susceptibility or
alter toxicokinetics, but did not identify
direct quantitative evidence. A 10* UF
was applied for human variability. The
potential impact of pre-existing diseases on
susceptibility to 1,4-dioxane is a remaining
source of uncertainty.
Page 174 of 570
-------
PESS
Categories
Potential Exposures Identified in Specific
Subpopulations and Incorporated into
Exposure Assessment
Potential Sources of Biological
Susceptibility Identified and
Incorporated into Hazard Assessment
Lifestyle
Activities
EPA did not identify specific lifestyle activities
that expected to increase 1,4-dioxane exposure.
This is a remaining source of uncertainty.
EPA did not identify lifestyle factors that
influence biological susceptibility to 1,4-
dioxane. This is a remaining source of
uncertainty.
Occupational
Exposures
EPA evaluated a range of occupational exposure
scenarios in manufacturing, hydraulic fracturing
and use of commercial products that increase
exposure to 1,4-dioxane present as a byproduct.
EPA evaluated risks for high-end exposure
estimates for each of these scenarios (Section
5.2.1).
EPA did not identify occupational factors
that increase biological susceptibility to
1,4-dioxane. This is a remaining source of
uncertainty.
Geographic
Factors
EPA evaluated risks to fenceline communities
from 1,4-dioxane in ambient air (Section 5.2.2.3)
and in drinking water downstream of release
sites (Section 5.2.2.1). EPA mapped tribal lands
in relation to air, surface water and ground water
releases of 1,4-dioxane to identify potential for
increased exposures for tribes due to geographic
proximity (Section 2.3).
EPA did not identify geographic factors
that increase biological susceptibility to
1,4-dioxane. This is a remaining source of
uncertainty.
Socio-
demographic
Factors
EPA did not identify specific sociodemographic
factors that influence exposure to 1,4-dioxane.
This is a remaining source of uncertainty.
EPA did not identify sociodemographic
factors that influence biological
susceptibility to 1,4-dioxane. This is a
remaining source of uncertainty.
Nutrition
EPA did not identify nutritional factors
influencing exposure to 1,4-dioxane. This is a
remaining source of uncertainty.
EPA did not identify nutritional factors
that influence biological susceptibility to
1,4-dioxane. This is a remaining source of
uncertainty.
Genetics
EPA did not identify genetic factors influencing
exposure to 1,4-dioxane. This is a remaining
source of uncertainty.
Indirect evidence that genetic variants may
increase susceptibility of the target organ
was addressed through a 10x UF for
human variability. The magnitude of the
impact of genetic variants is unknown and
is a source of uncertainty.
Unique
Activities
Some tribes may have increased exposure to
drinking water due to tribal activities such as
sweat lodges. EPA has identified upper bound
drinking water estimates of 2-4 L/day associated
with tribal lifewavs for some tribes (Hart>er.
2017; Harper and Ranco. 2009; Harper et aL
2007; Harper et aL, 2002). Risk calculations in
this supplement assume an acute adult drinking
water intake of 3.2 L/day and a chronic drinking
water intake of 0.88 L/day. Other potential
sources of increased exposure to 1,4-dioxane due
to specific tribal lifeways or other unique activity
patterns are a source of uncertainty.
EPA did not identify unique activities that
influence susceptibility to 1,4-dioxane.
This is a remaining source of uncertainty.
Page 175 of 570
-------
PESS
Categories
Potential Exposures Identified in Specific
Subpopulations and Incorporated into
Exposure Assessment
Potential Sources of Biological
Susceptibility Identified and
Incorporated into Hazard Assessment
Aggregate
Exposures
EPA evaluated risk from aggregate sources of
exposure contributing to 1,4-dioxane in water
(Section 5.2.2.1) or from multiple sources in
proximity releasing to air (Section 5.2.2.3,
Appendix J.4). Risks from aggregate exposures
across routes or pathways were evaluated
qualitatively and are a remaining source of
uncertainty.
EPA does not identify ways that aggregate
exposures would influence susceptibility to
1,4-dioxane. This is a remaining source of
uncertainty.
Other
Chemical and
Non-chemical
Stressors
EPA did not identify chemical and nonchemical
factors influencing exposure to 1,4-dioxane. This
is a remaining source of uncertainty.
EPA did not identify chemical or
nonchemical factors that influence
susceptibility to 1,4-dioxane. There is
insufficient data to quantitatively address
potential increased susceptibility due to
chemical or nonchemical stressors and this
is a remaining source of uncertainty.
5.2.4 Aggregate and Sentinel Exposures
In this supplement, EPA considers the combined 1,4-dioxane exposure an individual may experience due to
releases from multiple facilities in proximity releasing to air or multiple releases contributing to drinking
water concentrations in a particular location. For general population drinking water exposure scenarios,
EPA evaluated combined exposure and risks from multiple sources of 1,4-dioxane in surface water,
including direct and indirect industrial releases, DTD releases, and upstream background contamination
(Section 5.2.2.1). For general population air exposure scenarios, EPA evaluated combined exposure and
risk across multiple facilities in proximity releasing to air (Section 5.2.2.3 and Appendix J.4).
EPA considered aggregating cancer risks across inhalation, oral, and/or dermal routes of exposure.
There is uncertainty around the extent to which cancer risks across routes are additive for 1,4-dioxane.
Liver tumors are the primary site of cancer risk from oral exposures. Inhalation exposure in rats is
associated with multiple tumor types, including liver. The IUR used to calculate inhalation cancer risk
reflects combined risks from multiple tumor types. Although EPA concluded that nasal cavity lesions
are likely to be primarily the result of systematic delivery (as discussed on p. 192 of the 2020 RE), there
is uncertainty around the degree to which those effects could be partially due to portal of entry effects
following inhalation exposure. It is therefore unclear the extent to which it is appropriate to
quantitatively aggregate cancer risks based on the IUR with liver tumor risks associated with oral or
dermal exposures. EPA assessed the potential impact of aggregation across routes by summing risks
from dermal and inhalation exposures for each COU in the occupational risk calculator. Given the
uncertainty around the additive nature of cancer risk across routes, EPA is not relying on these
quantitative aggregate risk estimates as the basis for risk conclusions in this assessment. However, the
aggregate estimates illustrate the potential magnitude of the impact on risk estimates if risks are assumed
to be additive across routes. EPA considers the potential aggregate cancer risk across routes to be a
source of uncertainty for 1,4-dioxane cancer risk estimates.
EPA also considered aggregating cancer risks across dermal and oral exposures. The dermal cancer
slope factor is derived from the oral cancer slope factor by route-to-route extrapolation. Because the
systemic effect is assumed to be the same for both routes, the Agency determined that it could be
biologically appropriate to aggregate risk from dermal and oral exposures. General population scenarios
Page 176 of 570
-------
included inhalation and oral not dermal exposures and occupational and consumer exposure scenarios
included inhalation and dermal not oral exposures. However, this supplement does not include COUs or
pathways in which both oral and dermal exposure routes are considered.
EPA also considered potential for aggregate exposures across groups. For example, there may be some
individuals who are exposed at work as well as through general population air and drinking water
pathways or through consumer product use. Given the uncertainty around the degree to which
individuals may be exposed through multiple scenarios, the Agency did not further quantify aggregate
exposure across occupational, consumer and general population exposures. In most potential
combinations of exposures scenarios, the exposures and risks from one scenario are much greater than
from the other scenarios that may be aggregated with it (e.g., occupational risks for a particular COU
may be an order of magnitude greater than risks from 1,4-dioxane in drinking water in the community
where the worker lives). When this is the case, aggregate risk would be very similar to risk from the
scenario with the highest risk. In more rare cases where risks from a particular combination of exposure
scenarios are similar (e.g., occupational risks for a particular COU are equal to risks from drinking
water), aggregate risks could theoretically be double the risk from each pathway in isolation. These
types of aggregate risks were not quantified for specific combinations of scenarios and risks for
individual exposure scenarios should be interpreted with an appreciation for potential aggregate
exposures and risks.
Section 2605(b)(4)(F)(ii) of TSCA requires EPA, as a part of the risk evaluation, to describe whether
aggregate or sentinel exposures under the conditions of use were considered and the basis for their
consideration. EPA defines sentinel exposure as "the exposure to a single chemical substance that
represents the plausible upper bound of exposure relative to all other exposures within a broad category
of similar or related exposures (40 CFR § 702.33)." In this supplement, EPA considered sentinel
exposures by considering risks to populations who may have upper bound exposures. Where possible,
EPA focused on assessing exposure scenarios where the greatest exposures are likely to occur, including
workers and ONUs who perform activities with higher exposure potential and fenceline communities.
The Agency characterized high-end exposures in evaluating these exposure scenarios using both
monitoring data and modeling approaches. Where statistical data are available, EPA typically uses the
95th percentile value of the available dataset to characterize high-end exposure for a given COU.
Although the analysis is intended to capture the exposure scenarios and populations likely to result in the
greatest exposures, the Agency acknowledges that there may be additional groups with sentinel
exposures that are not captured in this analysis.
5,2,5 Summary of Overall Confidence and Remaining Uncertainties in Human Health
Risk Characterization
The overall level of confidence in each set of risk estimates depends on the level of confidence in the
underlying hazard values summarized in Section 4.3 and the level of confidence in exposure estimates
described in more detail in Section 3.3.
For all risk estimates, EPA has medium to high confidence in the underlying hazard PODs used as the
basis for this risk characterization. Sources of confidence in each of the hazard values were described in
the 2020 RE and are summarized in Section 4.3. Cancer risk is the primary risk driver for each of the
scenarios evaluated in this supplement and is therefore the basis of overall confidence levels described
herein. There is remaining uncertainty for all risk estimates around the potential impact of 1,4-dioxane
on potentially exposed or susceptible subpopulations (as discussed in Section 5.2.3). EPA applied an
intraspecies uncertainty factor of 10 to all non-cancer PODs to account for variation across gender, age,
health status, or genetic makeup, and other factors that may increase susceptibility, but the actual
Page 177 of 570
-------
magnitude of the impact of these factors on susceptibility is unknown. Similarly, EPA evaluated risks
from exposure scenarios that are intended to reflect the most highly exposed populations (including
fenceline communities and highly exposed workers), but the potential for other highly exposed
populations that were not identified in this analysis is a source of uncertainty. Potential for aggregate
risks across routes or pathways that are not quantitatively evaluated in this assessment is another source
of uncertainty.
5.2.5.1 Risks from Occupational Exposures
Overall confidence in risk estimates for occupational inhalation exposures ranges from low to high,
depending on the confidence in exposure assessment for each OES/COU. As described in Section 4.3,
overall confidence in the cancer inhalation unit risk underlying these risk estimates is high. As described
in Section 3.3.1.1, the measured and monitored inhalation exposure data are supported by moderate to
robust evidence. Additionally, the exposure modeling methodologies and underlying model input data is
supported by moderate to robust evidence. However, there is uncertainty in the representativeness of the
assessed exposure scenarios towards all potential exposures for the given OES/COU, limitations in the
amount and age of monitoring data, and limitations in the modeling approaches towards 1,4-dioxane-
specific use within the OES/COU. Therefore, while the underlying data and methods used to estimate
occupational inhalation risk is supported by moderate to robust evidence, the overall confidence of these
estimates ranges from low to high depending on the OES/COU. OES/COU-specific discussions of the
available inhalation exposure data and overall confidence are presented in Appendix F.6 and
summarized in Table 3-2.
Overall confidence in risk estimates for occupational dermal exposures is medium for all OES/COUs
because the same modeling approach was used for all OES/COUs. As described in Section 4.3 overall
confidence in the oral and dermal cancer slope factor underlying these risk estimates is medium to high.
As described in Section 3.3.1.2, the dermal exposure modeling methodology is supported by moderate
evidence, with model input parameters from literature sources, a European model, standard defaults
from the ChemSTEER User Guide ( 1015a). and 1,4-dioxane product concentration data from
process information. These sources range from slight to robust, depending on factors such as age and
applicability to OES/COU. The modeling is limited by the use of standard input parameters that are not
specific to 1,4-dioxane and a lack of variability in dermal exposure for different worker activities.
Therefore, EPA's overall confidence in the occupational dermal risk estimates is medium.
5.2.5.2 Risks from General Population Exposures through Drinking Water
Overall confidence in the overall distribution of risk estimates for drinking water exposures resulting
from facility releases is medium to high. Overall confidence in site-specific risk estimates for individual
facility releases varies both within and across OES, depending on the confidence in the source-specific
release data. As described in Section 4.3 overall confidence in the oral and dermal cancer slope factor
underlying these risk estimates is medium to high. As described in Section 3.3.2.1, the exposure
modeling methodology used for this analysis is supported by moderate evidence. It is designed to
estimate water concentrations expected at specific locations. Exposure estimates for this scenario are
based on some conservative assumptions about flow rates and release frequency and amount. A
summary of sources of flow and release data for facility release modeling is presented in Table 2-7.
Available monitoring data confirm that 1,4-dioxane is present in some surface water and drinking water,
though most of the available data were not collected near release sites are therefore not directly
comparable. The overall level of confidence depends on the source of OES/COU-specific release data,
as described in Appendix E.3 and summarized below:
Page 178 of 570
-------
• Overall confidence in drinking water exposure estimates is medium to high for OESs/COUs that
rely primarily on release data reported to DMR or to TRI via Form R. Most COUs/OESs are
included in this group.
• Overall confidence in drinking water exposure estimates is medium for OESs/COUs for which
release estimates are based on reporting to TRI via Form A. The Import and repackaging OES
releases used in this analysis are entirely based on Form A reporting of releases, and just under
half of the Industrial uses OES releases were reported via Form A.
• Overall confidence in drinking water exposure estimates is low to medium for OESs/COUs for
which release estimates are based on surrogate or modeled information.
Overall confidence in risk estimates for drinking water exposures resulting from DTD releases under
varying population and stream flow conditions is medium. As described in Section 4.3 overall
confidence in the oral and dermal cancer slope factor underlying these risk estimates is medium to high.
As described in Section 3.3.2.1, the exposure modeling methodology used for this analysis is supported
by robust evidence and is designed to provide a nationally representative distribution of estimated water
concentrations under varying conditions. Exposure estimates rely on estimated distributions of DTD
releases of consumer and commercial products for each COU.
Overall confidence in risk estimates for drinking water exposures resulting from hydraulic fracturing
releases is medium. As described in Section 4.3 overall confidence in the oral and dermal cancer slope
factor underlying these risk estimates is medium to high. As described in Section 3.3.2.1, the exposure
modeling methodology used for this analysis is supported by robust evidence and is designed to provide
a nationally representative distribution of estimated water concentrations under varying conditions.
Releases used as inputs in the model were estimated using Monte Carlo modeling that captures
variability across sites. However, the modeled exposure estimates are not directly tied to specific
releases at known locations, decreasing the strength of the evidence related to the representativeness of
the exposure estimates for actual exposures.
Overall confidence in risk estimates for drinking water exposures resulting from for aggregate surface
water concentrations predicted by probabilistic modeling varies across OES/COU. As described in
Section 4.3 overall confidence in the oral and dermal cancer slope factor underlying these risk estimates
is medium to high. As described in Section 3.3.2.1, the exposure modeling methodology used for this
analysis is supported by robust evidence and is designed to provide a nationally representative
distribution of estimated water concentrations under varying conditions. Available monitoring data
confirm that 1,4-dioxane is present in some surface water and drinking water—though most of the
available data were not collected near release sites and are therefore not directly comparable. In release
locations where monitoring data are available, case studies demonstrate general agreement between
modeled estimates and measured surface water concentrations. There is some uncertainty around the
proximity of releases to drinking water intake locations and the extent to which 1,4-dioxane is further
diluted prior to reaching intake locations. The characterization of downstream dilution presented in
5.2.2.1.2 for individual facility releases illustrates the extent to which downstream dilution may impact
overall risk estimates. The overall level of confidence in resulting exposure estimates depends on the
source of OES/COU-specific release data, as described in Appendix E.3:
• Overall confidence in drinking water exposure estimates is medium to high for OESs/COUs that
rely primarily on release data reported to DMR or to TRI via Form R. Most COUs/OESs are
included in this group.
• Overall confidence in drinking water exposure estimates is medium for OESs/COUs for which
release estimates are based on reporting to TRI via Form A. The Import and repackaging OES
Page 179 of 570
-------
releases used in this analysis are entirely based on Form A reporting of releases, and just under
half of the Industrial uses OES releases were reported via Form A.
• Overall confidence in drinking water exposure estimates is low to medium for OESs/COUs for
which release estimates are based on surrogate or modeled information.
5.2.5.3 Risks from General Population Exposures through Groundwater and Land
Disposal Pathways
Overall confidence in risk estimates for drinking water exposures resulting from disposal to landfills is
low to medium. As described in Section 4.3 overall confidence in the oral and dermal cancer slope
factor underlying these risk estimates is medium to high. As described in Section 3.3.2.2.1 the modeling
methodology is robust. However, the release information relied on as model input data is supported by
slight to moderate evidence. In addition, this drinking water exposure scenario relies on the assumption
that the estimated groundwater concentrations may occur in locations where groundwater is used as a
primary drinking water source. Although the substantial uncertainty around the extent to which these
exposures occur decreases overall confidence in the exposure scenario, this scenario represents a PESS
exposure.
Overall confidence in risk estimates for drinking waters resulting from disposal of hydraulic fracturing
waste is low to medium. As described in Section 4.3 overall confidence in the oral and dermal cancer
slope factor underlying these risk estimates is medium to high. As described in Section 3.3.2.2.2, the
modeling methodology is robust, and the release information relied on as model input data is supported
by moderate evidence. However, no monitoring data are available to confirm detection of 1,4-dioxane in
groundwater near hydraulic fracturing operations. This drinking water exposure scenario relies on the
assumption that the estimated groundwater concentrations may occur in locations where groundwater is
used as a primary drinking water source. Although the substantial uncertainty around the extent to which
these exposures occur decreases overall confidence in the exposure scenario, this scenario represents a
PESS exposure.
5.2.5.4 Risks from General Population Exposures through Air
Overall confidence in risk estimates for inhalation exposure resulting from industrial releases varies
across COUs. As described in Section 4.3, overall confidence in the cancer inhalation unit risk
underlying these risk estimates is high. As described in Section 3.3.3.1, the AERMOD modeling
methodology used for this analysis is robust. The exposure scenarios considered are most relevant to
long-term residents in fenceline communities. There is some uncertainty around the extent to which
people actually live and work around the specific facilities where risks are highest, decreasing overall
confidence in the exposure scenario. Overall confidence varies due to variable levels of confidence in
underlying release information used to estimate exposures. An OES-specific discussion of the
confidence in sources of release information is presented in Appendix E.5.4, but in general terms is
summarized below:
• Overall confidence in risk estimates is medium to high for OESs/COUs that rely primarily on
release data reported to TRI via Form R.
• Overall confidence in risk estimates is medium for OESs/COUs for which release estimates are
based on data reported to TRI via Form A.
• Overall confidence in risk estimates is low to medium for OESs/COUs for which release
estimates are based on surrogate or modeled information.
Overall confidence in risk estimates for inhalation exposures resulting for air concentrations modeled
based on releases from hydraulic fracturing operations is medium. As described in Section 4.3, overall
confidence in the cancer inhalation unit risk underlying these risk estimates is high. As described in
Page 180 of 570
-------
Section 3.3.3.2 the modeling methodologies used to estimate air concentrations are robust. The
distribution of air releases used as model input data were estimated using Monte Carlo modeling and
rely on assumptions. No air monitoring data were available to confirm detection of 1,4-dioxane is air
near hydraulic fracturing operations. Because the air concentrations underlying this analysis are based
on probabilistic modeling, they are not tied to specific locations that can be evaluated for land use
patterns. There is therefore substantial uncertainty around the extent to which people actually live and
work around the specific locations where risks are highest, decreasing overall confidence in the
exposure scenario.
Overall confidence in risk estimates from inhalation exposures resulting from industrial and institutional
laundries is medium. As described in Section 4.3, overall confidence in the cancer inhalation unit risk
underlying these risk estimates is high. As described in Section 3.3.3.2, the modeling methodologies are
robust. The distribution of air releases used as model input data were estimated using Monte Carlo
modeling and rely on assumptions. No air monitoring data were available to determine whether 1,4-
dioxane is detected near industrial and institutional laundry facilities. Because the air concentrations
underlying this analysis are based on probabilistic modeling, they are not tied to specific locations that
can be evaluated for land use patterns. There is therefore substantial uncertainty around the extent to
which people actually live and work around the specific locations where risks are highest, decreasing
overall confidence in the exposure scenario.
Page 181 of 570
-------
REFERENCES
ACC. (2023). Comments of the American Chemistry Council on Draft Supplement to the Risk
Evaluation for 1,4-Dioxane. Washington, DC. https://www.regulations.gov/comment/EPA-HQ-
QPPT-2022-0905-0052
Adamsoi lendra. S: Walker. KL. Jr.: Rauch. SR: Sengupta. S: Newell. CI. (2014). A multisite
survey to identify the scale of the 1,4-dioxane problem at contaminated groundwater sites.
Environ Sci Technol Lett 1: 254-258. http://dx.doi.( ;5QQQ92u
Adam soi i iH Pina. EA: Cartwright. AE: Rauch. SR: Anderson hU Uohr. T: Connor. JA. (2017). 1.4-
Dioxane drinking water occurrence data from the third unregulated contaminant monitoring rule.
Sci Total Environ 596-597: 236-245. http://dx.doic >/i.scitotenv.2017.04.085
Analvtica. (2015). Choosing an appropriate distribution [Website],
http://wiki.analytica.com/Choosine an appropriate distribution
(2012). Toxicological profile for 1,4 dioxane [ATSDR Tox Profile], Atlanta, GA: U.S.
Department of Health and Human Services, Public Health Service.
http://www.atsdr.cdc. gov/toxprofiles/tp.asp?id=955&tid=l 99
Baldwin. PE; Mavn (1998). A survey of wind speed in indoor workplaces. Ann Occup Hyg 42:
303-313. httdoi.org/10.1016/80003-4878(98)000 I '<
Black. RE: Hurley. FJ; Haverv. DC. (2001). Occurrence of 1,4-dioxane in cosmetic raw materials and
finished cosmetic products. J AOAC Int 84: 666-670.
Broughton. A: Sepulveda. A: Foster. K; Kruk. T; Nickel sen. M an. M; Mohr. TKG. (2019). 1.4-
Dioxane: Emerging technologies for an emerging contaminant. Remediation Journal 29: 49-63.
http://dx.doi.on 10 1002/rem.. I I
Brunswick County. (2022). Brunswick County, North Carolina: Annual reports/facilities information
[Website], https://www.bmnswickcountync.gov/utilities/annual-report.s-facilities-information/
CA Water Board. (2022). EDT library and water quality analyses data and download page [Website],
https://www.waterboards.ca.gov/drinking water/certlic/drinkingwater/EDTlibrary.html
Commonwealth of Massachusetts. (2018). Energy & Environmental Affairs Data Portal: Search for
drinking water [Website], https://eeaonline.eea.state.ma.us/portal#!/search/drinking-water
Cordone. L; Carlson t Haehn. W: Shangraw < Wilmoth 1 * (2016). Case study and retrospective:
Aerobic fixed film biological treatment process for 1,4-dioxane at the Lowry landfill superfund
site. Remediation Journal 27: 159-172. http://dx.doi.org/10.1002/rem.21502
nericas. (2023). Comments of the DAK Americas LLC on 1,4-Dioxane: Draft Supplement to
the TSCA Risk Evaluation Science Advisory Committee on Chemicals (SACC) Meeting.
Charlotte, NC: DAK Americas/Alpek Polyester. https://www.regulations.gov/comment/EPA-
HO-QPPT-2022-0905-0044
Davarani. SSH; Masoomi. L; Banitaba. IS ad. HRLZ; Sadeghi. O: Samiei. A. (2012). A New
Aluminium Hydroxide Coating on Fused Silica Fiber for the Determination of 1,4-Dioxane in
Surfactants and Detergents Using HS-SPME-GC. Chromatographia 75: 371-377.
http://dx.doi.on
Dow Chemical. (2023). Comments of Dow Chemical on 1,4-Dioxane: Draft Supplement to the TSCA
Risk Evaluation Science Advisory Committee on Chemicals (SACC) Meeting. Midland, MI.
https://www.regulations.gov/comment/EPA-HQ-OPPT-2022-0905-0059
Forkner. MW: Robson. J.H; Snellings. WM; Martin. [urphy. FH; Parsons. TE. (2004). Kirk-
Othmer Encyclopedia of Chemical Technology
Glycols. Hoboken, New Jersey: John Wiley & Sons, Inc.
http://dx.doi.on 10 1002/0 i i_ 8961.0520082: o s i-u »ol.pub2
Franklin Associates. (2020). Cradle-to-resin life cycle analysis of polyethylene terephthalate resin (final
revised report). Charlotte, NC: The National Association for PET Container Resources
(NAPCOR).
Page 182 of 570
-------
Fran; snnett. S; DeLeo. PC: Collatz. M; Kelly. K; Nekoomaram. J; Wieroniev. S. (2015).
Comments of the Adhesive and Sealant Council, the American Coatings Association, the
American Chemistry Council, the American Cleaning Institute, the Consumer Specialty Products
Association, and Waste Management on the 1,4-dioxane problem formulation and initial
assessment. (EPA-HQ-OPPT-2015-0078). Franz, C; Bennett, S; DeLeo, PC; Collatz, M; Kelly,
K; Nekoomaram, J; Wieroniey, S. https://arasp.americanchemistry.com/Media-Center/Industrv-
Comments-June-2015.pdf
Gillespie. BW: Chen. O: Reichert. H: Franzblau. A: Hedgemari «l 1 opkowski. J: Adriaens. P 1 K'mond.
A; Luksemburg. W: Garabrant DH. (2010). Estimating population distributions when some data
are below a limit of detection by using a reverse Kaplan-Meier estimator. Epidemiology 21: S64-
S70. http://dx.doi.org/10.1097/EDE.C ?f08
Godri Pollitt. KJ: Kim. JH: Peccia < «l limelech. M: Zhang N i iiarkoftaki. G: Hodges. <' u ;ker. I:
Huang. H: Deziel. NC: Murphy. K: Ishii. M: Johnson. CH: Boissevain. A; Q'Ke< .stas.
PT: Orlickv. 1): Thompsor iliou. V. (2019). 1,4-Dioxane as an emerging water
contaminant: State of the science and evaluation of research needs [Review], Sci Total Environ
690: 853-866. http://dx.doi.org/10.1016/i.scitotenv.i.O I" 0 I l'<
Harley. KG: Calderon. L: Nolan. JES: Maddalena. R: Russell. M: Roman. K: Mayo-Burgos. S: Cabrera.
J: Morga. N: Bradman. A. (2021). Changes in Latina women's exposure to cleaning chemicals
associated with switching from conventional to "green" household cleaning products: The
LUCIR intervention study. Environ Health Perspect 129: 97001.
http://dx.doi.oo 89/EHP8831
Harp (2017). Shoshone-Bannock exposure scenario for use in risk assessment. Harper, B.
https://siiperfimd.oregonstate.edu/sites/siiperfimd.oregonstate.edii/files/shoshone bannock scena
Harp-n Jlanco. I) (2009). Wabanaki traditional cultural lifeways exposure scenario. U.S.
Environmental Protection Agency and the Tribal Nations in Maine.
https://www.epa.gov/sites/default/files/2015-08/documents/ditca.pdf
Harp-'i ru U'Mi Tn i [arris. S: Abeyta. C: Kirschnei t (2002). The Spokane Tribe's multipathway
subsistence exposure scenario and screening level RME. Risk Anal 22: 513-526.
http://dx.doi.on 10 I I I I 0211 I '< 000 I
Harpci HI . Harding. AK: Waterhous < t tirris. SG. (2007). Traditional tribal subsistence exposure
scenario and risk assessment guidance manual. (Grant Number EPA-STAR-J1-R831046).
Corvallis, OR: Oregon State University Department of Public Health and Confederated Tribes of
the Umatilla Indian Reservation.
https://health.oregonstate.edu/sites/health.oregonstate.edu/files/research/pdf/tribal-
grant/exposure scenario and risk guidance manu;il J pdf
HCPA. (2023). Comments of the Household and Commercial Products Association (HCPA) on 1,4-
Dioxane: Draft Supplement to the TSCA Risk Evaluation; Science Advisory Committee on
Chemicals (SACC) Meeting. Washington, DC. https://www.regulations.gov/comment/EPA-HQ-
QPPT-2022-0905-0051
HHS. (2016). 1,4-Dioxane: CAS No. 123-91-1. In Report on Carcinogens (14th ed.). Research Triangle
Park, NC: National Toxicology Program, Department of Health and Human Services.
https://ntp.niehs.nih.gov/ntp/roc/content/profiles/dioxane.pdf
Hills. B: Klincewicz. S: Blade. LM; Sack. D (1989). Health hazard evaluation report no. HETA-87-
367-1987, BMY Corporation, A Division of Harsco Corporation, York, Pennsylvania. (HETA-
87-367-1987). Cincinnati, OH: National Institue for Occupational Safety and Health.
https://www.cdc.gov/niosh/hhe/reports/pdfs/1987-0367-1987.pdf
Huntsman. (2023). Comments of the Huntsman International, LLC on 1,4-Dioxane: Draft Supplement to
the TSCA Risk Evaluation Science Advisory Committee on Chemicals (SACC) Meeting. The
Page 183 of 570
-------
Woodlands, TX: Huntsman International, LLC. https://www.reeutations.gov/commentZ O-
OPPT-2022-0905-0046
Jackson. LE; Lemk (2019). Evidence for natural attenuation of 1,4-dioxane in a glacial aquifer
system. Hydrogeology Journal v. 27: 3009-3024. http://dx.doi. )28~6
Kano. H; LJmeda. Y; Kasai. T: Sasaki. T: Matsumoto. M; Yamazaki. K; Nagano. to. H;
Fukushima. S. (2009). Carcinogenicity studies of 1,4-dioxane administered in drinking-water to
rats and mice for 2 years. Food Chem Toxicol 47: 2776-2784.
http://dx.doi.on 10 101 i.fct.2009.0S 01J
Kasai i t ano. U. 1'ineda N 'vtsaki « U awa. N: Nishizawa. T: Nagano. K; Arito. H; Nagashima. H;
Fukushima. S. (2009). Two-year inhalation study of carcinogenicity and chronic toxicity of 1,4-
dioxane in male rats. Inhal Toxicol 21: 889-897. http://dx.doi.Hv 10 IQSQ/0895837080262' \ h1
Kasti iller. MA. (2006). Kinetics of finite dose absorption through skin 2: Volatile
compounds. J Pharm Sci 95: 268-280. http://dx.doi.org, '2/ips.20497
Lee. CS; Venkatesan \ U \\ qlkei t i\\ ^bler. CI. (2020). Impact of groundwater quality and
associated byproduct formation during UV/hydrogen peroxide treatment of 1,4-dioxane. Water
Res 173: 115534. http://dx.doi.org/10J016/i.watres.2010 II ¦;?; I
Lestei \ , rcttct L 1'imrman. EM; Sitterlev. i \ \ oral < \ \tU> Linden. KG. (2015).
Characterization of hydraulic fracturing flowback water in Colorado: Implications for water
treatment. Sci Total Environ 512-513: 637-644.
http://dx.doi.org/10.1016/i .scitotenv.l 01 01 0 l'<
Li. M; Mathh'u < \ \ i un^'nza. S: Desn \ rle. Z; Zhou j, \lvare (2013). Widespread
distribution of soluble di-iron monooxygenase (SDIMO) genes in Arctic groundwater impacted
by 1,4-dioxane. Environ Sci Technol 47: 9950-9958. http://dx.doi.org/10.102 l/es402228x
Li. M; Van Orden. ET; Devrieh 0,1. Xiong. Z; Hinchee. R; Alvare Pi (2015). Bench-scale
biodegradation tests to assess natural attenuation potential of 1,4-dioxane at three sites in
California. Biodegradation 26: 39-50. http://dx.doi.<
Lin. WT; Chen. WL; Cheng. \ iang. H.C; Tsai. SW. (2017). Determining the residual
characteristics of alkylphenols, arsenic, and lead as well as assessing the exposures of 1,4-
dioxane from household food detergents. J AOAC Int 100: 1086-1093.
http://dx.doi.on 10 ^ iO laoacint I 0 '>S0
Makino. R; Kawasaki. H; Kishimoto. no. M; Nakanishi. J. (2006). Estimating health risk from
exposure to 1,4-dioxane in Japan. Environ Sci 13: 43-58.
Marquart. H; Franken. R; Goede. H; Fransman. W; Schinkel. J. (2017). Validation of the dermal
exposure model in ECETOC TRA. Ann Work Expo Health 61: 854-871.
http://dx.doi.oo 93/annweh/wxx059
Mattie. PL ^uchei i \\ t .titer. AL; Stoffregen. DE; Reboulet. JE. (2012). Acute inhalation toxicity
study of 1, 4-dioxane in rats (Rattus norvegicus). GRA and I: 29.
McDaniel. MP; DesLaurie' (2015). Kirk-Othmer encyclopedia of chemical technology: Ethylene
polymers, HDPE. Hoboken, NJ: Wiley-Interscience.
http://dx.doi.on 10 1002/0 i i. 8961.0809070K I IO'M'M'\i01 ,pub3
Millard. SP. (2013). EnvStats: An R package for environmental statistics. New York, NY: Springer.
http://dx.doi.on 10 100 978-1 I I I s i -n I
Minitab. (2022). A comparison of the Pearson and Spearman correlation methods [Website],
https://support.minitab.com/en~us/minitab/21/help~and~how~to/statistics/basic~
statistics/supporting~topics/correlation~and~covariance/a~comparison~of~the~pearson~and~
spearman-correlation-methods/
Mohi seppi. WH. (2010). Case Studies of 1,4-Dioxane Releases, Treatment, and Drinking
Water Contamination. In T Mohr; B Diguiseppi; JK Anderson; JW Hatton (Eds.), (2nd ed., pp.
369-420). Boca Raton, FL: CRC Press, http://dx.doi.on 01/EBK1566706629~c8
Page 184 of 570
-------
K iciner. E. (2020). Kirk-Othmer Encyclopedia of Chemical Technology: Detergency and
detergents [Encyclopedia], New York, NY: John Wiley & Sons.
http://dx.doi.on 10 1002/0 i i. 8961.0405200512:^1 II U01.pub3
NIOSH. (1987). NIOSH Guide to Industrial Respiratory Protection (pp. 87-1 16). (NIOSH/00177545).
Anonymous.
NTP. (1986). NTP Toxicology and carcinogenesis studies of dichloromethane (methylene chloride)
(CAS No. 75-09-2) in F344/N rats and B6C3F1 mice (inhalation studies). 306: 1-208.
https://hero.epa.gov/hero/index.cfm?action=search.view&reference id=10368680NYDEC.
(2023). List of consumer products for which companies have submitted waivers to exceed 1 ppm
1,4-dioxane, including their concentrations. Albany, NY.
Oberle. W. (2015). Monte Carlo simulations: Number of iterations and accuracy. (ARL-TN-0684).
Aberdeen Proving Ground, MD: US Army Research Laboratory.
https://apps.dtic.mil/sti/pd: If
OECD. (2009). Emission scenario documents on coating industry (paints, lacquers and varnishes).
(JT03267833). Paris, France.
http://www.oecd. org/officialdocuments/publicdisplavdocumentpdf/?cote=env%20/im/mono(200
9)24& docl an guage=em
OECD. (201 la). Emission scenario document on coating application via spray-painting in the
automotive refinishing industry. In OECD Series on Emission Scenario Documents No 11.
(ENV/JM/MONO(2004)22/REV1). Paris, France: Organization for Economic Co-operation and
Development.
http://www.oecd.org/officialdocuments/publicdisplavdocumentpdf/?cote=env/im/mono(2004)22/
rev 1& docl an guage=em
OECD. (201 lb). Emission scenario document on the chemicals used in water-based washing operations
at industrial and institutional laundries. (JT03304937). Paris, France, https://www.oecd-
itibrary.org/docserver/9789264220973~
en.pdf?expi: 50&id=id&accname=guest&checksum=B8C714E510603EB565B89A
8AE6A.24C14
OECD. (2015). Emission scenario document (ESD) on industrial use of industrial cleaners.
(JT03373556). Paris, France.
http://www.oecd. org/officialdocuments/publicdisplaydocumentpdf/?cote=ENV/JM/MONO(2015
)3 & docl an guage=en
OECD. (2017). Emission scenario document (ESD) on the use of textile dyes. In Emission Scenario
Document, Series No 36. (ENV/JM/MONO(2015)50). Paris, France: OECD, Inter-Organization
Programme for the Sound Management of Chemicals (IOMC).
http://www.oecd.org/chemicalsafetv/risk-assessment/emissionscenariodocuments.htm
OECD. (2020). Emission scenario document on chemical additives used in automotive lubricants. Paris,
France.
Oracle. (2017). Oracle® Crystal Ball User's Guide; 2017 (Release 11.1.2.4): Selecting probability
distributions. Austin, TX. https://docs.oracle.com/cd/E57185 01/< \ < 51 GZapas03.html
Otto. M; Nagaraia. S. (2007). Treatment technologies for 1,4-Dioxane: Fundamentals and field
applications. Remediation Journal 17: 81-88. http://dx.doi.org/10.1002/rem.201 '< ?
P&G. (2023). Comments of The Procter & Gamble Company on 1,4-Dioxane: Draft Supplement to the
TSCA Risk Evaluation Science Advisory Committee on Chemicals (SACC) Meeting. Cincinnati,
OH. https://www.regulations.gov/comment/1 Q-QPPT-2022-0905-0053
Palisade. (2015a). Palisade Knowledge Base: 2.6. How many iterations do I need? [Website],
https://kb. palisade, com/index. php?pg=kb.page&id=125
Palisade. (2015b). Palisade Knowledge Base: 2.19. Random number generators [Website],
https://kb. palisade, com/index. php?pg=kb.page&id=25
Page 185 of 570
-------
Palisade. (2018). Palisade Knowledge Base: 2.18. Latin Hypercube versus Monte Carlo sampling
[Website], https://kb.palisade.com/index.php?pe=kb.paee&id=28
Palisade. (2022a). Define correlation matrix [Website], https://help.palisade.eom/v8/en/@RISK/l-
Define/5-Correlation/Define-Correlation-Matrix.htm
PEI Associates. (1988). Releases during cleaning of equipment. Washington, DC: U.S. Environmental
Protection Agency, Office of Pesticides and Toxic Substances.
https://ofmpiib.epa.eov/apex/euideme ext/guideme/file/releases%2Qduring%2Qcleaning%2Qof%
20equipmetrt.pdf
Putz. VR; Johnson. BL; Setzei (1979). A comparative study of the effects of carbon monoxide and
methylene chloride on human performance. J Environ Pathol Toxicol 2: 97-112.
https://hero.epa.eov/hero/index.cfm?action=search.view&reference id=l0626648Saraii. M;
Shirvani. N. (2017). Determination of residual 1,4-dioxane in surfactants and cleaning agents
using headspace single-drop microextraction followed by gas chromatography-flame ionization
detection. Int J Cosmet Sci 39: 36-41. http://dx.doi.ore/10.1 1 1 !/icv l. I
Stefl. ee. KF. (2014). Kirk-Othmer Encyclopedia of Chemical Technology: Antifreezes and
deicing fluids [Encyclopedia], New York, NY: John Wiley & Sons.
http://dx.doi.on 10 1002/0 I i: '4 x mm I i200919200506.a01.pub2
US JILS (2016). May 2016 Occupational Employment and Wage Estimates: National Industry-
Specific Estimates [Website], http://www.bls.eov/oes/tables.htm
(2022). Labor Force Statistics from the Current Population Survey: Household data, Annual
averages, 11. Employed persons by detailed occupation, sex, race, and Hispanic or Latino
ethnicity [Numbers in thousands] [Website], https://www.bls.eov/cps/cpsaatl 1 .htm
isus Bureau. (2015). Statistics of U.S. Businesses (SUSB).
https://www.census.eov/data/tables/2015/econ/susb/2015-susb-annual.html
(1988). National survey of solid waste (municipal) landfill facilities [EPA Report],
(EPA/530-SW88-034). Washington, DC: Office of Solid Waste and Emergency Response, U.S.
Environmental Protection Agency.
(1991). Chemical engineering branch manual for the preparation of engineering assessments.
Volume I. Ceb Engineering Manual. Washington, DC: Office of Pollution Prevention and
Toxics, US Environmental Protection Agency.
(1994a). Guidelines for Statistical Analysis of Occupational Exposure Data: Final. United
States Environmental Protection Agency :: U.S. EPA.
(1994b). Methods for derivation of inhalation reference concentrations and application of
inhalation dosimetry [EPA Report], (EPA600890066F). Research Triangle Park, NC.
https://cfpub.epa.gov/ncea/risk/recordisplay.cfiro?deid=yi993«
-------
https://cfpub.epa.gov/ncea/risk/recordisplav.cfrn?deid=24465Q&CFID=5^ CFT0KEN=
17139189
(2015a). ChemSTEERuser guide - Chemical screening tool for exposures and environmental
releases. Washington, D.C. https://www.epa.gov/sites/production/files/2Q15-
05/docum ents/ user gui de.pdf
(2015b). TSCA work plan chemical problem formulation and initial assessment. 1,4-
Dioxane. (740-R1-5003). Washington, DC: Environmental Protection Agency, Office of
Chemical Safety and Pollution Prevention.
http://nepis.epa. gov/Exe/ZyPURL. cei?Dockey=P 1OOMDC1. TXT
(2016a). Hydraulic fracturing for oil and gas: Impacts from the hydraulic fracturing water
cycle on drinking water resources in the United States [EPA Report], (EPA/600/R-16/236F).
Washington, DC. https://cfpub.epa.gov/ncea/hfstudy/recordisplay.cfm?deid=332990
(2016b). Public database 2016 chemical data reporting (May 2017 release). Washington, DC:
US Environmental Protection Agency, Office of Pollution Prevention and Toxics.
https://www.epa.gov/chemical-data-reporting
(2017a). Information on the various spray polyurethane foam products. U.S. Environmental
Protection Agency, Design for the Environment.
https://www.epa.gov/sites/prodiiction/files/2015-08/dociiments/spf product types.pdf
(2017b). Preliminary information on manufacturing, processing, distribution, use, and
disposal: 1,4-dioxane. https://www.regulations.gov/document?D=< " \ SIQ-C< JO h- 0 23-
0003
(2017c). Technical fact sheet 1,4-dioxane [EPA Report], (EPA-505/F-17/011). Office of
Land and Emergency Management, U.S. EPA.
https://nepis.epa. gov/Exe/ZyPIJRL.cgi?Dockey=P 100THWG.txt
(2018a). Application of spray polyurethane foam insulation - Generic scenario for estimating
occupational exposures and environmental releases -Draft. Washington, DC.
https://www.epa.gov/tsca-screening-tools/using-predictive-methods-assess-exposure-and-fate-
under-tsca
(2018b). Application of spray polyurethane foam insulation - Generic scenario for estimating
occupational exposures and environmental releases -Methodology review draft, revised.
https://www.epa.gov/tsca-screening-tools/chemsteer-chemical-screening-tool-exposures-and-
environmental-releases#genericscenarios
(2018c). Application of systematic review in TSCA risk evaluations. (740-P1-8001).
Washington, DC: U.S. Environmental Protection Agency, Office of Chemical Safety and
Pollution Prevention, https://www.epa.gov/sites/prodiiction/files/2018-
06/documents/final application of sr in ts«
(2018d). User's Guide for the AMS/EPA Regulatory Model (AERMOD). (EPA Document
Number: EPA-454/B-18-001). U.S. EPA.
(2019a). Exposure factors handbook chapter 3 (update): Ingestion of water and other select
liquids [EPA Report], (EPA/600/R-18/259F). Washington, DC.
https://cfpub. epa.gov/ncea/efp/recordisplay. cfm?deid=343661
(2019b). User's Guide: Integrated Indoor-Outdoor Air Calculator (IIOAC). Washington, DC:
U.S. EPA. https://hero.epa.gov/hero/index.cfm?action=search.view&reference id=6
EPA. (2020b). Final scope of the risk evaluation for 4,4'-(l-Methylethylidene)bis[2, 6-
dibromophenol] (TBBPA); CASRN 79-94-7. (EPA 740-R-20-008). Washington, DC: Office of
Chemical Safety and Pollution Prevention. https://www.regiilations.gov/dociiment/EPA-HQ-
OPPT-2018-0462-0047
Page 187 of 570
-------
(2020c). Risk evaluation for 1,4-dioxane (CASRN: 123-91-1) [EPA Report], (EPA-740-R1-
8007). Washington, DC: Office of Chemical Safety and Pollution Prevention.
https://www.reeiilations.eov/dociiment/EPA-HQ-OPPT-2019-0238-0092
(2021a). Draft systematic review protocol supporting TSCA risk evaluations for chemical
substances, Version 1.0: A generic TSCA systematic review protocol with chemical-specific
methodologies. (EPA Document #EPA-D-20-031). Washington, DC: Office of Chemical Safety
and Pollution Prevention, https://www.reeutations. gov/document/> " \ i IQ-OPPT-2Q. I Oil i •
0005
(2021b). Use of additives in plastic compounding - Generic scenario for estimating
occupational exposures and environmental releases (Revised draft) [EPA Report], Washington,
DC: Office of Pollution Prevention and Toxics, Risk Assessment Division.
(2022a). Chemicals used in furnishing cleaning products - Generic scenario for estimating
occupational exposures and environmental releases (revised draft). Washington, DC: Office of
Pollution Prevention and Toxics.
(2022b). Commercial use of automotive detailing products - Generic scenario for estimating
occupational exposures and environmental releases (Methodology review draft) [EPA Report],
Washington, DC: U.S. Environmental Protection Agency, Office of Pollution Prevention and
Toxics, Risk Assessment Division.
https://hero.epa.eov/hero/index.cfm?action=search.view&reference id=l048047211.8. EPA.
(2022d). Draft TSCA screening level approach for assessing ambient air and water exposures to
fenceline communities (version 1.0) [EPA Report], (EPA-744-D-22-001). Washington, DC:
Office of Chemical Safety and Pollution Prevention, U.S. Environmental Protection Agency.
https://www.epa.gov/svstem/files/documents/2022-01/draft-fenceline-report sacc.pdf
(2022e). Emission scenario document on chemicals used in hydraulic fracturing (draft). In
OECD Environmental Health and Safety Publications, Series on Emission Scenario Documents.
Paris, France: Organization for Economic Co-operation and Development.
https://hero.epa.eov/hero/index.cfm?action=search.view&reference id=10480474U.S. EPA.
(2022i). Use of laboratory chemicals - Generic scenario for estimating occupational exposures
and environmental releases (Revised draft generic scenario) [EPA Report], Washington, DC:
U.S. Environmental Protection Agency, Office of Pollution Prevention and Toxics, Existing
Chemicals Risk Assessment Division.
(2024a). Supplement to the Risk Evaluation for 1,4-Dioxane - Supplemental Information
File: 1,4-Dioxane Supplemental Information Files EWISRD-XL-R probabilistic model code.
Washington, DC: Office of Pollution Prevention and Toxics, Office of Chemical Safety and
Pollution Prevention, https://www.regulations.gov/docket/1 Q-OPPT-2016-0723
(2024b). Supplement to the Risk Evaluation for 1,4-Dioxane - Supplemental Information
File: Air Exposure and Risk Estimates for 1,4-Dioxane Emissions from Hydraulic Fracturing
Operations. Washington, DC: Office of Pollution Prevention and Toxics, Office of Chemical
Safety and Pollution Prevention, http s: //www, regul ati on s. gov/docket/< " \ S lQ-OPPT-201 0 '23
(2024c). Supplement to the Risk Evaluation for 1,4-Dioxane - Supplemental Information
File: Air Exposures and Risk Estimates for Institutional Laundry. Washington, DC: Office of
Pollution Prevention and Toxics, Office of Chemical Safety and Pollution Prevention.
https://www.reeiilations.eov/docket/EPA-HQ-0 23
(2024d). Supplement to the Risk Evaluation for 1,4-Dioxane - Supplemental Information
File: Air Exposures and Risk Estimates for Multi-Year Analysis. Washington, DC: Office of
Pollution Prevention and Toxics, Office of Chemical Safety and Pollution Prevention.
https://www.reeiilations.eov/docket/EPA-HQ-Q 23
(2024e). Supplement to the Risk Evaluation for 1,4-Dioxane - Supplemental Information
File: Air Exposures and Risk Estimates for Single Year Analysis. Washington, DC: Office of
Page 188 of 570
-------
Pollution Prevention and Toxics, Office of Chemical Safety and Pollution Prevention.
https://www.reeiilations.eov/docket/EPA-HQ-0 23
(2024f). Supplement to the Risk Evaluation for 1,4-Dioxane - Supplemental Information
File: Drinking Water Exposure and Risk Estimates for 1,4-Dioxane Land Releases to Landfills.
Washington, DC: Office of Pollution Prevention and Toxics, Office of Chemical Safety and
Pollution Prevention, https://www.reeulations.gov/docket/1 Q-OPPT-2016-0723
(2024g). Supplement to the Risk Evaluation for 1,4-Dioxane - Supplemental Information
File: Drinking Water Exposure and Risk Estimates for 1,4-Dioxane Land Releases to Surface
Impoundments. Washington, DC: Office of Pollution Prevention and Toxics, Office of Chemical
Safety and Pollution Prevention, https://www.reeutations.gov/docket/< " \ s IQ-P
-------
(2024q). Supplement to the Risk Evaluation for 1,4-Dioxane - Supplemental Information
File: EWISRDXL ColumbiaTN Case Study. Washington, DC: Office of Pollution Prevention
and Toxics, Office of Chemical Safety and Pollution Prevention.
https://www.reeiilations.eov/docket/EPA-HQ-0 13_
(2024r). Supplement to the Risk Evaluation for 1,4-Dioxane - Supplemental Information
File: EWISRDXL Liverpool OH Case Study. Washington, DC: Office of Pollution Prevention
and Toxics, Office of Chemical Safety and Pollution Prevention.
https://www.reeiilations.eov/docket/EPA-HQ-0 13_
(2024s). Supplement to the Risk Evaluation for 1,4-Dioxane - Supplemental Information
File: IIOAC Modeling and Results Files. Washington, DC: Office of Pollution Prevention and
Toxics, Office of Chemical Safety and Pollution Prevention.
https://www.reeiilations.eov/docket/EPA-HQ-0 13_
(2024t). Supplement to the Risk Evaluation for 1,4-Dioxane - Supplemental Information
File: Land Use Analysis for Ambient Air. Washington, DC: Office of Pollution Prevention and
Toxics, Office of Chemical Safety and Pollution Prevention.
https://www.reeiilations.eov/docket/EPA-HQ-0 13_
(2024u). Supplement to the Risk Evaluation for 1,4-Dioxane - Supplemental Information
File: Occupational Exposure and Risk Estimates. Washington, DC: Office of Pollution
Prevention and Toxics, Office of Chemical Safety and Pollution Prevention.
https://www.reeulations.eov/docket/EPA-H< 13_
(2024v). Supplement to the Risk Evaluation for 1,4-Dioxane - Supplemental Information
File: Probabilistic Surface Water Model (EWISRD-XL-R) Model Files. Washington, DC: Office
of Pollution Prevention and Toxics, Office of Chemical Safety and Pollution Prevention.
https://www.reeiilations.eov/docket/EPA-HQ-0 13_
(2024w). Supplement to the Risk Evaluation for 1,4-Dioxane - Supplemental Information
File: WQP Processed Surface Water Data. Washington, DC: Office of Pollution Prevention and
Toxics, Office of Chemical Safety and Pollution Prevention.
https://www.reeiilations.eov/docket/EPA-HQ-0 13_
(2024x). Supplement to the Risk Evaluation for 1,4-Dioxane - Systematic Review
Supplemental File: Data Quality Evaluation and Data Extraction Information for Environmental
Release and Occupational Exposure. Washington, DC: Office of Pollution Prevention and
Toxics, Office of Chemical Safety and Pollution Prevention.
https://www.reeiilations.eov/docket/EPA-HQ-0 23
(2024y). Supplement to the Risk Evaluation for 1,4-Dioxane - Systematic Review
Supplemental File: Data Quality Evaluation Information for General Population, Consumer, and
Environmental Exposure. Washington, DC: Office of Pollution Prevention and Toxics, Office of
Chemical Safety and Pollution Prevention. https://www.regulations.eov/docket/EPA-HQ-OPPT-
2 23
Wala-Jerzykiewicz. A; Szymanowski. J. (1998). Headspace gas chromatography analysis of toxic
contaminants in ethoxylated alcohols and alkylamines. Chromatographia 48: 299-304.
http://dx.doi.on !6
Youm < 1} \S l ehrim P < »rva1li Kc< Riniet. RL (1976). 1.4-Dioxane and beta-
hydroxyethoxyacetic acid excretion in urine of humans exposed to dioxane vapors. Toxicol Appl
Pharmacol 38: 643-646. http://dx.doi.ore/10.1016/00 11 008X(76)90195-2
Yoimv 11 * Braun. WH; Rampy. LW; Chenoweth.
-------
APPENDICES
Appendix A KEY ABBREVIATIONS AND ACRONYMS
30Q5
Lowest 30-day average flow that occurs in a 5-year period
7Q10
Lowest 7-day average flow that occurs in a 10-year period
AC
Acute concentrations
ACA
American Coatings Association
ACGM
American Conference of Governmental Industrial Hygienists
ADC
Average daily concentration
ADD
Average daily dose
ADR
Acute Dose Rate
AEC
Acute Exposure Concentration
APF
Assigned protection factor
ASTDR
Agency for Toxic Substances and Disease Registry
BHET
Bishydroxyethyl terephthalate
BLS
Bureau of Labor Statistics
BMD
Benchmark dose
BMDL
Benchmark dose level
CASRN
Chemical Abstracts Service Registry Number
CDR
Chemical Data Reporting
CEB
Chemical Engineering Branch
CEHD
Chemical Exposure Health Data
CERCLA
Comprehensive Environmental Response, Compensation and Liability Act
CFR
Code of Federal Regulations
COU
Condition of use
CSF
Cancer slope factor
CT
Central tendency
CWA
Clean Water Act
DAF
Dilution attenuation factor
DHHR
Department of Health and Human Services
DIY
Do-it-yourself
DMR
Discharge monitoring report
DRAS
Delisting Risk Assessment Software
DTD
Down-the-drain
DWI
Drinking water intake
DWT
Drinking water treatment
ECHA
European Chemicals Agency
ECHO
Environmental Compliance History Online database
EPA
Environmental Protection Agency
EPACMTP
Environmental Protection Agency Composite Model for Leachate Migration with
Transformation Products
EPCRA
Emergency Planning and Community Right-to-Know Act
EROM
Enhanced Runoff Method (database)
ESD
Emission Scenario Document
FRS
Facility Registry Service
FT
Full-text (screening)
GS
Generic Scenario
HAWC
Health Assessment Workplace Collaborative (tool)
Page 191 of 570
-------
HE
High-end
HEC
Human equivalent concentration
HED
Human equivalent dose
HERO
Health and Environmental Research Online (EPA Database)
HHE
Health hazard evaluation
HSDB
Hazardous Substances Data Bank
ICIS
Integrated Compliance Information System
IFC
Industrial Function Category
IIOAC
Integrated Indoor/Outdoor Air Calculator (EPA)
IRIS
Integrated Risk Information System
IUR
Inhalation unit risk
Koc
Soil organic carbon: water partitioning coefficient
Kow
Octanol: water partition coefficient
LADC
Lifetime Average Daily Concentration
LADD
Lifetime Average Daily Dose
LOAEC
Lowest-observed-adverse-effect-concentration
LOD
Limit of detection
Log Koc
Logarithmic organic carbon: water partition coefficient
Log Kow
Logarithmic octanol: water partition coefficient
LOQ
Limit of quantitation
MLD
Million liters per day
MOE
Margin of exposure
MRD
Methodology Review Draft (EPA)
MW
Molecular weight
NAICS
North American Industry Classification System
ND
Non-detect
NEI
National Emissions Inventory
NIOSH
National Institute for Occupational Safety and Health
NOAEL
No-observed-adverse-effect-level
NPDES
National Pollutant Discharge Elimination System
OAQPS
Office of Air Quality Planning and Standards
OCF
One-component foam
OCSPP
Office of Chemical Safety and Pollution Prevention
OD
Operating days
OECD
Organisation for Economic Co-operation and Development
OEHHA
Office of Environmental Health Hazard Assessment
OES
Occupational exposure scenario
ONU
Occupational non-user
OPPT
Office of Pollution Prevention and Toxics
OSHA
Occupational Safety and Health Administration
PBZ
Personal breathing zone
PECO
Population, exposure, comparator, and outcome
PEL
Permissible exposure limit
PESS
Potentially exposed or susceptible subpopulations
PET
Polyethylene terephthalate
PF
Protection factor
PNOR
Particulates not otherwise regulated
POD
Point of departure
POTW
Publicly owned treatment works (wastewater)
Page 192 of 570
-------
PPE
Personal protective equipment
PV
Production volume
PWS
Public water system
QA/QC
Quality assurance/quality control
QE
NHDPlus V2.1 flow values representing "the best EROM estimate of actual mean flow"
RE
(2020 RE) Risk Evaluation
RCRA
Resource Conservation and Recovery Act
REACH
Registration, Evaluation, Authorisation and Restriction of Chemicals (European Union)
RESO
Receptors, exposure, setting or scenario, and outcomes
SACC
Science Advisory Committee on Chemicals
SDS
Safety data sheet
SDWA
Safe Drinking Water Act
SHEDS-HT
Stochastic Human Exposure and Dose Simulation-High Throughput
SIC
Standard Industrial Classification
SOC
Standard Occupational Classification
SpERC
Specific Environmental Release Categories
SPF
Spray polyurethane foam
STORET
STOrage and RETrieval and Water Quality exchange
SUSB
Statistics of United States Businesses
SWIFT
Sciome Workbench for Interactive Computer-Facilitated Text-mining
TIAB
Title/abstract (screening)
TRI
Toxics Release Inventory
TSCA
Toxic Substances Control Act
TWA
Time-weighted average
UCMR
Unregulated Contaminant Monitoring Rule
U.S.
United States
USGS
U.S. Geological Survey
VOC
Volatile organic compound
VP
Vapor pressure
WQP
Water Quality Portal
WWT
Wastewater treatment
Page 193 of 570
-------
Appendix B LIST OF SUPPLEMENTAL DOCUMENTS
Associated Systematic Review Data Quality Evaluation and Data Extraction Documents - Provide
additional detail and information on individual study evaluations and data extractions including criteria
and data quality results.
Supplement to the Risk Evaluation for 1,4-Dioxane - Systematic Review Supplemental File: Data
Quality Evaluation and Data Extraction Information for Environmental Release and
Occupational Exposure - Provides a compilation of tables for the data extraction and data
quality evaluation information for 1,4-dioxane. Each table shows the data point, set, or
information element that was extracted and evaluated from a data source that has information
relevant for the evaluation of environmental release and occupational exposure (
2024x). This supplemental file may also be referred to as the 1,4-Dioxane Supplement to the
Risk Evaluation Data Quality Evaluation and Data Extraction Information for Environmental
Release and Occupational Exposure.
Supplement to the Risk Evaluation for 1,4-Dioxane - Systematic Review Supplemental File: Data
Quality Evaluation Information for General Population, Consumer, and Environmental
Exposure - Provides a compilation of tables for the data quality evaluation information for 1,4-
dioxane. Each table shows the data point, set, or information element that was evaluated from a
data source that has information relevant for the evaluation of general population, consumer, and
environmental exposure ( |u). This supplemental file may also be referred to as the
1,4-Dioxane Supplement to the Risk Evaluation Data Quality Evaluation Information for General
Population, Consumer, and Environmental Exposure.
Supplement to the Risk Evaluation for 1,4-Dioxane - Systematic Review Supplemental File: Data
Extraction Information for General Population, Consumer, and Environmental Exposure -
Provides a compilation of tables for the data extraction for 1,4-dioxane. Each table shows the
data point, set, or information element that was extracted from a data source that has information
relevant for the evaluation of general population, consumer, and environmental exposure (U.S.
24h). This supplemental file may also be referred to as the 1,4-Dioxane Supplement to
the Risk Evaluation Data Extraction Information for General Population, Consumer, and
Environmental Exposure.
Associated Supplemental Information Files - Provide additional details and information on exposure,
hazard, and risk assessments.
Supplement to the Risk Evaluation for 1,4-Dioxane - Supplemental Information File:
Environmental Releases to Air - Provides a summary of stack and fugitive air emissions for each
occupational exposure scenario (OES) in the 1,4-Dioxane Supplemental Risk Evaluation (Ij
24k).
Supplement to the Risk Evaluation for 1,4-Dioxane - Supplemental Information File:
Environmental Releases to Landfor all OES Except Disposal - Provides a summary of land
releases for each 1,4-dioxane OES except for the Disposal OES ( A. 20241).
Supplement to the Risk Evaluation for 1,4-Dioxane - Supplemental Information File:
Environmental Releases to Landfor the Disposal OES - This spreadsheet contains a summary of
land releases for the 1,4-dioxane occupational Disposal OES (U.S. EPA. 2024m).
Page 194 of 570
-------
Supplement to the Risk Evaluation for 1,4-Dioxane - Supplemental Information File:
Environmental Releases to Water for OES without TRI or DMR data - Provides a summary of
direct and indirect water releases for each 1,4-dioxane OES for which Toxics Release Inventory
(TRI) and Discharge Monitoring Report (DMR) data were not available ( ).
Supplement to the Risk Evaluation for 1,4-Dioxane - Supplemental Information File:
Environmental Releases to Water for OES with TRI and DMR - Provides a summary of direct
and indirect water releases for each 1,4-dioxane OES for which TRI or DMR data were available
( 2024n).
Supplement to the Risk Evaluation for 1,4-Dioxane - Supplemental Information File:
Occupational Exposure and Risk Estimates - Provides a summary of occupational exposures and
risks estimated for all conditions of use (COUs; (including those evaluated in this supplemental
evaluation as well as those previously evaluated in the 2020 RE) ( I4u).
Supplement to the Risk Evaluation for 1,4-Dioxane - Supplemental File Folder: Environmental
Release and Occupational Exposure Modeling ( 2024D.
Supplement to the Risk Evaluation for 1,4-Dioxane - Supplemental Information File: Drinking
Water Exposure and Risk Estimates for 1,4-Dioxane Release to Surface Water from Individual
Facilities - Provides water concentrations estimated from individual facility releases reported to
TRI and calculates corresponding drinking water exposures and risks ( 24h).
Supplement to the Risk Evaluation for 1,4-Dioxane - Supplemental Information File: Drinking
Water Exposure and Risk Estimates for 1,4-Dioxane Surface Water Concentrations Predicted
with Probabilistic Modeling - Provides water concentrations estimated by probabilistic modeling
for DTD releases, disposal of hydraulic fracturing waste to surface water, and for aggregate
concentrations estimated downstream of industrial release sites; calculates corresponding
drinking water exposures and risks. This file also calculates drinking water exposures and risks
estimated from drinking water monitoring data ( ).
Supplement to the Risk Evaluation for 1,4-Dioxane - Supplemental Information File:
EWISRDXL BrunswickCountyNC Case Study - Provides the Excel workbook file for the
Brunswick County, NC surface water case study ( 024p).
Supplement to the Risk Evaluation for 1,4-Dioxane - Supplemental Information File:
EWISRDXL ColumbiaTN Case Study - Provides the Excel workbook file for the Columbia, TN
surface water case study ( 2024q).
Supplement to the Risk Evaluation for 1,4-Dioxane - Supplemental Information File:
EWISRDXL LiverpoolOH Case Study - Provides the Excel workbook file for the Liverpool, OH
surface water case study ( 2024r).
Supplement to the Risk Evaluation for 1,4-Dioxane Supplemental Information File: EWISRD-
XL-Rprobabilistic model code - Provides the R script used to perform the probabilistic surface
water modeling by OES/COU ( )24a).
Page 195 of 570
-------
Supplement to the Risk Evaluation for 1,4-Dioxane - Supplemental Information File: WQP
Processed Surface Water Data - Provides the processed monitoring data in surface water
retrieved from the Water Quality Portal (U.S. EPA. 2024w).
Supplement to the Risk Evaluation for 1,4-Dioxane - Supplemental File Folder: Probabilistic
Surface Water Model (EWISRD-XL-R) Files ( I024v)
Supplement to the Risk Evaluation for 1,4-Dioxane - Supplemental Information File: Drinking
Water Exposure and Risk Estimates for 1,4-Dioxane Land Releases to Landfills - Provides
calculations of groundwater concentration derived from the waste adjusted dilution attenuation
factor extracted from Delisting Risk Assessment Software (DRAS) for Landfills and the
corresponding risk calculations ( I024f).
Supplement to the Risk Evaluation for 1,4-Dioxane - Supplemental Information File: Drinking
Water Exposure and Risk Estimates for 1,4-Dioxane Land Releases to Surface Impoundments -
Provides calculations of groundwater concentration derived from the waste adjusted dilution
attenuation factor extracted from DRAS for release of hydraulic fluid produced water to surface
impoundments and the corresponding risk calculations ( 24g).
Supplement to the Risk Evaluation for 1,4-Dioxane - Supplemental Information File: Air
Exposures and Risk Estimates for Single Year Analysis - Provides air concentrations estimated
by American Meteorological Society/Environmental Protection Agency Regulatory Model
(AERMOD) for air releases reported to TRI in 2019 and calculates corresponding exposure
concentrations and risk estimates ( 24e).
Supplement to the Risk Evaluation for 1,4-Dioxane - Supplemental Information File: Air
Exposure and Risk Estimates for 1,4-Dioxane Emissions from Hydraulic Fracturing Operations
- Provides air concentrations estimated by Integrated Indoor/Outdoor Air Calculator (IIOAC)
based on Monte Carlo modeling of air releases from hydraulic fracturing operations and
calculates corresponding exposure concentrations and risks ( 324b).
Supplement to the Risk Evaluation for 1,4-Dioxane - Supplemental Information File: Air
Exposures and Risk Estimates for Industrial Laundry - Provides air concentrations estimated by
IIOAC based on Monte Carlo modeling of air releases from industrial laundries and calculates
corresponding exposure concentrations and risks ( 024c).
Supplement to the Risk Evaluation for 1,4-Dioxane - Supplemental Information File: Air
Exposures and Risk Estimates for Institutional Laundry- Provides air concentrations estimated
by IIOAC based on Monte Carlo modeling of air releases from institutional laundries and
calculates corresponding exposure concentrations and risks ( 324c).
Supplement to the Risk Evaluation for 1,4-Dioxane - Supplemental Information File: Air
Exposures and Risk Estimates for Multi-Year Analysis - Provides air concentrations estimated by
IIOAC for 6 years (2015 to 2020) of air releases reported to TRI and calculates the
corresponding exposure concentrations and risk estimates (U.S. EPA. 2024d).
Supplement to the Risk Evaluation for 1,4-Dioxane - Supplemental Information File: Land Use
Analysis for Ambient Air - Provides documentation of land use analysis based on facilities
reporting air releases to TRI ( 2024t).
Page 196 of 570
-------
Supplement to the Risk Evaluation for 1,4-Dioxane - Supplemental File Folder: IIOAC
Modeling and Results Files ( 024s).
Page 197 of 570
-------
Appendix C
SYSTEMATIC REVIEW PROTOCOL FOR THE
DRAFT SUPPLEMENT TO THE RISK EVALUATION
FOR 1,4-DIOXANE
The U.S. EPA's Office of Pollution Prevention and Toxics (OPPT) applies systematic review principles
in the development of risk evaluations under the amended TSCA. TSCA section 26(h) requires EPA to
use scientific information, technical procedures, measures, methods, protocols, methodologies, and
models consistent with the best available science and base decisions under section 6 on the weight of
scientific evidence. Within the TSCA risk evaluation context, the weight of scientific evidence is
defined as "a systematic review method, applied in a manner suited to the nature of the evidence or
decision, that uses a pre-established protocol to comprehensively, objectively, transparently, and
consistently identify and evaluate each stream of evidence, including strengths, limitations, and
relevance of each study and to integrate evidence as necessary and appropriate based upon strengths,
limitations, and relevance" (40 CFR 702.33).
To meet the TSCA section 26(h) science standards, EPA used the TSCA systematic review process
described in the Draft Systematic Review Protocol Supporting TSCA Risk Evaluations for Chemical
Substances, Version 1.0: A Generic TSCA Systematic Review Protocol with Chemical-Specific
Methodologies (U.S. EPA. 2021a) (2021 Draft Systematic Review Protocol). Table Apx C-l. Section 3
of the 2021 Draft Systematic Review Protocol depicts the steps in which information is identified and
whether it undergoes the formal systematic review process (U.S. EPA 2021a). Information attained via
the systematic review process is integrated with information attained from sources of information that do
not undergo systematic review (e.g., EPA-generated model outputs) to support a weight of scientific
evidence analysis.
Scope
Risk Evaluation
Systematic Review
1
Literature Searching
and Screening
Data Evaluation
Data Extraction
(in scope chemicals)
Data Gap filling from sources
outside of the Systematic Review
process
(i.e., systematic approaches using
model outputs, analogue, quahtative
information on a COU)
Weight of the Scientific Evidence
Analysis1/o7' each discipline and across
disciplines
Evidence Integration of
Systematically Reviewed
data
Evidence Integration of data
obtained outside of
Systematic Review
Conclusions from
the Weight of the
Scientific Evidence
Analysis
Risk
Characterization
across disciplines
Legend
TSCA Process/Product
Systematic Review Step
Non-Systematic Review
Step (may encompass
systematic approaches)
Step utilizing Systematic
Review and non-
Systematic Review
Results
'Weight of the
Scientific Evidence
(WoSE) considers
the results of the
Systematic Review
method and
additional evidence
integrated from non-
systematic review
methods. Analysis of
the WoSE may
include evidence
integration across
disciplines.
FigureApx C-l. Overview of the TSCA Risk Evaluation Process with Identified Systematic
Review Steps
Page 198 of 570
-------
The process complements the risk evaluation process in that are used to develop the exposure and hazard
assessments based on reasonably available information. EPA defines "reasonably available information"
to mean information that EPA possesses or can reasonably obtain and synthesize for use in risk
evaluations, considering the deadlines for completing the evaluation (40 CFR 702.33).
C.l Clarifications and Updates to the 2021 Draft Systematic Review
Protocol
In 2021, EPA released the 2021 Draft Systematic Review Protocol ( ), a framework of
systematic review approaches under TSCA, to address comments received on a precursor systematic
review approaches framework, the Application of Systematic Review in TSCA Risk Evaluations (
18c). In April 2022, the SACC provided comments on the 2021 Draft Systematic Review
Protocol while additional comments on OPPT's systematic review approaches were garnered during the
public comment period. In lieu of an update to the 2021 Draft Systematic Review Protocol, this
systematic review protocol for the Supplement to the Risk Evaluation for 1,4-Dioxane describes some
clarifications and different approaches that were implemented than those described in the 2021 Draft
Systematic Review Protocol in response to (1) SACC comments, (2) public comments, or (3) to reflect
chemical-specific risk evaluation needs.
C.l.l Clarifications and Updates
Throughout the 2021 Draft Systematic Review Protocol, there were some terms used that were not
explicitly defined, resulting in their different uses within the document ( ,021a). Table Apx
C-l lists the terms that were updated to resolve some of the confusion expressed by the public and
SACC comments regarding the implementation of the respective systematic review-related step. One
main clarification is that all references that undergo systematic review are consideredfor use in the risk
evaluation—even those that do not meet the various discipline and sub-discipline screening criteria {i.e.,
RESO, PESO, PECO) or that are categorized as supplemental information at title and abstract (TIAB) or
full-text (FT) screening.
Section 4.2.5 of the 2021 Draft Systematic Review Protocol describes how data sources {e.g., individual
references, databases) may be tagged and linked in epidemiological cohort studies when information is
present in multiple studies ( >2la). References will generally undergo data quality evaluation
and extraction if there are data that pass screening criteria; however, to prevent the same data from being
represented multiple times and conflating the amount of available information on a subject area, EPA
selects the reference(s) that most appropriately describes the extractable results (indicated as the parent
reference in DistillerSR). For example, if two references portray the same information from the same
dataset, only one is counted in the overall dataset {i.e., deduplication). If two references contain
information about the same dataset, but only one provides additional contextual information or summary
statistics {e.g., mean), both data sources are linked but the extractable information from both may be
combined in DistillerSR. This allows the capture of key information while avoiding double counting the
data of interest, which may be the case whether or not one reference contains original or extractable data
that passes screening criteria.
The linked reference containing the majority of the data, which are evaluated and extracted, is identified
in DistillerSR as the parent reference; the "complementary child reference" in DistillerSR does not
undergo data evaluation and extraction. Linking the references in DistillerSR allows the reference with
more limited information or only contextual information to be tracked and utilized to evaluate the
extracted data in the other related studies. The child reference may undergo data quality evaluation and
extraction if there are additional unique and original data that pass screening criteria. One clarification is
that this procedure of identifying potential duplicative information applies to all information that is
Page 199 of 570
-------
considered in a risk evaluation under TSCA (not just epidemiological cohort studies). Also, this
procedure may apply when there is duplicative information in two references, even if it is more than just
"contextual."
Section 5 of the 2021 Draft Systematic Review Protocol describes how EPA conducts data quality
evaluation of data/information sources considered for a respective chemical risk evaluation, with Section
5.2 specifically explaining the terminology used to describe both metric and overall data/information
source quality determinations ( 21a). To respond to both SACC and public comments
regarding the inappropriate use of quantitative methodologies to calculate both "Metric Rankings" and
"Overall Study Rankings", EPA decided to not implement quantitative methodologies to attain either
metric and overall data/information source quality determinations and therefore updated the
terminology used for both metric ("Metric Ranking") and overall data/information source ("Overall
Study Ranking") quality determinations (TableApx C-l). Specifically, metric and overall
data/information source quality determination terminology have been updated to "Metric Rating" and
"Overall Quality Determination", respectively. The word "level" was also often used synonymously and
inconsistently with the word "ranking" in the 2021 Draft Systematic Review Protocol; that inconsistency
has been rectified, resulting in the word "level" no longer being used to indicate either metric or overall
data/information source quality determinations ( 2021a).
Sections 4.3.2.1.3 and 6 of the 2021 Draft Systematic Review Protocol describe when EPA may reach
out to authors of data/information sources to obtain raw data or missing elements that are important to
support the data evaluation and data integration steps ( ). In such cases, the request(s)
for additional data/information, number of contact attempts, and responses from the authors are
documented. EPA's outreach is considered unsuccessful if those contacted do not respond to email or
phone requests within 1 month of initial attempt(s) of contact. One important clarification to this
guidance is that EPA may reach out to authors anytime during the systematic review process for a given
data/information source or reference, and that contacting authors does not explicitly happen during the
data quality evaluation or extraction steps.
Table Apx C-l. Terminology Clarifications between the 2021 Draft Systematic Review Protocol
and the Systematic Review Protocol for the Supplement
to the Risk Evaluation for 1,4-Dioxane
2021 Draft Systematic
Review Protocol Term
Systematic Review Protocol
for the Supplement to the Risk
Evaluation for 1,4-Dioxane
Term Update
Clarification
"Title and abstract" or
"Title/abstract"
"Title and abstract"
To increase consistency, the term "title and
abstract" will be used to refer to information
specific to "title and abstract" screening.
Variations of how
"include," "on topic" or
"PEC07PESOfe/RESOc
relevant" implied a
reference was considered
for use in the risk
evaluation, whereas
"exclude," "off topic" or
"not
PEC07PES0fe/RES0c
relevant" implied a
Meets/does not meet
PEC07PES0fe/RES0c screening
criteria
The term "include" or "exclude" falsely
suggests that a reference was or was not,
respectively, considered in the risk evaluation.
There was also confusion regarding whether
"on topic" and "PEC07PESOfe/RESOc
relevant" were synonymous and suggested
those references were explicitly considered for
use in the risk evaluation (and by default, "off
topic" and "not PEC07PES07RES06
relevant" references were not). References that
meet the screening criteria proceed to the next
Page 200 of 570
-------
2021 Draft Systematic
Review Protocol Term
Systematic Review Protocol
for the Supplement to the Risk
Evaluation for 1,4-Dioxane
Term Update
Clarification
reference was not
considered for use in the
risk evaluation.
systematic review step; however, all references
that undergo systematic review at any time are
considered in the risk evaluation. Information
that is categorized as supplemental or does not
meet screening criteria are generally less
relevant for quantitative use in the risk
evaluation but may be considered if there is a
data need identified. For instance, mechanistic
studies are generally categorized as
supplemental information at either title and
abstract or full-text screening steps but may
undergo the remaining systematic review steps
if there is a relevant data need for the risk
evaluation (e.g., dose response, mode of
action).
Database source not
unique to a chemical
Database
Updated term and definition of "Database":
Data obtained from databases that collate
information for the chemical of interest using
methods that are reasonable and consistent
with sound scientific theory and/or accepted
approaches and are from sources generally
using sound methods and/or approaches (e.g.,
state or federal governments, academia).
Example databases include STORET and the
Massachusetts Energy and Environmental
Affairs Data Portal.
The term in the 2021 Draft Systematic Review
Protocol (Table_ApxN-l) incorrectly
suggested that databases that contain
information on a singular chemical are not
considered (U.S. EPA, 2021a). Furthermore,
the wording "large" was removed to prevent
confusion and the incorrect suggestion that
there is a data size requirement for databases
that contain information that may be
considered for systematic review.
Metric Ranking or Level
Metric Rating
As explained above, EPA is not implementing
quantitative methodologies to indicate metric
quality determinations, therefore the term
"ranking" is inappropriate. The term "level"
was inconsistently used to indicate metric
quality determinations previously; therefore,
the Agency is removing the use of this term to
reduce confusion when referring to metric
quality determinations. The term "Rating" is
more appropriate to indicate the use of
Page 201 of 570
-------
2021 Draft Systematic
Review Protocol Term
Systematic Review Protocol
for the Supplement to the Risk
Evaluation for 1,4-Dioxane
Term Update
Clarification
professional judgement to determine a quality
level for individual metrics.
Overall Study Ranking or
Level
Overall Quality Determination
(OQD)
As explained above, EPA is not implementing
quantitative methodologies to indicate overall
data/information source quality
determinations, therefore the term "ranking" is
inappropriate. The term "level" was
inconsistently used to indicate overall
data/information source quality determinations
previously; therefore, the Agency is removing
the use of this term to reduce confusion when
referring to overall data/information source
quality determinations. The term "Rating" is
more appropriate to indicate the use of
professional judgement to determine a quality
level for the overall data/information source
quality determination.
a "PECO" stands for Population, Exposure, Comparator or Scenario, and Outcomes.
b "PESO" stands for Pathways or Processes, Exposure, Setting or Scenario, and Outcomes.
c "RESO" stands for Receptors, Exposure, Setting or Scenario, and Outcomes.
C.2 Data Search
To expand upon the previous analysis conducted in the 2020 RE, this Supplement to the Risk Evaluation
for 1,4-Dioxane addresses additional COUs in which 1,4-dioxane is present as a byproduct of the
manufacturing process and evaluates risks from general population exposures to 1,4-dioxane released to
water, air, and land. This supplement focuses on evaluating additional exposure pathways that were not
addressed in the Final Risk Evaluation for 1,4-Dioxane ( 320c). Therefore, the data search
focused on prioritizing updated literature search results to characterize environmental releases and
occupational exposure, and general population, consumer, and environmental exposure information to
evaluate the exposure pathways in scope for this supplement. Data sources may also contain information
that may be used to evaluate exposure pathways already addressed in the 2020 RE ( 2020c)
{i.e., consumer exposure). Below are the additional exposure pathways being assessed in this
Supplement to the Risk Evaluation for 1,4-Dioxane (Section 1.2).
• Occupational exposure, including PESS, to
o 1,4-dioxane present as a byproduct in commercial products during ethoxylation
processing or polyethylene terephthalate (PET) manufacturing and in hydraulic fracturing
waste (Sections 3.1, 5.2.1)
• General population exposures, including PESS, to
o 1,4-dioxane present in drinking water sourced from surface water as a result of direct and
indirect industrial releases and DTD releases of consumer and commercial products
(Sections 2.3.1, 3.2.2 and 5.2.2.1);
o 1,4-dioxane present in drinking water sourced from groundwater contaminated as a result
of disposals (Sections 2.3.2, 3.2.2.2 and 5.2.2.1.6); and,
Page 202 of 570
-------
o 1,4-dioxane released to air from industrial and commercial sources (Sections 2.3.3, 3.2.3,
and 5.2.2.3).
C.2.1 Multi-disciplinary Updates to the Data Search
For this Supplement to the Risk Evaluation for 1,4-Dioxane, the updated literature search was conducted
as described in Section 4 of the 2021 Draft Systematic Review Protocol (U.S. EPA. 2021a). where the
peer-reviewed and gray literature updated search followed the approach outlined in Sections 4.2 and 4.3
of the 2021 Draft Systematic Review Protocol, respectively ( :021a). The updated search for
peer-reviewed and gray literature relevant references was completed in October 2021 and January 2022,
respectively, which also considered information found for the Final Risk Evaluation for 1,4-Dioxane
( lie). Occasionally additional data sources relevant for the risk evaluation may be
identified after the initial search for peer-reviewed and gray literature; these data sources will then
undergo systematic review for the relevant discipline(s). Additionally, each discipline utilizes different
strategies (e.g., search strings) to attain their discipline-specific pools of data sources that undergo
systematic review.
As mentioned in Section 4.2.2 of the 2021 Draft Systematic Review Protocol, a supplemental literature
search is conducted to fill data gaps, but in this supplement, the supplemental search was conducted to
update the literature search conducted to identify any potentially relevant environmental release and
occupation exposure and general population, consumer, and environmental exposure information (U.S.
>2la). Rather than utilizing positive and negative seed references as described in Section 4.2.4.2
of the 2021 Draft Systematic Review Protocol, search strings were used in SWIFT 14-Review to better
identify relevant references to evaluate exposure pathways addressed in this supplement (
2021a). The language describing the new exposure pathways and COUs that are in scope for this
supplement was used to derive the search strings listed below in Sections C.2.3.1 and C.2.3.2. When the
search strings are identified in the title, abstract, keyword, or Medical Subject Heading (MeSH) fields of
a given reference in SWIFT-Review, those references proceeded with TIAB screening.
The evaluation of physical and chemical properties, fate properties and environmental and human health
hazard information did not differ from the respective information provided in the Final Risk Evaluation
for 1,4-Dioxane ( 020c) to address the additional exposure pathways in this supplement,
therefore no additional references were identified for these respective topics or underwent systematic
review for these disciplines. One minor clarification to what was described in the 2021 Draft Systematic
Review Protocol is that the PECO statement used to screen general population, consumer, and
environmental exposure information considered for this supplement, currently resides in Appendix
Section H.5 (which was intended to encompass PECO statements regarding environmental and human
health hazard information), rather than in Appendix Section H.4 ( 2021a). Please see
Appendix C.3.2 below for additional updates specific to the implementation of the PECO statement.
C.2.2 Additional Data Sources Identified
As mentioned above in Appendix C.2, additional data sources containing potentially relevant
information for a respective risk evaluation may be identified. For this supplement, additional gray
literature data sources were identified for the characterization of environmental release and occupational
exposure and general population, consumer, and environmental exposure, as explained below in
Sections C.2.2.1 and C.2.2.2, respectively. Finally, during the public comment period and review of the
Draft Supplement to the Risk Evaluation for 1,4-Dioxane, additional data sources were identified and
14 SWIFT is an acronym for "Sciome Workbench for Interactive Computer-Facilitated Text-mining." SWIFT-Active
Screeneruses machine learning approaches to save screeners' time and effort.
Page 203 of 570
-------
considered for this supplement; those additional data sources were incorporated into the systematic
review process described below.
C.2.2.1 Additional Data Sources Identified for Environmental Release and
Occupational Exposure
As explained in Appendix E of the 2021 Draft Systematic Review Protocol ( ), generic
scenarios and emission scenarios documents are listed as part of the initial gray literature sources. Some
generic scenarios and a draft emission scenario document became available after the gray literature
search was completed in January 2022 and were considered for the environmental release and
occupational exposure assessment. This includes the Draft OECD ESD on Hydraulic Fracturing (
22e), Draft GS on Furnishing Cleaning Products ( 2022a), EPA Methodology Review
Draft (MRD) on Commercial Use of Automotive Detailing Products ( 2022b). and Draft GS
on Use of Laboratory Chemicals ( 12a). The updated sources were added to EPA's Health
and Environmental Research Online (HERO) database in 2022 as well as the systematic review process.
In addition to the gray literature sources listed above, an online database called FracFocus 3.0 (GWPC
and IPG- 22) was included in the pool of references EPA considered for environmental release and
occupational exposure through backward searching. These are described in Section 4.4 of the 2021 Draft
Systematic Review Protocol ( 2021a). Backward searching from the Draft OECD ESD on
Hydraulic Fracturing (U.S. EPA. 2022e) led to EPA's identification of the FracFocus data. The Agency
gathered the data directly from the source and only pulled data specific to sites that reported using 1,4-
dioxane in fracturing fluids. This source was added to the HERO database as well as the systematic
review process.
C.2.2.2 Additional Data Sources Identified for General Population, Consumer, and
Environmental Exposure
In addition to the gray literature sources listed in Appendix E of the 2021 Draft Systematic Review
Protocol ( ?21a), several other gray literature sources were considered for inclusion the
general population, consumer, and environmental exposure assessment and added to the HERO database
in 2022. The Water Quality Portal (WQP) database, the successor of EPA's STORET (STOrage and
RETrieval) database, was incorporated because it includes a large variety of chemical-specific data.
Also, WQP is a portal that combines data from multiple databases—not just STORET—such as the U.S.
Geological Survey's National Water Information System. Information from WQP was collected in July
2022.
A few additional gray literature sources (databases) were included in the pool of references EPA
considered on general population, consumer, and environmental exposure through backwards searching,
which is described in Section 4.3.3 of the 2021 Draft Systematic Review Protocol ( 1021a).
Backwards searching from the Third Unregulated Contaminant Monitoring Rule (UCMR3) database
(\ v << \ IVI i!) led to EPA's identification of data from a few states collecting data on 1,4-dioxane
for longer periods of time than reported in UCMR3. EPA was able to secure and incorporate data from
three state databases. In addition, elevated levels of 1,4-dioxane in samples from UCMR3 and a
reference found in the pool of peer-reviewed articles led to addition to databases with data on 1,4-
dioxane levels measured in drinking water and surface water in the Cape Fear Watershed in North
Carolina. Six of the highest concentrations of 1,4-dioxane in the UCMR3 database were reported in
water from this watershed, which were also discussed in one of the references found in the literature
search for peer-reviewed sources. All gray literature database sources added to the search, including the
datasets from North Carolina, were added to HERO and the systematic review process (see Section G. 1
for further information).
Page 204 of 570
-------
C.2.3 Search Strings
As explained above in Section C.2.3, the search strings below were used to identify references relevant
to evaluating environmental releases and occupational exposure, as well as general population,
consumer, and environmental exposure.
C.2.3.1 Environmental Release and Occupational Exposure Search Strings
Life Cycle:
TIAB: ("1,4-dioxane" AND ("MFG" OR "import" OR "processing" OR "manufactur*" OR "releases"
OR "waste disposal" OR "reaction product" OR "repackaging" OR "recycling" OR "throughput" OR
"operating days" OR "batch" OR "production speed"))
Treatment Efficiencies:
TIAB: ("1,4-dioxane" AND ("GAC" OR "granular activated carbon" OR "reverse osmosis" OR
"advanced oxidation" OR "hydrogen peroxide with ultraviolet" OR ("hydrogen peroxide" AND "UV")
OR "hydrogen peroxide with ozone" OR ("hydrogen peroxide" AND "ozone") OR "AOP" OR
"Fenton's reagent" OR "bioremediation"))
Occupational Workers:
TIAB: ("1,4-dioxane" AND ("janitor*" OR "mechanic" OR "laborer" OR "custodia*" OR "painter*"
OR "laboratory technician" OR "laboratory employee*" OR ("pharmaceutical" AND ("employee" OR
"worker" OR "technician")) OR "residential construction" OR "industrial construction"))
General:
TIAB: ("1,4-dioxane" AND ("surfactant" OR "ethoxylat*" OR "nonylphenol ethoxylate" OR
"alkylphenol ethoxylate" OR "sulfated" OR "industrial laundr*" OR "commercial laundr*" OR
"institutional laundr*" OR "institutional laundr*" OR "advanced oxidation" OR "ozone-peroxide
advanced oxidation" OR "low dioxane" OR "low dioxane ether sulfates" OR "low dioxane ethoxylated
surfactants" OR "low 1,4 dioxane ether sulfates" OR "low 1,4 dioxane ethoxylated surfactants" OR
"safety data sheet" OR "material safety data sheet"))
Process Uses:
TIAB: ("1,4-dioxane" AND ("stabilizer" AND ("chlorinated solvents" OR "degreasing" OR
"electronics manufacturing" OR "metal finishing")) OR ("solvent" AND ("histology" OR "cellulose
acetate membrane" OR "microscopy" OR "organic chemical manufacturing" OR "organic chemical"))
OR ("textile" AND ("wetting" OR "dispersing")) OR ("esterification" AND ("by-product" OR
"byproduct")))
Product Uses:
TIAB: ("1,4-dioxane" AND ("solvent" AND ("paint*" OR "lacquer*" OR "varnish remover" OR
"stain" OR "printing" OR "scintillation" OR "resin*" OR "oil*" OR "rubber chemicals" OR "rubber"
OR "sealant*" OR "adhesive*" OR "wax*" OR "cement*")))
TIAB: ("1,4-dioxane" AND ("artificial leather" OR "purifying agent" OR "antifreeze" OR "de-icing"
OR "pesticide*" OR "fumigant*"))
CASRNs of Ethoxylated Chemicals:
TIAB: ("9005-65-6" OR "3088-31-1" OR "68081-98-1" OR "68439-50-9" OR "68551-12-2" OR
"68439-49-6" OR "9043-30-5" OR "26183-52-8" OR "9002-92-0" OR "9004-82-4" OR "9005-64-5"
OR "68131-40-8" OR "68991-48-0" OR "37251-67-5" OR "5274-68-0" OR "864529-51-1" OR
"84133-50-6" OR "68439-45-2" OR "68987-81-5" OR "9003-11-6" OR "61791-29-5" OR "9005-08-7"
Page 205 of 570
-------
OR "61791-13-7" OR "166736-08-9" OR "3055-99-0" OR "66455-14-9" OR "68131-39-5" OR
"68213-23-0" OR "68951-67-7" OR "66455-15-0" OR "61791-26-2" OR "9004-95-9" OR "9005-00-9"
OR "61827-42-7" OR "68081-91-4" OR "68585-40-0" OR "68815-56-5" OR "61788-85-0" OR "3055-
97-8" OR "120313-48-6" OR "68439-46-3" OR "69227-22-1" OR "68002-97-1")
C.2.3.2 General Population, Consumer, and Environmental Exposure Search Strings
Population:
TIAB: ("general population" OR "bystanders" OR "near-facility" OR "industrial facilit*" OR
"commercial facilit*" OR "employee" OR "employees" OR "worker*" OR "manufacturer" OR "near-
disposal" OR "near surface disposal" OR "child*" OR "teenage*" OR "susceptible population" OR
"immunocompromised" OR "preschool" OR "senior*" OR "older adults" OR "elderly" OR "pregnant
women" OR "preexisting condition*" OR "lactating women" OR "childbearing" OR "prenatal" OR
"infant*" OR "adolescen*")
Landfills:
TIAB: ("dioxane" AND ("landfill" OR "leach*" OR "incineration" OR "wastewater" OR"GAC" OR
"granular activated carbon" OR "reverse osmosis" OR "waste site" OR "land disposal" OR "waste
disposal" OR "landfill leach*"))
Indoor Air and Water:
TIAB: ("dioxane" AND ("inhal*" OR "tap water" OR "water well" OR "indoor air" OR "surface water"
OR "groundwater" OR "outdoor air" OR "ambient air" OR "drinking water" OR ("biomonitoring" OR
"monitoring" AND ("air" OR "water")) OR "drinking" OR "aquifer" OR "leach*" OR "municipal
water")) NOT ("spill")
Consumer and Industrial Use:
TIAB: ("1,4-dioxane" AND ("ingest*" OR "swallow*" OR "showering" OR "bathing" OR "swimming"
OR "wading" OR "inhal*" OR "paint*" OR "industrial manufactur*" OR "residential construction" OR
"commercial construction" OR "cleaning" OR "dishwasher" OR "printing" OR "food supplement*" OR
"packaging" OR "breast milk" OR "human milk" OR "intake rates" OR "launder*" OR "surface
cleaner" OR "automotive"))
Concentration and Dose:
TIAB: ("reference concentration" OR "RfC" OR "NOAEL" OR "LOAEL" OR "benchmark
concentration" OR "reference dose" OR "RfD" OR "chronic oral" OR "chronic inhalation" OR "oral
slope factor" OR "soil screening level" OR "PEL" OR "permissible exposure limit" OR "weighted
average" OR "weight fraction" OR "emission rate*" OR "inhalation unit risk" OR "IUR" OR "dose-
response" OR "reverse dosimetry" OR "biomonitoring" OR "media concentration*" OR ("estimate*"
AND ("acute" OR "subchronic" OR "chronic")) OR "single-dose" OR "repeated-dose" OR "daily
intake")
C.3 Data Screening
Sections 4.2.5 and 4.3.2 of the 2021 Draft Systematic Review Protocol describe how title and abstract
(TIAB) and full-text (FT) screening, respectively, are conducted to identify references that may contain
relevant information for use in risk evaluations under TSCA using discipline-specific screening criteria
(defined below in Sections C.3.1.1 and C.3.2.1 (U.S. EPA. 2021a). Specifically, TIAB screening efforts
Page 206 of 570
-------
may be conducted using the specialized web-based software programs DistillerSR15 and SWIFT-Active-
Screener16; however, for this supplement, EPA used SWIFT-Active-Screener exclusively. Additional
details on how SWIFT Active-Screener utilizes a machine-learning algorithm to automatically compute
which unscreened documents are most likely to be relevant17 are available in Section 4.2.5 of the 2021
Draft Systematic Review Protocol ( 021a). During TIAB screening, if it was unclear whether
a reference met the screening criteria (e.g., PECO/RESO/PESO statements) without having the full
reference to review, or if a reference was determined to meet the screening criteria, that reference
advanced to full-text screening if the full reference could be retrieved and generated into a Portable
Document Format (PDF).
Literature inventory trees were introduced in the scoping process for the risk evaluations that began
systematic review in 2019 in response to comments received from the SACC and public to better
illustrate how references underwent various systematic review steps (e.g., TIAB and full-text screening).
As explained in various final scope documents (e.g., Section 2.1.2 in the Final Scope of the Risk
Evaluation for 4,4'-(l-Methylethylidene)bis[2, 6-dibromophenol] ( 20b)). literature
inventory trees demonstrate how references that meet screening criteria progress to the next systematic
review step. EPA used the Health Assessment Workplace Collaborative (HAWC) tool to develop web-
based literature inventory trees to enhance the transparency of the decisions resulting from the screening
processes. Updates made to the available literature considered for the supplement that are made between
publishing the draft and final Supplement to the Risk Evaluation for 1,4-Dioxane (e.g., additional
references may be provided to the EPA through public comment) will be reflected in HAWC (see also
hyperlinks to HAWC in the figure captions below for each respective literature inventory tree).
The web-based literature inventory trees in HAWC also allow users to directly access the references in
the HERO database (more details available in Section 1 of the 2021 Draft Systematic Review Protocol)
by selecting appropriate nodes, which indicate whether a reference has met screening criteria at different
screening steps and/or types of content that may be discerned at that respective systematic review step
( 2021a). Furthermore, as mentioned in the various final scope documents, the sum of the
numbers for the various nodes in the literature inventory trees may be smaller or larger than the
preceding node because some studies may have unclear relevance or be relevant for many categories of
information. The screening process for each discipline varies and the nodes in the literature inventory
tree indicate the screening decisions determined for each reference and whether specific content could
be determined; if no references had a specific screening decision and/or contained specific content
relevant for a respective discipline, a node will not be present on the literature inventory tree to depict
this.
In the literature inventory trees below, which depict systematic review search results used to evaluate the
new exposure pathways in this supplement, some references were unattainable for full-text screening.
The "PDF unavailable" node refers to references or sources of information for which EPA was unable to
15 As noted on the DistillerSR web page, this systematic review software "automates the management of literature collection,
triage, and assessment using AI and intelligent workflows...to produce transparent, audit ready, and compliant literature
reviews." EPA uses DistillerSR to manage the workflow related to screening and evaluating references; the literature search
is conducted external to DistillerSR.
16 SWIFT-Active Screener is another systematic review software that EPA is adopting in the TSCA systematic review
process. From Sciome's SWIFT-Active Screener web page: "As screening proceeds, reviewers include or exclude articles
while an underlying statistical model in SWIFT-Active Screener automatically computes which of the remaining unscreened
documents are most likely to be relevant. This 'Active Learning' model is continuously updated during screening, improving
its performance with each reference reviewed. Meanwhile, a separate statistical model estimates the number of relevant
articles remaining in the unscreened document list."
17 Description comes from the SWIFT-Active Screener web page.
Page 207 of 570
-------
obtain the entire reference or source of data/information but were identified in the literature search
because of the availability of the title and abstract. For the references considered to evaluate
environmental release and occupational exposure, all references that passed screening criteria were
found and underwent full-text screening. For the references considered to evaluate general population,
consumer, and environmental exposure, one PDF could not be obtained through interlibrary loan or
through other channels available to EPA to obtain reprints of published sources.
As mentioned in Section C.l, although all information contained in references that enter systematic
review are considered for use in the risk evaluation, the references that satisfy the screening criteria are
generally deemed to contain the most relevant and useful information for characterizing the uses,
exposure, and hazard of a chemical of interest and are generally utilized in the risk evaluation (and can
be used later on to identify further data needs). On the other hand, data or information sources that do
not satisfy the screening criteria outlined below may undergo data quality evaluation and extraction
should a data need arise for the risk evaluation.
C.3.1 Environmental Release and Occupational Exposure
During data screening, EPA followed the process described in Appendix H, Section H-3 of the 2021
Draft Systematic Review Protocol ( 021a) to conduct title and abstract and full-text
screening for 1,4-dioxane literature search results guided by the RESO statement. RESO stands for
Receptors, Exposure, Setting or Scenario, and Outcomes. The same RESO statement was used during
title and abstract, and full-text screening for references considered for the evaluation of environmental
release and occupational exposure information for 1,4-dioxane. TIAB were performed using SWIFT
Active-Screener. Data or information sources that comply with the screening criteria specified in the
RESO statement then undergo data quality evaluation and extraction. Figure Apx C-2 presents the
number of references that report general engineering data, environmental release, and occupational
exposure data that passed RESO screening criteria at TIAB and full-text screening.
Page 208 of 570
-------
C.3.1.1 Environmental Release and Occupational Exposure Literature Inventory Tree
Figure Apx C-2. Literature Inventory Tree - Environmental Releases and Occupational
Exposure Search Results for 1,4-Dioxane
View the interactive literature inventory tree in HAWC. Data in this figure represent all references obtained from
the publicly available databases and gray literature references searches that were included in systematic review as
of March 25, 2024. Additional data may be added to the interactive version as they become available.
C.3.2 General Population, Consumer, and Environmental Exposure
The TIAB and full-text screening process was consistent with what EPA previously outlined in Sections
4.2.5 and 4.3.2 of the 2021 Draft Systematic Review Protocol (U.S. EPA. 2021a). PECO stands for
Population, Exposure, Comparator or Scenario, and Outcomes for Exposure Concentration or Dose. The
PECO statement, as depicted in Appendix H.5.14 of the 2021 Draft Systematic Review Protocol (U.S.
EPA. 202la), was refined to better identify references that may contain information relevant for this
supplement. Specifically, data that are relevant for characterizing exposure to 1,4-dioxane in food,
including biota that humans consume, was not evaluated and extracted because 1,4-dioxane is not
expected to bioaccumulate in organisms likely to be consumed by humans. During TIAB screening, if it
is unclear if a reference will meet the PECO screening criteria without having the full reference to
review, or if a reference is determined to meet the PECO screening criteria, that reference will advance
to full-text screening. Studies containing potentially relevant supplemental material were also tracked
and categorized during the literature screening process. Relevant supplemental material may be
reviewed, evaluated for data quality, and incorporated into risk evaluations, as needed. For example,
references were considered supplemental if they contained data from countries outside of North America
Page 209 of 570
-------
on 1,4-dioxane levels associated with landfills because different countries have very different waste
management policies (including requirements for landfills), and local hydrogeology in other regions
may not be relevant to sites in the United States. FigureApx C-3 presents the number of references that
report general population, consumer, and environmental exposure data that passed PECO screening
criteria at TIAB and full-text screening.
C.3.2.1 General Population, Consumer, and Environmental Exposure Literature
Inventory Tree
Figure Apx C-3. Literature Inventory Tree - General Population, Consumer, and Environmental
Exposure Search Results for 1,4-Dioxane
View the interactive literature inventory tree in HAWC. Data in this figure represent all references obtained from
the publicly available databases and gray literature references searches that were included in systematic review as
of April 23, 2024. Additional data may be added to the interactive version as they become available.
C.4 Data Evaluation and Data Extraction
Data evaluation and extraction for this supplement are as described in Sections 5 and 6 of the 2021 Draft
Systematic Review Protocol (U.S. EPA. 2021a). Data evaluation is the systematic review step in which
Page 210 of 570
-------
EPA assesses quality of the individual data sources using the evaluation strategies and criteria for each
discipline (e.g., physical and chemical property data, fate and transport data, occupational exposure and
environmental release data, general population, consumer, and environmental exposure data). The
evaluation method uses a structured framework with predefined criteria for each type of
data/information source. The goal of the method used by EPA is to provide transparency, consistency,
and as much objectivity as possible to the evaluation process along with meeting the TSCA science
standards. Data extraction is the systematic review step in which EPA identifies quantitative and
qualitative information from data sources that meet screening criteria and extract the data/information
using structured forms or templates.
As explained above in Section C.l, terminology updates were made regarding the description of both
metric and overall data/information source quality determinations from what was originally described in
the 2021 Draft Systematic Review Protocol (U.S. EPA. 2021a). Specifically, metric and overall
data/information source quality determination terminology have been updated to "Metric Rating" and
"Overall Quality Determination", respectively. For additional clarifications regarding these updates,
please see TableApx C-l.
Although data sources that meet screening criteria following full-text screening will generally proceed to
data quality evaluation and extraction steps, one clarification to the procedures outlined in Section 6 of
the 2021 Draft Systematic Review Protocol is that in situations where EPA is unable to extract
data/information from sources that meet screening criteria (e.g., formatting prohibits accurate
extraction), such sources may not have extracted data to present in the risk evaluation or the respective
supplemental documents. Systematic review support documents for the supplement contain results from
the data quality evaluation and extraction systematic review steps. Also, the template used to display the
data may be modified from those that were provided in the 2021 Draft Systematic Review Protocol
( ) because the purpose of these supplemental documents is to accommodate the data
needs for each respective risk evaluation. The following sections provide specific information about the
data quality and extraction process followed to address the exposure pathways in scope for this
supplement and any clarifications or updates regarding these systematic review steps as described in the
2021 Draft Systematic Review Protocol ( 21a).
C.4.1 Environmental Release and Occupational Exposure
As described in the 2021 Draft Systematic Review Protocol, evaluation and extraction followed the
steps outlined in Sections 5, 6, and 6.2 ( 2021a). The data extraction and data quality results
are summarized in Table Apx E-8 for air, Table Apx E-4 for water, Table Apx E-6 for land, and
TableApx F-35 for occupational exposure. The Data Quality Evaluation and Extraction Information
for Environmental Release and Occupational Exposure for 1,4-Dioxane (1,4-D) provides the results
from the data extraction and quality evaluation, including metric rating and the overall quality
determination for each data source ( |x)
C.4.2 General Population, Consumer, and Environmental Exposure
As described in the 2021 Draft Systematic Review Protocol, evaluation and extraction generally
followed the steps outlined in Section 5 and 6 ( 2021a). However, a few updates were made to
the data quality evaluation metrics for a few evidence streams since the metrics were published in the
2021 Draft Systematic Review Protocol. Most of the changes were editorial or minor clarifications,
including the standardization of some metrics that apply to multiple evidence streams, where
appropriate. For example, in the Quality Assurance/Quality Control (QA/QC) metric for evaluating
monitoring and experimental evidence streams, the acronym QA/QC was defined and replaced all
Page 211 of 570
-------
references to quality assurance and quality control when occurring separately or together, and the term
QA/QC techniques was changed to QA/QC measures, which already appeared in the metrics.
A few metrics applicable to multiple evidence streams were modified slightly to better fit some of the
unique situations that frequently arise for a certain type of evidence stream (e.g., databases). For
example, some metrics were updated to clarify the intent of the metric and better account for variation in
types of evidence included in one grouping (e.g., experiments involving chamber studies vs. product
concentration assessments). The domains did not change; however, see below for the changes and
updates made to the data evaluation metrics for the respective evidence types (i.e., monitoring,
experimental studies and databases) as presented in Sections C.4.2.2, C.4.2.3, and C.4.2.4. No changes
were made to the data evaluation metrics for modeling data, as described in Appendix N Section N.6.2
in the 2021 Draft Systematic Review Protocol. The Data Quality Evaluation Information for General
Population, Consumer, and Environmental Exposure for 1,4-Dioxane (1,4-D) provides details of the
data quality evaluation results, including metric rating and the overall quality determination for each
data source (U.S. EPA. 2024y).
Data extraction is the process in which quantitative and qualitative data/information are identified from
each relevant data/information source and extracted using structured forms or templates. Data extraction
was conducted as described in Section 6 of the 2021 Draft Systematic Review Protocol for all evidence
streams relevant for this supplement. However, with respect to information stored within databases, EPA
does not conduct a separate data extraction because the data are more accessible and have additional
context in the original database format. Both the date and data present in the database when the database
underwent full-text screening are available in the HERO database (HERO IDs: 10365582, 10365609,
10365665, 10365667, 10365696, 10365698, 10368680, 10410586, 10501014, and 11414335). If a
reference or data/information source (e.g., a peer-reviewed reference) presents data from a database that
did not undergo systematic review (e.g., a foreign database that is not publicly accessible), the data
would be extracted from the reference or data/information source to the extent possible; this did not
apply to references or sources of data or information that underwent systematic review for this
supplement.
As mentioned above in Section C.4, references may not undergo data extraction, regardless of data
quality rating, if they contain no extractable data points (e.g., values are contained in a non-digitizable
figure or are representative of unspecified media or treatment processes). This constitutes an update to
Section 6 of the 2021 Draft Systematic Review Protocol (U.S. EPA. 2021a). Extraction forms and
templates are tailored to fit the data extraction needs for each risk evaluation.
The types of fields extracted vary by evidence stream and generally followed Section 6.3 of the 2021
Draft Systematic Review Protocol with regard to the data characteristics captured ( :021a).
Examples of types of data extracted and the extraction formats for the four evidence streams identified
through systematic review to evaluate environmental, general population, and consumer exposure data
are listed in the extraction tables provided in the Data Extraction Information for General Population,
Consumer, and Environmental Exposure for 1,4-Dioxane (1,4-D) ( 24x).
C.4.2.1 Data Quality Evaluation Metric Updates
Shown below are the data evaluation metrics for three evidence streams, presenting which data
evaluation metrics changed since the publication of the 2021 Draft Systematic Review Protocol (]j _S
21a). For evidence streams not listed below, there were no changes to the data evaluation
metrics since the 2021 Draft Systematic Review Protocol was published. Other data quality criteria for
studies on consumer, general population, and environmental exposure appear in Appendix N of the 2021
Draft Systematic Review Protocol ( 021a). For example, the criteria for modeling studies
Page 212 of 570
-------
appear in Table Apx N9. Data quality criteria for other types of studies (e.g., environmental release and
occupational exposure assessment) are published in other appendices to the 2021 Draft Systematic
Review Protocol ( ,021aY
For the below tables in Sections C.4.2.2, C.4.2.3, and C.4.2.4, in order to make it easier for the reader to
see what the changes were to the data evaluation metrics, the following convention is used: text inserted
is underlined, and text deleted is in strikeout.
C.4.2.2 Data Evaluation Criteria for Monitoring Data, as Revised
Table Apx C-2.
Evaluation Criteria for Sources of Monitoring Data
Data Quality
Rating
Description
Domain 1 Rcliubilil\
\lclnc 1 Samnlinu nvlhoi.loloij\
High
Samples were collected according to publicly available SOPs that are scientifically sound
and widely accepted (i.e., from a source generally known to use sound methods and/or
approaches) for the chemical and media of interest. Example SOPs include USGS' "National
Field Manual for the Collection of Water-Quality Data," EPA's "Ambient Air Sampling"
(SESDPROC-303-R5), etc.
OR
The sampling protocol used was not a publiclv available SOP from a source aenerallv known
to use using sound methods and/or approaches, but the sampling methodoloav is clear,
appropriate (i.e., scientifically sound), and similar to widely accepted protocols for the
chemical and media of interest. All pertinent sampling information is provided in the data
source or companion source. Examples include:
1. sampling equipment
2. sampling procedures/regime
3. sample storage conditions/duration
4. performance/calibration of sampler
5. study site characteristics
6. matrix characteristics
Medium
Sampling methodology is discussed in the data source or companion source and is generally
appropriate (i.e., scientifically sound) for the chemical and media of interest; however, one
or more pieces of sampling information is not described. The missing information is unlikely
to have a substantial impact on results.
OR
Standards, methods, protocols, or test guidelines may not be widely accepted, but a
successful validation study for the new/unconventional procedure was conducted prior to the
sampling event and is consistent with sound scientific theory and/or accepted approaches. Or
a review of information indicates the methodology is acceptable and differences in methods
are not expected to lead to lower quality data.
Low
Sampling methodology is only briefly discussed; therefore, most sampling information is
missing and likely to have a substantial impact on results.
AND/OR
Page 213 of 570
-------
Data Quality
Rating
Description
The sampling methodology does not represent best sampling methods, protocols, or
guidelines for the chemical and media of interest (e.g., outdated [but still valid] sampling
equipment or procedures, long storage durations).
AND/OR
There are some inconsistencies in the reporting of sampling information (e.g., differences
between text and tables in data source, differences between standard method and actual
procedures reported to have been used, etc.) that led to a low confidence in the sampling
methodology used.
Critically
Deficient
The sampling methodology is not discussed in the data source or companion source.
AND/OR
Sampling methodology is not scientifically sound or is not consistent with widely accepted
methods/approaches for the chemical and media being analyzed (e.g., inappropriate sampling
equipment, improper storage conditions).
AND/OR
There are numerous inconsistencies in the reporting of sampling information, resulting in
high uncertainty in the sampling methods used.
Not rated/not
applicable
Reviewer's
comments
[Document concerns, uncertainties, limitations, and deficiencies and any additional
comments that may highlight study strengths or important elements such as relevance]
\lclnc 2 AikiK Heal nvlhodolotj\
High
Samples were analyzed according to publicly available analytical methods that are
scientifically sound and widelv accepted (i.e., from a source aenerallv using known to use
sound methods and/or approaches) and are appropriate for the chemical and media of
interest. Examples include EPA SW-846 Methods, NIOSH Manual of Analytical Methods
5th Edition, etc.
OR
The analytical method used was not a publicly available method from a source generally
known to use sound methods and/or approaches, but the methodology is clear and
appropriate (i.e., scientifically sound) and similar to widely accepted protocols for the
chemical and media of interest. All pertinent sampling information is provided in the data
source or companion source. Examples include:
1. extraction method
2. analytical instrumentation (required)
3. instrument calibration
4. limit of quantitation (LOQ), LOD, detection limits, and/or reporting limits
5. recovery samples
6. biomarker used (if applicable)
matrix-adjustment method (i.e., creatinine, lipid, moisture)
Medium
Analytical methodology is discussed in detail and is clear and appropriate (i.e., scientifically
sound) for the chemical and media of interest; however, one or more pieces of analytical
information is not described. The missing information is unlikely to have a substantial
impact on results.
Page 214 of 570
-------
Data Quality
Rating
Description
AND/OR
The analytical method may not be standard/widely accepted, but a method validation study
was conducted prior to sample analysis and is expected to be consistent with sound scientific
theory and/or accepted approaches.
AND/OR
Samples were collected at a site and immediately analyzed using an on-site mobile
laboratory, rather than shipped to a stationary laboratory.
Low
Analytical methodology is only briefly discussed. Analytical instrumentation is provided and
consistent with accepted analytical instrumentation/methods. However, most analytical
information is missing and likely to have a substantial impact on results.
AND/OR
Analytical method is not standard/widely accepted, and method validation is limited or not
available.
AND/OR
Samples were analyzed using field screening techniques.
AND/OR
LOQ, LOD, detection limits, and/or reporting limits not reported.
AND/OR
There are some inconsistencies or possible errors in the reporting of analytical information
(e.g., differences between text and tables in data source, differences between standard
method and actual procedures reported to have been used, etc.) which leads to a lower
confidence in the method used.
Critically
Deficient
Analytical methodology is not described, including analytical instrumentation (i.e., HPLC,
GC).
AND/OR
Analytical methodology is not scientifically appropriate for the chemical and media being
analyzed (e.g., method not sensitive enough, not specific to the chemical, out of date).
AND/OR
There are numerous inconsistencies in the reporting of analytical information, resulting in
high uncertainty in the analytical methods used.
Not rated/not
applicable
Reviewer's
communis
[Document concerns, uncertainties, limitations, and deficiencies and any additional
comments that may highlight study strengths or important elements such as relevance]
\lclnc 3 Sckvlion of lnoniaiker of c\|insuiv
High
Biomarker in a specified matrix is known to have an accurate and precise quantitative
relationship with external exposure, internal dose, or target dose (e.g., previous studies (or
the current study) have indicated the biomarker of interest reflects external exposures).
AND
Biomarker (parent chemical or metabolite) is derived from exposure to the chemical of
interest.
Page 215 of 570
-------
Data Quality
Rating
Description
Medium
Biomarker in a specified matrix has accurate and precise quantitative relationship with
external exposure, internal dose, or target dose.
AND
Biomarker is derived from multiple parent chemicals, not only the chemical of interest, but
there is a stated method to apportion the estimate to only the chemical of interest
Low
Biomarker in a specified matrix has accurate and precise quantitative relationship with
external exposure, internal dose, or target dose.
AND
Biomarker is derived from multiple parent chemicals, not only the chemical of interest, and
there is NOT an accurate method to apportion the estimate to only the chemical of interest.
OR
Biomarker in a specified matrix is a poor surrogate (low accuracy and precision) for
exposure/dose.
Critically
Deficient
Not applicable. A study will not be deemed critically deficient based on the use of biomarker
of exposure.
Not rated/not
applicable
Metric is not applicable to the data source.
Reviewer's
communis
[Document concerns, uncertainties, limitations, and deficiencies and any additional
comments that may highlight study strengths or important elements such as relevance]
Domain 2 Rquvscnlalixc
Mclnc 4 (.icouranhic aiva
High
Geographic location(s) is reported, discussed, or referenced.
Medium
Not applicable. This metric is dichotomous (i.e., high vs. critically deficient).
Low
Not applicable. This metric is dichotomous (i.e., high vs. critically deficient).
Critically
Deficient
Geographic location is not reported, discussed, or referenced.
Not rated/not
applicable
Reviewer's
communis
[Document concerns, uncertainties, limitations, and deficiencies and any additional
comments that may highlight study strengths or important elements such as relevance]
\lelnc 5 Tcmnomlih
High
Timing of sample collection for monitoring data is consistent with current or recent
exposures (within 5 years) may be expected.
Medium
Timing of sample collection for monitoring data is less consistent with current or recent
exposures (>5 to 15 years) may be expected.
Page 216 of 570
-------
Data Quality
Rating
Description
Low
Timing of sample collection for monitoring data is not consistent with when current
exposures (>15 years old) may be expected and likely to have a substantial impact on results.
Critically
Deficient
Timing of sample collection for monitoring data is not reported, discussed, or referenced.
Not rated/not
applicable
Reviewer's
comments
[Document concerns, uncertainties, limitations, and deficiencies and any additional
comments that may highlight study strengths or important elements such as relevance]
\lclnc h Snalial and leninoial \aruihilil\
High
Sampling approach accurately captures variability of environmental contamination in
population/scenario/media of interest based on the heterogeneity/homogeneity and
dynamic/static state of the environmental system. For example:
1. Large sample size (i.e., >10 samples for a single scenario).
2. Use of replicate samples.
3. Use of systematic or continuous monitoring methods.
4. Sampling over a sufficient period of time to characterize trends.
5. For urine, 24-hour samples are collected (vs. first morning voids or spot).
For biomonitoring studies, the timing of sample collected is appropriate based on chemical
properties (e.g., half-life), the pharmacokinetics of the chemical (e.g., rate of uptake and
elimination), and when the exposure event occurred.
Medium
Sampling approach likely captures variability of environmental contamination in
population/scenario/media of interest based on the heterogeneity/homogeneity and
dynamic/static state of the environmental system. Some uncertainty may exist, but it is
unlikely to have a substantial impact on results. For example:
1. Moderate sample size (i.e., 5-10 samples for a single scenario), or
2. Use of judgmental (non-statistical) sampling approach, or
3. No replicate samples.
For urine, first morning voids or pooled spot samples.
Low
Sampling approach poorly captures variability of environmental contamination in
population/scenario/media of interest. For example:
1. Small sample size (i.e., <5 samples), or
2. Use of haphazard sampling approach, or
3. No replicate samples, or
4. Grab or spot samples in single space or time, or
5. Random sampling that does not include all periods of time or locations, or
For urine, un-pooled spot samples.
Critically
Deficient
Sample size is not reported.
Single sample collected per data set.
For biomonitoring studies, the timing of sample collected is not appropriate based on
chemical properties (e.g., half-life), the pharmacokinetics of the chemical (e.g., rate of
uptake and elimination), and when the exposure event occurred.
Page 217 of 570
-------
Data Quality
Rating
Description
Not rated/not
applicable
Reviewer's
comments
[Document concerns, uncertainties, limitations, and deficiencies and any additional
comments that may highlight study strengths or important elements such as relevance]
\lclnc 7 Iaihisuiv scenario
High
The data closely represent relevant exposure scenario (i.e., the population/scenario/media of
interest). Examples include:
1. amount and type of chemical/product used
2. source of exposure
3. method of application or by-stander exposure
4. use of exposure controls
microenvironment (location, time, climate)
Medium
The data likely represent the relevant exposure scenario (i.e., population/scenario/media of
interest). One or more key pieces of information may not be described but the deficiencies
are unlikely to have a substantial impact on the characterization of the exposure scenario.
AND/OR
If surrogate data, activities seem similar to the activities within scope.
Low
The data lack multiple key pieces of information, and the deficiencies are likely to have a
substantial impact on the characterization of the exposure scenario.
AND/OR
There are some inconsistencies or possible errors in the reporting of scenario information
(e.g., differences between text and tables in data source, differences between standard
method and actual procedures reported to have been used, etc.) which leads to a lower
confidence in the scenario assessed.
AND/OR
If surrogate data, activities have lesser similarity but are still potentially applicable to the
activities within scope.
Critically
Deficient
If reported, the exposure scenario discussed in the monitored study does not represent the
exposure scenario of interest for the chemical.
Not rated/not
applicable
Reviewer's
commcnls
[Document concerns, uncertainties, limitations, and deficiencies and any additional
comments that may highlight study strengths or important elements such as relevance]
Domain 3. Acccssilnlih clanl\
Mcli'ic S kqiorlin^ nl'ivsulls
High
Supplementary or raw data (i.e., individual data points) are reported, allowing summary
statistics to be calculated or reproduced.
AND
Summary statistics are detailed and complete. Example parameters include:
Page 218 of 570
-------
Data Quality
Rating
Description
1. Description of data set summarized (i.e., location, population, dates, etc.)
2. Range of concentrations or percentiles
3. Number of samples in data set
4. Frequency of detection
5. Measure of variation (coefficient of variation [CV], standard deviation)
6. Measure of central tendency (mean, geometric mean, median)
7. Test for outliers (if applicable)
AND
Both adjusted and unadjusted results are provided (i.e., correction for void completeness in
urine biomonitoring, whole-volume or lipid adjusted for blood biomonitoring, wet or dry
weight for environmental tissue samples or soil samples) [only if applicable].
Medium
Supplementary or raw data (i.e., individual data points) are not reported, and therefore
summary statistics cannot be reproduced.
AND/OR
Summary statistics are reported but are missing one or more parameters (see description for
high).
AND/OR
Only adjusted or unadjusted results are provided, but not both [only if applicable].
Low
Supplementary data are not provided, and summary statistics are missing most parameters
(see description for high).
AND/OR
There are some inconsistencies or errors in the results reported, resulting in low confidence
in the results reported (e.g., differences between text and tables in data source, less
appropriate statistical methods).
Critically
Deficient
There are numerous inconsistencies or errors in the calculation and/or reporting of results,
resulting in highly uncertain reported results.
Not Rated/not
Applicable
Reviewer's
comments
[Document concerns, uncertainties, limitations, and deficiencies and any additional
comments that may highlight study strengths or important elements such as relevance]
\ lei lie Oualil\ assurance
High
The study QA/QC measures and all pertinent quality assurance QA/QC information is
provided in the data source or companion source. Examples include:
1. Field, laboratory, and/or storage recoveries
2. Field and laboratory control samples
3. Baseline (pre-exposure) samples
4. Biomarker stability
5. Completeness of sample (i.e., creatinine, specific gravity, osmolality for urine
samples)
AND
No QA/QC issues were identified, or any identified issues were minor and adequately
addressed (i.e., correction for low recoveries, correction for completeness).
Page 219 of 570
-------
Data Quality
Rating
Description
Medium
The study applied and documented QA/QC measures; however, one or more pieces of
QA/QC information is not described. Missing information is unlikely to have a substantial
impact on results.
AND
No QA/QC issues were identified, or any identified issues were minor and addressed (i.e.,
correction for low recoveries, correction for completeness).
Low
OA/OC measures and results were not directlv discussed but are implied throuah the studv's
use of standard field and laboratory protocols.
AND/OR
Deficiencies were noted in QA/QC control measures that are likely to have a substantial
impact on results.
AND/OR
There are some inconsistencies in the QA/QC measures reported, resulting in low confidence
in the QA/QC measures taken and results (e.g., differences between text and tables in data
source).
Critically
Deficient
QA/QC issues have been identified which significantly interfere with the overall reliability
of the study.
Not Rated/not
Applicable
Reviewer's
comments
[Document concerns, uncertainties, limitations, and deficiencies and any additional
comments that may highlight study strengths or important elements such as relevance]
Domain 4 \ analnlih and unccrlainh
Melnc 1" \ ariiihi 111\ and unccilaiim
High
The study characterizes variability in the population/media studied.
AND
Key uncertainties, limitations, and data gaps have been identified.
AND
The uncertainties are minimal and have been characterized.
Medium
The study has limited characterization of variability in the population/media studied.
AND/OR
The study has limited discussion of key uncertainties, limitations, and data gaps.
AND/OR
Multiple uncertainties have been identified but are unlikely to have a substantial impact on
results.
Low
The characterization of variability is absent.
AND/OR
Key uncertainties, limitations, and data gaps are not discussed.
AND/OR
Uncertainties identified may have a substantial impact on the exposure the exposure
assessment
Page 220 of 570
-------
Data Quality
Rating
Description
Critically
Deficient
Estimates are highly uncertain based on characterization of variability and uncertainty.
Not Rated/not
Applicable
C.4.2.3 Data Evaluation Criteria for Experimental Data, as Revised
Table Apx C-3
. Evaluation Criteria for Sources of Experimental Data
Data Quality
Rating
Description
Domain 1 Rcliahilil\
\lclnc 1 Sampling \lclhodoloij\ and ( oiulilions
High
Samples were collected according to publicly available SOPs, methods, protocols, or test
guidelines that are scientifically sound and widely accepted from a source generally known to
use sound methods and/or approaches such as EPA, NIST, American Society for Testing and
Materials, ISO, and ACGIH.
OR
The sampling protocol used was not a publicly available SOP from a source generally known
to use sound methods and/or approaches, but the sampling methodology is clear, appropriate
(/.
-------
Data Quality
Rating
Description
There are some inconsistencies in the reporting of sampling information (e.g., differences
between text and tables in data source, differences between standard method and actual
procedures reported to have been used, etc.) which lead to a low confidence in the sampling
methodology used.
Critically
Deficient
The sampling methodology is not discussed in the data source or companion source.
AND/OR
Sampling methodology is not scientifically sound or is not consistent with widely accepted
methods/approaches for the chemical and media being analyzed (e.g., inappropriate sampling
equipment, improper storage conditions).
AND/OR
There are numerous inconsistencies in the reporting of sampling information, resulting in high
uncertainty in the sampling methods used.
Not rated/not
applicable
Reviewer's
comments
[Document concerns, uncertainties, limitations, and deficiencies and any additional
comments that may highlight study strengths or important elements such as relevance]
\lelnc 2 AikiK lical nvlhodoloij\
High
Samples were analyzed according to publicly available analytical methods that are
scientifically sound and widely accepted (i.e., from a source generally using sound methods
and/or approaches) and are appropriate for the chemical and media of interest. Examples
include EPA SW-846 Methods, NIOSH Manual of Analytical Methods 5th Edition, etc.
OR
The analytical method used was not a publicly available method from a source generally
known to use sound methods and/or approaches, but the methodology is clear and appropriate
(/. e., scientifically sound) and similar to widely accepted protocols for the chemical and media
of interest. All pertinent sampling information is provided in the data source or companion
source. Examples include:
1. extraction method
2. analytical instrumentation (required)
3. instrument calibration
4. LOQ, LOD, detection limits, and/or reporting limits
5. recovery samples
6. biomarker used (if applicable)
7. matrix-adjustment method (/'.e., creatinine, lipid, moisture)
Medium
Analytical methodology is discussed in detail and is clear and appropriate (/.
-------
Data Quality
Rating
Description
Samples were collected at a site and immediately analyzed using an on-site mobile laboratory,
rather than shipped to a stationary laboratory.
Low
Analytical methodology is only briefly discussed. Analytical instrumentation is provided and
consistent with accepted analytical instrumentation/methods. However, most analytical
information is missing and likely to have a substantial impact on results.
AND/OR
Analytical method is not standard/widely accepted, and method validation is limited or not
available.
AND/OR
Samples were analyzed using field screening techniques.
AND/OR
LOQ, LOD, detection limits, and/or reporting limits not reported.
AND/OR
There are some inconsistencies or possible errors in the reporting of analytical information
(e.g., differences between text and tables in data source, differences between standard method
and actual procedures reported to have been used, etc.) which leads to a lower confidence in
the method used.
Critically
Deficient
Analytical methodology is not described, including analytical instrumentation (/.nsuiv
High
Biomarker in a specified matrix is known to have an accurate and precise quantitative
relationship with external exposure, internal dose, or target dose (e.g., previous studies (or the
current study) have indicated the biomarker of interest reflects external exposures).
AND
Biomarker (parent chemical or metabolite) is derived from exposure to the chemical of
interest.
Medium
Biomarker in a specified matrix has accurate and precise quantitative relationship with
external exposure, internal dose, or target dose.
AND
Biomarker is derived from multiple parent chemicals, not only the chemical of interest, but
there is a stated method to apportion the estimate to only the chemical of interest
Page 223 of 570
-------
Data Quality
Rating
Description
Low
Biomarker in a specified matrix has accurate and precise quantitative relationship with
external exposure, internal dose, or target dose.
AND
Biomarker is derived from multiple parent chemicals, not only the chemical of interest, and
there is NOT a stated method to apportion the estimate to only the chemical of interest.
Critically
Deficient
Biomarker in a specified matrix is a poor surrogate (low accuracy and precision) for
exposure/dose.
Not rated/not
applicable
Metric is not applicable to the data source.
Reviewer's
communis
[Document concerns, uncertainties, limitations, and deficiencies and any additional
comments that may highlight study strengths or important elements such as relevance]
Domain 2 Rcnivscnlali\c
Melnc 4 Tcslin
g scenario
High
Testing conditions closely represent relevant exposure scenarios (i.e.,
population/scenario/media of interest). Examples include:
1. amount and type of chemical/product used
2. source of exposure/test substance
3. method of application or by-stander exposure
4. use of exposure controls
5. microenvironment (location, time, climate, temperature, humidity, pressure, airflow)
AND
Testing conducted under a broad range of conditions for factors such as temperature,
humidity, pressure, airflow, and chemical mass/weight fraction (if appropriate).
Medium
The data likely represent the relevant exposure scenario (i.e., population/scenario/media of
interest). One or more key pieces of information may not be described but the deficiencies are
unlikely to have a substantial impact on the characterization of the exposure scenario.
AND/OR
If surrogate data, activities seem similar to the activities within scope.
Low
The data lack multiple key pieces of information and the deficiencies are likely to have a
substantial impact on the characterization of the exposure scenario.
AND/OR
There are some inconsistencies or possible errors in the reporting of scenario information
(e.g., differences between text and tables in data source, differences between standard method
and actual procedures reported to have been used, etc.) which leads to a lower confidence in
the scenario assessed.
AND/OR
If surrogate data, activities have lesser similarity but are still potentially applicable to the
activities within scope.
AND/OR
Testing conducted under a single set of conditions, except for experiments to determine a
weight fraction or concentration in a product.
Page 224 of 570
-------
Data Quality
Rating
Description
Critically
Deficient
Testing conditions are not relevant to the exposure scenario of interest for the chemical.
Not rated/not
applicable
Reviewer's
comments
[Document concerns, uncertainties, limitations, and deficiencies and any additional
comments that may highlight study strengths or important elements such as relevance]
\lclnc 5 Sample si/.e and \ariahi 1 Us
High
Sample size is reported and large enough (i.e., >10 samples) to be reasonably assured that the
samples represent the scenario of interest.
AND
Replicate tests performed and variability across tests is characterized (if appropriate).
Medium
Sample size is moderate (i.e., 5 to <10_samples), thus the data are likely to represent the
scenario of interest.
AND
Replicate tests performed and variability across tests is characterized (if appropriate).
Low
Sample size is small (i.e., <5 samples for most types of experiments or 1 per product for
experiments to determine a weight fraction or concentration in a product), thus the data are
likely to poorly represent the scenario of interest.
AND/OR
Replicate tests were not performed.
Critically
Deficient
Sample size is not reported.
AND/OR
Single sample collected per data set, except for experiments to determine a weight fraction or
concentration in a product.
AND/OR
For biomonitoring studies, the timing of sample collected is not appropriate based on
chemical properties (e.g., half-life), the pharmacokinetics of the chemical (e.g., rate of uptake
and elimination), and when the exposure event occurred.
Not rated/not
applicable
Reviewer's
comments
[Document concerns, uncertainties, limitations, and deficiencies and any additional
comments that may highlight study strengths or important elements such as relevance]
\lclnc <•> Tcni|H)ralil\
High
Source(s) of tested items appears to be current (within 5 years).
Medium
Source(s) of tested items is less consistent with when current or recent exposures (>5 to 15
years) are expected.
Page 225 of 570
-------
Data Quality
Rating
Description
Low
Source(s) of tested items is not consistent with when current or recent exposures (>15 years)
are expected or is not identified.
Critically
Deficient
Temporality of tested items is not reported, discussed, or referenced.
Not rated/not
applicable
Reviewer's
comments
[Document concerns, uncertainties, limitations, and deficiencies and any additional
comments that may highlight study strengths or important elements such as relevance]
Domain 3 Aa.vssilnlil\ ckuil\
Mclnc 7 RenmlintJ ol'ivsulls
High
Supplementary or raw data (i.e.. individual data points) are reported, allowing summary
statistics to be calculated or reproduced.
AND
Summary statistics are detailed and complete. Example parameters include:
1. Description of data set summarized (i.e., location, population, dates, etc.)
2. Range of concentrations or percentiles
3. Number of samples in data set
4. Frequency of detection
5. Measure of variation (CV, standard deviation)
6. Measure of central tendency (mean, geometric mean, median)
7. Test for outliers (if applicable)
AND
Both adjusted and unadjusted results are provided (i.e., correction for void completeness in
urine biomonitoring, whole-volume or lipid adjusted for blood biomonitoring) [only if
applicable].
Medium
Supplementary or raw data (i.e.. individual data points) are not reported, and therefore
summary statistics cannot be reproduced.
AND/OR
Summary statistics are reported but are missing one or more parameters (see description for
high).
AND/OR
Only adjusted or unadjusted results are provided, but not both [only if applicable].
Low
Supplementary data are not provided, and summary statistics are missing most parameters
(see description for high).
AND/OR
There are some inconsistencies or errors in the results reported, resulting in low confidence in
the results reported (e.g., differences between text and tables in data source, less appropriate
statistical methods).
Critically
Deficient
There are numerous inconsistencies or errors in the calculation and/or reporting of results,
resulting in highly uncertain reported results.
Page 226 of 570
-------
Data Quality
Rating
Description
Not rated/not
applicable
Reviewer's
comments
[Document concerns, uncertainties, limitations, and deficiencies and any additional
comments that may highlight study strengths or important elements such as relevance]
Melnc S Oualil\ assurance
High
The study applied quality assurance/quality control (QA/QC) measures and all pertinent
QA/QC information is provided in the data source or companion source. Examples include:
1. Laboratory, and/or storage recoveries.
2. Laboratory control samples.
3. Baseline (pre-exposure) samples.
4. Biomarker stability
5. Completeness of sample (i.e., creatinine, specific gravity, osmolality for urine
samples)
AND
No QA/QC issues were identified, or any identified issues were minor and adequately
addressed (i.e., correction for low recoveries, correction for completeness).
Medium
The study applied and documented QA/QC measures; however, one or more pieces of
QA/QC information is not described. Missing information is unlikely to have a substantial
impact on results.
AND
No QA/QC issues were identified, or any identified issues were minor and addressed (i.e.,
correction for low recoveries, correction for completeness).
Low
QA/QC measures and results were not directly discussed but are implied through the study's
use of standard field and laboratory protocols.
AND/OR
Deficiencies were noted in QA/QC measures that are likely to have a substantial impact on
results.
AND/OR
There are some inconsistencies in the QA/QC measures reported, resulting in low confidence
in the QA/QC measures taken and results (e.g., differences between text and tables in data
source).
Critically
Deficient
QA/QC issues have been identified which significantly interfere with the overall reliability of
the study.
Not rated/not
applicable
Reviewer's
comments
[Document concerns, uncertainties, limitations, and deficiencies and any additional
comments that may highlight study strengths or important elements such as relevance]
Domain 4 Variabilis and unceilainl\
\lclnc Variabilis and unceilaum
Page 227 of 570
-------
Data Quality
Rating
Description
High
The study characterizes variability in the population/media studied.
AND
Key uncertainties, limitations, and data gaps have been identified.
AND
The uncertainties are minimal and have been characterized.
Medium
The study has limited characterization of variability in the population/media studied.
AND/OR
The study has limited discussion of key uncertainties, limitations, and data gaps.
AND/OR
Multiple uncertainties have been identified but are unlikely to have a substantial impact on
results.
Low
The characterization of variability is absent.
AND/OR
Key uncertainties, limitations, and data gaps are not discussed.
AND/OR
Uncertainties identified may have a substantial impact on the exposure the exposure
assessment
Critically
Deficient
Estimates are highly uncertain based on characterization of variability and uncertainty.
Not rated/not
applicable
Reviewer's
comments
[Document concerns, uncertainties, limitations, and deficiencies and any additional
comments that may highlight study strengths or important elements such as relevance]
C.4.2.4 Data Evaluation Criteria for Databases, as Revised
Table Apx C-4. Evaluation Criteria for Sources of Database Data
Data Quality
Rating
Description
Domain 1 Rcliubilil\
Mel l ie 1 Sam pi i n
g nvlhodolog\
High
Widelv accepted sampling methodologies (i.e., from a source aenerallv known to use using
sound methods and/or approaches) were used to generate the data presented in the database.
Example SOPs include USGS's "National Field Manual for the Collection of Water-Quality
Data," EPA's "Ambient Air Sampling" (SESDPROC-303-R5), etc.
Medium
One or more pieces of sampling methodology information is not described, but missing
information is unlikely to have a substantial impact on results.
OR
The sampling methodologies were consistent with sound scientific theory and/or accepted
approaches based on the reported sampling information but may not have followed
Page 228 of 570
-------
Data Quality
Rating
Description
published procedures from a source generally known to use sound methods and/or
approaches.
Low
The sampling methodology was not reported in data source or readily available companion
data source.
Critically
Deficient
The sampling methodologies used were not appropriate for the chemical/media of interest
in the database (e.g., inappropriate sampling equipment, improper storage conditions).
Not rated/not
applicable
Reviewer's
comments
[Document concerns, uncertainties, limitations, and deficiencies and any additional
comments that may highlight study strengths or important elements such as relevance]
Metric 2 Analwical methodoloij\
High
Widely accepted analytical methodologies (i.e., from a source generally using sound
methods and/or approaches) were used to generate the data presented in the database.
Example SOPs include EPA SW-846 Methods, NIOSH Manual of Analytical Methods 5th
Edition, etc.
Medium
The analytical methodologies were consistent with sound scientific theory and/or accepted
approaches based on the reported analytical information but may not have followed
published procedures from a source generally known to use sound methods and/or
approaches.
Low
The analytical methodology was not reported in data source or companion data source.
Critically
Deficient
The analytical methodologies used were not appropriate for the chemical/media of interest
in the database (e.g., method not sensitive enough, not specific to the chemical, out of date).
Not rated/not
applicable
Reviewer's
comments
[Document concerns, uncertainties, limitations, and deficiencies and any additional
comments that may highlight study strengths or important elements such as relevance]
Domain 2 Representative
Metric 3 ( icol:rapine area
High
Geographic location(s) is reported, discussed, or referenced.
Medium
Not applicable. This metric is dichotomous (i.e., high vs. critically deficient).
Low
Not applicable. This metric is dichotomous (i.e., high vs. critically deficient).
Critically
Deficient
Geographic location is not reported, discussed, or referenced.
Not rated/not
applicable
Reviewer's
comments
[Document concerns, uncertainties, limitations, and deficiencies and any additional
comments that may highlight study strengths or important elements such as relevance]
Metric 4. Temporal
High
The data reflect current conditions (within 5 years)
Page 229 of 570
-------
Data Quality
Rating
Description
AND/OR
Database contains robust historical data for spatial and temporal analyses (if applicable).
Medium
The data are less consistent with current or recent exposures (>5 to 15 years)
AND/OR
Database contains sufficient historical data for spatial and temporal analyses (if applicable).
Low
Data are not consistent with when current exposures (>15 years old) may be expected
AND/OR
Database does not contain enough historical data for spatial and temporal analyses (if
applicable).
Critically
Deficient
Timing of sample data is not reported, discussed, or referenced.
Not rated/not
applicable
Reviewer's
comments
[Document concerns, uncertainties, limitations, and deficiencies and any additional
comments that may highlight study strengths or important elements such as relevance]
Melnc 5 Lxnosuiv scenario
High
The data closely represent relevant exposure scenario (i.e., the population/scenario/media of
interest). Examples include:
1. Amount and type of chemical/product used
2. Source of exposure
3. Method of application or by-stander exposure
4. Use of exposure controls
Microenvironment (location, time, climate)
Medium
The data likely represent the relevant exposure scenario (i.e., population/scenario/media of
interest). One or more key pieces of information may not be described but the deficiencies
are unlikely to have a substantial impact on the characterization of the exposure scenario.
AND/OR
If surrogate data, activities seem similar to the activities within scope.
Low
The data lack multiple key pieces of information and the deficiencies are likely to have a
substantial impact on the characterization of the exposure scenario.
AND/OR
There are some inconsistencies or possible errors in the reporting of scenario information
(e.g., differences between text and tables in data source, differences between standard
method and actual procedures reported to have been used, etc.) which leads to a lower
confidence in the scenario assessed.
AND/OR
If surrogate data, activities have lesser similarity but are still potentially applicable to the
activities within scope.
Critically
Deficient
If reported, the exposure scenario discussed in the monitored study does not represent the
exposure scenario of interest for the chemical.
Not rated/not
applicable
Reviewer's
comments
[Document concerns, uncertainties, limitations, and deficiencies and any additional
comments that may highlight study strengths or important elements such as relevance]
Page 230 of 570
-------
Data Quality
Rating
Description
Domain 3 Acccssilnlih clanl\
\k-lnc (¦> \\ailahi 1 il\ of database and sunnoilniij documcnls
High
Database is widely accepted and/or from a source generally known to use sound methods
and/or approaches (e.g., raw data from NHANES, STORET).
Medium
The database may not be widely known or accepted (e.g., state-maintained databases), but
the database is adequately documented with most or all of the following information:
1. Within the database, metadata is present (sample identifiers, annotations, flags,
units, matrix descriptions, etc.) and-data fields are generally clear and defined.
2. A user manual and other supporting documentation is available, or there is
sufficient documentation in the data source or companion source.
Database quality assurance and data quality control measures are defined and/or a QA/QC
protocol was followed.
Low
The database may not be widely known or accepted, and only limited database
documentation is available (see the medium rating).
Critically
Deficient
No information is provided on the database source or availability to the public.
Not rated/not
applicable
Reviewer's
comments
[Document concerns, uncertainties, limitations, and deficiencies and any additional
comments that may highlight study strengths or important elements such as relevance]
\lclnc 7 Rcnoi'lniu ol'ivsulls
High
The database or information source reporting the analysis of the database data is well
organized and understandable by the target audience.
AND
Summary statistics in the data source are detailed and complete. Example parameters
include:
1. Description of data set summarized (i.e., location, population, dates, etc.)
2. Range of concentrations or percentiles
3. Number of samples in data set
4. Frequency of detection
5. Measure of variation (CV, standard deviation)
6. Measure of central tendency (mean, geometric mean, median)
Test for outliers (if applicable)
Medium
The database or information source reporting the analysis of the database data is well
organized and understandable by the target audience.
AND/OR
Summary statistics are missing one or more parameters (see description for high).
Low
The database or information source reporting the analysis of the database data is unclear or
not well organized.
AND/OR
Summary statistics are missing most parameters (see description for high)
AND/OR
There are some inconsistencies or errors in the results reported, resulting in low confidence
in the results reported (e.g., differences between text and tables in data source, less
appropriate statistical methods).
Page 231 of 570
-------
Data Quality
Rating
Description
Critically
Deficient
There are numerous inconsistencies or errors in the calculation and/or reporting of results,
resulting in highly uncertain reported results.
AND/OR
The information source reporting the analysis of the database data is missing key sections or
lacks enough organization and clarity to locate and extract necessary information.
Not rated/not
applicable
Reviewer's
comments
[Document concerns, uncertainties, limitations, and deficiencies and any additional
comments that may highlight study strengths or important elements such as relevance]
Domain 4. Yanalnlih and uiKviUnnh
\lclnc N Viiruibi 111\ and uiKvilainl\
High
Variability, key uncertainties, limitations, and/or data gaps have been identified.
AND/OR
The uncertainties are minimal and have been characterized.
Medium
The study has limited discussion of variability, key uncertainties, limitations, and/or data
gaps.
AND/OR
Multiple uncertainties have been identified but are unlikely to have a substantial impact on
results.
Low
Variability, key uncertainties, limitations, and data gaps are not discussed.
AND/OR
Uncertainties identified may have a substantial impact on the exposure the exposure
assessment
Critically
Deficient
Estimates are highly uncertain based on characterization of variability and uncertainty.
Not rated/not
applicable
Reviewer's
comments
[Document concerns, uncertainties, limitations, and deficiencies and any additional
comments that may highlight study strengths or important elements such as relevance]
C.5 Evidence Integration
As described in Section 7 of the 2021 Draft Systematic Review Protocol ( ), evidence
integration refers to the consideration of evidence obtained from systematic review and scientific
information obtained from sources that did not undergo systematic review to implement a weight of
scientific evidence approach. The weight of scientific evidence is defined as "a systematic review
method, applied in a manner suited to the nature of the evidence or decision, that uses a pre-established
protocol to comprehensively, objectively, transparently, and consistently identify and evaluate each
stream of evidence, including strengths, limitations, and relevance of each study and to integrate
evidence as necessary and appropriate based upon strengths, limitations, and relevance" (40 CFR
702.33).The consideration of the quality and relevance of the data, while taking into account the
strengths and limitations of the data, to appropriately evaluate the evidence for this supplement, is
described in Section 7 of the 2021 Draft Systematic Review Protocol ( 21a).
Page 232 of 570
-------
TableApx C-5 and TableApx C-6, originally from Section 7.3 of the 2021 Draft Systematic Review
Protocol, provide general considerations and examples of factors that contribute to the strength of
evidence for each evidence stream and example weight of scientific evidence judgments based on these
general considerations, respectively, when evaluating potentially relevant exposure data for this
supplement (U.S. EPA. 2021a).
Table Apx C-5. Considerations that Inform Evaluations of the Strength of the Evidence
Considerations
Factors that
Increase Strength
Factors that
Decrease Strength
The overall weight of scientific evidence judgment considers the general considerations below as well as
chemical-specific considerations to designate each exposure scenario as robust, moderate, slight, or
indeterminate. The designation is a measure of the weight of the evidence supporting the representativeness of
the exposure estimates toward the true distribution of exposure (and releases) for the scenario.
Relevance to
exposure scenario
• Directly relevant to evaluated
exposure scenario
• Data used is for an alternative or surrogate
scenario
For modeled
estimates
• Model used has been peer-
reviewed and is being applied in a
manner appropriate to its design
and objective
• Evidence demonstrating implausibility
• Model has not been peer-reviewed and no
ground-truthing has been performed
• Parameterization is not well described,
documented or is not appropriate to the
evaluated scenario
Data quality
• Medium or high data quality
rating (via Data Evaluation)
• Low data quality rating (via Data Evaluation)
• Imprecision or inaccuracy
Data points
• High number of data points
• Low number of data points
• High proportion of data sampled prior to
changes in industry or other relevant
conditions (e.g., OSHA PEL)
Representative of
the whole industry
(for occupational
scenarios)
• Large proportion of sites included
within the exposure scenario were
measured
• Evidence may not be sufficiently
representative of all of the sites for the
exposure scenario
Representative of
the sub-population
• Applicable to most or all of the
different population groups
included within the exposure
scenario
• Information was not available to sufficiently
cover most or all of population groups
included within the exposure scenarios
Consistency
• Consistency and replication
within a study and across studies
• Inexplicable contradictory findings across
studies
Variability
• Variability is accounted in
estimates
• Full distributions of input
parameters
• Variability unaccounted in estimates
Uncertainties
• Uncertainties are low and the
uncertainties are unlikely to
significantly impact exposure
estimates
• Uncertainties that are likely to over- or under-
estimate exposure from the actual exposures
for the exposure scenario
Page 233 of 570
-------
Table Apx C-6. Evaluation of the Weight of Scientific Evidence for Exposure Assessments
Category
Robust
Moderate
Slight
Indeterminate
Overall Weight of
Scientific Evidence
Exposure Scenario
Factors
(e.g., habits, worker
activities, exposure
factors)
• Directly relevant to
evaluated exposure scenario
• Applicable to most or all of
the different population
groups included within the
exposure scenario
• Full distributions of input
parameters
• High or medium quality
data ratings
• The habits, worker
activities, and/or use
patterns are accounted for,
are current
• Uncertainties are low and
the uncertainties are
unlikely to significantly
impact exposure estimates
• Surrogate scenarios from
similar chemicals are used
to infer similar exposures
or emissions.
• Some distribution of input
parameters
• High or medium quality
data ratings
• There is some, but not
complete, documentation
or description of
assumptions, limitations
and uncertainties
• Surrogate scenarios from
similar uses are used to
infer similar use patterns
or habits and practices
• Medium or low
quality data ratings
• Partially supported by
assumptions
• Uncertainties are not
fully known or
documented
• Habits and practices
are not fully known
and there is a high
degree of uncertainty
in defining use
patterns
• Qualitative
descriptions of
exposure without
additional
context.
• No supporting
data on habits and
practices are
available
The consideration
factors and the
categories to the left
result in an overall
weight of scientific
evidence judgment
as one of the
following:
• Robust
• Moderate
• Slight
• Indeterminate
Measured/
Monitored Data
There is measured information
and the temporal and spatial
aspect of the measurements are
well described, relevant and
reflect current conditions
• Medium or high data
quality rating (via Data
Evaluation)
• High number of data points
• Multiple studies or a large
number of data points
which indicate similar
findings
• Large proportion of sites
included within the
exposure scenario were
measured
• Consistency and replication
within a study and across
studies
There is measured
information which does not
reflect current environmental
conditions or does not
correspond to current
activities but provides
evidence of exposure
• Limited number of studies
or limited number of data
points which indicate
similar findings
• Information was not
available to sufficiently
cover most or all of
population groups included
within the exposure
scenarios
• There is some, but not
complete, documentation
or description of
There is limited measured
information and
information and does not
reflect exposure
conditions and does not
correspond to known
activities
• Information was not
available to sufficiently
cover most or all of
population groups
included within the
exposure scenarios
• Assumptions and
uncertainties are not
known or documented
No measured or
monitored data are
available
Page 234 of 570
-------
Category
Robust
Moderate
Slight
Indeterminate
Overall Weight of
Scientifie Evidcnee
• Uncertainties are low and
the uncertainties are
unlikely to significantly
impact exposure estimates
• Sensitivity of the exposure
estimates has been
described and quantified
incorporating assumptions,
limitations, and
uncertainties
assumptions, limitations,
and uncertainties
The consideration
factors and the
categories to the left
result in an overall
weight of scientific
evidence judgment
as one of the
following:
• Robust
• Moderate
• Slight
• Indeterminate
Estimation
Methodology/Data
• The methodology for
deriving the estimate is well
described and the
underlying computational
and/or scientific basis is
robust, has an empirical
basis or well documented
mathematical basis and
considers chemical
specificity (e.g., physical
and chemical properties and
fate)
• Applicable to most or all of
the different population
groups included within the
exposure scenario
(representative)
• Sensitivity of the exposure
estimates has been
described and quantified
incorporating assumptions,
limitations, and
uncertainties
• The methodology for
deriving the estimate is
well described and the
underlying computational
and/or scientific basis is
robust, however there is
uncertainty in the
parameterization or
applicability
• There is some, but not
complete, documentation
or description of
assumptions, limitations
and uncertainties.
• Modeling approach
used to estimate
exposures is not rooted
in scientific rigor or
does not
mathematically
represent the exposure
scenario;
parameterization is not
complete or does not
utilize the best
available science.
• Assumptions and
uncertainties are not
known or documented
• Modeling
approach is not
available for the
scenario or lack
of information on
parameters
prohibits use of
available models.
Comparison of
Estimated and
Measured
Exposures {if both
estimated and
• There are comparable
estimates using alternate
approaches
• There is concordance
between measured and/or
reported and modeled
• Modeled estimates and
measured exposure
values are comparable,
however differences in
methodology, collection,
or context make it
• There is a lack of
correspondence
between measured
exposures and modeled
exposure estimates
even when uncertainty
• Category does
not have
indeterminate
criterion.
Page 235 of 570
-------
Category
Robust
Moderate
Slight
Indeterminate
Overall Weight of
Scientifie Evidcnee
measured estimates
are used)
estimates/predictions for the
same exposure scenario
• Sensitivity of the exposure
estimates has been
described and quantified
incorporating assumptions,
limitations, and
uncertainties
difficult to arrive at full
concordance
• There is some, but not
complete, documentation
or description of
assumptions, limitations
and uncertainties
and variability are
accounted for.
• Assumptions and
uncertainties are not
known or documented
Page 236 of 570
-------
C.5.1 Environmental Release and Occupational Exposure
EPA evaluated environmental releases based on reported release data, modeling approaches, and
industry sector information from standard engineering sources such as TRI and DMR. As described in
Appendix E, EPA estimated COU-specific releases where supporting data existed and documented
uncertainties where an absence of such data required a broader application of release estimates.
EPA evaluated occupational exposures based on monitoring data, modeling approaches, and worker
activity information from standard engineering sources and systematic review as described in Appendix
F. EPA used COU-specific assessment approaches where supporting data existed and documented
uncertainties where supporting data were only applicable for broader assessment approaches.
Through public comment and peer review, EPA identified additional sources of information that were
also incorporated into the assessment. Specifically, public commenters shared data on occupational
exposure monitoring, product concentration, process descriptions, and environmental releases.
C.5.2 General Population
General population exposures were evaluated for each exposure pathway based on environmental
release data identified as described above in Section C.4.1, environmental monitoring data identified
through available databases or as described in Section C.4.2, and any other relevant information
identified through systematic review. As described in Section 1, all physical and chemical and fate
properties evaluated in the 2020 RE were used to evaluate the in-scope exposure pathways of the
supplement.
C.5.2.1 General Population: Surface Water
To evaluate the surface water pathway, EPA relied on modeled surface water concentrations based on
environmental release data reported to TRI and DMR (Appendix E.3.1) and releases modeled for other
release types, including DTD and hydraulic fracturing (Appendix E.3.2).
EPA identified ambient surface water monitoring data through the WQP, drinking water monitoring
from PWSs through the UCMR3 database and three state-specific databases (Section 2.3.1.1). EPA used
available surface water monitoring data to confirm the accuracy of model predictions in location-
specific case-studies (Appendix G.2.3.2). In addition, available drinking water monitoring data (see The
Data Quality Evaluation Information for General Population, Consumer, and Environmental Exposure
for 1,4-Dioxane (1,4-D)) were used to provide context and a point of reference for modeled drinking
water concentrations and risk estimates (Section 5.2.2.1.5) ( ?24y).
C.5.2.2 General Population: Groundwater
To evaluate the land pathway (groundwater) releases, EPA relied on environmental release data reported
to TRI (Section 2.2.1.1 and Appendix E.4.1) and releases modeled for hydraulic fracturing operations
(Appendix E.4.2).
EPA identified groundwater monitoring data for 1,4-dioxane through the WQP as presented in Section
2.3.2.1 and described in Appendix H.l. Furthermore, EPA contextualized potential groundwater
concentrations identified in the literature through systematic review (see The Data Quality Evaluation
Information for General Population, Consumer, and Environmental Exposure for 1,4-Dioxane (1,4-D))
using search terms identified in Appendix C.2 ( 2024y).
Page 237 of 570
-------
C.5.2.3 General Population Exposure: Ambient Air
EPA did not identify quantitative outdoor air monitoring data for 1,4-dioxane through systematic
review. To evaluate the air pathway, EPA relied on modeled air concentrations based on industrial
releases reported to TRI (Section 2.3.3.2.2 and Appendix E.5.1), releases modeled for laundry facilities
(Section 2.3.3.2.4 and Appendix E.12), and releases modeled for hydraulic fracturing operations
(Section 2.3.3.2.4 and Appendix E.13).
Page 238 of 570
-------
Appendix D COU-OES MAPPING AND CROSSWALK
This appendix contains additional information about the relationship between the COUs and OESs
determined for 1,4-dioxane.
Condition of Use (COU): TSCA section 3(4) defines COUs as "the circumstances, as determined by the
Administrator, under which a chemical substance is intended, known, or reasonably foreseen to be
manufactured, processed, distributed in commerce, used, or disposed of'. COUs included in the scope of
EPA's risk evaluations are typically tabulated in scope documents and risk evaluation documents as
summaries of life cycle stages, categories, and subcategories of use. Therefore, a COU is composed of a
combination of life cycle stage, category, and subcategory. COU development may include Chemical
Data Reporting (CDR) information, market profile information, and literature sources. Early in the risk
evaluation process, EPA maps each COU to an occupational exposure scenario for the environmental
release and occupational exposure assessment.
Occupational Exposure Scenario (OES): This term is intended to describe the grouping or segmenting
of COUs for assessment of releases and exposures. For example, EPA may assess a group of multiple
COUs together as one OES due to similarities in release and exposure sources, worker activities, and use
patterns. Alternatively, EPA may assess multiple OES for one COU because there are different release
and exposure potentials for a given COU. OES determinations are also largely driven by the availability
of data and modeling approaches to assess occupational releases and exposures. For example, even if
there are similarities between multiple COUs, if there is sufficient data to separately assess releases and
exposures for each COU, EPA would not group them into the same OES.
D.l COU-OES Mapping
The details of an identified COU will determine the number of associated OES(s). Mapping OES to
COUs may come in many forms, including a direct one-to-one mapping of a single OES to a single
COU, mapping of one OES to multiple COUs, or mapping of multiple OES to a single COU, as shown
in Figure Apx D-l. The OES mapping is driven by similarities and differences in the expected
occupational exposures and releases for a COU and the reasonably available data to estimates such
exposures and releases, as discussed in Section 2.1.1. Further, there may be differences in the name of
an OES from the name of the COU to which it is mapped. This is because OES names are intended to be
succinct, capture all COUs where one OES is mapped to multiple COUs, and distinct enough to
represent the specific occupational exposure and release scenario.
Page 239 of 570
-------
Condition of Use
OES
Life Cycle
Stage
Category
Subcategory
Processing
Byproduct
Byproduct produced during
production of PET
PET Byproduct
Multiple COUs may be mapped to the same OES
Multiple COUs may be mapped to one OES when the COUs have similar activities and exposure
potentials, and exposures and releases can be assessed for the COUs using a single approach
There may be differences between the COU and OES names because the OES is name is intended to
be succinct and encompass all COUs grouped therein under a general name
For example, the 1,4-dioxane COUs for "Industrial wastewater treatment", "underground injection",
"municipal landfill", and others were assessed together under the OES named "disposal" (see excerpt
from crosswalk Table 2-1 and Apx D-l below)
COU liCOU2lCOU3
OES1
Condition of Use
Life Cycle
Stage
Category
Subcategory
OES
Industrial pre-treatment
Industrial wastewater treatment
Publicly owned treatment
works (POTW)
Underground injection
Disposal
Disposal
Municipal landfill
Disposal
Hazardous landfill
Other land disposal
Municipal waste incinerator
Hazardous waste incinerator
Off-site waste transfer
COU1
OESliOES2iOES3
One COU may be mapped to multiple OES
Mapping a COU to multiple OES allows for the assessment of distinct scenarios that are not
expected to result in similar releases and exposures
There may be differences between the COU and OES names because the OES capture more distinct
scenarios of occupational release and exposure than the COU
For example, the 1,4-dioxane COU for "dish soap, dishwasher detergent, laundry detergent" (which
is a single COU) was assessed as separate OES named "dish soap", "dishwasher detergent",
"laundry detergent (industrial)", and "laundry detergent (institutional)" (see excerpt from crosswalk
Table 2-1 and Apx D-l below)
Condition of Use
OES
Life Cycle
Stage
Category
Subcategory
Consumer use,
commercial
use
Laundry and
Dishwashing
Products
Dish soap
Dishwasher detergent
Laundry detergent
Dish Soap
Dishwasher Detergent
Laundry Detergent
(Industrial)
Laundry Detergent
(Institutional)
Figure Apx D-l. COU and OES Mapping
Page 240 of 570
-------
D.2 COU-OES Crosswalk
A crosswalk of the COU with the OES assessed is provided in TableApx D-l. As discussed in Section
2.1.1, a COU is a combination of life cycle stage, category, and subcategory and EPA mapped each
COU to an OES. The purpose of an OES is to group, where appropriate, COUs based on similarity of
the operations and data availability for each COU. EPA assessed environmental releases (air, water, and
land) and occupational exposures (inhalation and dermal) to 1,4-dioxane for each of the OES listed in
TableApx D-l. As noted in this table, some of these OESs were in scope of the Final Risk Evaluation
for 1,4-Dioxane (U.S. EPA. 2020c) while others were in scope of this supplemental risk evaluation.
Table Apx D-l. Categories and Subcategories of Conditions of Use Included in the Scope of the
Risk Evaluation
Condition of Use
OES
Risk Evaluation in Which
Occupational Exposures
Were Assessed
Life Cycle
Stage
Category"
Subcategory''
Manufacturing
Domestic
Manufacture
Domestic Manufacture
Manufacturing
2020 RE
Import
Import
Repackaging
Import and repackaging
2020 RE
Processing
Processing as a
Reactant
Polymerization catalyst
Industrial uses
2020 RE
Non-
incorporative
Basic organic chemical
manufacturing (process
solvent)
Byproduct
Byproduct produced
during processes
Ethoxylation process
byproduct
Supplemental RE
Byproduct produced
during production of PET
PET byproduct
Supplemental RE
Recycling
Recycling
Disposal
2020 RE
Distribution in
Commerce
Distribution
Distribution
Distribution activities
(e.g., loading, unloading)
considered throughout life
cycle, rather than using a
single distribution
scenario17
N/A
Industrial Use
Intermediate use
Plasticizer intermediate
Catalysts and reagents for
anhydrous acid reactions,
brominations, and
sulfonations
Industrial uses
2020 RE
Processing aids,
not otherwise
listed
Wood pulping
Extraction of animal and
vegetable oils
Wetting and dispersing
agent in textile processing
Polymerization catalyst
Purification of process
intermediates
Etching of fluoropolymers
Page 241 of 570
-------
Condition of Use
OES
Risk Evaluation in Which
Occupational Exposures
Were Assessed
Life Cycle
Stage
Category"
Subcategory''
Functional
fluids (open and
closed system)
Polyalkylene glycol
lubricant
Functional fluids (open-
system)
2020 RE
Synthetic metalworking
fluid
Cutting and tapping fluid
Hydraulic fluid
Functional fluids (closed-
system)
2020 RE
Industrial Use,
Commercial
Use
Laboratory
chemicals
Chemical reagent
Laboratory chemicals
2020 RE
Reference material
Spectroscopic and
photometric measurement
Liquid scintillation
counting medium
Stable reaction medium
Cryoscopic solvent for
molecular mass
determinations
Preparation of histological
sections for microscopic
examination
Adhesives and
Sealants
Film cement
Film cement
2020 RE
Other Uses
Spray polyurethane foam;
Printing and printing
compositions, including
3D printing; dry film
lubricant; Hydraulic
fracturing
Spray foam application
2020 RE
Printing inks (3D)
2020 RE
Dry film lubricant
2020 RE
Hydraulic Fracturing
Supplemental RE
Consumer
Use,
Commercial
Use
Paints and
Coatings
Latex wall paint or floor
lacquer
Paint and floor lacquer
Supplemental RE
Cleaning and
Furniture Care
Products
Surface cleaner
Surface Cleaner
Supplemental RE
Laundry and
Dishwashing
Products
Dish soap
Dishwasher detergent
Laundry detergent
Dish soap
Dishwasher detergent
Laundry detergent
(industrial)
Laundry detergent
(institutional)
Supplemental RE
Arts, Crafts, and
Hobby
Materials
Textile dye
Textile dye
Supplemental RE
Automotive
Care Products
Antifreeze
Antifreeze
Supplemental RE
Other Consumer
Uses
Spray polyurethane foam
Spray foam application
2020 RE
Page 242 of 570
-------
Condition of Use
OES
Risk Evaluation in Which
Occupational Exposures
Were Assessed
Life Cycle
Stage
Category"
Subcategory''
Disposal
Disposal
Industrial pre-treatment
Industrial wastewater
treatment
Publicly owned treatment
works (POTW)
Underground injection
Municipal landfill
Hazardous landfill
Other land disposal
Municipal waste
incinerator
Hazardous waste
incinerator
Off-site waste transfer
Disposal
2020 RE
" These categories of conditions of use reflect CDR rule codes and broadly represent conditions of use for 1,4-dioxane in
industrial and/or commercial settings.
h These subcategories reflect more specific uses of 1,4-dioxane.
c Potential releases and exposures from loading and unloading are considered throughout life cycle, for each OES. This
includes handling of both neat 1,4-dioxane and product formulations containing 1,4-dioxane.
Page 243 of 570
-------
Appendix E INDUSTRIAL AND COMMERCIAL
ENVIRONMENTAL RELEASES
This appendix contains additional information relevant to the assessment of industrial and commercial
environmental releases.
E.l Estimates of the Number of Industrial and Commercial Facilities with
Environmental Releases
As a part of the assessment of industrial and commercial environmental releases, EPA estimated the
number of facilities with releases for each OES. Where available, EPA used 2013 to 2019 TRI (U.S.
22h) and 2013 to 2019 DMR ( )22c) data to provide a basis to estimate the number
of sites using 1,4-dioxane within an OES. Additional information on how EPA utilized TRI and DMR to
estimate the number of sites using 1,4-dioxane within a COU can be found in Section 2.2.1.2.2 of the
December 2020 Final Risk Evaluation for 1,4-Dioxane ( s20c).
Where the number of sites could not be determined using TRI or DMR or where these data were
determined to not capture the entirety of sites within an OES, EPA supplemented the available data with
U.S. economic data using the following methods:
• Identify the North American Industry Classification System (NAICS) codes for the industry
sectors associated with these uses.
• Estimate total number of sites using the U.S. Census" Statistics of US Businesses (SUSB) (U.S.
Census Bureau. 2015) data on total establishments by 6-digit NAICS.
• Review available ESDs and GSs for established facility estimates for each occupational exposure
scenario.
• Combine the data generated in bullets 1 through 3 to produce an estimate of the number of sites
using 1,4-dioxane in each 6-digit NAICS code and sum across all applicable NAICS codes for
the COU, augmenting as needed with data from the ESDs and GSs, to arrive at a total estimate of
the number of sites within the COU.
A summary of the number of facilities EPA determined for each OES and each type of release is shown
in Table Apx E-l. The number of facilities may be different for each type of release within the same
OES if sufficient data were available to make this differentiation.
Table Apx E-l. Summary of EPA's Estimates for the Number of Facilities for Each OES
OES
Type of Release
Number of
Facilities
Notes
Manufacturing
Air, Land
1
Based on 2019 TRI reDortina (U.S. EPA. 2022M
Surface Water
1
Based on 2013-2019 DMR and TRI reporting (U.S.
EPA. 2022c. M
POTW or Non-
POTWWWT
1
Based on 2013-2019 TRI reoortina (U.S. EPA. 2022M
Import and
repackaging
Air, Land
1
Based on 2019 TRI reoortina (U.S. EPA. 2022h)
Surface Water
6
Based on 2013-2019 DMR and TRI reporting (U.S.
EPA. 2022c. h)
POTW or Non-
POTWWWT
6
Based on 2013-2019 TRI reoortina (U.S. EPA. 2022M
Page 244 of 570
-------
OES
Type of Release
Number of
Facilities
Notes
Industrial uses
Air, Land
12
Based on 2019 TRI reDortina (U.S. EPA. 2022M
Surface Water
24
Based on 2013-2019 DMR and TRI reporting (U.S.
EPA. 2022c. h)
POTW or Non-
POTWWWT
17
Based on 2013-2019 TRI reDortine (TJ.S. EPA. 2022M
Functional fluids
(open-system)
Air, Land
2
Based on 2019 TRI reDortina (U.S. EPA. 2022M
Surface Water
6
Based on 2013-2019 DMR and TRI reporting (U.S.
EPA. 2022c. h)
POTW or Non-
POTWWWT
1
Based on 2013-2019 TRI reDortina ("U.S. EPA. 2022M
Functional fluids
(closed-system)
All
N/A
Assessed as a part of Industrial Uses OES
Laboratory
chemical
All
132
Calculated using the GS on Use of Laboratory
Chemicals (U.S. EPA, 2022i) and the amount of 1.4-
dioxane used in laboratory uses per the December 2020
Final Risk Evaluation for 1,4-Dioxane ( A,
2020c)
Film cement
All
211
Based on the number of sites for this OES in the
December 2020 Final Risk Evaluation for 1,4-Dioxane
(U.S. EPA, 2020c). which is a boundina estimate based
on U.S. Census Bureau data forNAICS code 512199,
Other Motion Picture and Video Industries
Spray foam
application
All
1,553,559
Based on the number of sites for this OES in the
December 2020 Final Risk Evaluation for 1,4-Dioxane
(U.S. EPA, 2020c). which is a boundina estimate based
on U.S. Census Bureau data forNAICS code 238310,
Drywall and Insulation Contractors
Printing inks
(3D)
Air, Land
N/A
Assessed as a part of Industrial Uses OES
Surface Water,
POTW, non-POTW
WWT
1
Based on 2013-2019 DMR and TRI reporting (U.S.
EPA. 2022c. hV
Dry film
lubricant
All
8
Based on the number of sites for this OES in the
December 2020 Final Risk Evaluation for 1,4-Dioxane
(U.S. EPA. 2020c). which is based on conversations
with the Kansas City National Security Campus
(manufacturer and uses of dry film lubricants)
Disposal
Air
15
Based on 2019 TRI reportina (U.S. EPA. 2022hV
Surface Water
24
Based on 2013-2019 DMR and TRI reporting (U.S.
EPA. 2022c. h)
POTW or Non-
POTW WWT,
Land
4
Based on 2013-2019 TRI reDortina (U.S. EPA. 2022M
Textile dye
All
783
Bounding estimate based on U.S. Census Bureau data
forNAICS code 313310, Textiles and Fabric Finishing
Mills
Page 245 of 570
-------
OES
Type of Release
Number of
Facilities
Notes
Antifreeze
All
84,383
Bounding estimate based on U.S. Census Bureau data
for NAICS codes 811111, General Automotive Repair,
and 811198, All Other Automotive Repair and
Maintenance
Surface cleaner
All
Unknown
within
Liverpool OH
Land release estimates for this OES were developed for
the Liverpool, OH case study and the number of sites
within this locality is unknown (e.g., the release
estimates are not per site but for the entire locality)
Dish soap
All
773,851
(industry
bounding
estimate)
Bounding estimate for the industry is based on U.S.
Census Bureau data for NAICS codes 623300, 713900,
721100, 721300, 722300, 722400, and 722500
Dishwasher
detergent
All
773,851
(industry
bounding
estimate)
Bounding estimate for the industry is based on U.S.
Census Bureau data for NAICS codes 623300, 713900,
721100, 721300, 722300, 722400, and 722500
Laundry
detergent
(institutional)
All
95,533
Bounding estimate based on industry information as
described in the ESD on Water Based Washing
Operations at Industrial and Institutional Laundries
(OECD.: )
Laundry
detergent
(industrial)
All
2,453
Bounding estimate based on U.S. Census Bureau data
for NAICS code 812330, Linen and Uniform Supply
Paints and floor
lacquer
All
33,648
Bounding estimate based on U.S. Census Bureau data
for NAICS code 811121, Automotive Body, Paint, and
Interior Repair and Maintenance
Polyethylene
terephthalate
(PET) byproduct
Air, Land
13
Based on 2019 TRI reDortina (U.S. EPA. 2022h)
Surface Water
19
Based on 2013-2019 DMR and TRI reporting (U.S.
EPA. 2022c. M
POTW or Non-
POTWWWT
14
Based on 2013-2019 TRI reDortina ("U.S. EPA. 2022h)
Ethoxylation
process
byproduct
Air, Land
8
Based on 2019 TRI reDortina (U.S. EPA. 2022h)
Surface Water
7
Based on 2013-2019 DMR and TRI reporting (U.S.
EPA. 2022c. M
POTW or Non-
POTWWWT
6
Based on 2013-2019 TRI reDortina (U.S. EPA. 2022M
Hydraulic
fracturing
All
411
Based on the number of sites that reported using 1,4-
dioxane to FracFocus 3.0 (GWPC and IOGCC, 2022)
E.2 Estimates of Number of Release Days for Industrial and Commercial
Releases
As a part of the assessment of industrial and commercial environmental releases, EPA also estimated the
number of release days for each OES. EPA referenced the December 2020 Final Risk Evaluation for
1,4-Dioxane ( 320c), GSs, ESDs, or made assumptions when estimating release days for
each OES. In summary, EPA estimated the number of operating days using the below sources of data:
Page 246 of 570
-------
1.
Facility-Specific Data: Use facility-specific data if available. If facility-specific data is not
available, estimate the days/year using one of the following approaches:
a. If facilities have known or estimated average daily use rates, calculate the days/year as:
Days/year = Estimated Annual Use Rate for the Site (kg/year) / average daily use rate
from sites with available data (kg/day).
b. If sites with days/year data do not have known or estimate average daily use rates, use the
average number of days/year from the sites with such data.
2. Industry-Specific Data: Industry-specific data may be available in the form of GSs, ESDs, trade
publications, or other relevant literature. In such cases, these estimates should take precedent
over other approaches, unless facility-specific data are available.
3. Manufacture of Lower-PV Specialty Chemicals: For the manufacture of lower-PV specialty
chemicals like 1,4-dioxane, the chemical is not expected to be manufactured continuously
throughout the year. Therefore, a value of 250 days/year should be used. This assumes the plant
manufactures the chemical 5 days/week and 50 weeks/year (with 2 weeks down for turnaround).
For the manufacture of 1,4-dioxane as a byproduct (e.g., ethoxylation process, PET
manufacturing), 250 days/year is also used, assuming these industrial manufacturing facilities
have a similar operating schedule of 5 days/week and 50 weeks/year.
4. Processing as Reactant (Intermediate Use) in the Manufacture of Specialty Chemicals:
Similar to #3, the manufacture of specialty chemicals is not expected to occur continuously
throughout the year. Therefore, a value of 250 days/year can be used.
5. Other Chemical Plant OES (e.gIndustrial Uses): For these OESs, it is reasonable to assume
that 1,4-dioxane is not always in use at the facility, even if the facility operates 24/7. Therefore,
in general, a value of 300 days/year can be used based on the "SpERC [Specific Environmental
Release Categories] fact sheet - Formulation & (re)packing of substances and mixtures -
Industrial (Solvent-borne)" that uses a default of 300 days/year for the chemical industry.
However, in instances where the OES uses a low volume of the chemical of interest, 250
days/year can be used as a lower estimate for the days/year.
6. POTWs: Although POTWs are expected to operate continuously over 365 days/year, the
discharge frequency of 1,4-dioxane from a POTW will be dependent on the discharge patterns of
the chemical from the upstream facilities discharging to the POTW. However, there can be
multiple upstream facilities (possibly with different OES) discharging to the same POTW and
information to determine when the discharges from each facility occur on the same day or
separate days is typically not available. Therefore, an exact number of days/year the 1,4-dioxane
is discharged from the POTW cannot be determined and a value of 365 days/year should be
used.
7. All Other OESs: Regardless of the facility operating schedule, other OESs are unlikely to use
1,4-dioxane every day. Therefore, a value of 250 days/year should be used for these OESs.
A summary along with a brief explanation is presented in Table Apx E-2 below. These estimates of
release days are applicable to the air and water release estimates for each OES; however, there is a high
level of variability and uncertainty associated with the number of days of release associated with land
releases. Therefore, EPA could not estimate the number of days of release for land releases.
Page 247 of 570
-------
TableApx E-2. Summary of EPA's Estimates for Air and Water Release Days Expected for Each
OES
OES
Release
Days
Notes
Manufacturing
250
Assumed 5 days per week and 50 weeks per year with 2 weeks per year for
shutdown activities, consistent with the 2020 RE ("U.S. EPA. 2020c)
Import and
repackaging
250
Assumed 5 days per week and 50 weeks per year with 2 weeks per year for
shutdown activities, consistent with the December 2020 Final Risk
Evaluation for 1,4-Dioxane ("U.S. EPA. 2020c)
Industrial uses
250
Assumed 5 days per week and 50 weeks per year with 2 weeks per year for
shutdown activities, consistent with the December 2020 Final Risk
Evaluation for 1,4-Dioxane (U.S. EPA, 2020c)
Functional fluids
(open-system)
247
Per the 2011 OECD Emission Scenario Document on the Use of
Metalworking Fluids, consistent with the December 2020 Final Risk
Evaluation for 1,4-Dioxane (U.S. EPA. 2020c).
Functional fluids
(closed-system)
N/A
Assessed as a part of Industrial Uses OES.
Laboratory chemical
250
Assumed 5 days per week and 50 weeks per year with 2 weeks per year for
shutdown activities, consistent with the December 2020 Final Risk
Evaluation for 1,4-Dioxane (U.S. EPA. 2020c).
Film cement
250
Assumed 5 days per week and 50 weeks per year with 2 weeks per year for
shutdown activities, consistent with the December 2020 Final Risk
Evaluation for 1,4-Dioxane (U.S. EPA, 2020c).
Spray foam
application
3
Per the 2018 EPA generic scenario Application of Spray Polyurethane Foam
Insulation, consistent with the December 2020 Final Risk Evaluation for 1,4-
Dioxane (U.S. EPA, 2020c). Releases occur at the residence or a site where
the SPF is applied and not at the SPF application company location. Each
SPF application job takes 3 days; however, employees may apply SPF at
multiple locations throughout a year, resulting in an overall number of
exposure days higher than 3 days/yr.
Printing inks (3D)
250
Assumed 5 days per week and 50 weeks per year with 2 weeks per year for
shutdown activities, consistent with the December 2020 Final Risk
Evaluation for 1,4-Dioxane (U.S. EPA. 2020c).
Dry film lubricant
48
Per process description information provided in the December 2020 Final
Risk Evaluation for 1,4-Dioxane (U.S. EPA, 2020c).
Disposal
250
Assumed 5 days per week and 50 weeks per year with 2 weeks per year for
shutdown activities, consistent with the December 2020 Final Risk
Evaluation for 1,4-Dioxane (U.S. EPA. 2020c).
Textile dye
31-295
Based on the 2015 OECD on Textile Dves (OECD. 2017) and Monte Carlo
Modeling.
Antifreeze
250
Assumed 5 days per week and 50 weeks per year with 2 weeks per year for
shutdown activities.
Surface cleaner
350
Assumed 7 days per week and 50 weeks per year with 2 weeks per year for
shutdown activities.
Page 248 of 570
-------
OES
Release
Days
Notes
Dish soap
350
Assumed 7 days per week and 50 weeks per year with 2 weeks per year for
shutdown activities.
Dishwasher
detergent
250
Assumed 5 days per week and 50 weeks per year with 2 weeks per year for
shutdown activities.
Laundry detergent
(institutional)
250-
365
Based on the 2011 OECD ESD on Industrial and Institutional Laundries
(¦ M CD. .V1 1and Monte Carlo Modeling.
Laundry detergent
(industrial)
20-365
Based on the 2011 OECD ESD on Industrial and Institutional Laundries
(¦ M CD. .V1 11») and Monte Carlo Modeling.
Paints and floor
lacquer
250
Based on the 2011 OECD ESD on Coating Application via Spray Painting in
the Automotive Refinishine Industry (OECD, 2011a).
Polyethylene
terephthalate (PET)
byproduct
250
Assumed 5 days per week and 50 weeks per year with 2 weeks per year for
shutdown activities.
Ethoxylation
process byproduct
250
Assumed 5 days per week and 50 weeks per year with 2 weeks per year for
shutdown activities.
Hydraulic fracturing
1-72
Based on the reported number of days for sites that use 1,4-dioxane in
FracFocus 3.0 (GWPC and IOGCC. 2022). This ranse of release davs refers
to only the hydraulic fracturing and not post-fracturing production stages.
EPA's estimates for flowback and produced water releases during production
staaes occur over 350 davs/vear (U.S. EPA. 2022e).
E.3 Water Release Assessment
This section describes EPA's methodology for estimating daily wastewater discharges from industrial
and commercial facilities manufacturing, processing, or using 1,4-dioxane. Facilities report wastewater
discharges either via Discharge Monitoring Reports (DMRs) under the NPDES or TRI. EPA used 2013
to 2019 DMR (U.S. EPA. 2022c) and 2013 to 2019 TRI ( 022h) data to estimate daily
wastewater discharges for the OES where available; however, EPA did not have these data for every
OES. For OES without DMR and TRI data, EPA used alternate assessment approaches to estimate
wastewater discharges. Both approaches—one for OESs with DMR and TRI data and the other for OESs
without these data—are described below.
E.3.1 Assessment Using TRI and DMR
EPA found 2013 to 2019 DMR and/or 2013 to 2019 TRI data for facilities within the following OESs:
• Manufacturing;
• Import and repackaging;
• Industrial uses;
• Functional fluids (open-system);
• 3D printing;
• Disposal;
• PET byproduct; and
• Ethoxylation byproduct.
Page 249 of 570
-------
The 2013 to 2019 TRI data were rated medium in EPA's systematic review process and the 2013 to
2019 DMR were rated "medium." EPA estimated daily discharges using TRI and DMR data for these
OESs, with the following general stages as described in the rest of this section:
1. Collect wastewater discharge data from 2013 to 2019 DMR and TRI data,
2. Map wastewater discharge data to occupational exposure scenarios,
3. Estimate the number of facility operating days per year, and
4. Estimate daily wastewater discharges and summarize wastewater discharges for each OES.
Note that EPA compared the TRI and DMR data used to estimate water releases for the PET byproduct
OES in this risk evaluation to information from a life cycle analysis on the PET manufacturing process
in Appendix E.6.
Step 1: Collect Wastewater Discharge Data from DMR and TRI
The first step in estimating daily releases was to obtain 2013 through 2019 DMR and TRI data. Under
the CWA, EPA regulates the discharge of pollutants into receiving waters through NPDES. A NPDES
permit authorizes discharging facilities to discharge pollutants to specified limits. NPDES permits apply
pollutant discharge limits to each outfall at a facility. For risk evaluation purpose, EPA is interested only
in the outfalls to surface water body. NPDES permits also include internal outfalls, but they are not
included in this analysis. This is because these outfalls are internal monitoring points within the facility
wastewater collection or treatment system, so they do not represent discharges from the facility. The
permits require facilities to monitor their discharges and report the results to EPA and the state
regulatory agency. Facilities report these results in DMRs. EPA makes these reported data publicly
available via EPA's ECHO system and EPA's Water Pollutant Loading Tool (Loading Tool). The
Loading Tool is a web-based tool that obtains DMR data through ECHO, presents data summaries and
calculates pollutant loading (mass of pollutant discharged). EPA queried the ECHO Loading Tool to
pull data for each of years 2013 through 2019. EPA removed facilities reporting zero discharges for 1,4-
dioxane in DMR from the analysis because EPA cannot confirm if the pollutant is present at the facility.
Each facility subject to the TRI reporting rule must report annually the volume of chemical released to
the environment and/or managed through recycling, energy recovery, and treatment. Unlike DMR, TRI
includes both reports of annual direct discharges to surface water and annual indirect discharges to off-
site publicly owned treatment works (POTW) and wastewater treatment (WWT) facilities (non-POTW
WWT). Similar to the air release assessment, EPA included both TRI reporting Form R and TRI
reporting Form A submissions in the water release assessment. Where sites reported to TRI with Form
A, EPA used the Form A threshold for total releases of 500 lb/year. EPA used the entire 500 lb/year for
both direct and indirect wastewater discharges; however, since this threshold is for total site releases,
these 500 lb/year are attributed either to direct discharges or indirect discharges for this analysis, not
both (since that would double count the releases and exceed the total release threshold for Form A). EPA
pulled the TRI Basic Plus Data Files for each of years 2013 through 2019.
In summary, wastewater discharges reported to DMR and TRI include the following:
• DMR:
o
• TRI:
o
o
o
Page 250 of 570
On-site releases to surface water (direct discharges).
On-site releases to surface water (direct discharges),
Off-site transfers to POTWs (indirect discharges), and
Off-site transfers to non-POTW WWT (indirect discharges).
-------
Note that the two datasets are not updated concurrently. The Loading Tool automatically and
continuously checks ICIS-NPDES for newly submitted DMRs. The Loading Tool processes the data
weekly and calculates pollutant loading estimates; therefore, water discharge data (DMR data) are
available on a continual basis. Although the Loading Tool process data weekly, each permitted
discharging facility is only required to report their monitoring results for each pollutant at a frequency
specified in the permit (e.g., monthly, every 2 months, quarterly). TRI data is reported annually for the
previous calendar year and is typically released in October (i.e., 2020 TRI data is released in October
2021).
Step 2: Map Wastewater Discharge Data to Occupational Exposure Scenarios
The next step in estimating daily releases was to map 2013 through 2019 DMR and TRI data to the 1,4-
dioxane OES. EPA used the same mapping methodology for the water assessment as that described in
Appendix E.5.1. EPA ensured consistency in the OES mapping for sites that reported to both TRI and
DMR. EPA also ensured consistency in the OES mapping between the air, water, and land assessments.
Step 3: Estimate the Number of Facility Operating Days per Year
EPA then estimated the number of operating days (days/year) for each facility reporting wastewater
discharges to DMR and TRI. EPA generally used the same number of operating days for the same OES
for both the air and water analysis, which is based on the general methodology described previously in
Appendix E.2.
Step 4: Estimate Daily Wastewater Discharges and Summarize Wastewater Discharges for each OES
After the initial steps of selecting and mapping of the water discharge data and estimating the number of
facility operating days/year were completed, the next step was to summarize annual and daily
wastewater discharges for each OES. EPA summarized annual wastewater discharges reported in DMR
and TRI for each facility. EPA estimated daily wastewater discharges separately for direct and indirect
discharges, as discussed below.
EPA estimated the median and maximum daily direct wastewater discharges at each facility, using the
steps below. EPA presented the calculated median and maximum daily direct wastewater discharged
separately for the DMR and TRI datasets because these data do not always agree/match.
1. Obtained total annual loads calculated from the Loading Tool and reported annual surface water
discharges in TRI for years 2013 through 2019.
2. Divided the annual direct discharge over the number of estimated operating days for the OES to
which the facility has been mapped. The number of operating days differ for each OES, as
summarized in Appendix E.2.
3. Calculated the median daily direct wastewater discharge across all years of data for each facility,
separately for both DMR and TRI data.
4. Identified the maximum daily direct wastewater discharge across all years of data for each
facility. EPA also noted which reporting year had this maximum daily direct wastewater
discharge, separately for both DMR and TRI data.
For indirect discharges to POTW or non-POTW WWT, EPA estimated the average daily indirect
discharges for each facility and each reporting year (2013-2019) in TRI using steps #1 and #2 above.
DMR data do not include indirect discharges. EPA did not estimate the median or maximum daily
indirect discharges across all years.
Page 251 of 570
-------
A summary of the estimated daily discharges using 2013 to 2019 DMR and TRI is included in 1,4-
Dioxane Supplemental Information File: Environmental Releases to Water for OES with TRI and DMR
(U.S. EPA. 2024n).
E.3.2 Assessment for OES Without TRI and DMR
EPA did not find DMR or TRI data for any of the years included in this analysis for the following OESs:
• Functional fluids (closed-systems);
• Laboratory chemicals;
• Film cement;
• Spray polyurethane foam;
• Dry film lubricant;
• Textile dye;
• Antifreeze;
• Surface cleaner;
• Dish soap;
• Dishwasher detergent;
• Laundry detergent;
• Paints and floor lacquer; and
• Hydraulic fracturing.
For these OESs, EPA estimated daily wastewater discharges by using various modeling approaches—
including the use of surrogate TRI and DMR data and modeling using data from literature, GSs, and
ESDs. EPA's assessment of daily wastewater discharges for each of these OESs is described below.
Functional Fluids (Closed-Systems)
Wastewater discharge data were not available for this OES and EPA did not find any information to
model wastewater discharges for this OES using literature, GSs, or ESDs. EPA expects that the sources
of release for this OES to be similar to those for the Industrial Uses OES, based on the process
information in the Final Risk Evaluation for 1,4-Dioxane ( 2020c). Therefore, EPA grouped
the water release assessment for Functional Fluids (Closed-Systems) into that for Industrial Uses.
However, there is uncertainty in this assumption of similar release sources between these OESs.
Laboratory Chemicals
EPA estimated daily wastewater discharges for facilities within the Laboratory chemicals OESs using
the Draft GS on Use of Laboratory Chemicals ( 2022i). The GS on Use of Laboratory
Chemicals was rated high during EPA's systematic review process.
Per the GS on Use of Laboratory Chemicals, water releases are not expected for hazardous chemicals.
Because 1,4-dioxane is considered a hazardous substance under CERCLA (40 CFR Part 302.4) and the
PubChem Hazardous Substances Data Bank (HSDB), there are no water releases for this OES. This is
consistent with the water release assessment for this OES in the Final Risk Evaluation for 1,4-Dioxane
( 2020c). which indicates that water releases are not expected for laboratory uses of 1,4-
dioxane.
Film Cement
EPA estimated daily wastewater discharges for facilities within the Film cement OES using process
information from the Final Risk Evaluation for 1,4-Dioxane ( )c). The underlying process
information for this assessment was rated high during EPA's systematic review process.
Page 252 of 570
-------
Per the risk evaluation, EPA does not expect water releases of 1,4-dioxane for this OES. EPA expects
the glue bottles to be disposed of as solid waste without. There is some uncertainty as to whether and
how much 1,4-dioxane may remain in the glue bottles when disposed. However, due to the small
quantities of the glue and high volatility of the 1,4-dioxane, EPA expects any residual 1,4-dioxane to
evaporate to the air or remain in the solid waste stream ( 20c).
Spray Polyurethane Foam
EPA estimated daily wastewater discharges for facilities within the Spray polyurethane foam OES using
the same approach described for this OES in Appendix E.5.2, which is the use of the GS on Application
of Spray Polyurethane Foam Insulation ( 2020c). The GS on the Application of Spray
Polyurethane Foam Insulation was rated medium during EPA's systematic review process.
The GS indicates that there are six release points:
1. Releases to fugitive air for volatile chemicals during unloading of raw materials from transport
containers;
2. Releases to water, incineration, or landfill from cleaning or disposal of transport containers;
3. Releases to fugitive air for volatile chemicals during transport container cleaning;
4. Releases to incineration or landfill from spray polyurethane foam application equipment
cleaning;
5. Releases to fugitive air for volatile chemicals during equipment cleaning; and
6. Releases to landfill of scrap foam from trimming applied foam.
Based on the GS, only release point #2 has the potential for wastewater discharges. To estimate this
release, EPA used the equations specified in the GS ( )20c). Apart from weight fraction in
spray polyurethan foam, EPA did not find any data specific to 1,4-dioxane in this OES. Therefore, the
calculation of releases using this GS are for a "generic site," using the default input parameter values
from the GS. Specifically, EPA used the same input parameter values that were used in the original risk
evaluation for estimates of occupational exposure; see Appendix G of the Final Risk Evaluation for 1,4-
Dioxane (U.S. EPA. 2020c).
Using this methodology, EPA calculated a range of wastewater releases for this OES. For the low-end,
EPA assumed there are no water releases, which is consistent with the GS explanation that containers
may be disposed of without rinsing. For the high-end, EPA assumed the containers may be rinsed /
poured down drains such that the entire release point #2 is to POTW. Direct water discharges are not
likely given the setting (construction/ renovation sites).
EPA's calculation of wastewater discharges for this OES, including all calculation inputs, can be found
in 1,4-Dioxane Supplemental Information File: Environmental Releases to Water for OES without TRI
or DMR data (11. S. EPA. 2024n).
Dry Film Lubricant
EPA estimated daily wastewater discharges for facilities within the Dry film lubricant OES using
process information from the Final Risk Evaluation for 1,4-Dioxane ( 2020c). The underlying
process information for this assessment was rated high during EPA's systematic review process.
Per the risk evaluation, EPA does not expect water releases of 1,4-dioxane for this OES. Based on
conversations the with only known user, EPA expects wastes to be drummed and sent to a waste handler
with residual wastes releasing to air or being disposed to landfill. ( 20c).
Page 253 of 570
-------
Textile Dye
EPA estimated daily wastewater discharges for facilities within the Textile dye OES using the OECD
ESD on Textile Dyes ( ) and Monte Carlo modeling. The ESD on Textile Dyes was rated
medium during EPA's systematic review process. The use of Montel Carlo modeling allows for
variation of calculation input parameters such that a distribution of environmental releases can be
calculated, from which EPA can estimate the 50th and 95th percentile releases. An explanation of this
modeling approach is included in Appendix E. 11.
Antifreeze
EPA did not find any information to model wastewater discharges for this OES using literature, GSs, or
ESDs, nor does EPA expect this OES to be similar to other OES such that surrogate data may be used.
EPA evaluated the potential for releases using the OECD ESD on Chemical Additives used in
Automotive Lubricants (OECD. 2020) and the EPA MRD on Commercial Use of Automotive Detailing
Products ( BP A. 2022b). The ESD and MRD were both rated high during EPA's systematic review
process.
For the use of antifreeze, EPA expects releases may occur from volatilizations of 1,4-dioxane, disposal
or cleaning of empty antifreeze containers, and spent antifreeze. Both the ESD and MRD indicate that
containers of automotive maintenance fluids are typically small and are not rinsed, but rather disposed of
as solid waste (• c. \ V \ 2022b; OEt O . '20). Additionally, the ESD on Chemical Additives used in
Automotive Lubricants indicates that spent lubricants are disposed of via incineration, which EPA
expects is similarly done for spent antifreeze (OECD. 2020). Therefore, based on this information, EPA
does not expect water releases of 1,4-dioxane for this OES.
Surface Cleaner
EPA estimated daily wastewater discharges for facilities within the Surface cleaner OES using the
SHEDs-HT model, which is described in Section 2.1.1.2. This modeling was completed for one case
study location (Liverpool OH) and only estimates indirect wastewater discharges. EPA does not expect
direct wastewater discharges to surface water from the types of commercial facilities within this OES
(e.g., restaurants, office buildings, other locations with janitorial services).
Dish Soap
EPA estimated daily wastewater discharges for facilities within the Dish Soap OES using data from a
public comment, EPA/OPPT models, and Monte Carlo modeling. The public comment was rated high
during EPA's systematic review process (P&G. 2023). The use of Monte Carlo modeling allows for
variation of calculation input parameters such that a distribution of environmental releases can be
calculated, from which EPA can estimate the 50th and 95th percentile releases. An explanation of this
modeling approach is included in Appendix E. 14.
Dishwasher Detergent
EPA estimated daily wastewater discharges for facilities within the Dishwasher detergent OES using
data from a public comment, EPA/OPPT models, and Monte Carlo modeling. The public comment was
rated high during EPA's systematic review process (P&G. 2023). The use of Monte Carlo modeling
allows for variation of calculation input parameters such that a distribution of environmental releases
can be calculated, from which EPA can estimate the 50th and 95th percentile releases. An explanation of
this modeling approach is included in Appendix E.14.
Page 254 of 570
-------
Laundry Detergent
EPA estimated daily wastewater discharges for facilities within the Laundry detergent OES using the
OECD ESD on Industrial and Institutional Laundries (( ) and Monte Carlo modeling. The
ESD on Industrial and Institutional Laundries was rated medium during EPA's systematic review
process. The use of Montel Carlo modeling allows for variation of calculation input parameters such that
a distribution of environmental releases can be calculated, from which EPA can estimate the 50th and
95th percentile releases. An explanation of this modeling approach is included in Appendix E.12.
Paints and Floor Lacquer
EPA estimated daily wastewater discharges for facilities within the Paints and floor lacquers OES using
the OECD ESD on Coating Application via Spray-Painting in the Automotive Refinishing Industry
(( ). The ESD was rated medium during EPA's systematic review process.
As described in the process description in Appendix F.4.7, 1,4-dioxane was identified by a public
comment as present in automotive refinishing products, thereby allowing EPA to identify the above
ESD as the most applicable GS/ESD available. This ESD indicates that releases are expected from
1. Releases to incineration or landfill from container cleaning/disposal,
2. Releases to incineration or landfill from equipment cleaning,
3. Releases to incineration or landfill from over sprayed coating that is captured by emission
controls, and
4. Releases to air from over sprayed coating that is not captured by emission controls.
None of these releases are expected to water (OECD. 201 la). Therefore, based on this ESD, EPA does
not expect water releases of 1,4-dioxane for this OES.
Hydraulic Fracturing
EPA estimated daily wastewater discharges for facilities within the Hydraulic fracturing OES using the
Draft OECD ESD on Hydraulic Fracturing ( 2022e) and Monte Carlo modeling. The Revised
ESD on Hydraulic Fracturing was rated high during EPA's systematic review process. The use of
Montel Carlo modeling allows for variation of calculation input parameters such that a distribution of
environmental releases can be calculated, from which EPA can estimate the 50th and 95th percentile
releases. An explanation of this modeling approach is included in Appendix E.13.
E.3.3 Water Release Estimates Summary
A summary of industrial and commercial water releases estimated using the above methods is presented
in Table Apx E-3 below. Specifically, this table presents the range of daily water releases per site for
each OES.
Page 255 of 570
-------
Table Apx E-3. Summary of Daily Industrial and Commercial Water Release Estimates for 1,4-Dioxane
OES
Type of Water
Discharge
Number of
Facilities with
Releases"
Range of Estimated Daily
Release Per Site
(kg/site-day)
Estimated
Release
Frequency
Range
(days)''
Overall Data
Quality
Determination
Sources'
Min
Max
Manufacturing
Surface Water
1
1.21
21.4
250
Medium
TRI, DMR
POTW or Industrial
WWT
1
0
6.69
Medium
TRI
Import and
repackaging
Surface Water
6
0.91™
250
Medium
TRI, DMR
POTW or Industrial
WWT
6
0
0.91
Medium
TRI
Industrial uses
Surface Water
24
0
24.5
250
Medium
TRI, DMR
POTW or Industrial
WWT
17
0
105
Medium
TRI
Functional fluids
(open-system)
Surface Water
6
0
0.67
247
Medium
TRI, DMR
POTW or Industrial
WWT
1
4.67
70.9
Medium
TRI
Functional fluids
(closed-system)
All
Assessed as a part of Industrial Uses OES
N/A
N/A
Laboratory
chemical
Surface Water, POTW,
or Industrial WWT
132
0 (water releases not expected)
250
High
GSd
Film cement
Surface Water, POTW,
or Industrial WWT
211
0 (water releases not expected)
250
High
Process
information®
Spray foam
application
Surface Water
1,553,559
0 (surface water releases not
expected)
3
Medium
GS^
POTW
1,553,559
0
0.0036
Medium
GS^
Printing inks
(3D)
Surface Water
1
0.018
0.022
250
Medium
TRI, DMR
POTW or Industrial
WWT
1
0 (no indirect releases per TRI)
Medium
Medium
Dry film
lubricant
Surface Water, POTW,
or Industrial WWT
8
0 (water releases not expected)
48
High
Process
information®
Page 256 of 570
-------
OES
Type of Water
Discharge
Number of
Facilities with
Releases"
Range of Estimated Daily
Release Per Site
(kg/site-day)
Estimated
Release
Frequency
Range
(days)''
Overall Data
Quality
Determination
Sources'
Min
Max
Disposal
Surface Water
24
0
31.8
250
Medium
TRI, DMR
POTW or Industrial
WWT
4
0
0.91
Medium
TRI
Textile dye
(draft RE
estimates)0
POTW
783
1.50E-05
0.001
31 to 295
Medium
ESDg and Monte
Carlo Modeling''
Land (unknown landfill
type) or POTW
(unknown partitioning)
783
2.09E-07
9.72E-05
Medium
ESDg and Monte
Carlo Modeling''
Textile dye
(updated
estimates)0
POTW
783
1.3E-05
9.9E-04
10 to 312
Medium
ESDg and Monte
Carlo Modeling''
Land (unknown landfill
type) or POTW
(unknown partitioning)
783
1.9E-07
9.6E-05
Medium
ESDg and Monte
Carlo Modeling''
Antifreeze
Surface water, POTW,
or Industrial WWT
84,383
0 (water releases not expected)
250
High
Process
information® and
Modeling''
Surface cleaner
POTW
Unknown
0.072 (single daily release value
for all sites combined in
Liverpool OH case study)
250
N/A
SHEDS-HT'
Land (unknown
landfill) or POTW
Unknown
18" (single daily release value for
all sites combined in Liverpool
OH case study)
High
SHEDS-HT,
Process
information®
Modeling''
Dish soap (draft
RE estimates)0
POTW
Unknown
0.064"
250
N/A
SHEDS-HT
Page 257 of 570
-------
OES
Type of Water
Discharge
Number of
Facilities with
Releases"
Range of Estimated Daily
Release Per Site
(kg/site-day)
Estimated
Release
Frequency
Range
(days)''
Overall Data
Quality
Determination
Sources'
Min
Max
Dish soap
(updated
estimates)0
POTW or Fugitive Air
(unknown partitioning)
773,851
8.4E-08
1.5E-03
350
High
fP&G. 2023) and
Monte Carlo
Modeling**
Dishwasher
detergent (draft
RE estimates)0
POTW
Unknown
0.00144"
250
N/A
SHEDS-HT
Dishwasher
detergent
(updated
estimates)0
POTW or Fugitive Air
(unknown partitioning)
773,851
1.2E-06
3.7E-04
350
High
(P&G. 2023) and
Monte Carlo
Modeling**
Laundry
detergent
(institutional) -
Liquid
detergents (draft
RE estimates)0
Fugitive air, stack air,
or POTW (unknown
partitioning)
95,533
1.51 E— 10
0.00714
250 to 365
Medium
ESD' and Monte
Carlo Modeling**
Land (unknown
landfill), incineration,
or POTW (unknown
partitioning)
95,533
4.05E-12
3.95E-05
Medium
ESD' and Monte
Carlo Modeling**
Laundry
detergent
(institutional) -
Liquid
detergents
(updated
estimates)0
Fugitive air, stack air,
or POTW (unknown
partitioning)
95,533
3.0E-013
6.6E-02
250 to 365
Medium
ESD' and Monte
Carlo Modeling**
Land (unknown
landfill), incineration,
or POTW (unknown
partitioning)
95,533
8.1E-13
3.8E-04
Medium
ESD' and Monte
Carlo Modeling**
Page 258 of 570
-------
OES
Type of Water
Discharge
Number of
Facilities with
Releases"
Range of Estimated Daily
Release Per Site
(kg/site-day)
Estimated
Release
Frequency
Range
(days)''
Overall Data
Quality
Determination
Sources'
Min
Max
Laundry
detergent
(institutional) -
Powder
detergents (draft
RE estimates)0
Fugitive air, stack air,
or POTW (unknown
partitioning)
95,533
3.05E-08
2.10E-04
20 to 365
Medium
ESD' and Monte
Carlo Modeling''
Land (unknown
landfill), incineration,
or POTW (unknown
partitioning)
95,533
5.36E-08
0.0018
Medium
ESD' and Monte
Carlo Modeling''
Laundry
detergent
(institutional) -
Powder
detergents
(updated
estimates)0
Fugitive air, stack air,
or POTW (unknown
partitioning)
95,533
2.1E-08
1.9E-03
250 to 365
Medium
ESD' and Monte
Carlo Modeling''
Land (unknown
landfill), incineration,
or POTW (unknown
partitioning)
95,533
1.4E-08
1.8E-02
Medium
ESD' and Monte
Carlo Modeling''
Laundry
detergent
(industrial) -
Liquid
detergents (draft
RE estimates)0
Fugitive air, stack air,
or POTW (unknown
partitioning)
2,453
5.48E-12
0.011
20 to 365
Medium
ESD' and Monte
Carlo Modeling''
Land (unknown
landfill), incineration,
or POTW (unknown
partitioning)
2,453
4.78E-12
1.46E-04
Medium
ESD' and Monte
Carlo Modeling''
Laundry
detergent
(industrial) -
Liquid
detergents
(updated
estimates)0
Fugitive air, stack air,
or POTW (unknown
partitioning)
2,453
3.1E-11
0.11
20 to 365
Medium
ESD' and Monte
Carlo Modeling''
Land (unknown
landfill), incineration,
or POTW (unknown
partitioning)
2,453
6.6E-13
1.4E-03
Medium
ESD' and Monte
Carlo Modeling''
Page 259 of 570
-------
OES
Type of Water
Discharge
Number of
Facilities with
Releases"
Range of Estimated Daily
Release Per Site
(kg/site-day)
Estimated
Release
Frequency
Range
(days)''
Overall Data
Quality
Determination
Sources'
Min
Max
Laundry
detergent
(industrial) -
Fugitive air, stack air,
or POTW (unknown
partitioning)
2,453
1.76E-09
0.0112
Medium
ESD' and Monte
Carlo Modeling''
Land (unknown
2,453
2.92E-11
3.92E-04
20 to 365
Medium
ESD' and Monte
detergents (draft
RE estimates)0
landfill), incineration,
or POTW (unknown
partitioning)
Carlo Modeling''
Laundry
detergent
(industrial) -
Fugitive air, stack air,
or POTW (unknown
partitioning)
2,453
1.8E-11
0.10
Medium
ESD' and Monte
Carlo Modeling''
Powder
Land (unknown
2,453
1.5E-11
3.8E-03
20 to 365
Medium
ESD1 and Monte
detergents
(updated
estimates)0
landfill), incineration,
or POTW (unknown
partitioning)
Carlo Modeling''
Paints and floor
Surface water, POTW,
33,648
0 (water releases not expected)
250
Medium
ESE>> and process
lacquer
or Industrial WWT
information®
Surface water
19
0
10,050
Medium
TRI, DMR
PET byproduct
POTW or Industrial
WWT
14
0
682
250
Medium
TRI
Ethoxylation
Surface water
7
0
0.25
Medium
TRI, DMR
process
byproduct
POTW or Industrial
WWT
6
0
448
250
Medium
TRI
Surface water,
411
3.61E-10
4.59
Medium
ESD'' and Monte
incineration, or landfill
Carlo Modeling''
Hydraulic
(unknown partitioning)
fracturing (draft
RE estimates)"
Recycle/Reuse (48%),
underground injection
(43%), Surface water
(6%), or land (3%)
411
1.85E-10
1.12
1 to 72
Medium
ESD'' and Monte
Carlo Modeling''
Page 260 of 570
-------
OES
Type of Water
Discharge
Number of
Facilities with
Releases"
Range of Estimated Daily
Release Per Site
(kg/site-day)
Estimated
Release
Frequency
Range
(days)''
Overall Data
Quality
Determination
Sources'
Min
Max
Hydraulic
fracturing
(updated
estimates)0
Surface water,
incineration, or landfill
(unknown partitioning)
411
4.3E-10
5.6
1 to 72
Medium
ESD' and Monte
Carlo Modeling''
Recycle/reuse (5%),
underground injection
(70%), Surface water
(19%), or land
(evaporation ponds,
percolation ponds,
irrigation, road
treatment) (6%)
411
2.8E-09
14
Medium
ESD'' and Monte
Carlo Modeling''
Surface water (13%),
Land (soil) (64%), and
Landfill or Incineration
(23%)
411
4.9E-11
0.64
Medium
ESD'' and Monte
Carlo Modeling''
a See Appendix E.l for explanation of how EPA determined the number of sites for each OES.
* Where available. EPA used the December 2020 Final Risk Evaluation for 1,4-Dioxane (U.S. EPA, 2020c). generic scenarios, and emission scenario
documents to provide a basis to estimate the number of release days of 1,4-dioxane within a COU.
c Narrative descriptions of all release estimate sources are provided in Appendix E.3.2.
d The generic scenario used for this COU is the GS on Use of Laboratory Chemicals (U.S. EPA, 20220.
'' For this COU, EPA used process information, which is further described in Appendix E.3.2.
' The generic scenario used for this COU is the GS on Application of Sorav Polvurethane Foam Insulation (U.S. EPA. 2018b).
8 The emission scenario document used for this COU is the ESD on Textile Dves fOECD. 2017).
h For this COU, EPA used various models and literature for model input parameters as described in Appendix E.3.2.
' The emission scenario document used for this COU is the ESD on Industrial and Institutional Laundries fOECD. 201 lb).
1 The emission scenario document used for this COU is the ESD on Coating Application via Sprav Painting in the Automotive Refinishing Industry fOECD.
2011a).
k The emission scenario document used for this COU is the Revised ESD on Hydraulic Fracturing (U.S. EPA. 2022e).
' This value is the Commercial Upstream POTW releases estimated from the SHEDS-HT Down the Drain Model for the Liverpool OH case study (see Section
2.1.1.2).
m All sites for this OES reported under Form A in TRI.
" A single annual value was provided for all sites in the Liverpool OH case study.
° For select OESs, updates to release estimates were made via information provided by the SACC and public comments.
Page 261 of 570
-------
E.3.4 Summary of Weight of Scientific Evidence Conclusions in Water Release Estimates
Table Apx E-4 provides a summary of EPA's weight of scientific evidence conclusions in its water release estimates for each of the OES.
Detailed descriptions of non-OES specific strengths, limitations, assumptions, and uncertainties (e.g., general limitations for TRI, DMR, etc.)
are provided in Appendix E.6.
Table Apx E-4. Summary of Weight of Scientific Evidence Conclusions in Water Release Estimates by OES
OES
Weight of Scientific Evidence Conclusion in Release Estimates
Manufacturing
Wastewater discharges are assessed using reported discharges from 2013-2019 TRI and DMR. Factors that increase the
strength of evidence for this OES are that the release data are directly relevant to the OES (as opposed to surrogate), that TRI
and DMR have medium overall data quality determinations, and consistency within the dataset (all reporters use the same or
similar reporting forms). EPA included 7 years of TRI and DMR data in the analysis, which increases the variability of the
dataset. A strength of TRI data is that TRI compiles the best reasonably available release data for all reporting facilities.
Strengths of DMR data are that it is based on monitoring data collected by facilities and the annual pollutant load is
calculated by integrating release reports over shorter timeframes (e.g., monthly, quarterly) and extrapolating over the course
of the year. Factors that decrease the strength of the evidence for this OES include the low number of data points, uncertainty
in the accuracy of reported releases, and the limitations in representativeness to all sites because TRI may not capture all
relevant sites. Additionally, EPA made assumptions on the number of operating days to estimate daily releases, which
introduces additional uncertainty. Based on this information, EPA has concluded that the weight of scientific evidence for
this assessment is moderate to robust and provides a plausible estimate of releases in consideration of the strengths and
limitations of reasonably available data.
Import and repackaging
Wastewater discharges are assessed using reported discharges from 2013-2019 TRI and DMR. Factors that increase the
strength of evidence for this OES are that the release data are directly relevant to the OES (as opposed to surrogate), that TRI
and DMR have medium overall data quality determinations, and consistency within the dataset (all reporters use the same or
similar reporting forms). EPA included seven years of TRI and DMR data in the analysis, which increases the variability of
the dataset. A strength of TRI data is that TRI compiles the best reasonably available release data for all reporting facilities.
Strengths of DMR data are that it is based on monitoring data collected by facilities and the annual pollutant load is
calculated by integrating release reports over shorter timeframes (e.g., monthly, quarterly) and extrapolating over the course
of the year. Factors that decrease the strength of the evidence for this OES include the low number of data points, uncertainty
in the accuracy of reported releases, uncertainty in EPA's use of Form A submissions, and the limitations in
representativeness to all sites because TRI may not capture all relevant sites. Some facilities within this OES reported to TRI
using a Form A, which does not include any details on chemical release quantities. When a facility has submitted a Form A,
there is no way to discern the quantity released. Therefore, where facilities reported to TRI with a Form A, EPA used the
Form A threshold for total releases of 500 lb/year for each release media; however, there is uncertainty in this because the
actual release quantity is unknown. Furthermore, the threshold represents an upper limit on total releases from the facility;
therefore, assessing releases at the threshold value may overestimate actual releases from the facility. Additionally,
uncertainty is introduced from EPA's assumptions on the number of operating days to estimate daily releases and in the
mapping of DMR-reporting facilities to this OES. Based on this information, EPA has concluded that the weight of scientific
Page 262 of 570
-------
OES
Weight of Scientific Evidence Conclusion in Release Estimates
evidence for this assessment is moderate to robust and provides a plausible estimate of releases in consideration of the
strengths and limitations of reasonably available data.
Industrial Uses
Wastewater discharges are assessed using reported discharges from 2013-2019 TRI and DMR. Factors that increase the
strength of evidence for this OES are that the release data are directly relevant to the OES (as opposed to surrogate), that TRI
and DMR have medium overall data quality determinations, and consistency within the dataset (all reporters use the same or
similar reporting forms). EPA included seven years of TRI and DMR data in the analysis, which increases the variability of
the dataset. A strength of TRI data is that TRI compiles the best reasonably available release data for all reporting facilities.
Strengths of DMR data are that it is based on monitoring data collected by facilities and the annual pollutant load is
calculated by integrating release reports over shorter timeframes (e.g., monthly, quarterly) and extrapolating over the course
of the year. Factors that decrease the strength of the evidence for this OES include uncertainty in the accuracy of reported
releases, uncertainty in EPA's use of Form A submissions, and the limitations in representativeness to all sites because TRI
may not capture all relevant sites. Some facilities within this OES reported to TRI using a Form A, which does not include
any details on chemical release quantities. When a facility has submitted a Form A, there is no way to discern the quantity
released. Therefore, where facilities reported to TRI with a Form A, EPA used the Form A threshold for total releases of 500
lb/year for each release media; however, there is uncertainty in this because the actual release quantity is unknown.
Furthermore, the threshold represents an upper limit on total releases from the facility; therefore, assessing releases at the
threshold value may overestimate actual releases from the facility. Additionally, uncertainty is introduced from EPA's
assumptions on the number of operating days to estimate daily releases and in the mapping of DMR-reporting facilities to
this OES. Based on this information, EPA has concluded that the weight of scientific evidence for this assessment is
moderate to robust and provides a plausible estimate of releases in consideration of the strengths and limitations of
reasonably available data.
Functional fluids (open-
system)
Wastewater discharges are assessed using reported discharges from 2013-2019 TRI and DMR. Factors that increase the
strength of evidence for this OES are that the release data are directly relevant to the OES (as opposed to surrogate), that TRI
and DMR have medium overall data quality determinations, and consistency within the dataset (all reporters use the same or
similar reporting forms). EPA included seven years of TRI and DMR data in the analysis, which increases the variability of
the dataset. A strength of TRI data is that TRI compiles the best reasonably available release data for all reporting facilities.
Strengths of DMR data are that it is based on monitoring data collected by facilities and the annual pollutant load is
calculated by integrating release reports over shorter timeframes (e.g., monthly, quarterly) and extrapolating over the course
of the year. Factors that decrease the strength of the evidence for this OES include the low number of data points, uncertainty
in the accuracy of reported releases, and the limitations in representativeness to all sites because TRI may not capture all
relevant sites. The assessment includes data from only two sites that reported to TRI (one of which reported zero water
releases) and four that reported to DMR. Additionally, uncertainty is introduced from EPA's assumptions on the number of
operating days to estimate daily releases and in the mapping of DMR-reporting facilities to this OES. Based on this
information, EPA has concluded that the weight of scientific evidence for this assessment is moderate to robust and provides
a plausible estimate of releases in consideration of the strengths and limitations of reasonably available data.
Functional fluids (closed-
system)
No data was available to estimate releases for this OES. For the water release assessment, EPA grouped this OES with the
Industrial Uses OES because the sources of release are expected to be similar between these OESs. Factors that increase the
Page 263 of 570
-------
OES
Weight of Scientific Evidence Conclusion in Release Estimates
strength of evidence for this OES are that TRI and DMR have medium overall data quality determinations and consistency
within the dataset (all reporters use the same or similar reporting forms). Additionally, EPA included seven years of TRI and
DMR data in the analysis, which increases the variability of the dataset. Factors that decrease the strength of evidence for
this OES are that the Industrial Releases OES release data are use as surrogate for this OES, uncertainty in the accuracy of
reported releases, and the limitations in representativeness to all sites because TRI may not capture all relevant sites.
Refer to the Industrial Uses OES discussion for additional discussion. Based on this information, EPA has concluded that the
weight of scientific evidence for this assessment is slight and provides a plausible estimate of releases in consideration of the
strengths and limitations of reasonably available data.
Laboratory chemicals
Wastewater discharges are assessed using the Draft GS on Use of Laboratory Chemicals. Per the GS, water releases are not
expected for hazardous chemicals. Because 1,4-dioxane is considered a hazardous chemical under CERCLA, no water
releases are expected for this OES according to the GS. Factors that increase the strength of evidence for this OES are that
the release estimates are directly relevant to the OES (as opposed to surrogate), the Draft GS on Use of Laboratory
Chemicals has a high overall data quality determination, and there is a low level of uncertainty in the data. Factors that
decrease the strength of the evidence for this OES include the that the GS has not been peer-reviewed, uncertainty in the
representativeness of the GS towards all sites in this OES, and a lack of variability in the analysis. Specifically, because the
default values in the ESD are generic, there is uncertainty in the representativeness of generic site estimates of actual releases
from real-world sites that use 1,4-dioxane. Another uncertainty is lack of consideration for release controls. The ESD
assumes that all activities occur without any release controls. Actual releases may be less than estimated if facilities utilize
pollution control methods, contributing to uncertainty. Based on this information, EPA has concluded that the weight of
scientific evidence for this assessment is slight to moderate and provides a plausible estimate of releases in consideration of
the strengths and limitations of reasonably available data.
Film cement
Wastewater discharges are assessed using process information from the Final Risk Evaluation for 1,4-Dioxane. Per the
process information, EPA does not expect water releases of 1,4-dioxane for this OES because 1,4-dioxane volatilizes to air
after application of the film cement and empty cement bottles are disposed of as solid waste without rinsing. Factors that
increase the strength of evidence for this OES are that the release estimates are directly relevant to the OES (as opposed to
surrogate), the underlying data sources for the process information have a high overall data quality determination, and there
is a low level of uncertainty in the data because the process information comes directly from actual users of 1,4-dioxane in
film cement. Factors that decrease the strength of the evidence for this OES include the uncertainty in the representativeness
of evidence to all sites in this OES and a lack of variability. Specifically, the process information for the production and use
of film cement is based on information from three use sites, one from Australia and two from the United States. Based on
this information, EPA has concluded that the weight of scientific evidence for this assessment is slight to moderate and
provides a plausible estimate of releases in consideration of the strengths and limitations of reasonably available data.
Spray foam application
Wastewater discharges are assessed using the GS on Application of Spray Polyurethane Foam Insulation. Factors that
increase the strength of evidence for this OES are that the release estimates are directly relevant to the OES (as opposed to
surrogate), the underlying data sources for the process information have a medium overall data quality determination, and
there is a low level of uncertainty in the data. Factors that decrease the strength of the evidence for this OES include
uncertainty in the representativeness of the GS to all sites since it is generic and not specific to sites that use 1,4-dioxane and
Page 264 of 570
-------
OES
Weight of Scientific Evidence Conclusion in Release Estimates
a lack of variability. Based on this information, EPA has concluded that the weight of scientific evidence for this assessment
is slight to moderate and provides a plausible estimate of releases in consideration of the strengths and limitations of
reasonably available data.
Printing inks (3D)
Wastewater discharges are assessed using reported discharges from 2013-2019 DMR. Factors that increase the strength of
evidence for this OES are that the release data are directly relevant to the OES (as opposed to surrogate), that DMR has a
medium overall data quality determination, and consistency within the dataset (all reporters use the same or similar reporting
forms). Additionally, EPA used DMR data for seven years, which increases the variability of the dataset. Strengths of DMR
data are that it is based on monitoring data collected by facilities and the annual pollutant load is calculated by integrating
release reports over shorter timeframes (e.g., monthly, quarterly) and extrapolating over the course of the year. Factors that
decrease the strength of the evidence for this OES include the low number of data points, uncertainty in the accuracy of
reported releases, and the limitations in representativeness to all sites. Additionally, no TRI data is available for this OES,
EPA made assumptions on the number of operating days, and there is uncertainty in the mapping of DMR-reporting facilities
to this OES. Based on this information, EPA has concluded that the weight of scientific evidence for this assessment is
moderate to robust and provides a plausible estimate of releases in consideration of the strengths and limitations of
reasonably available data.
Dry film lubricant
Wastewater discharges are assessed using process information from the Final Risk Evaluation for 1,4-Dioxane. Based on
conversations the with only known user who supplied this process information, EPA expects wastes to be drummed and sent
to a waste handler with residual wastes releasing to air or being disposed to landfill, such that there are no water releases.
Factors that increase the strength of evidence for this OES are that the release estimates are directly relevant to the OES (as
opposed to surrogate), the underlying data sources for the process information have a high overall data quality determination,
and there is a low level of uncertainty in the data. Additionally, the process information comes directly from an actual user of
1,4-dioxane in dry film lubricants. Factors that decrease the strength of the evidence for this OES include uncertainty in the
representativeness of evidence to all sites and a lack of variability. Based on this information, EPA has concluded that the
weight of scientific evidence for this assessment is slight to moderate and provides a plausible estimate of releases in
consideration of the strengths and limitations of reasonably available data.
Disposal
Wastewater discharges are assessed using reported discharges from 2013-2019 TRI and DMR. Factors that increase the
strength of evidence for this OES are that the release data are directly relevant to the OES (as opposed to surrogate), that TRI
and DMR have medium overall data quality determinations, and consistency within the dataset (all reporters use the same or
similar reporting forms). EPA included seven years of TRI and DMR data in the analysis, which increases the variability of
the dataset. A strength of TRI data is that TRI compiles the best reasonably available release data for all reporting facilities.
Strengths of DMR data are that it is based on monitoring data collected by facilities and the annual pollutant load is
calculated by integrating release reports over shorter timeframes (e.g., monthly, quarterly) and extrapolating over the course
of the year. Factors that decrease the strength of the evidence for this OES include uncertainty in the accuracy of reported
releases, uncertainty in EPA's use of Form A submissions, and the limitations in representativeness to all sites because TRI
may not capture all relevant sites. Some facilities within this OES reported to TRI using a Form A, which does not include
any details on chemical release quantities. When a facility has submitted a Form A, there is no way to discern the quantity
released. Therefore, where facilities reported to TRI with a Form A, EPA used the Form A threshold for total releases of 500
Page 265 of 570
-------
OES
Weight of Scientific Evidence Conclusion in Release Estimates
lb/year for each release media; however, there is uncertainty in this because the actual release quantity is unknown.
Furthermore, the threshold represents an upper limit on total releases from the facility; therefore, assessing releases at the
threshold value may overestimate actual releases from the facility. Additionally, uncertainty is introduced from EPA's
assumptions on the number of operating days to estimate daily releases and in the mapping of DMR-reporting facilities to
this OES. Based on this information, EPA has concluded that the weight of scientific evidence for this assessment is
moderate to robust and provides a plausible estimate of releases in consideration of the strengths and limitations of
reasonably available data.
Textile dye
Wastewater discharges are assessed using Monte Carlo modeling with information from the ESD on Textile Dyes. Factors
that increase the strength of evidence for this OES are that the release estimates are directly relevant to the OES (as opposed
to surrogate), that the ESD on Textile Dyes has a medium overall data quality determination and was peer reviewed, the high
number of data points (simulation runs), consistency within the dataset, and full distributions of input parameters. The Monte
Carlo modeling accounts for the entire distribution of input parameters, calculating a distribution of potential release values
that represents a larger proportion of sites than a discrete value. Factors that decrease the strength of the evidence for this
OES include uncertainties and limitations in the representativeness of the estimates for sites that specifically use 1,4-dioxane
because the default values in the ESD are generic. Another uncertainty is lack of consideration for release controls. The ESD
assumes that all activities occur without any release controls. Actual releases may be less than estimated if facilities utilize
pollution control methods, contributing to uncertainty. Based on this information, EPA has concluded that the weight of
scientific evidence for this assessment is moderate and provides a plausible estimate of releases in consideration of the
strengths and limitations of reasonably available data.
Antifreeze
Wastewater discharges are assessed using the OECD ESD on Chemical Additives used in Automotive Lubricants and the
EPA MRD on Commercial Use of Automotive Detailing Products. Factors that increase the strength of evidence for this
OES are that the ESD and MRD used have high overall data quality determinations, consistency within the sources used, and
a low number of uncertainties. Both sources indicate that containers of automotive maintenance fluids are not typically
rinsed, but rather disposed of as solid waste or incinerated, such that there are no water releases, contributing to consistency
and a low level of uncertainty in the data. Factors that decrease the strength of the evidence for this OES include that the
ESD and MRD are not directly applicable to antifreeze uses (used as surrogate), uncertainty in the representativeness of the
ESD and MRD to all sites and sites that specifically use 1,4-dioxane since these documents contain generic values, and a
lack of variability. Based on this information, EPA has concluded that the weight of scientific evidence for this assessment is
slight to moderate and provides a plausible estimate of releases in consideration of the strengths and limitations of
reasonably available data.
Surface cleaner
Wastewater discharges are assessed using the SHEDS-HT model. Factors that increase the strength of evidence for this OES
include that the release estimates are directly relevant to the OES (as opposed to surrogate) and variability in the model input
parameters. Factors that decrease the strength of the evidence for this OES include uncertainty in the representativeness to all
sites because the estimate is based on one case study for Liverpool, OH and because the estimate is not site-specific (the
release estimate is a total for all sites in Liverpool, OH). Based on this information, EPA has concluded that the weight of
scientific evidence for this assessment is slight and provides a plausible estimate of releases in consideration of the strengths
and limitations of reasonably available data.
Page 266 of 570
-------
OES
Weight of Scientific Evidence Conclusion in Release Estimates
Dish Soap
Wastewater discharges are assessed using Monte Carlo modeling with information from a public comment and standard
EPA/OPPT models. Factors that increase the strength of evidence for this OES are that the release estimates are directly
relevant to the OES (as opposed to surrogate), that the public comment has a hieh overall data aualitv determination (P&G.
2023). there are a hieh number of data points (simulation runs), and full distributions of input parameters. Monte Carlo
modeling accounts for the entire distribution of input parameters, calculating a distribution of potential release values that
represents a larger proportion of sites than a discrete value. The major factor that decreases the strength of the evidence for
this OES include the uncertainties and limitations in the representativeness of the data from the public comment towards all
sites that use dish soaps containing 1,4-dioxane. Another uncertainty is the lack of a GS or ESD describing this scenario;
EPA used standard EPA/OPPT models for each of the expected release points to build the model. Based on this information,
EPA has concluded that the weight of scientific evidence for this assessment is moderate and provides a plausible estimate of
releases in consideration of the strengths and limitations of reasonably available data.
Dishwasher detergent
EPA used the same approach to estimate wastewater discharges for this OES as the Dish Soap OES. Therefore, the same
rationale and overall weight of scientific evidence apply to this OES.
Laundry detergent
Wastewater discharges are assessed using Monte Carlo modeling with information from the ESD on Industrial and
Institutional Laundries. Factors that increase the strength of evidence for this OES are that the release estimates are directly
relevant to the OES (as opposed to surrogate), that the ESD on Industrial and Institutional Laundries has a medium overall
data quality determination and was peer reviewed, there are a high number of data points (simulation runs), consistency
within the dataset, and full distributions of input parameters. Monte Carlo modeling accounts for the entire distribution of
input parameters, calculating a distribution of potential release values that represents a larger proportion of sites than a
discrete value. Additionally, EPA was able to separately estimate releases for industrial and institutional laundry settings.
Factors that decrease the strength of the evidence for this OES include uncertainties and limitations in the representativeness
of the estimates for sites that specifically use 1,4-dioxane because the default values in the ESD are generic. Another
uncertainty is lack of consideration for release controls. The ESD assumes that all activities occur without any release
controls. Actual releases may be less than estimated if facilities utilize pollution control methods. Based on this information,
EPA has concluded that the weight of scientific evidence for this assessment is moderate and provides a plausible estimate of
releases in consideration of the strengths and limitations of reasonably available data.
Paint and floor lacquer
Wastewater discharges are assessed using OECD ESD on Coating Application via Spray-Painting in the Automotive
Refinishing Industry. According to the ESD, no releases are expected to water. Factors that increase the strength of evidence
for this OES are that the release estimates are directly relevant to the OES (as opposed to surrogate), the ESD has a medium
overall data quality determination, and a low number of uncertainties. F Factors that decrease the strength of the evidence for
this OES include a lack of variability and uncertainty in the representativeness of the ESD to all sites and sites that
specifically use 1,4-dioxane since the ESD is generic. Based on this information, EPA has concluded that the weight of
scientific evidence for this assessment is slight to moderate and provides a plausible estimate of releases in consideration of
the strengths and limitations of reasonably available data.
Polyethylene terephthalate
(PET) byproduct
Wastewater discharges are assessed using reported discharges from 2013-2019 TRI and DMR. Factors that increase the
strength of evidence for this OES are that the release data are directly relevant to the OES (as opposed to surrogate), that TRI
and DMR have medium overall data quality determinations, consistency within the dataset (all reporters use the same or
Page 267 of 570
-------
OES
Weight of Scientific Evidence Conclusion in Release Estimates
similar reporting forms), and consistency with the emission data from the related life cycle analysis discussed in Appendix
E.6. EPA included seven years of TRI and DMR data in the analysis, which increases the variability of the dataset. A
strength of TRI data is that TRI compiles the best reasonably available release data for all reporting facilities. Strengths of
DMR data are that it is based on monitoring data collected by facilities and the annual pollutant load is calculated by
integrating release reports over shorter timeframes (e.g., monthly, quarterly) and extrapolating over the course of the year.
Factors that decrease the strength of the evidence for this OES include the uncertainty in the accuracy of reported releases
and the limitations in representativeness to all sites because TRI may not capture all relevant sites. Additionally, EPA made
assumptions on the number of operating days to estimate daily releases, which introduces additional uncertainty. Based on
this information, EPA has concluded that the weight of scientific evidence for this assessment is moderate to robust and
provides a plausible estimate of releases in consideration of the strengths and limitations of reasonably available data.
Ethoxylation process
byproduct
Wastewater discharges are assessed using reported discharges from 2013-2019 TRI and DMR. Factors that increase the
strength of evidence for this OES are that the release data are directly relevant to the OES (as opposed to surrogate), that TRI
and DMR have medium overall data quality determinations, and consistency within the dataset (all reporters use the same or
similar reporting forms). EPA included seven years of TRI and DMR data in the analysis, which increases the variability of
the dataset. A strength of TRI data is that TRI compiles the best reasonably available release data for all reporting facilities.
Strengths of DMR data are that it is based on monitoring data collected by facilities and the annual pollutant load is
calculated by integrating release reports over shorter timeframes (e.g., monthly, quarterly) and extrapolating over the course
of the year. Factors that decrease the strength of the evidence for this OES include uncertainty in the accuracy of reported
releases and the limitations in representativeness to all sites because TRI may not capture all relevant sites. Additionally,
EPA made assumptions on the number of operating days to estimate daily releases, which introduces additional uncertainty.
Based on this information, EPA has concluded that the weight of scientific evidence for this assessment is moderate to robust
and provides a plausible estimate of releases in consideration of the strengths and limitations of reasonably available data.
Hydraulic fracturing
Wastewater discharges are assessed using Monte Carlo modeling with information from the Revised ESD on Hydraulic
Fracturing and FracFocus 3.0. Factors that increase the strength of evidence for this OES are that the release estimates are
directly relevant to the OES (as opposed to surrogate), that the Revised ESD on Hydraulic Fracturing and FracFocus 3.0
have medium overall data quality determinations, that the Revised ESD has undergone peer review by OECD, the high
number of data points (simulation runs), and the full distributions of input parameters. Monte Carlo modeling accounts for
the entire distribution of input parameters, calculating a distribution of potential release values that represents a larger
proportion of sites than a discrete value. Factors that decrease the strength of the evidence for this OES include uncertainties
that may result in over-estimates of releases and limitations in the representativeness of the estimates for all sites.
Specifically, EPA used some input values from the Revised ESD; because the default values in the ESD are generic, there is
uncertainty in the representativeness of the generic site estimates of real-world sites that use 1,4-dioxane. Another
uncertainty is lack of consideration for release controls. The ESD assumes that all activities occur without any release
controls. Actual releases may be less than estimated if facilities utilize pollution control methods, contributing to uncertainty.
Based on this information, EPA has concluded that the weight of scientific evidence for this assessment is moderate to robust
and provides a plausible estimate of releases in consideration of the strengths and limitations of reasonably available data.
Page 268 of 570
-------
E.4 Land Release Assessment
This section describes EPA's methodology for estimating annual land releases from industrial and
commercial facilities manufacturing, processing, or using 1,4-dioxane. EPA did not estimate daily land
releases due to the high level of uncertainty in the number of release days associated with land releases.
Facilities report annual land releases to the Toxics Release Inventory (TRI), which include a variety of
release mechanisms, including but not limited to underground injection, RCRA Subtitle C landfills,
other landfills, surface impoundments, and land treatment. EPA used 2019 TRI ( 2022h) data
to estimate annual land releases for the OES where available; however, EPA did not have these data for
every OES. For OES without TRI data, EPA used alternate assessment approaches to estimate annual
land releases.
In addition, EPA did a more in-depth analysis of TRI for sites within the Disposal OES. Specifically,
EPA did an analysis of 2013 to 2019 TRI data for this OES. Operations at disposal sites are expected to
be more complex than those at sites in other OES, which typically generate waste for land disposal off
site. Additionally, the disposal OES includes the sites of ultimate disposal {i.e., they are the landfills
themselves) and EPA considered the impact of continuous years of land releases of 1,4-dioxane at these
sites on general population and ecological exposures.
E.4.1 Assessment Using TRI
EPA found 2019 TRI data for facilities within the following OESs:
• Manufacturing;
• Import and repackaging;
• Industrial uses;
• Functional fluids (open-system);
• Disposal;
• PET byproduct; and
• Ethoxylation byproduct.
The TRI data were rated medium in EPA's systematic review process. EPA estimated annual land
releases using TRI for these OESs, with the following general stages as described in the rest of this
section.
1. Collect land release data from the 2013 to 2019 TRI for the Disposal OES and 2019 TRI data for
all other OES,
2. Map land release data to occupational exposure scenarios,
3. Analyze 2013 to 2019 TRI data for the disposal OES, and
4. Summarize 2019 annual land releases for the other OES.
Step 1: Collect Land Release Data from TRI
The first step in estimating land releases was to obtain TRI data. As previously discussed in Appendix
E.3.1, each facility subject to the TRI reporting rule must report annually the volume of chemical
released to the environment and/or managed through recycling, energy recovery, and treatment. Similar
to the air release assessment, EPA included both TRI reporting Form R and TRI reporting Form A
submissions in the land release assessment. Where sites reported to TRI with Form A, EPA used the
Form A threshold for total releases of 500 lb/year. EPA used the entire 500 lb/year for each type of land
release; however, since this threshold is for total site releases, these 500 lb/year are attributed one type
of land release at a time (since assessing it for more than one land release media at a time would double
count the releases and exceed the total release threshold for Form A). EPA pulled the TRI Basic Plus
Data Files for each of years 2013 through 2019.
Page 269 of 570
-------
TRI data include both on- and off-site land releases. In summary, TRI includes the following land
release media:
• On-site releases:
o Underground injection
o RCRA subtitle C landfills
o Other landfills
o Land treatment
o RCRA surface impoundments
o Other surface impoundments
o Other land disposal
o Waste rock
• Off-site releases:
o Underground injection
o RCRA subtitle C landfills
o Other landfills
o Land treatment
o RCRA surface impoundments
o Other surface impoundments
o Other land disposal
o Transfer to waste broker for disposal
o Solidification/stabilization
Step 2: Map Land Release Data to Occupational Exposure Scenarios
The next step in estimating land releases was to map the 2013 to 2019 TRI data to the 1,4-dioxane OES.
EPA used the same mapping methodology as that used for both the air and water assessments, which is
described in Appendix E.5.1. EPA ensured consistency in the OES mapping between the air, water, and
land assessments.
Step 3: Analyze and Summarize 2013 to 2019 TRI Data for the Disposal OES
For the sites that EPA mapped to the disposal OES in the 2013 to 2019 TRI data, EPA analyzed and
summarized the land release data as follows:
• EPA summarized which of the reporting years that each disposal facility submitted data to TRI.
This summary allows for visualization of which sites report recurring land disposal of 1,4-
dioxane between 2013 and 2019.
• EPA differentiated between disposal sites that transferred 1,4-dioxane to other sites for disposal
and the receiving sites that disposed of 1,4-dioxane on site. For the receiving sites at which 1,4-
dioxane was disposed of to land, EPA summarized the number of unique sites from which the
receiving sites received 1,4-dioxane for land disposal and the total amount of 1,4-dioxane
received for land disposal between 2013 and 2019.
• EPA summarized the total amount of 1,4-dioxane released to each land release media between
2013 and 2019. In summary, 1,4-dioxane was disposed of from disposal OES sites via on-site
and off-site RCRA subtitle C landfills, on-site and off-site underground injection, and off-site
other landfills between 2013 and 2019.
EPA's analysis and summary of land releases for 2013 to 2019 TRI sites in the disposal OES can be
found in 1,4-Dioxane Supplemental Information File: Environmental Releases to landfor the Disposal
OES ( 04m).
Page 270 of 570
-------
Step 4: Summarize Annual Land Releases for Other OES with 2019 TRI data
For the remaining OESs for which 2019 TRI data were available, EPA summarized the annual land
releases by media type (e.g., underground injection, RCRA subtitle C landfills, other landfills, land
treatment) and site information, including site identity, city, state, zip code, TRI facility ID, and FRS ID.
EPA did not estimate daily land releases due to the high level of uncertainty in the number of release
days associated with land releases.
EPA's summary of land release for these OESs is included in 1,4-Dioxane Supplemental Information
File: Environmental Releases to Landfor all OES Except Disposal ( 0241).
E.4.2 Assessment for OES Without TRI
EPA did not find 2019 TRI data for the following OESs:
• Functional fluids (closed-systems);
• Laboratory chemicals;
• Film cement;
• Spray polyurethane foam;
• 3D Printing;
• Dry film lubricant;
• Textile dye;
• Antifreeze;
• Surface cleaner;
• Dish soap;
• Dishwasher detergent;
• Laundry detergent;
• Paints and floor lacquer; and
• Hydraulic fracturing.
For these OESs, EPA estimated land releases by using various modeling approaches, including the use
of surrogate TRI data and modeling using data from literature, GSs, and ESDs. EPA's assessment of
land releases for each of these OESs is described below.
Functional Fluids (Closed-Systems)
Land release data were not available for this OES and EPA did not find any information to model land
release for this OES using literature, GSs, or ESDs. EPA expects that the sources of release for this OES
to be similar to those for the Industrial Uses OES, based on the process information in the Final Risk
Evaluation for 1,4-Dioxane (U, S. EPA. 2020c). Therefore, EPA grouped the land release assessment for
Functional Fluids (Closed-Systems) into that for Industrial Uses. However, there is uncertainty in this
assumption of similar release sources between these OESs.
Laboratory Chemicals
EPA estimated land releases for facilities within the Laboratory chemicals OES using the Draft GS on
Use of Laboratory Chemicals ( 221). The GS on Use of Laboratory Chemicals was rated
high during EPA's systematic review process.
The GS indicates that there are eight release points:
1. Release to air from transferring volatile chemicals from transport containers.
2. Release to air, water, incineration, or landfill from transferring solid powders.
3. Release to water, incineration, or land from cleaning or disposal of transport containers.
Page 271 of 570
-------
4. Release to air from cleaning containers used for volatile chemicals.
5. Labware equipment cleaning residuals released to water, incineration, or landfill.
6. Release to air during labware equipment cleaning for volatile chemicals.
7. Release to air from laboratory analyses for volatile chemicals.
8. Release to water, incineration, or landfill from laboratory waste disposal.
Based on the GS, release points #2, 3, 5, and 8 have the potential for land releases; however, release
point #2 is not applicable because 1,4-dioxane is not a solid powder. To estimate the remaining land
releases, EPA used the equations specified in the Draft GS (U.S. EPA. 20221). EPA did not find any data
specific to 1,4-dioxane in this OES. Therefore, the calculation of releases using this GS are for a
"generic site," using the default input parameter values from the GS.
Using this methodology, EPA calculated high-end and low-end annual land releases for this OES. The
low- and high-end estimates are based on the low-end or typical and high-end or worst-case calculation
input parameter defaults from the GS. EPA's calculation of land releases for this OES, including all
calculation inputs, can be found in 1,4-Dioxane Supplemental Information File: Environmental Releases
to Landfor all OES Except Disposal (U.S. EPA. 20241).
Film Cement
EPA estimated land releases for facilities within the Film cement OES using process information from
the Final Risk Evaluation for 1,4-Dioxane ( 2020c). The underlying process information for
this assessment was rated high during EPA's systematic review process.
The process of using film cement involves applying the cement onto edges of photographic films by
hand using a small brush, then joining the pieces of film by pressing and heating to dry the cement.
Based on this process information, EPA expects land releases may result from disposal of empty film
cement bottles that contain residual amounts of film cement containing 1,4-dioxane. EPA estimated this
land release as a range, using a film cement use rate of 2.5 to 12 L/site-year and a concentration of 1,4-
dioxane in the film cement of 45 to 50 percent from the process information in the Final Risk Evaluation
for 1,4-Dioxane (U.S. EPA. 2020c). and the EPA/OPPT Small Container Residual Model central
tendency loss fraction of 0.3 percent and high-end loss fraction of 0.6 percent. EPA is uncertain of the
specific type of land disposal for the empty film cement bottles.
EPA's calculation of land releases for this OES, including all calculation inputs and assumptions, can be
found in 1,4-Dioxane Supplemental Information File: Environmental Releases to Landfor all OES
Except Disposal ( EPA. 20241).
Spray Polyurethane Foam
EPA estimated land releases for facilities within the Spray polyurethane foam OES using the GS on
Application of Spray Polyurethane Foam Insulation (U.S. EPA. 2020c). The GS on the Application of
Spray Polyurethane Foam Insulation was rated medium during EPA's systematic review process.
The GS indicates that there are six release points:
1. Releases to fugitive air for volatile chemicals during unloading of raw materials from transport
containers;
2. Releases to water, incineration, or landfill from cleaning or disposal of transport containers;
3. Releases to fugitive air for volatile chemicals during transport container cleaning;
4. Releases to incineration or landfill from spray polyurethane foam application equipment
cleaning;
Page 272 of 570
-------
5. Releases to fugitive air for volatile chemicals during equipment cleaning; and
6. Releases to landfill of scrap foam from trimming applied foam.
Based on the GS, release points #2, 4, and 6 have the potential for land releases. To estimate these
releases, EPA used the equations specified in the GS ( )c). Apart from weight fraction in
spray polyurethan foam, EPA did not find any data specific to 1,4-dioxane in this OES. Therefore, the
calculation of releases using this GS are for a "generic site," using the default input parameter values
from the GS. Specifically, EPA used the same input parameter values that were used in the original risk
evaluation for estimates of occupational exposure; see Appendix G of the Final Risk Evaluation for 1,4-
Dioxane ( 320c).
Using this methodology, EPA calculated high-end and low-end annual land releases for this OES. The
low- and high-end estimates are based on the low-end or typical and high-end or worst-case calculation
input parameter defaults from the GS. EPA's calculation of land releases for this OES, including all
calculation inputs, can be found 1,4-Dioxane Supplemental Information File: Environmental Releases to
Land for all OES Except Disposal ( 0241).
3D Printing
Land release data were not available for this OES and EPA did not find any information to model land
releases for this OES using literature, GSs, or ESDs. EPA expects that industrial applications of this
OES to be accounted for in the Industrial Uses TRI data. Per the December 2020 Final Risk Evaluation
for 1,4-Dioxane (U.S. EPA. 2020c). 3D printing ink containing 1,4-dioxane is used in research labs to
print biomedical products. Because the 2019 TRI data for the Industrial Uses OES include medicinal and
pharmaceutical manufacturing NAICS codes, medical research labs that conduct 3D printing with 1,4-
dioxane inks may be captured in that OES. Therefore, EPA grouped the land release assessment for 3D
Printing into that for Industrial Uses. However, there is uncertainty in whether 3D printing sites are truly
captured in the Industrial Uses TRI data.
Dry Film Lubricant
EPA estimated land releases for facilities within the Dry film lubricant OES using process information
from the Final Risk Evaluation for 1,4-Dioxane ( )20c). The underlying process information
for this assessment was rated high during EPA's systematic review process.
The process for the production and use of dry film lubricant is described in the 2020 RE and is based on
information provided to EPA by the one known user. In summary, the process entails first producing the
concentrated dry film lubricant by mixing 1,4-dioxane with other additives, followed by the dilution of
the concentrated dry film lubricant with additional 1,4-dioxane and the use of the dry film lubricant. The
use involves spray application onto substrates in a vented paint booth and the subsequent curing in a
vented oven and cleaning of the dried parts in a 1,4-dioxane bath ( 20c). Based on this
process description, EPA expects land releases may result from
1. Residuals in empty containers of pure 1,4-dioxane used for mixing of the concentrated dry film
lubricant,
2. Cleaning residuals for equipment used for mixing of the concentrated dry film lubricant,
3. Residuals in empty containers of pure 1,4-dioxane used for diluting the concentrated dry film
lubricant,
4. Residuals in empty containers of concentrated dry film lubricant, and
5. Waste from cleaning spray application equipment and the parts onto which the dry film lubricant
was applied.
Page 273 of 570
-------
EPA estimated land releases using 1,4-dioxane use rates derived from the process information and
standard EPA models. Specifically, EPA estimated land releases from release points #1, 3 and 4 using
the EPA/OPPT Small Container Residual Model, which has a central tendency loss fraction of 0.3
percent and a high-end loss fraction of 0.6 percent of the container volume. EPA used container volumes
specified in the process information, which are either 1-gallon or 0.5-pint containers ( 320c).
EPA estimated releases from release point #2 using the EPA/OPPT Single Process Vessel Residual
Model, which has a central tendency loss fraction of 0.2 percent and a high-end loss fraction of 1 percent
of the 1,4-dioxane throughput. EPA estimated land releases from release point #5 by assuming the entire
volume of the cleaning bath used for equipment and parts is released to landfill. This is consistent with
the process information, which indicates that spent 1,4-dioxane is disposed of as chemical waste, which
EPA assumes may be to either incineration or RCRA subpart C landfills ( 20c).
EPA's calculation of land releases for this OES, including all calculation inputs and assumptions, can be
found in 1,4-Dioxane Supplemental Information File: Environmental Releases to Landfor all OES
Except Disposal (U.S.. 0241).
Textile Dye
EPA estimated land releases for facilities within the Textile Dye OES using the OECD ESD on Textile
Dyes (OECD. 2017) and Monte Carlo modeling. The ESD on Textile Dyes was rated medium during
EPA's systematic review process. The use of Montel Carlo modeling allows for variation of calculation
input parameters such that a distribution of environmental releases can be calculated, from which EPA
can estimate the 50th and 95th percentile releases. An explanation of this modeling approach is included
in Appendix E. 11.
Antifreeze
EPA did not find any directly applicable GS/ESD or literature sources for this OES; however, EPA
evaluated the potential for releases using the OECD ESD on Chemical Additives used in Automotive
Lubricants (OBt «* -°20) and the EPA MRD on Commercial Use of Automotive Detailing Products
( 'A. 2022b). The ESD and MRD were both rated high during EPA's systematic review process.
For the use of antifreeze, EPA expects releases may occur from volatilizations of 1,4-dioxane during
unloading/ pouring antifreeze into vehicles, disposal or cleaning of empty antifreeze containers, and
disposal of spent antifreeze. Both the ESD and MRD indicate that containers of automotive maintenance
fluids are typically small and are not rinsed, but rather disposed of as solid waste ( 322b;
OECD. 2020). Additionally, the ESD on Chemical Additives used in Automotive Lubricants indicates
that spent lubricants are disposed of via incineration by blending with fuel oil (OECD. 2020). However,
since spent antifreeze is unlikely to be blended with fuel oil, EPA expects spent antifreeze may be
disposed of via incineration or landfills that take chemical waste. Therefore, EPA expects land releases
result from disposal of empty antifreeze containers and spent antifreeze.
To estimate the use rate of 1,4-dioxane for this OES, EPA used the consumer use rate of antifreeze (0.15
kg antifreeze/job) from the Final Risk Evaluation for 1,4-Dioxane ( )20c) and scaled this
value up to a commercial use rate based on a range of the number of cars serviced at auto shops from the
Near-Field/Far-Field Brake Model and Automotive Detailing MRD (1 to 9 jobs/day). EPA used a range
of concentration of 1,4-dioxane in antifreeze from the process description in Appendix F.4.2 and
assumed antifreeze container sizes ranging from 16 ounces to 5 gallons per the default container sizes in
the MRD and ESD, respectively (U.S. EPA. 202:h; UUCP. 2020).
Page 274 of 570
-------
To estimate the land release from container disposal, EPA used the calculated 1,4-dioxane throughput
based on the above batch parameters and the EPA/OPPT Small Container Residual Model, which has a
central tendency loss fraction of 0.3 percent and a high-end loss fraction of 0.6 percent. To estimate the
land release from spent antifreeze, EPA used the 1,4-dioxane throughput and a mass balance assuming
100 percent release minus upstream losses from container disposal and volatilizations during unloading
(estimated with the EPA/OAQPS AP-42 Loading Model).
EPA's calculation of land releases for this OES, including all calculation inputs and assumptions, can be
found in the supplemental attachment 1,4-Dioxane Supplemental Information File: Environmental
Releases to Landfor all OES Except Disposal (U.S. EPA. 20241).
Surface Cleaner
EPA did not find any directly applicable GS/ESD or literature sources for this OES; however, EPA
estimated land releases using the SHEDs-HT modeling conducted for the one case study location
(Liverpool OH) and the assumptions described here. EPA expects that the main release points from the
use of surface cleaners are from
1. Disposal of empty containers containing residual cleaning solution,
2. Application of the cleaning solution, and
3. Disposal of cleaning solution by rinsing or wiping.
Because EPA did not find any directly applicable GSs or ESDs, EPA used the Draft GS on Furnishing
Cleaning ( 022a) to inform these releases due to the similarities in surface cleaning and
furnishing cleaning. The Draft GS on Furnishing Cleaning was rated high during EPA's systematic
review process. Per this Draft GS, empty containers may be rinsed out in sinks or disposed of without
rinsing, such that releases may be to wastewater or landfill; the GS uses the EPA/OPPT Small Container
Residual Model to estimate this release. Application losses are to fugitive air from spray application; the
GS uses literature data to estimate this release. Once applied, the cleaner may be rinsed off or wiped off
with rags or towels, such that releases may be to wastewater or landfill; the GS assumes 100 percent
release scenario, estimating this release by subtracting the upstream losses from the cleaner use rate
( 2022a).
The SHEDs-HT modeling estimated wastewater discharges of 72 g of 1,4-dioxane per day for
commercial uses of surface cleaners containing 1,4-dioxane in Liverpool OH. As described previously,
because both release point #1 and #3 may also be to either wastewater or landfills, EPA assumes the
same quantity of 72 g of 1,4-dioxane per day from the SHEDs-HT model may be released to unknown
landfills for this OES. EPA notes that these 72 g is either entirely to wastewater or landfill or some split
between the two media. The 72 g is not to both wastewater and landfill because that would double count
the release,
EPA's calculation of land releases for this OES, including all calculation inputs and assumptions, can be
found in 1,4-Dioxane Supplemental Information File: Environmental Releases to Landfor all OES
Except Disposal (U.S.. 0241).
Page 275 of 570
-------
Dish Soap
EPA estimated land releases for facilities within the Dish soap OES using data from a public comment,
EPA/OPPT models, and Monte Carlo modeling. The public comment was rated high during EPA's
systematic review process (P&G. 2023). The use of Monte Carlo modeling allows for variation of
calculation input parameters such that a distribution of environmental releases can be calculated, from
which EPA can estimate the 50th and 95th percentile releases. An explanation of this modeling approach
is included in Appendix E. 14.
Dishwasher Detergent
EPA estimated land releases for facilities within the Dishwasher detergent OES using data from a public
comment, EPA/OPPT models, and Monte Carlo modeling. The public comment was rated high during
EPA's systematic review process (P&G. 2023). The use of Monte Carlo modeling allows for variation
of calculation input parameters such that a distribution of environmental releases can be calculated, from
which EPA can estimate the 50th and 95th percentile releases. An explanation of this modeling approach
is included in Appendix E. 14.
Laundry Detergent
EPA estimated land releases for facilities within the Laundry detergent OES using the OECD ESD on
Industrial and Institutional Laundries (01 ) and Monte Carlo modeling. The ESD on
Industrial and Institutional Laundries was rated medium during EPA's systematic review process. The
use of Montel Carlo modeling allows for variation of calculation input parameters such that a
distribution of environmental releases can be calculated, from which EPA can estimate the 50th and 95th
percentile releases. An explanation of this modeling approach is included in Appendix E.11.16.
Paints and Floor Lacquer
EPA estimated land releases for facilities within the Paints and floor lacquers OES using the OECD
ESD on Coating Application via Spray-Painting in the Automotive Refinishing Industry (
201 la). The ESD was rated medium during EPA's systematic review process.
As described in the process description in Appendix F.4.7, 1,4-dioxane was identified by a public
comment as present in automotive refinishing products, thereby allowing EPA to identify the above
ESD as the most applicable GS/ESD available. This ESD indicates that releases are expected from
1. Releases to incineration or landfill from container cleaning/disposal,
2. Releases to incineration or landfill from equipment cleaning,
3. Releases to incineration or landfill from over sprayed coating that is captured by emission
controls, and
4. Releases to stack air from over sprayed coating that is not captured by emission controls.
Based on the GS, release points #1 through 3 have the potential for land releases. To estimate these
releases, EPA used the equations specified in the ESD (OE ). Apart from weight fraction in
coatings (see Appendix F.4.7), EPA did not find any data specific to 1,4-dioxane in this OES. Therefore,
the calculation of releases using this GS are for a "generic site," using the default input parameter values
from the ESD.
Using this methodology, EPA calculated the low-end and high-end land releases for this OES, which are
expected to be to unknown landfills per the ESD (OECD. 201 la). The low- and high-end estimates are
based on the low- and high-end calculation input parameter defaults from the ESD. EPA's calculation of
land releases for this OES, including all calculation inputs and assumptions, can be found in 1,4-
Page 276 of 570
-------
Dioxane Supplemental Information File: Environmental Releases to Landfor all OES Except Disposal
(U.S. EPA. 202411
Hydraulic Fracturing
EPA estimated land releases for facilities within the Hydraulic fracturing OES using the Draft OECD
ESD on Hydraulic Fracturing (U.S. EPA. 2022e) and Monte Carlo modeling. The Revised ESD on
Hydraulic Fracturing was rated high during EPA's systematic review process. The use of Montel Carlo
modeling allows for variation of calculation input parameters such that a distribution of environmental
releases can be calculated, from which EPA can estimate the 50th and 95th percentile releases. An
explanation of this modeling approach is included in Appendix E.13.
E.4.3 Land Release Estimates Summary
A summary of industrial and commercial land releases estimated using the above methods is presented
in Table Apx E-5 below. Specifically, this table presents the range of daily or annual land releases per
site for each OES.
Page 277 of 570
-------
Table Apx E-5. Summary of Daily Int
ustrial and Commercial Land Release Estimates for 1,4-Dioxane
OES
Type of Land Release
Number of
Facilities with
Releases"
Range of Estimated Daily
Release Per Site
(kg/site-day)
Estimated
Release
Frequency
Range
(days)''
Overall Data
Quality
Determination
Sources'
Min
Max
Manufacturing
Land (all types)
1
0
250
Medium
TRI
Import and
repackaging
Land (all types)
1
0
250
Medium
TRI
Industrial uses
Land (all types)
12
0
227 (annually)
250
Medium
TRI
Functional fluids
(open-system)
Land (all types)
2
0
0
247
Medium
TRI
Functional fluids
(closed-system)
All
Assessed as a part of Industrial uses OES
N/A
N/A
Laboratory
chemical
Land (unknown type)
132
0
489 (annually)
250
High
GSd
Film cement
Land (unknown type)
211
0.0035
(annually)
0.037
(annually)
250
High
Process
information®
Spray foam
application
Land (unknown type)
1,553,559
0.032 (annually)
0.047
(annually)
3
Medium
GS^
Printing inks (3D)
Fugitive Air, Stack
Air, and Land (all
types)
Assessed as a part of Industrial uses OES
250
N/A
N/A
Dry film lubricant
Land (hazardous waste
landfill)
8
0
188 (annually)
48
High
Process
information®
Page 278 of 570
-------
OES
Type of Land Release
Number of
Facilities with
Releases"
Range of Estimated Daily
Release Per Site
(kg/site-day)
Estimated
Release
Frequency
Range
(days)''
Overall Data
Quality
Determination
Sources'
Min
Max
Disposal
Land (RCRA Sub C
landfill)
18
0
7,307
(annually)
250
Medium
TRI
Land (Underground
injection)
18
0
331,980
(annually)
Medium
TRI
Land (Non-RCRA
landfills)
18
0
890 (annually)
Medium
TRI
Land (all other types)
18
0
0
Medium
TRI
Textile dye (draft
RE estimates)"
Land (unknown
landfill type) or
POTW (unknown
partitioning)
783
2.09E-07
9.72E-05
31 to 295
Medium
ESDg and Monte
Carlo Modeling''
Textile dye
(updated
estimates)"
Land (unknown
landfill type) or
POTW (unknown
partitioning)
783
1.9E-07
9.6E-05
31 to 295
Medium
ESDg and Monte
Carlo Modeling''
Antifreeze
Land (unknown
landfill)
84,383
3.75E-07
(annually)
0.029
(annually)
250
High
Process
information® and
Modeling''
Surface cleaner
Land (unknown
landfill) or POTW
Unknown
18"' (single daily release value for
all sites combined in Liverpool
OH case study)
250
High
SHEDS-HT',
Process
information®
Modeling''
Dish soap (draft
RE estimates)"
Land (unknown
landfill)
Unknown
0.048 (annual
value for all
sites in
Liverpool OH
case study)
0.097 (annual
value for all
sites in
Liverpool OH
case study)
250
High
SHEDS-HT,
Process
information
Modeling
Dish soap
(updated
estimates)"
Land (unknown
landfill)
773,851
7.0E-11
7.4E-05
350
High
fP&G. 2023) and
Monte Carlo
Modeling''
Page 279 of 570
-------
OES
Type of Land Release
Number of
Facilities with
Releases"
Range of Estimated Daily
Release Per Site
(kg/site-day)
Estimated
Release
Frequency
Range
(days)''
Overall Data
Quality
Determination
Sources'
Min
Max
Dishwasher
detergent (draft
RE estimates)"
Land (unknown
landfill)
Unknown
1.08E-03
(annual value for
all sites in
Liverpool, OH
case study)
2.17E-03
(annual value
for all sites in
Liverpool, OH
case study)
250
High
SHEDS-HT,
Process
information
Modeling
Dishwasher
detergent (updated
estimates)"
Land (unknown
landfill)
773,851
7.6E-10
2.2E-05
350
High
2023) and
Monte Carlo
Modeling72
Laundry detergent
(institutional) -
liquid detergents
(draft RE
estimates)"
Land (unknown
landfill), incineration,
or POTW (unknown
partitioning)
95,533
4.05E-12
3.95E-05
250 to 365
Medium
ESD2 and Monte
Carlo Modeling72
Laundry detergent
(institutional) -
liquid detergents
(updated
estimates)"
Land (unknown
landfill), incineration,
or POTW (unknown
partitioning)
95,533
8.1E-13
3.8E-04
250 to 365
Medium
ESD' and Monte
Carlo Modeling'2
Laundry detergent
(institutional) -
powder detergents
(draft RE
estimates)"
Land (unknown
landfill), incineration,
or POTW (unknown
partitioning)
95,533
5.36E-08
0.0018
250 to 365
Medium
ESD' and Monte
Carlo Modeling'2
Laundry detergent
(institutional) -
powder detergents
(updated
estimates)"
Land (unknown
landfill), incineration,
or POTW (unknown
partitioning)
95,533
1.4E-08
1.8E-02
250 to 365
Medium
ESD' and Monte
Carlo Modeling'2
Laundry detergent
(industrial) -
liquid detergents
Land (unknown
landfill), incineration,
2,453
4.78E-12
1.46E-04
250 to 365
Medium
ESD' and Monte
Carlo Modeling'2
Page 280 of 570
-------
OES
Type of Land Release
Number of
Facilities with
Releases"
Range of Estimated Daily
Release Per Site
(kg/site-day)
Estimated
Release
Frequency
Range
(days)''
Overall Data
Quality
Determination
Sources'
Min
Max
(draft RE
estimates)"
or POTW (unknown
partitioning)
Laundry detergent
(industrial) -
liquid detergents
(updated
estimates)"
Land (unknown
landfill), incineration,
or POTW (unknown
partitioning)
2,453
6.6E-13
1.4E-03
20 to 365
Medium
ESD' and Monte
Carlo Modeling''
Laundry detergent
(industrial) -
powder detergents
(draft RE
estimates)"
Land (unknown
landfill), incineration,
or POTW (unknown
partitioning)
2,453
2.92E-11
3.92E-04
250 to 365
Medium
ESD' and Monte
Carlo Modeling''
Laundry detergent
(industrial) -
powder detergents
(updated
estimates)"
Land (unknown
landfill), incineration,
or POTW (unknown
partitioning)
2,453
1.5E-11
3.8E-03
20 to 365
Medium
ESD' and Monte
Carlo Modeling''
Paints and floor
lacquer
Land (unknown
landfill)
33,648
3.04E-06
(annually)
0.010
(annually)
250
Medium
ESD'
Polyethylene
terephthalate
(PET) byproduct
Land (Land treatment)
13
0
45.4 (annually)
Medium
TRI
Land (Non-RCRA
landfills)
13
0
0.10 (annually)
250
Medium
TRI
Land (all other types)
13
0
0
Medium
TRI
Ethoxylation
process byproduct
Land (underground
injection)
8
0
197,714
(annually)
250
Medium
TRI
Land (all other types)
8
0
0
Medium
TRI
Page 281 of 570
-------
OES
Type of Land Release
Number of
Facilities with
Releases"
Range of Estimated Daily
Release Per Site
(kg/site-day)
Estimated
Release
Frequency
Range
(days)''
Overall Data
Quality
Determination
Sources'
Min
Max
Surface water,
incineration, or landfill
(unknown
partitioning)
411
3.61E-10
4.59
1 to 72
Medium
ESD' and Monte
Carlo Modeling11
Hydraulic
fracturing (draft
RE estimates)"
Land (underground
injection)
411
5.35E-09
108
Medium
ESD' and Monte
Carlo Modeling11
Recycle/Reuse (48%),
underground injection
(43%), Surface water
(6%), or land (3%)
411
1.85E-10
1.12
Medium
ESD' and Monte
Carlo Modeling11
Surface water,
incineration, or landfill
(unknown
partitioning)
411
4.3E-10
5.6
1 to 72
Medium
ESD' and Monte
Carlo Modeling11
Land (underground
injection)
411
1.2E-08
179
Medium
ESD' and Monte
Carlo Modeling''
Hydraulic
fracturing
(updated
estimates)"
Recycle/Reuse (5%),
underground injection
(70%), Surface water
(19%), or land
(evaporation ponds,
percolation ponds,
irrigation, road
treatment) (6%)
411
2.8E-09
14
Medium
ESD'' and Monte
Carlo Modeling''
Surface water (13%),
Land (soil) (64%), and
Landfill or
Incineration (23%)
411
4.9E-11
0.64
Medium
ESD'' and Monte
Carlo Modeling''
Page 282 of 570
-------
OES
Type of Land Release
Number of
Facilities with
Releases"
Range of Estimated Daily
Release Per Site
(kg/site-day)
Min
Max
Estimated
Release
Frequency
Range
(days)''
Overall Data
Quality
Determination
Sources'
a See Appendix E.l for explanation of how EPA determined the number of sites for each OES.
b Where available. EPA used the December 2020 Final Risk Evaluation for 1,4-Dioxane (U.S. EPA. 2020c). generic scenarios, and emission scenario documents
to provide a basis to estimate the number of release days of 1,4-dioxane within a COU.
c Narrative descriptions of all release estimate sources are provided in Appendix E.4.2.
d The generic scenario used for this COU is the GS on Use of Laboratory Chemicals (U.S. EPA. 20220.
'' For this COU, EPA used process information, which is further described in Appendix E.4.2.
' The generic scenario used for this COU is the GS on Application of Spray PoK urethane Foam Insulation (U.S. EPA. 2018b).
g The emission scenario document used for this COU is the ESD on Textile Dyes (OECD. 2017).
h For this COU, EPA used various models and literature for model input parameters as described in Appendix E.4.2.
1 The emission scenario document used for this COU is the ESD on Industrial and Institutional Laundries (OECD. 201 lb).
1 The emission scenario document used for this COU is the ESD on Coating Application via Spray Painting in the Automotive Refinishing Industry (OECD.
la).
k The emission scenario document used for this COU is the Revised ESD on Hydraulic Fracturing (U.S. EPA. 2022e).
' EPA used the down the drain water release estimates from the SHEDs-HT model for the Liverpool OH case study (see Section 2.1.1.2) to estimate air and land
releases by back calculating 1,4-dioxane use rates and applying loss fractions for air and land releases using literature and standard models described in Appendix
E.4.2.
A single annual value was provided for all sites in the Liverpool, OH case study.
" For select OESs, updates to release estimates were made via information provided by the SACC and public comments.
Page 283 of 570
-------
E.4.4 Summary of Weight of Scientific Evidence Conclusions in Land Release Estimates
Table Apx E-6 provides a summary of EPA's weight of scientific evidence conclusions in its land release estimates for each of the
Occupational Exposure Scenarios assessed. Detailed descriptions of non-OES specific strengths, limitations, assumptions, and uncertainties
(e.g., general limitations for TRI, DMR, etc.) are provided in Appendix E.6.
Table Apx E-6. Summary of Weight of Scientific Evidence Conclusions in Land Release Estimates by PES
OES
Weight of Scientific Evidence Conclusion in Release Estimates
Manufacturing
Land releases are assessed using reported releases from 2019 TRI. Factors that increase the strength of evidence for this
OES are that the release data are directly relevant to the OES (as opposed to surrogate), that TRI has a medium overall data
quality determination, and consistency within the dataset (all reporters use the same or similar reporting forms). A strength
of TRI data is that TRI compiles the best reasonably available release data for all reporting facilities. Factors that decrease
the strength of the evidence for this OES include the low number of data points, lack of variability (only 1 year of data
used), uncertainty in the accuracy of reported releases, and the limitations in representativeness to all sites because TRI
may not capture all relevant sites. Additionally, EPA could not estimate the number of release days per year associated
with land releases. Based on this information, EPA has concluded that the weight of scientific evidence for this assessment
is moderate to robust and provides a plausible estimate of releases in consideration of the strengths and limitations of
reasonably available data.
Import and repackaging
Land releases are assessed using reported releases from 2019 TRI. Factors that increase the strength of evidence for this
OES are that the release data are directly relevant to the OES (as opposed to surrogate), that TRI has a medium overall data
quality determination, and consistency within the dataset (all reporters use the same or similar reporting forms). A strength
of TRI data is that TRI compiles the best reasonably available release data for all reporting facilities. Factors that decrease
the strength of the evidence for this OES include the low number of data points, lack of variability (only 1 year of data
used), uncertainty in the accuracy of reported releases, and the limitations in representativeness to all sites because TRI
may not capture all relevant sites. Additionally, the land release assessment is based on one reporting site that reported no
land releases and EPA did not have additional sources to estimate land releases for other sites in this OES. Additionally,
EPA could not estimate the number of release days per year associated with land releases. Based on this information, EPA
has concluded that the weight of scientific evidence for this assessment is moderate to robust and provides a plausible
estimate of releases in consideration of the strengths and limitations of reasonably available data.
Industrial uses
Land releases are assessed using reported releases from 2019 TRI. Factors that increase the strength of evidence for this
OES are that the release data are directly relevant to the OES (as opposed to surrogate), that TRI has a medium overall data
quality determination, and consistency within the dataset (all reporters use the same or similar reporting forms). A strength
of TRI data is that TRI compiles the best reasonably available release data for all reporting facilities. Factors that decrease
the strength of the evidence for this OES include the lack of variability (only 1 year of data used), uncertainty in the
accuracy of reported releases, uncertainty in EPA's use of Form A submissions, and the limitations in representativeness to
all sites because TRI may not capture all relevant sites. Some facilities within this OES reported to TRI using a Form A,
which does not include any details on chemical release quantities. When a facility has submitted a Form A, there is no way
to discern the quantity released. Therefore, where facilities reported to TRI with a Form A, EPA used the Form A threshold
Page 284 of 570
-------
OES
Weight of Scientific Evidence Conclusion in Release Estimates
for total releases of 500 lb/year for each release media; however, there is uncertainty in this because the actual release
quantity is unknown. Furthermore, the threshold represents an upper limit on total releases from the facility; therefore,
assessing releases at the threshold value may overestimate actual releases from the facility. Additionally, EPA could not
estimate the number of release days per year associated with land releases. Based on this information, EPA has concluded
that the weight of scientific evidence for this assessment is moderate to robust and provides a plausible estimate of releases
in consideration of the strengths and limitations of reasonably available data.
Functional fluids (open-
system)
Land releases are assessed using reported releases from 2019 TRI. Factors that increase the strength of evidence for this
OES are that the release data are directly relevant to the OES (as opposed to surrogate), that TRI has a medium overall data
quality determination, and consistency within the dataset (all reporters use the same or similar reporting forms). A strength
of TRI data is that TRI compiles the best reasonably available release data for all reporting facilities. Factors that decrease
the strength of the evidence for this OES include the low number of data points, lack of variability (only 1 year of data
used), uncertainty in the accuracy of reported releases, and the limitations in representativeness to all sites because TRI
may not capture all relevant sites. The land release assessment is based on two reporting sites that both reported no land
releases and EPA did not have additional sources to estimate land releases for sites in this OES. Additionally, EPA could
not estimate the number of release days per year associated with land releases. Based on this information, EPA has
concluded that the weight of scientific evidence for this assessment is moderate to robust and provides a plausible estimate
of releases in consideration of the strengths and limitations of reasonably available data.
Functional fluids (closed-
system)
No data was available to estimate releases for this OES. For the land release assessment, EPA grouped this OES with the
Industrial uses OES because the sources of release are expected to be similar between these OESs. Factors that increase the
strength of evidence for this OES are that TRI has a medium overall data quality determination and consistency within the
dataset (all reporters use the same or similar reporting forms). Factors that decrease the strength of evidence for this OES
are that the Industrial Releases OES release data are use as surrogate for this OES, uncertainty in the accuracy of reported
releases, limitations in representativeness to all sites because TRI may not capture all relevant sites, and lack of variability
(only 1 year of data used). Refer to the Industrial uses OES discussion for additional discussion. Based on this information,
EPA has concluded that the weight of scientific evidence for this assessment is slight and provides a plausible estimate of
releases in consideration of the strengths and limitations of reasonably available data.
Laboratory chemicals
Land releases are assessed using the Draft GS on Use of Laboratory Chemicals. Factors that increase the strength of
evidence for this OES are that the release estimates are directly relevant to the OES (as opposed to surrogate), the Draft GS
on Use of Laboratory Chemicals has a high overall data quality determination, and the low level of uncertainty in the data.
Factors that decrease the strength of the evidence for this OES include the that the GS has not been peer-reviewed,
uncertainty in the representativeness of the GS towards all sites in this OES, and a lack of variability in the analysis.
Specifically, because the default values in the ESD are generic, there is uncertainty in the representativeness of generic site
estimates of actual releases from real-world sites that use 1,4-dioxane. Another uncertainty is lack of consideration for
release controls. The ESD assumes that all activities occur without any release controls. Actual releases may be less than
estimated if facilities utilize pollution control methods, contributing to uncertainty. Based on this information, EPA has
concluded that the weight of scientific evidence for this assessment is slight to moderate and provides a plausible estimate
of releases in consideration of the strengths and limitations of reasonably available data.
Page 285 of 570
-------
OES
Weight of Scientific Evidence Conclusion in Release Estimates
Film cement
Land releases are assessed using process information from the Final Risk Evaluation for 1,4-Dioxane and EPA/OPPT
models. Factors that increase the strength of evidence for this OES are that the release estimates are directly relevant to the
OES (as opposed to surrogate), the underlying data sources for the process information have a high overall data quality
determination, and the low level of uncertainty in the data because the process information comes directly from actual
users of 1,4-dioxane in film cement. Factors that decrease the strength of the evidence for this OES include uncertainty in
the representativeness of evidence to all sites in this OES and a lack of variability in the input parameters for the used
models. Specifically, the process information for the production and use of film cement is based on information from three
use sites, one from Australia and two from the U.S. Based on this information, EPA has concluded that the weight of
scientific evidence for this assessment is slight to moderate and provides a plausible estimate of releases in consideration
of the strengths and limitations of reasonably available data.
Spray foam application
Land releases are assessed using the GS on Application of Spray Polyurethane Foam Insulation. Factors that increase the
strength of evidence for this OES are that the release estimates are directly relevant to the OES (as opposed to surrogate),
the underlying data sources for the process information have a medium overall data quality determination, and a low level
of uncertainty in the data. Factors that decrease the strength of the evidence for this OES include uncertainty in the
representativeness of the GS to all sites since it is generic and not specific to sites that use 1,4-dioxane and a lack of
variability. Based on this information, EPA has concluded that the weight of scientific evidence for this assessment is
slight to moderate and provides a plausible estimate of releases in consideration of the strengths and limitations of
reasonably available data.
Printing inks (3D)
No data was available to estimate releases for this OES. For the land release assessment, EPA grouped this OES with the
Industrial uses OES because the sources of release are expected to be similar between these OESs. Factors that increase the
strength of evidence for this OES are that TRI has a medium overall data quality determination and consistency within the
dataset (all reporters use the same or similar reporting forms). Factors that decrease the strength of evidence for this OES
are that the Industrial Releases OES release data are use as surrogate for this OES, uncertainty in the accuracy of reported
releases, limitations in representativeness to all sites because TRI may not capture all relevant sites or smaller commercial
3D printing uses, and lack of variability (only 1 year of data used). Refer to the Industrial uses OES discussion for
additional discussion. Based on this information, EPA has concluded that the weight of scientific evidence for this
assessment is slight and provides a plausible estimate of releases in consideration of the strengths and limitations of
reasonably available data.
Dry film lubricant
Land releases are assessed using process information from the Final Risk Evaluation for 1,4-Dioxane. Factors that increase
the strength of evidence for this OES are that the release estimates are directly relevant to the OES (as opposed to
surrogate), that the underlying data sources for the process information have a high overall data quality determination, and
a low level of uncertainty in the data because the process information comes directly from an actual user of 1,4-dioxane in
dry film lubricants. Factors that decrease the strength of the evidence for this OES include uncertainty in the
representativeness of evidence to all sites and a lack of variability. Based on this information, EPA has concluded that the
weight of scientific evidence for this assessment is slight to moderate and provides a plausible estimate of releases in
consideration of the strengths and limitations of reasonably available data.
Page 286 of 570
-------
OES
Weight of Scientific Evidence Conclusion in Release Estimates
Disposal
Land releases are assessed using reported discharges from 2013-2019 TRI. Factors that increase the strength of evidence
for this OES are that the release data are directly relevant to the OES (as opposed to surrogate), that TRI has a medium
overall data quality determination, high number of data points, and consistency within the dataset (all reporters use the
same or similar reporting forms). Additionally, EPA included seven years of TRI data in the analysis, which increases the
variability of the dataset. A strength of TRI data is that TRI compiles the best reasonably available release data for all
reporting facilities. Factors that decrease the strength of the evidence for this OES include uncertainty in the accuracy of
reported releases, uncertainty in EPA's use of Form A submissions, and the limitations in representativeness to all sites
because TRI may not capture all relevant sites. Some facilities within this OES reported to TRI using a Form A, which
does not include any details on chemical release quantities. When a facility has submitted a Form A, there is no way to
discern the quantity released. Therefore, where facilities reported to TRI with a Form A, EPA used the Form A threshold
for total releases of 500 lb/year for each release media; however, there is uncertainty in this because the actual release
quantity is unknown. Furthermore, the threshold represents an upper limit on total releases from the facility; therefore,
assessing releases at the threshold value may overestimate actual releases from the facility. Based on this information, EPA
has concluded that the weight of scientific evidence for this assessment is moderate to robust and provides a plausible
estimate of releases in consideration of the strengths and limitations of reasonably available data.
Textile dye
Land releases are assessed using Monte Carlo modeling with information from the ESD on Textile Dyes. Factors that
increase the strength of evidence for this OES are that the release estimates are directly relevant to the OES (as opposed to
surrogate), the ESD on Textile Dyes has a medium overall data quality determination and was peer reviewed, the high
number of data points (simulation runs), consistency within the dataset, and full distributions of input parameters. The
Monte Carlo modeling accounts for the entire distribution of input parameters, calculating a distribution of potential
release values that represents a larger proportion of sites than a discrete value. Factors that decrease the strength of the
evidence for this OES include uncertainties and limitations in the representativeness of the estimates for sites that
specifically use 1,4-dioxane because the default values in the ESD are generic. Another uncertainty is lack of consideration
for release controls. The ESD assumes that all activities occur without any release controls. Actual releases may be less
than estimated if facilities utilize pollution control methods, contributing to uncertainty. Based on this information, EPA
has concluded that the weight of scientific evidence for this assessment is moderate and provides a plausible estimate of
releases in consideration of the strengths and limitations of reasonably available data.
Antifreeze
Land releases are assessed using the OECD ESD on Chemical Additives used in Automotive Lubricants, the EPA MRD on
Commercial Use of Automotive Detailing Products, and EPA/OPPT models. Factors that increase the strength of evidence
for this OES are that the ESD and MRD used have high overall data quality determinations and consistency within the
sources used. Factors that decrease the strength of the evidence for this OES include that the ESD and MRD are not
directly applicable to antifreeze uses (used as surrogate), uncertainty in the representativeness of the ESD and MRD to all
sites and sites that specifically use 1,4-dioxane since these documents contain generic values, and a lack of variability.
Additionally, EPA scaled up a consumer antifreeze use rate to a commercial use rate based on information in the ESD and
MRD, which increases uncertainty. Based on this information, EPA has concluded that the weight of scientific evidence
for this assessment is slight to moderate and provides a plausible estimate of releases in consideration of the strengths and
limitations of reasonably available data.
Page 287 of 570
-------
OES
Weight of Scientific Evidence Conclusion in Release Estimates
Surface cleaner
Land releases are assessed using SHEDS-HT modeled water releases in conjunction with the Draft GS on Furnishing
Cleaning. Factors that increase the strength of evidence for this OES include that the release estimates are directly relevant
to the OES (as opposed to surrogate), that the Draft GS used has a high overall data quality determination, and variability
in the model input parameters. Factors that decrease the strength of the evidence for this OES include uncertainty in the
representativeness to all sites because the SHEDS-HT estimate is based on one case study for Liverpool, OH and because
the estimate is not site-specific (the release estimate is a total for all sites in Liverpool, OH). Additionally, the Draft GS
describes potential release points for this OES, identifying releases that may be to either water or land depending on site
practices (e.g., surface cleaner may be rinsed down drains or wiped off with rags that are disposed of as trash). Because
there is no information to determine the quantity released specifically to land, EPA assumed that the entire quantity
modeled to water with the SHEDS-HT model may also be released to land, which introduces uncertainty. Based on this
information, EPA has concluded that the weight of scientific evidence for this assessment is slight and provides a plausible
estimate of releases in consideration of the strengths and limitations of reasonably available data.
Dish soap
Land releases are assessed using Monte Carlo modeling with information from a public comment and standard EPA/OPPT
models. Factors that increase the strength of evidence for this OES are that the release estimates are directly relevant to the
OES (as opposed to surrogate), that the public comment has a hiah overall data qualitv determination (P&G. 2023). there
are a high number of data points (simulation runs), and full distributions of input parameters. Monte Carlo modeling
accounts for the entire distribution of input parameters, calculating a distribution of potential release values that represents
a larger proportion of sites than a discrete value. The major factor that decreases the strength of the evidence for this OES
include the uncertainties and limitations in the representativeness of the data from the public comment towards all sites that
use dish soaps containing 1,4-dioxane. Another uncertainty is the lack of a GS or ESD describing this scenario; EPA used
standard EPA/OPPT models for each of the expected release points to build the model. Based on this information, EPA has
concluded that the weight of scientific evidence for this assessment is moderate and provides a plausible estimate of
releases in consideration of the strengths and limitations of reasonably available data.
Dishwasher detergent
EPA used the same approach to estimate land releases for this OES as the Dish soap OES. Therefore, the same rationale
and overall weight of scientific evidence apply to this OES.
Laundry detergent
Land releases are assessed using Monte Carlo modeling with information from the ESD on Industrial and Institutional
Laundries. Factors that increase the strength of evidence for this OES are that the release estimates are directly relevant to
the OES (as opposed to surrogate), that the ESD on Industrial and Institutional Laundries has a medium overall data
quality determination and was peer reviewed, there are high number of data points (simulation runs), consistency within
the dataset, and full distributions of input parameters. The Monte Carlo modeling accounts for the entire distribution of
input parameters, calculating a distribution of potential release values that represents a larger proportion of sites than a
discrete value. Additionally, EPA was able to separately estimate releases for industrial and institutional laundry settings.
Factors that decrease the strength of the evidence for this OES include uncertainties and limitations in the
representativeness of the estimates for sites that specifically use 1,4-dioxane because the default values in the ESD are
generic. Another uncertainty is lack of consideration for release controls. The ESD assumes that all activities occur without
any release controls. Actual releases may be less than estimated if facilities utilize pollution control methods. Based on this
Page 288 of 570
-------
OES
Weight of Scientific Evidence Conclusion in Release Estimates
information, EPA has concluded that the weight of scientific evidence for this assessment is moderate and provides a
plausible estimate of releases in consideration of the strengths and limitations of reasonably available data.
Paint and floor lacquer
Land releases are assessed using OECD ESD on Coating Application via Spray-Painting in the Automotive Refinishing
Industry. Factors that increase the strength of evidence for this OES are that the release estimates are directly relevant to
the OES (as opposed to surrogate), the ESD has a medium overall data quality determination and has been peer reviewed,
consistency within the sources used, and a low amount of uncertainties. Factors that decrease the strength of the evidence
for this OES include a lack of variability and uncertainty in the representativeness of the ESD to all sites and sites that
specifically use 1,4-dioxane since the ESD is generic. Based on this information, EPA has concluded that the weight of
scientific evidence for this assessment is slight to moderate and provides a plausible estimate of releases in consideration
of the strengths and limitations of reasonably available data.
PET byproduct
Land releases are assessed using reported discharges from 2019 TRI. Factors that increase the strength of evidence for this
OES are that the release data are directly relevant to the OES (as opposed to surrogate), that TRI has a medium overall data
quality determination, and consistency within the dataset (all reporters use the same or similar reporting forms). A strength
of TRI data is that TRI compiles the best reasonably available release data for all reporting facilities. Factors that decrease
the strength of the evidence for this OES include uncertainty in the accuracy of reported releases, lack of variability (only 1
year of data used), and the limitations in representativeness to all sites because TRI may not capture all relevant sites. The
land release assessment is based on 13 reporting sites, 11 of which reported no land releases. EPA did not have additional
sources to estimate land releases for site in this OES that may not report to TRI. Based on this information, EPA has
concluded that the weight of scientific evidence for this assessment is moderate to robust and provides a plausible estimate
of releases in consideration of the strengths and limitations of reasonably available data.
Ethoxylation process
byproduct
Land releases are assessed using reported discharges from 2019 TRI. Factors that increase the strength of evidence for this
OES are that the release data are directly relevant to the OES (as opposed to surrogate), that TRI has a medium overall data
quality determination, and consistency within the dataset (all reporters use the same or similar reporting forms). A strength
of TRI data is that TRI compiles the best reasonably available release data for all reporting facilities. Factors that decrease
the strength of the evidence for this OES include uncertainty in the accuracy of reported releases, lack of variability (only 1
year of data used), and the limitations in representativeness to all sites because TRI may not capture all relevant sites. The
land release assessment is based on eight reporting sites, seven of which reported no land releases. EPA did not have
additional sources to estimate land releases for site in this OES that may not report to TRI. Based on this information, EPA
has concluded that the weight of scientific evidence for this assessment is moderate to robust and provides a plausible
estimate of releases in consideration of the strengths and limitations of reasonably available data.
Hydraulic fracturing
Land releases are assessed using Monte Carlo modeling with information from the Revised ESD on Hydraulic Fracturing
and FracFocus 3.0. Factors that increase the strength of evidence for this OES are that the release estimates are directly
relevant to the OES (as opposed to surrogate), that the Revised ESD on Hydraulic Fracturing and FracFocus 3.0 have
medium overall data quality determinations, that the Revised ESD has undergone peer review by OECD, the high number
of data points (simulation runs), consistency within the dataset, and full distributions of input parameters. The Monte Carlo
modeling accounts for the entire distribution of input parameters, calculating a distribution of potential release values that
represents a larger proportion of sites than a discrete value. Factors that decrease the strength of the evidence for this OES
Page 289 of 570
-------
OES
Weight of Scientific Evidence Conclusion in Release Estimates
include the uncertainties and limitations in the representativeness of the estimates for sites that specifically use 1,4-dioxane
because the default values from the Revised ESD on Hydraulic Fracturing. Another uncertainty is lack of consideration for
release controls. The ESD assumes that all activities occur without any release controls. Actual releases may be less than
estimated if facilities utilize pollution control methods, contributing to uncertainty. Based on this information, EPA has
concluded that the weight of scientific evidence for this assessment is moderate to robust and provides a plausible estimate
of releases in consideration of the strengths and limitations of reasonably available data.
Page 290 of 570
-------
E.5 Air Release Assessment
This section describes EPA's methodology for estimating daily fugitive and stack air emissions from
industrial and commercial facilities manufacturing, processing, or using 1,4-dioxane. Facilities report air
emissions to the TRI. EPA used 2019 TRI (I v « « A. 2022h) data to estimate daily air emissions for
each OES where available; however, EPA did not have these data for every OES. For OES without TRI
data, EPA used alternate assessment approaches to estimate air emissions. These approaches are
described below.
E.5.1 Assessment Using TRI
EPA found 2019 TRI data for facilities within the following OESs:
• Manufacturing;
• Import and repackaging;
• Industrial uses;
• Functional fluids (open-system);
• Disposal;
• PET byproduct; and
• Ethoxylation byproduct.
The 2019 TRI data were rated medium in EPA's systematic review process. EPA estimated daily air
emissions using TRI data for these OESs, with the following general steps as described in the rest of this
section.
1. Collect air emission data from 2019 TRI data,
2. Map air emission data to occupational exposure scenarios,
3. Estimate the number of facility operating days per year, and
4. Estimate daily air emissions and prepare a summary of the air emissions for each OES.
Note that EPA compared the TRI data used to estimate air releases for the PET byproduct OES in this
risk evaluation to information from a life cycle analysis on the PET manufacturing process in Appendix
E.6.
Step 1: Collect Air Emission Data TRI
The first step in the methodology for estimating air emissions was to obtain 2019 TRI data for the
chemical from EPA's Basic Plus Data Files. TRI requires U.S. facilities in various industry sectors to
report the annual release volumes to the environment through air emissions, water discharges, and land,
and/or managed through recycling, energy recovery, and treatment, including by off-site transfers. TRI
reporters may report either with a Form R or a Form A. Facilities must report with a Form R, which
requires reporting of release quantities and uses/sub-uses of the chemical, among other information,
unless they meet the alternate threshold requirements for submitting a Form A. Specifically, facilities
may submit a Form A if the volume of chemical manufactured, processed, or otherwise used does not
exceed 1,000,000 lb per year (lb/year) and the total annual reportable releases do not exceed 500 lb/year.
Facilities do not need to report release quantities or uses/sub-uses on Form A. EPA included both TRI
reporting Form R and TRI reporting Form A submissions in the air release assessment.
Air emissions in TRI are reported separately for stack air and fugitive air and always occur on-site at the
facility. Where sites reported to 2019 TRI with Form A, EPA used the Form A threshold for total
releases of 500 lb/year. EPA used the entire 500 lb/year for both the fugitive and stack air release
estimates; however, since this threshold is for total site releases, these 500 lb/year are attributed either to
Page 291 of 570
-------
fugitive air or stack air for this analysis—not both (to avoid double counting the releases and exceeding
the total release threshold for Form A).
Step 2: Map Air Emission Data to Occupational Exposure Scenarios
In the next step of air release assessment, EPA mapped the chemical's 2019 TRI data to the 1,4-dioxane
OES. EPA used the following procedure to map 2019 TRI data to OES:
1. Compile TRI uses/sub-uses: EPA first compiled all the reported TRI uses/sub-uses for each
facility into one column.
2. Map TRI uses/sub-uses to Chemical Data Reporting (CDR) IFC codes: EPA then mapped the
TRI uses/sub-uses for each facility to one or more 2016 CDR Industrial Function Category (IFC)
codes using the TRI-to-CDR Use Mapping crosswalk (see Appendix E.9).
3. Map OES to CDR IFC codes: EPA then mapped each COU/OES combination to a 2016 CDR
IFC code and description. The basis for this mapping was generally the COU category and
subcategory.
4. Map TRIfacilities to an OES: Using the CDR IFC codes from Step 2 and the COU-CDR
Mapping from Step 3, EPA mapped each TRI facility to an OES. EPA's rationale for the OES
determination is generally described below.
o In some cases, the facility mapped to only one OES and the mapping appeared to be
correct given the facility name and NAICS code. For these, the OES as mapped from
Steps 2 and 3 was used without adjustment,
o There were instances where a facility mapped to multiple OESs which required some
engineering judgement to identify a primary OES. EPA documented the rationale for
these determinations for each facility in 1,4-Dioxane Supplemental Information File:
Environmental Releases to Air ( 1024k). In summary, these determinations
were made with the following considerations:
• Industry and NAICS codes reported in TRI (e.g., for a facility that reported TRI
uses for both waste treatment and ancillary use, EPA assigned the Disposal OES
if the NAICS code was 562211, Hazardous Waste Treatment and Disposal);
• Internet research of the types of products manufactured at the facility (e.g., if a
facility's website indicates the facility manufactures PET, the facility is likely to
produce 1,4-dioxane as a byproduct in PET manufacturing); and
• Grouping of similar OES (e.g., for facilities that reported the sub-use of chemical
processing aid, process solvent, or processing as a reactant), EPA assigned the
Industrial uses OES because this includes multiple processes as described in the
2020 RE ( 1020c)).
o In some cases, EPA identified that there were instances where the preliminary OES
mapping from the TRI use/sub-use - CDR IFC code required re-mapping. This re-
mapping is a result of limitations of the TRI-to-CDR Use Mapping crosswalk. For
example, the crosswalk maps the TRI use/sub-use of "Otherwise Use as Manufacturing
Aid (Other)" to only CDR IFC codes U013 (closed-system functional fluids) and U023
(plating agents and surface treating agents); however, this TRI use/sub-use may
encompass multiple other uses that are not captured in these CDR IFC codes. In these
cases, EPA reviewed the reported NAICS codes and conducted internet research on the
types of products manufactured at the facility to determine the likely OES.
o Additionally, EPA reviewed 2016 CDR ( ) for sites that reported
manufacturing (including importing) of 1,4-dioxane. If the sites that reported to 2016
CDR also reported in 2019 TRI, EPA assigned the OES according to 2016 CDR. Note
that some sites that reported to 2019 TRI may not be in 2016 CDR (e.g., sites that
Page 292 of 570
-------
manufacture the chemical as a byproduct). In these cases, EPA determined the OES using
only the above bulleted steps.
5. Form A's: For Form A submissions, there were no reported TRI uses/sub-uses. To determine the
COU for these facilities, EPA used 2016 CDR as described above, the NAICS codes, and
internet searches to determine the type of products and operations at the facility.
The specific rationale for the OES mapping for each facility is described in 1,4-Dioxane Supplemental
Information File: Environmental Releases to Air (U.S. EPA. 2024k).
Step 3: Estimate the Number of Facility Operating Days per Year
EPA then estimated the number of operating days (days/year) for each facility reporting air emissions to
TRI. For the OES that were included in the Final Risk Evaluation for 1,4-Dioxane ( 2020c),
EPA used the number of operating days from that risk evaluation. For the additional OES included in
this supplemental risk evaluation, EPA estimated the number of operating days using the methodology
described in Appendix E.2.
Step 4: Estimate Daily Air Emissions and Summarize Air Emissions for each OES
The final step was to prepare a summary of the fugitive and stack releases. For each OES and facility
mapped to that OES, EPA summarized the annual fugitive and stack air emissions reported in 2019 TRI
and daily fugitive and stack air emissions that EPA estimated by dividing the annual emissions by the
number of operating days determined for the OES in Step 3. EPA also summarized site information,
including site identity, city, state, zip code, TRI facility ID, and Facility Registry Service (FRS) ID
because the subsequent exposure modeling is site and location specific. Latitude and longitude
coordinates are included in 1,4-Dioxane Supplemental Information File: Environmental Releases to Air
(I v i i \ 2024k) but not in the summary tables.
To accompany the summary table for each OES, EPA also provided any reasonably available
information on the release duration and pattern, which are needed for the exposure modeling. Release
duration is the expected amount of time per day during which the air releases may occur. Release pattern
is the temporal variation of the air release, such as over consecutive days throughout the year, over
cycles that occur intermittently throughout the year, or in a puff/instantaneous release that occurs over a
short duration. The TRI dataset does not include release pattern or duration; therefore, EPA used
information from models or literature, where available. For release pattern, EPA provided the number of
release days with the associated basis as described in Step 3. However, for most OES, no information
was found on release duration and pattern. In such cases, EPA listed the release duration and pattern as
"unknown."
EPA's summary of air releases for each OES is included in 1,4-Dioxane Supplemental Information File:
Environmental Releases to Air ( 024k).
E.5.2 Assessment for OESs Without TRI
EPA did not find 2019 TRI data for the following OESs:
• Functional fluids (closed-systems);
• Laboratory chemicals;
• Film cement;
• Spray polyurethane foam;
• 3D printing;
• Dry film lubricant;
• Textile dye;
Page 293 of 570
-------
• Antifreeze;
• Surface cleaner;
• Dish soap;
• Dishwasher detergent;
• Laundry detergent;
• Paints and floor lacquer; and
• Hydraulic fracturing.
For these OESs, EPA estimated air emissions by using various modeling approaches, including the use
of surrogate TRI data and data from literature, GSs, and ESDs. EPA's assessment of air emissions for
each of these OESs is described below.
Functional Fluids (Closed-Systems)
Air emission data were not available for this OES and EPA did not find any information to model air
emissions for this OES using literature, GSs, or ESDs. EPA expects that the sources of release for this
OES to be similar to those for the Industrial uses OES, based on the process information in the Final
Risk Evaluation for 1,4-Dioxane (U. 2020c). Therefore, EPA grouped the air release assessment
for Functional Fluids (Closed-Systems) into the OES for Industrial uses. However, there is uncertainty
in the assumption of similar release sources between these OESs.
Laboratory Chemicals
EPA estimated air emissions for facilities within the Laboratory chemicals OES using the Draft GS on
Use of Laboratory Chemicals ( 221). The Draft GS on Use of Laboratory Chemicals was
rated high during EPA's systematic review process.
The GS indicates that there are eight release points:
1. Release to air from transferring volatile chemicals from transport containers.
2. Release to air, water, incineration, or landfill from transferring solid powders.
3. Release to water, incineration, or land from cleaning or disposal of transport containers.
4. Release to air from cleaning containers used for volatile chemicals.
5. Labware equipment cleaning residuals released to water, incineration, or landfill.
6. Release to air during labware equipment cleaning for volatile chemicals.
7. Release to air from laboratory analyses for volatile chemicals.
8. Release to water, incineration, or landfill from laboratory waste disposal.
Based on the GS, release points #1, 2, 4, 6, and 7 have the potential for air emissions; however, release
point #2 is not applicable because 1,4-dioxane is not a solid powder. To estimate the remaining air
releases, EPA used the equations specified in the Draft GS (U.S. EPA. 20221). EPA did not find any data
specific to 1,4-dioxane in this OES. Therefore, the calculation of releases using this GS are for a
"generic site," using the default input parameter values from the GS.
Using this methodology, EPA calculated the "typical" and "worst-case" air emissions for this OES.
These characterizations are based on the GS, which provides default "typical" and "worst-case" input
parameters for the release calculations. EPA's calculation of air emissions for this OES, including all
calculation inputs, can be found in 1,4-Dioxane Supplemental Information File: Environmental Releases
to Air ( As_2024k).
Page 294 of 570
-------
Film Cement
EPA estimated air emissions for facilities within the Film cement OES using process information from
the Final Risk Evaluation for 1,4-Dioxane ( E020c). The underlying process information for
this assessment was rated high during EPA's systematic review process.
The process of using film cement involves applying the cement onto edges of photographic films by
hand using a small brush, then joining the pieces of film by pressing and heating to dry the cement.
Based on this process information, EPA assumes that the 1,4-dioxane within film cement is volatilized
to air during the drying process and that 1,4-dioxane residual within empty film cement bottles may also
be volatilized to air. EPA estimated these air releases for this OES as a range, using a film cement use
rate of 2.5 to 12 L/site-year and a concentration of 1,4-dioxane in the film cement of 45 to 50 percent,
from the process information in the Final Risk Evaluation for 1,4-Dioxane ( '020c). These
releases may be to fugitive air or stack air depending on site-specific engineering controls.
EPA's calculation of air emissions for this OES, including all calculation inputs and assumptions, can be
found in 1,4-Dioxane Supplemental Information File: Environmental Releases to Air (
2024k).
Spray Polyurethane Foam
EPA estimated air emissions for facilities within the spray polyurethane foam OES using the GS on
Application of Spray Polyurethane Foam Insulation (U.S. EPA. 2020c). The GS on the Application of
Spray Polyurethane Foam Insulation was rated medium during EPA's systematic review process.
The GS indicates that there are six release points:
1. Releases to fugitive air for volatile chemicals during unloading of raw materials from transport
containers.
2. Releases to water, incineration, or landfill from cleaning or disposal of transport containers.
3. Releases to fugitive air for volatile chemicals during transport container cleaning.
4. Releases to incineration or landfill from spray polyurethane foam application equipment
cleaning.
5. Releases to fugitive air for volatile chemicals during equipment cleaning.
6. Releases to landfill of scrap foam from trimming applied foam.
Based on the GS, release points #1, 3, and 5 have the potential for air emissions. To estimate these
releases, EPA used the equations specified in the GS ( :020c). Apart from weight fraction in
spray polyurethan foam, EPA did not find any data specific to 1,4-dioxane in this OES. Therefore, the
calculation of releases using this GS are for a "generic site," using the default input parameter values
from the GS. Specifically, EPA used the input parameter values that were presented in the original risk
evaluation for estimates of occupational exposure; see Appendix G of the Final Risk Evaluation for 1,4-
Dioxane (U.S. EPA. 2020c).
Using this methodology, EPA calculated the "typical" and "worst-case" air emissions for this OES.
These characterizations are based on the GS, which provides default "typical" and "worst-case" input
parameters for the release calculations. EPA's calculation of air emissions for this OES, including all
calculation inputs and assumptions, can be found in 1,4-Dioxane Supplemental Information File:
Environmental Releases to Air (\ c< < i1 \ J024k).
Page 295 of 570
-------
3D Printing
Air emission data were not available for this OES and EPA did not find any information to model air
emissions for this OES using literature, GSs, or ESDs. EPA expects that industrial applications of this
OES to be accounted for in the Industrial uses TRI data. Per the December 2020 Final Risk Evaluation
for 1,4-Dioxane (U.S. EPA. 2020c). 3D printing ink containing 1,4-dioxane is used in research labs to
print biomedical products. Because the 2019 TRI data for the Industrial uses OES include medicinal and
pharmaceutical manufacturing NAICS codes, medical research labs that conduct 3D printing with 1,4-
dioxane inks may be captured in that OES. Therefore, EPA grouped the air release assessment for 3D
Printing into that for Industrial uses. However, there is uncertainty in whether 3D printing sites are truly
captured in the Industrial uses TRI data.
Dry Film Lubricant
EPA estimated air emissions for facilities within the Dry film lubricant OES using process information
from the Final Risk Evaluation for 1,4-Dioxane ( )20c). The underlying process information
for this assessment was rated high during EPA's systematic review process.
The process for the production and use of dry film lubricant is described in the 2020 RE and is based on
information provided to EPA by the one known user. In summary, the process entails producing the
concentrated dry film lubricant by mixing 1,4-dioxane with other additives, followed by the dilution of
the concentrated dry film lubricant with additional 1,4-dioxane and finally the use of the dry film
lubricant. The use involves spray application onto substrates in a vented paint booth and the subsequent
curing in a vented oven and cleaning of the dried parts in a 1,4-dioxane bath ( s20c). Based
on this process description, EPA assumes that 100 percent of the 1,4-dioxane in the applied dry film
lubricant is released to stack air from the paint booth and the oven. EPA estimated this release quantity
using batch parameters from the process description, including 5 percent 1,4-dioxane in the dry film
lubricant, 48 dry film lubricant applications per year, 0.5-pints of concentrated dry film lubricant, and
1.5-pints of pure 1,4-dioxane per application (U.S. EPA. 2020c).
EPA's calculation of air emissions for this OES, including all calculation inputs and assumptions, can be
found in 1,4-Dioxane Supplemental Information File: Environmental Releases to Air (U.S. EPA.
2024k).
Textile Dye
EPA used the OECD ESD on Textile Dyes (OE ) to estimate water and land releases for this
OES; however, this ESD does not include approaches for estimating air releases. EPA did not find any
other GS/ESD, literature sources, or process information to model air releases for this OES. In addition,
EPA does not expect this OES to be sufficiently similar to other OES such that surrogate TRI data can
be used to estimate air emissions for this OES. Therefore, EPA was not able to estimate air releases for
these OESs.
Antifreeze
EPA did not find any directly applicable GS/ESD or literature sources for this OES; however, EPA
evaluated the potential for releases using the OECD ESD on Chemical Additives used in Automotive
Lubricants (OE 20) and the EPA MRD on Commercial Use of Automotive Detailing Products
( 'A. 2022b). The ESD and MRD were both rated high during EPA's systematic review process.
For the use of antifreeze, EPA expects releases may occur from volatilizations of 1,4-dioxane during
unloading/pouring antifreeze into vehicles, disposal or cleaning of empty antifreeze containers, and
disposal of spent antifreeze. Both the ESD and MRD indicate that containers of automotive maintenance
Page 296 of 570
-------
fluids are typically small and are not rinsed, but rather disposed of as solid waste ( 322b;
OECD. 2020). Additionally, the ESD on Chemical Additives used in Automotive Lubricants indicates
that spent lubricants are disposed of via incineration, which EPA expects is similarly done for spent
antifreeze (OECD. 2020). Therefore, EPA expects the main source of air emissions to be from
volatilizations of 1,4-dioxane during unloading/ pouring antifreeze into vehicles. EPA estimated this
release using the EPA/OAQPS AP-42 Loading Model and batch parameters from the ESD, MRD, and
other sources.
Specifically, EPA used the consumer use rate of antifreeze (0.15 kg antifreeze/job) from the Final Risk
Evaluation for 1,4-Dioxane ( 1020c) and scaled this value up to a commercial use rate based
on a range of the number of cars serviced at auto shops from the Near-Field/Far-Field Brake Model and
Automotive Detailing MRD (1 to 9 jobs/day). EPA used a range of concentrations of 1,4-dioxane in
antifreeze from the process description in Appendix F.4.2 and assumed antifreeze container sizes
ranging from 16 ounces to 5 gallons per the default container sizes in the MRD and ESD, respectively
0 C" < \ 202 J., ( n :020). Using these batch parameters and the default parameters for the
EPA/OAQPS AP-42 Loading Model, EPA estimated low-end and high-end air emissions. EPA expects
these air emissions to be to fugitive air based on the use setting (e.g., outdoors, maintenance garages).
EPA's calculation of air emissions for this OES, including all calculation inputs and assumptions, can be
found in 1,4-Dioxane Supplemental Information File: Environmental Releases to Air (
2024k).
Surface Cleaner
EPA did not find any directly applicable GS/ESD or literature sources for this OES; however, EPA
estimated air releases using the SHEDs-HT modeling conducted for the one case study location
(Liverpool OH) and the assumptions described herein. EPA expects that the main release points from the
use of surface cleaners are from
1. Disposal of empty containers containing residual cleaning solution,
2. Application of the cleaning solution, and
3. Disposal of cleaning solution by rinsing or wiping.
Because EPA did not find any directly applicable GSs or ESDs, EPA used the Draft GS on Furnishing
Cleaning ( 022a) to inform these releases due to the similarities in surface cleaning and
furnishing cleaning. The Draft GS on Furnishing Cleaning was rated high during EPA's systematic
review process. Per this Draft GS, empty containers may be rinsed out in sinks or disposed of without
rinsing, such that releases may be to wastewater or landfill; the GS uses the EPA/OPPT Small Container
Residual Model to estimate this release. Application losses are to fugitive air from spray application; the
GS uses literature data to estimate this release. Once applied, the cleaner may be rinsed off or wiped off
with rags or towels, such that releases may be to wastewater or landfill; the GS assumes 100 percent
release scenario, estimating this release by subtracting the upstream losses from the cleaner use rate
( 2022a).
The SHEDs-HT modeling estimated wastewater discharges of 72 g of 1,4-dioxane per day for
commercial uses of surface cleaners containing 1,4-dioxane in Liverpool OH. EPA used this quantity
and the above release information and models from the Draft GS on Furnishing Cleaning to back-
calculate a 1,4-dioxane use rate. EPA then applied the loss fraction to fugitive air from release point #2
to estimate air releases for this OES. EPA's calculation of air releases for this OES, including all
calculation inputs and assumptions, can be found in 1,4-Dioxane Supplemental Information File:
Environmental Releases to Air Q x \ \\ ,024k).
Page 297 of 570
-------
Dish Soap
EPA estimated air emissions for facilities within the Dish soap OES using data from a public comment,
EPA/OPPT models, and Monte Carlo modeling. The public comment was rated high during EPA's
systematic review process (P&G. 2023). The use of Monte Carlo modeling allows for variation of
calculation input parameters such that a distribution of environmental releases can be calculated, from
which EPA can estimate the 50th and 95th percentile releases. An explanation of this modeling approach
is included in Appendix E. 14.
Dishwasher Detergent
EPA estimated air emissions for facilities within the Dishwasher detergent OES using data from a public
comment, EPA/OPPT models, and Monte Carlo modeling. The public comment was rated high during
EPA's systematic review process (P&G. 2023). The use of Monte Carlo modeling allows for variation
of calculation input parameters such that a distribution of environmental releases can be calculated, from
which EPA can estimate the 50th and 95th percentile releases. An explanation of this modeling approach
is included in Appendix E. 14.
Laundry Detergent
EPA estimated air emissions for facilities within the Laundry detergent OES using the OECD ESD on
Industrial and Institutional Laundries (01 ) and Monte Carlo modeling. The ESD on
Industrial and Institutional Laundries was rated medium during EPA's systematic review process. The
use of Monte Carlo modeling allows for variation of calculation input parameters such that a distribution
of environmental releases can be calculated, from which EPA can estimate the 50th and 95th percentile
releases. An explanation of this modeling approach is included in Appendix E. 11.16.
Paints and Floor Lacquer
EPA estimated air emissions for facilities within the Paints and floor lacquers OES using the OECD
ESD on Coating Application via Spray-Painting in the Automotive Refinishing Industry (
201 la). The ESD was rated medium during EPA's systematic review process.
As described in the process description in Appendix F.4.7, 1,4-dioxane was identified by a public
comment as present in automotive refinishing products, thereby allowing EPA to identify the above
ESD as the most applicable GS/ESD available. This ESD indicates that releases are expected from
1. Releases to incineration or landfill from container cleaning/disposal,
2. Releases to incineration or landfill from equipment cleaning,
3. Releases to incineration or landfill from over sprayed coating that is captured by emission
controls, and
4. Releases to stack air from over sprayed coating that is not captured by emission controls.
Based on the GS, release point #4 has the potential for air emissions. To estimate this release, EPA used
the equations specified in the ESD (OECD. 201 la). Apart from weight fraction in coatings (see
Appendix F.4.7), EPA did not find any data specific to 1,4-dioxane in this OES. Therefore, the
calculation of releases using this GS are for a "generic site," using the default input parameter values
from the ESD.
Using this methodology, EPA calculated the low-end and high-end air emissions for this OES, which are
expected to be to stack air per the ESD (( ). The low- and high-end estimates are based on
the low- and high-end calculation input parameter defaults from the ESD. EPA's calculation of air
emissions for this OES, including all calculation inputs and assumptions, can be found in 1,4-Dioxane
Supplemental Information File: Environmental Releases to Air ( 324k).
Page 298 of 570
-------
Hydraulic Fracturing
EPA estimated air emissions for facilities within the Hydraulic fracturing OES using the Draft OECD
ESD on Hydraulic Fracturing (U.S. EPA. 2022e) and Monte Carlo modeling. The Revised ESD on
Hydraulic Fracturing was rated high during EPA's systematic review process. The use of Monte Carlo
modeling allows for variation of calculation input parameters such that a distribution of environmental
releases can be calculated, from which EPA can estimate the 50th and 95th percentile releases. An
explanation of this modeling approach is included in Appendix 0.
E.5.3 Air Release Estimates Summary
A summary of industrial and commercial air releases estimated using the above methods is presented in
Table Apx E-7 below. Specifically, this table presents the range of daily air releases per site for each
OES.
Page 299 of 570
-------
Table Apx E-7. Summary of Daily Industrial and Commercial Air Release Estimates for 1,4-Dioxane
Number of
Range of Estimated Daily
Release Per Site
Estimated
Release
Overall Data
OES
Type of Air Release
Facilities with
Releases"
(kg/site-day)
Frequency
Range
(days)''
Quality
Determination
Sources'
Min
Max
Manufacturing
Fugitive Air
1
2.62
250
Medium
TRI
Stack Air
1
0.0018
Medium
TRI
Import and
Fugitive Air
1
0
250
Medium
TRI
repackaging
Stack Air
1
0.091
Medium
TRI
Industrial uses
Fugitive Air
12
0
0.91
250
Medium
TRI
Stack Air
12
0
8.14
Medium
TRI
Functional fluids
Fugitive Air
2
0
0.009
247
Medium
TRI
(open-system)
Stack Air
2
0.19
1.38
Medium
TRI
Functional fluids
All
Assessed as a part of Industrial uses OES
N/A
N/A
(closed-system)
Laboratory
chemical
Fugitive Air or Stack
Air (Unknown)
132
0.11 (typical)
0.41 (worst-
case)
250
High
GSd
Film cement
Fugitive Air or Stack
211
0.0046
0.025
250
High
Process
Air (Unknown)
information®
Fugitive Air
1,553,559
0.0024
0.012 (worst-
Medium
GS^
Spray foam
(typical)
case)
3
application
Stack Air
1,553,559
0 (all air releases assessed to
Medium
GS^
fugitive)
Printing inks (3D)
Fugitive Air, Stack
Assessed as a part of Industrial uses OES
N/A
N/A
Air, and Land (all
250
types)
Dry film lubricant
Fugitive Air
8
0 (no fugitive releases per
process information)
48
High
Process
information®
Stack Air
8
0.75 (single value estimated
from process information)
High
Process
information®
Page 300 of 570
-------
OES
Type of Air Release
Number of
Facilities with
Releases"
Range of Estimated Daily
Release Per Site
(kg/site-day)
Estimated
Release
Frequency
Range
(days)''
Overall Data
Quality
Determination
Sources'
Min
Max
Disposal
Fugitive Air
15
0
0.91
250
Medium
TRI
Stack Air
15
0
0.91
Medium
TRI
Textile dye
Fugitive Air and
Stack Air
Not assessed
N/A
N/A
N/A
Antifreeze
Fugitive Air and
Stack Air
84,383
7.26E-16
1.80E-07
250
High
Process
information® and
Modeling''
Surface cleaner
Fugitive Air
Unknown
0.0071
(typical; daily
release value
for all sites
combined in
Liverpool
OH, case
study)
0.013 (worst
case; daily
release value
for all sites
combined in
Liverpool, OH,
case study)
250
High
SHEDS-HT,'
Process
information®
Modeling''
Dish soap (draft RE
estimates)™
Fugitive air and
stack air
Not assessed
250
N/A
N/A
Dish soap (updated
estimates)™
Fugitive air
773,851
8.8E-12
3.9E-07
350
High
(P&G. 2023)
and Monte Carlo
Modeling''
POTW or fugitive
air (unknown
partitioning)
773,851
8.4E-08
1.5E-03
High
(P&G. 2023)
and Monte Carlo
Modeling''
Page 301 of 570
-------
OES
Type of Air Release
Number of
Facilities with
Releases"
Range of Estimated Daily
Release Per Site
(kg/site-day)
Estimated
Release
Frequency
Range
(days)''
Overall Data
Quality
Determination
Sources'
Min
Max
Dishwasher
detergent (draft RE
estimates)™
Fugitive air and
stack air
Not assessed
250
N/A
N/A
Dishwasher
detergent (updated
estimates)™
Fugitive air
773,851
1.3E-10
9.3E-08
350
High
(P&G. 2023)
and Monte Carlo
Modeling**
POTW or fugitive
air (unknown
partitioning)
773,851
1.2E-06
3.7E-04
High
(P&G. 2023)
and Monte Carlo
Modeling**
Fugitive air
95,533
1.83E-10
6.52E-07
Medium
ESD' and Monte
Carlo Modeling**
Laundry detergent
(institutional) -
liquid detergents
(draft RE
estimates)™
Fugitive air, stack
air, or POTW
(unknown
partitioning)
95,533
1.5 IE—10
0.00714
250 to 365
Medium
ESD' and Monte
Carlo Modeling**
Land (unknown
landfill),
incineration, or
POTW (unknown
partitioning)
95,533
4.05E-12
3.95E-05
Medium
ESD' and Monte
Carlo Modeling**
Page 302 of 570
-------
OES
Type of Air Release
Number of
Facilities with
Releases"
Range of Estimated Daily
Release Per Site
(kg/site-day)
Estimated
Release
Frequency
Range
(days)''
Overall Data
Quality
Determination
Sources'
Min
Max
Fugitive air
95,533
9.1E-10
2.5E-05
Medium
ESD' and Monte
Carlo Modeling''
Laundry detergent
(institutional) -
liquid detergents
Fugitive air, stack
air, or POTW
(unknown
partitioning)
95,533
3.0E-13
6.6E-02
250 to 365
Medium
ESD' and Monte
Carlo Modeling''
(updated
estimates)™
Land (unknown
landfill),
incineration, or
POTW (unknown
partitioning)
95,533
8.1E-13
3.8E-04
Medium
ESD' and Monte
Carlo Modeling''
Fugitive air
95,533
3.42E-12
2.77E-07
Medium
ESD' and Monte
Carlo Modeling''
Stack air
95,533
1.40E-11
3.75E-06
Medium
ESD' and Monte
Carlo Modeling''
Laundry detergent
(institutional) -
powder detergents
(draft RE
estimates)™
Fugitive air, stack
air, or POTW
(unknown
partitioning)
95,533
3.05E-08
2.10E-04
250 to 365
Medium
ESD' and Monte
Carlo Modeling''
Land (unknown
landfill),
incineration, or
POTW (unknown
partitioning)
95,533
5.36E-08
0.0018
Medium
ESD' and Monte
Carlo Modeling''
Page 303 of 570
-------
OES
Type of Air Release
Number of
Facilities with
Releases"
Range of Estimated Daily
Release Per Site
(kg/site-day)
Estimated
Release
Frequency
Range
(days)''
Overall Data
Quality
Determination
Sources'
Min
Max
Laundry detergent
(institutional) -
powder detergents
(updated
estimates)™
Fugitive air
95,533
7.8E-10
2.2E-05
250 to 365
Medium
ESD' and Monte
Carlo Modeling''
Stack air
95,533
3.4E-12
4.1E-05
Medium
ESD' and Monte
Carlo Modeling''
Fugitive air, stack
air, or POTW
(unknown
partitioning)
95,533
2.1E-08
1.9E-03
Medium
ESD' and Monte
Carlo Modeling''
Land (unknown
landfill),
incineration, or
POTW (unknown
partitioning)
95,533
1.4E-08
1.8E-02
Medium
ESD' and Monte
Carlo Modeling''
Laundry detergent
(industrial) - liquid
detergents (draft
RE estimates)™
Fugitive air
2,453
6.25E-10
1.93E-06
20 to 365
Medium
ESD' and Monte
Carlo Modeling''
Fugitive air, stack
air, or POTW
(unknown
partitioning)
2,453
5.48E-12
0.011
Medium
ESD' and Monte
Carlo Modeling''
Land (unknown
landfill),
incineration, or
POTW (unknown
partitioning)
2,453
4.78E-12
1.46E-04
Medium
ESD' and Monte
Carlo Modeling''
Page 304 of 570
-------
OES
Type of Air Release
Number of
Facilities with
Releases"
Range of Estimated Daily
Release Per Site
(kg/site-day)
Estimated
Release
Frequency
Range
(days)''
Overall Data
Quality
Determination
Sources'
Min
Max
Laundry detergent
(industrial) - liquid
detergents (updated
estimates)™
Fugitive air
2,453
1.2E-09
3.7E-05
20 to 365
Medium
ESD' and Monte
Carlo Modeling''
Fugitive air, stack
air, or POTW
(unknown
partitioning)
2,453
3.1E-11
0.11
Medium
ESD' and Monte
Carlo Modeling''
Land (unknown
landfill),
incineration, or
POTW (unknown
partitioning)
2,453
6.6E-13
1.4E-03
Medium
ESD' and Monte
Carlo Modeling''
Laundry detergent
(industrial) -
powder detergents
(draft RE
estimates)™
Fugitive air
2,453
3.13E-13
1.47E-05
20 to 365
Medium
ESD' and Monte
Carlo Modeling''
Stack air
2,453
1.68E-12
1.82E-04
Medium
ESD' and Monte
Carlo Modeling''
Fugitive air, stack
air, or POTW
(unknown
partitioning)
2,453
1.76E-09
0.0112
Medium
ESD' and Monte
Carlo Modeling''
Land (unknown
landfill),
incineration, or
POTW (unknown
partitioning)
2,453
2.92E-11
3.92E-04
Medium
ESD' and Monte
Carlo Modeling''
Page 305 of 570
-------
OES
Type of Air Release
Number of
Facilities with
Releases"
Range of Estimated Daily
Release Per Site
(kg/site-day)
Estimated
Release
Frequency
Range
(days)''
Overall Data
Quality
Determination
Sources'
Min
Max
Laundry detergent
(industrial) -
powder detergents
(updated
estimates)™
Fugitive air
2,453
1.1E-09
1.6E-04
20 to 365
Medium
ESD' and Monte
Carlo Modeling''
Stack air
2,453
7.7E-14
2.6E-03
Medium
ESD' and Monte
Carlo Modeling''
Fugitive air, stack
air, or POTW
(unknown
partitioning)
2,453
1.8E-11
0.10
Medium
ESD' and Monte
Carlo Modeling''
Land (unknown
landfill),
incineration, or
POTW (unknown
partitioning)
2,453
1.5E-11
3.8E-03
Medium
ESD' and Monte
Carlo Modeling''
Paints and floor
lacquer
Stack air
33,648
4.68E-10
1.60E-06
250
Medium
ESD'
Polyethylene
terephthalate (PET)
byproduct
Fugitive Air
13
0
1.57
250
Medium
TRI
Stack Air
13
0.0049
13.8
Medium
TRI
Ethoxylation
process byproduct
Fugitive Air
8
0
7.4
250
Medium
TRI
Stack Air
8
0
32
Medium
TRI
Hydraulic
fracturing (draft RE
estimates)™
Fugitive air
411
1.99E-07
5482
1 to 72
Medium
ESD'' and Monte
Carlo Modeling''
Stack air
411
0 (all air releases assessed to
fugitive)
Medium
ESD'' and Monte
Carlo Modeling''
Surface water,
incineration, or
landfill (unknown
partitioning)
411
3.61E-10
4.59
Medium
ESD'' and Monte
Carlo Modeling''
Page 306 of 570
-------
OES
Type of Air Release
Number of
Facilities with
Releases"
Range of Estimated Daily
Release Per Site
(kg/site-day)
Estimated
Release
Frequency
Range
(days)''
Overall Data
Quality
Determination
Sources'
Min
Max
Fugitive air
411
3.2E-12
1.3E-02
Medium
ESD' and Monte
Carlo Modeling''
Stack air
411
0 (all air releases assessed to
fugitive)
Medium
ESD" and Monte
Carlo Modeling''
Hydraulic
fracturing (updated
estimates)™
Surface water,
incineration, or
landfill (unknown
partitioning)
411
4.3E-10
5.6
1 to 72
Medium
ESD'' and Monte
Carlo Modeling''
Surface water
(13%), Land (soil)
(64%), and Landfill
or Incineration
(23%)
411
4.9E-11
0.64
Medium
ESD'' and Monte
Carlo Modeling''
a See Appendix E.l for explanation of how EPA determined the number of sites for each OES.
* Where available. EPA used the December 2020 Final Risk Evaluation for 1,4-Dioxane (U.S. EPA, 2020c). generic scenarios, and emission scenario documents
to provide a basis to estimate the number of release days of 1,4-dioxane within a COU.
c Narrative descriptions of all release estimate sources are provided in Appendix E.5.
d The generic scenario used for this COU is the GS on Use of Laboratory Chemicals (U.S. EPA. 2022i).
e For this COU, EPA used process information, which is further described in Appendix E.5.2.
' The generic scenario used for this COU is the GS on Application of Spray PoK urethane Foam Insulation (U.S. EPA, 2018b).
g The emission scenario document used for this COU is the ESD on Textile Dyes (OECD, 2017).
h For this COU, EPA used various models and literature for model input parameters as described in Appendix E.5.2.
1 The emission scenario document used for this COU is the ESD on Industrial and Institutional Laundries (OECD, 201 lb).
1 The emission scenario document used for this COU is the ESD on Coating Application via Sprav Painting in the Automotive Refinishing Industry (OECD,
2011a).
k The emission scenario document used for this COU is the Revised ESD on Hvdraulic Fracturing (U.S. EPA. 2022e).
' EPA used the down the drain water release estimates from the SHEDs-HT model for the Liverpool OH case study (see Section 2.1.1.2) to estimate air releases
by back calculating 1,4-dioxane use rates and applying loss fractions for air releases using literature and standard models described in Appendix E.5.2.
For select OESs, updates to release estimates were made via information provided by the SACC and public comments.
Page 307 of 570
-------
E.5.4 Summary of Weight of Scientific Evidence Conclusions in Air Release Estimates
Table Apx E-8 provides a summary of EPA's weight of scientific evidence conclusions in its air release estimates for each of the
Occupational Exposure Scenarios assessed. Detailed descriptions of non-OES specific strengths, limitations, assumptions, and uncertainties
(e.g., general limitations for TRI, DMR, etc.) are provided in Appendix E.6.
Table Apx E-8 Summary of Weight of Scientific Evidence Conclusions in Air Release Estimates by PES
OES
Weight of Scientific Evidence Conclusion in Release Estimates
Manufacturing
Air releases are assessed using reported releases from 2019 TRI. Factors that increase the strength of evidence for this OES
are that the release data are directly relevant to the OES (as opposed to surrogate), that TRI has a medium overall data
quality determination, and consistency within the dataset (all reporters use the same or similar reporting forms). A strength
of TRI data is that TRI compiles the best reasonably available release data for all reporting facilities. Factors that decrease
the strength of the evidence for this OES include the low number of data points, lack of variability (only 1 year of data
used), uncertainty in the accuracy of reported releases, and the limitations in representativeness to all sites because TRI
may not capture all relevant sites. Additionally, EPA made assumptions on the number of operating days to estimate daily
releases. Based on this information, EPA has concluded that the weight of scientific evidence for this assessment is
moderate to robust and provides a plausible estimate of releases in consideration of the strengths and limitations of
reasonably available data.
Import and repackaging
Air releases are assessed using reported releases from 2019 TRI. Factors that increase the strength of evidence for this OES
are that the release data are directly relevant to the OES (as opposed to surrogate), that TRI has a medium overall data
quality determination, and consistency within the dataset (all reporters use the same or similar reporting forms). A strength
of TRI data is that TRI compiles the best reasonably available release data for all reporting facilities. Factors that decrease
the strength of the evidence for this OES include the low number of data points, lack of variability (only 1 year of data
used), uncertainty in the accuracy of reported releases, uncertainty in EPA's use of Form A submissions, and the
limitations in representativeness to all sites because TRI may not capture all relevant sites. Some facilities within this OES
reported to TRI using a Form A, which does not include any details on chemical release quantities. When a facility has
submitted a Form A, there is no way to discern the quantity released. Therefore, where facilities reported to TRI with a
Form A, EPA used the Form A threshold for total releases of 500 lb/year for each release media; however, there is
uncertainty in this because the actual release quantity is unknown. Furthermore, the threshold represents an upper limit on
total releases from the facility; therefore, assessing releases at the threshold value may overestimate actual releases from
the facility. Additionally, EPA made assumptions on the number of operating days to estimate daily releases. Based on this
information, EPA has concluded that the weight of scientific evidence for this assessment is moderate to robust and
provides a plausible estimate of releases in consideration of the strengths and limitations of reasonably available data.
Industrial uses
Air releases are assessed using reported releases from 2019 TRI. Factors that increase the strength of evidence for this OES
are that the release data are directly relevant to the OES (as opposed to surrogate), that TRI has a medium overall data
quality determination, and consistency within the dataset (all reporters use the same or similar reporting forms). A strength
of TRI data is that TRI compiles the best reasonably available release data for all reporting facilities. Factors that decrease
the strength of the evidence for this OES include a lack of variability (only 1 year of data used), uncertainty in the accuracy
Page 308 of 570
-------
OES
Weight of Scientific Evidence Conclusion in Release Estimates
of reported releases, uncertainty in EPA's use of Form A submissions, and the limitations in representativeness to all sites
because TRI may not capture all relevant sites. Some facilities within this OES reported to TRI using a Form A, which
does not include any details on chemical release quantities. When a facility has submitted a Form A, there is no way to
discern the quantity released. Therefore, where facilities reported to TRI with a Form A, EPA used the Form A threshold
for total releases of 500 lb/year for each release media; however, there is uncertainty in this because the actual release
quantity is unknown. Furthermore, the threshold represents an upper limit on total releases from the facility; therefore,
assessing releases at the threshold value may overestimate actual releases from the facility. Based on this information, EPA
has concluded that the weight of scientific evidence for this assessment is moderate to robust and provides a plausible
estimate of releases in consideration of the strengths and limitations of reasonably available data.
Functional fluids (open-
system)
Air releases are assessed using reported releases from 2019 TRI. Factors that increase the strength of evidence for this OES
are that the release data are directly relevant to the OES (as opposed to surrogate), that TRI has a medium overall data
quality determination, and consistency within the dataset (all reporters use the same or similar reporting forms). A strength
of TRI data is that TRI compiles the best reasonably available release data for all reporting facilities. Factors that decrease
the strength of the evidence for this OES include the low number of data points (only two reporting sites), lack of
variability (only 1 year of data used), uncertainty in the accuracy of reported releases, and the limitations in
representativeness to all sites because TRI may not capture all relevant sites. Additionally, EPA made assumptions on the
number of operating days to estimate daily releases. Based on this information, EPA has concluded that the weight of
scientific evidence for this assessment is moderate to robust and provides a plausible estimate of releases in consideration
of the strengths and limitations of reasonably available data.
Functional fluids (closed-
system)
No data was available to estimate releases for this OES. For the air release assessment, EPA grouped this OES with the
Industrial uses OES because the sources of release are expected to be similar between these OESs. Factors that increase the
strength of evidence for this OES are that TRI has a medium overall data quality determination and consistency within the
dataset (all reporters use the same or similar reporting forms). Factors that decrease the strength of evidence for this OES
are that the Industrial Releases OES release data are use as surrogate for this OES, uncertainty in the accuracy of reported
releases, limitations in representativeness to all sites because TRI may not capture all relevant sites, and lack of variability
(only 1 year of data used). Refer to the Industrial uses OES discussion for additional discussion. Based on this information,
EPA has concluded that the weight of scientific evidence for this assessment is slight and provides a plausible estimate of
releases in consideration of the strengths and limitations of reasonably available data.
Laboratory chemicals
Air releases are assessed using the Draft GS on Use of Laboratory Chemicals. Factors that increase the strength of
evidence for this OES are that the release estimates are directly relevant to the OES (as opposed to surrogate), the Draft GS
on Use of Laboratory Chemicals has a high overall data quality determination, and the low level of uncertainty in the data.
Factors that decrease the strength of the evidence for this OES include the that the GS has not been peer-reviewed,
uncertainty in the representativeness of the GS towards all sites in this OES, and a lack of variability in the analysis.
Specifically, because the default values in the ESD are generic, there is uncertainty in the representativeness of generic site
estimates of actual releases from real-world sites that use 1,4-dioxane. Another uncertainty is lack of consideration for
release controls. The ESD assumes that all activities occur without any release controls. Actual releases may be less than
estimated if facilities utilize pollution control methods, contributing to uncertainty. Based on this information, EPA has
Page 309 of 570
-------
OES
Weight of Scientific Evidence Conclusion in Release Estimates
concluded that the weight of scientific evidence for this assessment is slight to moderate and provides a plausible estimate
of releases in consideration of the strengths and limitations of reasonably available data.
Film cement
Air releases are assessed using process information from the Final Risk Evaluation for 1,4-Dioxane. Factors that increase
the strength of evidence for this OES are that the release estimates are directly relevant to the OES (as opposed to
surrogate), the underlying data sources for the process information have a high overall data quality determination, and the
low level of uncertainty in the data because the process information comes directly from actual users of 1,4-dioxane in film
cement. Factors that decrease the strength of the evidence for this OES include uncertainty in the representativeness of
evidence to all sites in this OES and a lack of variability. Specifically, the process information for the production and use
of film cement is based on information from three use sites, one from Australia and two from the U.S. Based on this
information, EPA has concluded that the weight of scientific evidence for this assessment is slight to moderate and
provides a plausible estimate of releases in consideration of the strengths and limitations of reasonably available data.
Spray foam application
Air releases are assessed using the GS on Application of Spray Polyurethane Foam Insulation. Factors that increase the
strength of evidence for this OES are that the release estimates are directly relevant to the OES (as opposed to surrogate),
the underlying data sources for the process information have a medium overall data quality determination, and the low
level of uncertainty in the data. Factors that decrease the strength of the evidence for this OES include uncertainty in the
representativeness of the GS to all sites since it is generic and not specific to sites that use 1,4-dioxane and a lack of
variability. Based on this information, EPA has concluded that the weight of scientific evidence for this assessment is
slight to moderate and provides a plausible estimate of releases in consideration of the strengths and limitations of
reasonably available data.
Printing inks (3D)
No data was available to estimate releases for this OES. For the air release assessment, EPA grouped this OES with the
Industrial uses OES because the sources of release are expected to be similar between these OESs. Factors that increase the
strength of evidence for this OES are that TRI has a medium overall data quality determination and consistency within the
dataset (all reporters use the same or similar reporting forms). Factors that decrease the strength of evidence for this OES
are that the Industrial Releases OES release data are use as surrogate for this OES, uncertainty in the accuracy of reported
releases, limitations in representativeness to all sites because TRI may not capture all relevant sites or smaller commercial
3D printing uses, and lack of variability (only 1 year of data used). Refer to the Industrial uses OES discussion for
additional discussion. Based on this information, EPA has concluded that the weight of scientific evidence for this
assessment is slight and provides a plausible estimate of releases in consideration of the strengths and limitations of
reasonably available data.
Dry film lubricant
Air releases are assessed using process information from the Final Risk Evaluation for 1,4-Dioxane. Factors that increase
the strength of evidence for this OES are that the release estimates are directly relevant to the OES (as opposed to
surrogate), that the underlying data sources for the process information have a high overall data quality determination, and
a low level of uncertainty in the data because the process information comes directly from an actual user of 1,4-dioxane in
dry film lubricants. Factors that decrease the strength of the evidence for this OES include uncertainty in the
representativeness of evidence to all sites and a lack of variability. Based on this information, EPA has concluded that the
weight of scientific evidence for this assessment is slight to moderate and provides a plausible estimate of releases in
consideration of the strengths and limitations of reasonably available data.
Page 310 of 570
-------
OES
Weight of Scientific Evidence Conclusion in Release Estimates
Disposal
Air releases are assessed using reported discharges from 2019 TRI. Factors that increase the strength of evidence for this
OES are that the release data are directly relevant to the OES (as opposed to surrogate), that TRI has a medium overall data
quality determination, high number of data points, and consistency within the dataset (all reporters use the same or similar
reporting forms). A strength of TRI data is that TRI compiles the best reasonably available release data for all reporting
facilities. Factors that decrease the strength of the evidence for this OES include lack of variability (only 1 year of data
used), uncertainty in the accuracy of reported releases, uncertainty in EPA's use of Form A submissions, and the
limitations in representativeness to all sites because TRI may not capture all relevant sites. Some facilities within this OES
reported to TRI using a Form A, which does not include any details on chemical release quantities. When a facility has
submitted a Form A, there is no way to discern the quantity released. Therefore, where facilities reported to TRI with a
Form A, EPA used the Form A threshold for total releases of 500 lb/year for each release media; however, there is
uncertainty in this because the actual release quantity is unknown. Furthermore, the threshold represents an upper limit on
total releases from the facility; therefore, assessing releases at the threshold value may overestimate actual releases from
the facility. Additionally, uncertainty is introduced from EPA's assumptions on the number of operating days to estimate
daily releases. Based on this information, EPA has concluded that the weight of scientific evidence for this assessment is
moderate to robust and provides a plausible estimate of releases in consideration of the strengths and limitations of
reasonably available data.
Textile dye
EPA used the ESD on Textile Dyes to estimate land and water releases; however, this ESD does not include approaches for
estimating air releases. EPA did not find any other GS/ESD, literature sources, or process information to model air releases
for this OES. Furthermore, EPA does not expect this OES to be sufficiently similar to other OES such that surrogate TRI
data can be used to estimate air emissions for this OES. Therefore, EPA was not able to estimate air releases for this OES
and concluded that the weight of scientific evidence is indeterminant.
Antifreeze
Air releases are assessed using the OECD ESD on Chemical Additives used in Automotive Lubricants, the EPA MRD on
Commercial Use of Automotive Detailing Products, and EPA/OPPT models. Factors that increase the strength of evidence
for this OES are that the ESD and MRD used have high overall data quality determinations and consistency within the
sources used. Factors that decrease the strength of the evidence for this OES include that the ESD and MRD are not
directly applicable to antifreeze uses (used as surrogate), uncertainty in the representativeness of the ESD and MRD to all
sites and sites that specifically use 1,4-dioxane since these documents contain generic values, and a lack of variability.
Additionally, EPA scaled up a consumer antifreeze use rate to a commercial use rate based on information in the ESD and
MRD, which increases uncertainty. Based on this information, EPA has concluded that the weight of scientific evidence
for this assessment is slight to moderate and provides a plausible estimate of releases in consideration of the strengths and
limitations of reasonably available data.
Surface cleaner
Air releases are assessed using the SHEDS-HT model and the Draft GS on Furnishing Cleaning. To estimate air releases,
EPA used loss fractions for water releases from the GS and the modeled water release from SHEDS-HT to back-calculate a
1,4-dioxane use rate. EPA then applied loss fractions for air releases from the GS to estimate air releases for this OES.
Factors that increase the strength of evidence for this OES include that the release estimates are directly relevant to the
OES (as opposed to surrogate), that the Draft GS used has a high overall data quality determination, and variability in the
model input parameters. Factors that decrease the strength of the evidence for this OES include uncertainty in the
Page 311 of 570
-------
OES
Weight of Scientific Evidence Conclusion in Release Estimates
representativeness to all sites because the SHEDS-HT estimate is based on one case study for Liverpool, OH and because
the estimate is not site-specific (the release estimate is a total for all sites in Liverpool, OH). Based on this information,
EPA has concluded that the weight of scientific evidence for this assessment is slight and provides a plausible estimate of
releases in consideration of the strengths and limitations of reasonably available data.
Dish soap
Air releases are assessed using Monte Carlo modeling with information from a public comment and standard EPA/OPPT
models. Factors that increase the strength of evidence for this OES are that the release estimates are directly relevant to the
OES (as opposed to surrogate), that the public comment has a hinh overall data aualitv determination (P&G. 2023). there
are a high number of data points (simulation runs), and full distributions of input parameters. Monte Carlo modeling
accounts for the entire distribution of input parameters, calculating a distribution of potential release values that represents
a larger proportion of sites than a discrete value. The major factor that decreases the strength of the evidence for this OES
include the uncertainties and limitations in the representativeness of the data from the public comment towards all sites that
use dish soaps containing 1,4-dioxane. Another uncertainty is the lack of a GS or ESD describing this scenario; EPA used
standard EPA/OPPT models for each of the expected release points to build the model. Based on this information, EPA has
concluded that the weight of scientific evidence for this assessment is moderate and provides a plausible estimate of
releases in consideration of the strengths and limitations of reasonably available data.
Dishwasher detergent
EPA used the same approach to estimate air releases for this OES as the Dish soap OES. Therefore, the same rationale and
overall weight of scientific evidence apply to this OES.
Laundry detergent
Air releases are assessed using Monte Carlo modeling with information from the ESD on Industrial and Institutional
Laundries. Factors that increase the strength of evidence for this OES are that the release estimates are directly relevant to
the OES (as opposed to surrogate), that the ESD on Industrial and Institutional Laundries has a medium overall data
quality determination and was peer reviewed, high number of data points (simulation runs), consistency within the dataset,
and full distributions of input parameters. The Monte Carlo modeling accounts for the entire distribution of input
parameters, calculating a distribution of potential release values that represents a larger proportion of sites than a discrete
value. Additionally, EPA was able to separately estimate releases for industrial and institutional laundry settings. Factors
that decrease the strength of the evidence for this OES include uncertainties and limitations in the representativeness of the
estimates for sites that specifically use 1,4-dioxane because the default values in the ESD are generic. Another uncertainty
is lack of consideration for release controls. The ESD assumes that all activities occur without any release controls. Actual
releases may be less than estimated if facilities utilize pollution control methods. Based on this information, EPA has
concluded that the weight of scientific evidence for this assessment is moderate and provides a plausible estimate of
releases in consideration of the strengths and limitations of reasonably available data.
Paint and floor lacquer
Air releases are assessed using OECD ESD on Coating Application via Spray-Painting in the Automotive Refinishing
Industry. Factors that increase the strength of evidence for this OES are that the release estimates are directly relevant to
the OES (as opposed to surrogate), the ESD has a medium overall data quality determination, consistency within the
sources used, and a low amount of uncertainties. Factors that decrease the strength of the evidence for this OES include a
lack of variability and uncertainty in the representativeness of the ESD to all sites and sites that specifically use 1,4-
dioxane since the ESD is generic. Based on this information, EPA has concluded that the weight of scientific evidence for
Page 312 of 570
-------
OES
Weight of Scientific Evidence Conclusion in Release Estimates
this assessment is slight to moderate and provides a plausible estimate of releases in consideration of the strengths and
limitations of reasonably available data.
Polyethylene terephthalate
(PET) byproduct
Air releases are assessed using reported discharges from 2019 TRI. Factors that increase the strength of evidence for this
OES are that the release data are directly relevant to the OES (as opposed to surrogate), that TRI has a medium overall data
quality determination, consistency within the dataset (all reporters use the same or similar reporting forms), and
consistency with the emission data from the related life cycle analysis discussed in Appendix E.6. A strength of TRI data is
that TRI compiles the best reasonably available release data for all reporting facilities. Factors that decrease the strength of
the evidence for this OES include uncertainty in the accuracy of reported releases, lack of variability (only 1 year of data
used), and the limitations in representativeness to all sites because TRI may not capture all relevant sites. Additionally,
EPA made assumptions on the number of operating days to estimate daily releases, which introduces additional
uncertainty. Based on this information, EPA has concluded that the weight of scientific evidence for this assessment is
moderate to robust and provides a plausible estimate of releases in consideration of the strengths and limitations of
reasonably available data.
Ethoxylation process
byproduct
Air releases are assessed using reported discharges from 2019 TRI. Factors that increase the strength of evidence for this
OES are that the release data are directly relevant to the OES (as opposed to surrogate), that TRI has a medium overall data
quality determination, and consistency within the dataset (all reporters use the same or similar reporting forms). A strength
of TRI data is that TRI compiles the best reasonably available release data for all reporting facilities. Factors that decrease
the strength of the evidence for this OES include the uncertainty in the accuracy of reported releases, lack of variability
(only 1 year of data used), and the limitations in representativeness to all sites because TRI may not capture all relevant
sites. Additionally, EPA made assumptions on the number of operating days to estimate daily releases, which introduces
additional uncertainty. Based on this information, EPA has concluded that the weight of scientific evidence for this
assessment is moderate to robust and provides a plausible estimate of releases in consideration of the strengths and
limitations of reasonably available data.
Hydraulic fracturing
Air releases are assessed using Monte Carlo modeling with information from the Revised ESD on Hydraulic Fracturing
and FracFocus 3.0. Factors that increase the strength of evidence for this OES are that the release estimates are directly
relevant to the OES (as opposed to surrogate), that the Revised ESD on Hydraulic Fracturing and FracFocus 3.0 have
medium overall data quality determinations, that the Revised ESD has undergone peer review by OECD, the high number
of data points (simulation runs), consistency within the dataset, and full distributions of input parameters. The Monte Carlo
modeling accounts for the entire distribution of input parameters, calculating a distribution of potential release values that
represents a larger proportion of sites than a discrete value. Factors that decrease the strength of the evidence for this OES
include the uncertainties and limitations in the representativeness of the estimates for sites that specifically use 1,4-dioxane
because the default values from the Revised ESD on Hydraulic Fracturing. Another uncertainty is lack of consideration for
release controls. The ESD assumes that all activities occur without any release controls. Actual releases may be less than
estimated if facilities utilize pollution control methods, contributing to uncertainty. Based on this information, EPA has
concluded that the weight of scientific evidence for this assessment is moderate to robust and provides a plausible estimate
of releases in consideration of the strengths and limitations of reasonably available data.
Page 313 of 570
-------
E,6 Comparison to PET Life Cycle Analysis
EPA identified a relevant life cycle analysis (LCA) published by the National Association for PET
Container Resources (NAPCOR) that included 1,4-dioxane emission factors from PET resin production
(Franklin Associates. 2020). EPA did not use these emission factors to estimate releases in this Risk
Evaluation because there were site-specific releases reported in TRI and DMR. This LCA only provided
generic emission factors for air and surface water releases aggregated across seven unspecified sites; the
LCA did not provide 1,4-dioxane emission factors for land releases. The emission factors in the LCA
were reported by three producers (seven sites) that account for 50 percent of the 2015 U.S. PET
production in a survey, and the basis of the emission factors is not provided. However, the survey states
that the release data is primary data {i.e., the data were provided by directly by the surveyed PET
producers). As opposed to conventional emission factors, the report only provides the order of
magnitude of the average amount of 1,4-dioxane released per amount of PET produced. Discrete, site-
specific emission factors are not provided. As a result, the variability of 1,4-dioxane releases from site to
site is unknown. EPA prefers the use of site-specific release data as opposed to generic emission factors.
Therefore, a comparison between total annual air and water releases from the LCA and from the TRI
and DMR data used in this Risk Evaluation is provided below for context.
The LCA estimated that 4.7 million tons of PET capacity was available in 2015 in North America
(Franklin Associates. 2020). To obtain total annual air and water release estimates from the LCA, EPA
multiplied this production volume by the reported 1,4-dioxane emission factors of 0.001 kg 1,4-dioxane
emitted per 1,000 kg PET for air releases, and 0.01 kg 1,4-dioxane emitted per 1000 kg PET for surface
water releases. To obtain the total annual air and water releases from the TRI and DMR used in this Risk
Evaluation, EPA summed all reported annual site-specific air emissions and surface water discharges
that were mapped to the "PET manufacturing" OES (see Appendix E.3 and E.5 for additional
information on the use of TRI and DMR). The total annual releases from the LCA and from TRI and
DMR is compiled in Table Apx E-9. The Agency did this comparison with 2019 TRI/DMR because
EPA's Risk Evaluation largely uses 2019 data, as well as 2015 TRI/DMR data because the releases
estimated with the LCA data are based on 2015 PET manufacturing data.
For air emissions, the LCA estimate and EPA's estimates from the 2019 and 2015 TRI are comparable,
being within an order of magnitude. Differences in the estimates likely arise since EPA's analysis
accounted for emissions from 13 PET manufacturing facilities compared to the seven facilities in the
LCA. Additionally, the LCA is an aggregate of releases across sites whereas EPA's analysis accounts
for variability by using data from individual sites.
For surface water discharges, the LCA estimate and EPA's estimates from the 2019 TRI and DMR show
a larger discrepancy, with EPA's estimate being two orders of magnitude larger than the LCA estimate.
However, over 2.51 million kg of the approximately 2.53 million kg (99.2%) of surface water discharges
in EPA's estimate comes from a single facility's 2019 DMR report (DAK Americas LLC, Columbia).
The other 12 facilities with non-zero surface water discharges in EPA's analysis had site-specific annual
discharges ranging between 5.37 and 8,922 kgs for 2019. EPA reviewed the DAK Americas LLC,
Columbia DMR reports from other years for comparison, which indicated approximately 14,000 kg of
1,4-dioxane were discharged in 2022, 8,800 kg in 2021, 6.8 million kg in 2020, and 2,300 kg in 2018.
DAK Americas LLC, Columbia did not include 1,4-dioxane in their DMRs in 2016 or 2017 (the two
earliest reporting years EPA looked at for this analysis). It is unclear why DAK Americas LLC,
Columbia's discharges were significantly higher in 2019 and 2020 or why these discharges were
different than other PET manufacturers in EPA's analysis. However, it is more likely that the facilities
analyzed in the LCA were more similar to the other PET manufacturing facilities in EPA's analysis,
Page 314 of 570
-------
with the total surface water emissions estimated from the LCA data being within one order of magnitude
of EPA's estimate when excluding the discharges from the DAK Americas LLC, Columbia facility, as
shown in Table Apx E-9. The LCA estimate and EPA's estimates from the 2015 TRI are comparable.
As indicated above, estimates from the LCA analysis and EPA's analysis with TRI and DMR data may
differ for several reasons. First, the LCA provides a single emission factor aggregated from data across
seven sites, reducing the impact of site-specific variability in releases. Whereas EPA's analysis uses site-
specific release data from 13 sites (for air emissions) and 19 sites (for surface water discharges, 6 of
which reported 0 surface water discharges). EPA also does not have access to site identities, or the
underlying data/methodologies used to estimate emission factors in the LCA, which limits EPA's ability
to do a direct site-to-site comparison of results between the two analyses. Additionally, the LCA study
states that some emissions are reported only by the order of magnitude of the average to protect the
confidentiality of individual companies, introducing further uncertainty in the emission factors presented
in the study. Lastly, the LCA data is from 2015 whereas EPA used data from 2019.
Table Apx E-9. Comparison of TRI/DMR Release Data to LCA Study for PET Byproduct
Data Source
Total Release for All Sites
(kg/vr)
Air emissions
EPA Estimate in this Risk Evaluation -
Based on 2019 TRI (including DAK Americas LLC, Columbia)
10,695
EPA Estimate in this Risk Evaluation -
Based on 2015 TRI (including DAK Americas LLC, Columbia)
12,407
LCA Estimate (Franklin Associates. 2020) - Based on 2015 data
4,264
Surface water discharges
EPA Estimate in this Risk Evaluation -
Based on 2019 DMR and TRI (including DAK Americas LLC,
Columbia)
2,531,730
EPA Estimate in this Risk Evaluation -
Based on 2019 DMR and TRI (excluding DAK Americas LLC,
Columbia)
19,296
EPA Estimate in this Risk Evaluation -
Based on 2015 DMR and TRI
20,511
LCA Estimate (Franklin Associates. 2020) - Based on 2015 data
42,638
E.7 Detailed Strengths, Limitations, Assumptions and Key Sources of
Uncertainties for the Environmental Release Assessment
This section includes detailed strengths, limitations, assumptions, and uncertainties associated with
EPA's approaches for estimating air, water, and land releases in this supplemental risk evaluation. This
section is intended to supplement the summary of strengths, limitations, assumptions, and uncertainties
discussed in Section 2.2.1.3 with additional details.
Page 315 of 570
-------
Facilities Not Captured by TRI or DMR Reporting
When using TRI data to analyze chemical releases, it is important to acknowledge that TRI reporting
does not include all releases of the chemical and therefore, the number of sites for a given OES may be
underestimated. For each OES that had TRI or DMR data, the analysis of releases for those OES was
limited to the facilities that reported releases to TRI and DMR. Therefore, it is uncertain the extent to
which sites not captured in these databases have air, water, or land releases of 1,4-dioxane and what the
exact media of release for those releases would be (e.g., stack vs. fugitive air, surface water vs POTW,
RCRA or another type of landfill). To the extent additional sites are not captured, releases may be
underestimated; however, the magnitude of this underestimation is unknown. TRI data do not include:
• Releases from any facility that used the chemical in quantities below the applicable annual
chemical activity threshold (e.g., 25,000 lb manufactured or processed, or 10,000 lb otherwise
used, for most chemicals);
• Releases from any facility that is not in a TRI covered sector; and
• Releases from any facility that does not meet the TRI employment threshold of greater than 10
full-time employee equivalents (20,000 labor hours) for the year.
Due to these TRI reporting thresholds, estimated releases using TRI data may not be representative of all
sites, particularly those sites that handle 1,4-dioxane at quantities below the TRI reporting threshold.
DMR Release Data
For facilities that reported having zero pollutant loads to DMR, the ECHO Pollutant Loading Tool
Advanced Search applies a hybrid method to analyze non-detects. The EZ Search Load Module uses a
combination of setting non-detects equal to zero and as one half the detection limit to calculate the
annual pollutant loadings. Specifically, if the pollutant was measured as non-detect for all monitoring
periods in a reporting year, then the EZ Search Load Module sets the annual pollutant load to zero. If the
pollutant was detected for at least one monitoring period in a reporting year, then the EZ Search Load
Module calculates the annual pollutant load by setting the non-detects equal to one half the detection
limit. This method could cause overestimation or underestimation of annual and daily pollutant loads.
However, EPA uses this method for handling non-detects as it is consistent with the established
procedures for the EZ Search Load Module.
TRI Release Data
EPCRA section 313 states that facilities may estimate their release quantities using "readily available
data," including monitoring data, collected for other purposes. When data are not readily available,
EPCRA section 313 states that "reasonable estimates" may be used. The facility is not required to
monitor or measure the quantities, concentration, or frequency of any toxic chemical release for TRI
reporting. TRI guidance states that not using readily available information, such as relevant monitoring
data collected for compliance with other regulations, could result in enforcement and penalties.
For each release quantity reported, TRI facilities select a "Basis of estimate" code indicating the
principal method used to determine the amount of the release. TRI provides six basis of estimate codes
to choose from: continuous monitoring, periodic monitoring, mass balance, published emissions factors,
site-specific emissions factors, or engineering calculations/best engineering judgment. In facilities where
a chemical is used in multiple operations, the facility may use a combination of methods to calculate the
release reported. In such cases, TRI instructs the facility to enter the basis of estimate code of the
method that applies to the largest portion of the release quantity. Additional details on the basis of
estimate, such as any calculations and underlying assumptions, are not reported. Depending on the
inputs and/or monitoring methods used by each facility, any of the methods used to estimate releases
may over or underestimate releases. The magnitude of this uncertainty is unknown.
Page 316 of 570
-------
For any release quantity that is less than 1,000 lb, facilities may report either the estimated quantity or a
range code. The 1,000-pound limit for range code reporting applies to each type of release reported to
TRI - fugitive air emissions, stack air emissions, water discharges, each type of land disposal, and each
type of off-site transfer. There are three TRI range codes: 1 to 10; 11 to 499; and 500 to 999 lb. TRI data
tools display the approximate midpoint of the range {i.e., 5, 250, or 750 lb). Using this midpoint value
may be either an over or an underestimate of the true value, depending where on the range the true value
lies. Although analyses using data that was reported as a range code may add uncertainty, it is not clear
that the uncertainty associated with a range code is greater than that associated with any other estimated
release value. Range code reporting is not permitted for chemicals of special concern.
TRI guidance states that release estimates need not be reported to more than two significant figures.
However, the guidance also states that facilities should report release quantities at a level of precision
supported by the accuracy of the underlying data and the estimation techniques on which the estimate
was based. If a facility's release calculations support reporting an amount that is more precise than two
significant digits, then the facility should report that more precise amount. The facility makes the
determination of the accuracy of their estimate and the appropriate significant digits to use.
For chemicals that meet certain criteria, facilities have the option of submitting a TRI Form A
Certification Statement instead of a TRI Form R. The Form A does not include any details on the
chemical release or waste management quantities. The criteria for a Form A are that during the reporting
year, the chemical (1) did not exceed 500 lb for the total annual reportable amount (including the sum of
on- and off-site quantities released, treated, recycled, and used for energy recovery); (2) amounts
manufactured, processed, or otherwise used do not exceed 1 million lb; and (3) the chemical is not a
chemical of special concern. When conducting analyses of chemical releases and a facility has submitted
a Form A for the chemical, there is no way to discern the quantity released to each medium or even if
there were any releases. Where facilities reported to TRI with a Form A, EPA used the Form A
threshold for total releases of 500 lb/year for each release media {e.g., fugitive air, stack air, surface
water, POTW, non-POTW WWT, RCRA landfill, other landfill). EPA used the entire 500 lb/year for
each release media; however, since this threshold is for total site releases, these 500 lb/year are to only
one of these media at a time (since assessing 500 lb/year for all media at once would double count the
releases and exceed the total release threshold for Form A). Furthermore, the threshold represents an
upper limit on total releases to all environmental media from the facility; therefore, assessing releases at
the threshold value likely overestimates actual releases from the facility.
Differences between TRI and DMR
There is uncertainty when the reported surface water discharges for a given site differs between DMR
and TRI for the same year. In these instances, EPA uses the higher of the reported discharge quantities.
This uncertainty is particularly prevalent for the PET manufacturing site, DAK Americas LLC.
Specifically, this site reported the discharge of millions of pounds of 1,4-dioxane in 2019 DMR but only
16 pounds in 2019 TRI. See Appendix E.6 for additional discussion of this site and comparison to other
PET manufacturing sites and a life cycle analysis on PET manufacturing.
Mapping TRI and DMR Facilities to OES
EPA used a crosswalk between TRI uses/sub-uses and CDR Industrial Function Category (IFC) codes
(see Appendix E.9) along with a mapping of CDR IFC codes to OES to assign the OES for each facility
that reported to TRI. However, there are limitations to this approach. For example, this approach may
result in the mapping of multiple OES for one facility. Additionally, there are limitations to the TRI -
CDR crosswalk. For example, a TRI use/sub-use may encompass multiple uses that are not captured in
the crosswalked CDR IFC codes. In these instances, EPA determined the primary OES using the NAICS
Page 317 of 570
-------
codes that the facility reported in TRI, information on types of products that the facility manufactures
found from internet searches, and which OES is most likely to result in release. OES are related to
COUs as defined in the crosswalk in Table 2-1 and Appendix D.2.This approach requires some level of
engineering judgment to determine which OES is the most applicable to the facility, which introduces
uncertainty in the OES mapping. Additionally, this approach assumes only one OES is applicable to the
facility, which may be incorrect if the facility uses 1,4-dioxane for multiple purposes. If facilities were
categorized under a different OES, the annual releases for each site would remain unchanged; however,
average daily releases may change depending on the release days expected for the different OES.
Additional uncertainty is present in the OES mapping for TRI sites that reported using a Form A and
DMR sites because there is no reported use/sub-use information. EPA used a similar procedure as
described above to map these sites to an OES, involving the use of NAICS and Standard Industrial
Classification (SIC) codes reported to TRI and DMR, internet searches on the types of products made at
the facility, and which OES is most likely to result in release. Since this approach involves engineering
judgment to determine which OES is the most applicable to the facility, there is uncertainty in the OES
mapping.
There is also uncertainty in the NAICS codes and SIC codes reported in TRI and DMR. TRI facilities
enter the facility's primary NAICS code indicating the primary economic activity at the facility.
Facilities can also enter secondary NAICS codes. When using TRI chemical release data for a facility
that also reported secondary NAICS codes, there may be uncertainty as to which NAICS is associated
with the use of the chemical. Additionally, NAICS codes and SIC codes are reported for the facility as a
whole and are not chemical specific.
Estimating Daily Releases from Annual TRI and DMR Release Data
Facilities reporting to TRI and DMR only report annual releases; to assess daily air and water releases,
EPA estimated the release days and averaged the annual releases over these days. There is some
uncertainty that all facilities for a given OES operate for the assumed duration; therefore, the average
daily release may be higher if sites have fewer release days or lower if they have greater release days.
Furthermore, chemical concentrations in air emissions and wastewater streams at each facility may vary
from day to day such that on any given day the actual daily releases may be higher or lower than the
estimated average daily discharge. Thus, this approach minimizes spikes and drops in emissions and
discharges from day to day.
EPA did not estimate daily land releases due to the high level of uncertainty in the number of release
days associated with land releases; instead, EPA estimated annual land releases.
Representativeness of TRI and DMR for an OES as a Whole
The representativeness of TRI and DMR data for an OES as a whole is dependent on (1) the extent to
which these reporting mechanisms capture all potential sites within the OES and, (2) the extent to which
the release quantities provided by reporting sites reflect releases from non-reporting sites.
For some OES, the total number of sites was determined from TRI. For these OESs, there is uncertainty
in if there may be additional sites using the chemical within the OES that did not report to TRI (e.g., due
to being below reporting thresholds). For some OES, such as manufacturing and other OES involving
larger industrial sites, TRI is more likely to capture the majority of potential sites because these sites
typically meet the reporting threshold. For other OES, such as functional fluids (open-systems), 3D
printing, and other OES that may be performed at a range of different scales, the extent to which TRI
captures all potential sites is more uncertain because not all sites may meet the reporting threshold. This
Page 318 of 570
-------
uncertainty may result in an underestimate in the total number of sites using the chemical within the
OES. To the extent additional sites are not captured, releases may be underestimated; however, the
magnitude of this underestimation is unknown for each OES. In some cases, the total number of
facilities for a given OES was estimated using data from the U.S. Census. This may result in an
overestimate of the actual number of sites that use the chemical for that OES.
Additionally, it is unknown how representative release estimates from TRI and DMR reporting sites
accurately reflect all releases from within an OES since releases from non-reporting sites cannot be
quantified. Specifically, where the number of sites was estimated from U.S. Census data, the average
daily release calculated from sites reporting to TRI or DMR was applied to the total number of sites
reported in (U.S. Census Bureau.: ). It is uncertain how accurate this average release is to actual
releases at these sites; therefore, releases may be higher or lower than the calculated amount.
The estimates presented use TRI and DMR data from 2013 to 2019 for water releases and just 2019 TRI
data for land and air releases. There is uncertainty in the representativeness of past years TRI and DMR
data towards current conditions. Pollution control technologies, production rates, and other factors may
change from year-to-year.
Estimating Emissions for OES Without TRI Data
For release estimates developed for an OES when directly applicable TRI or DMR data were not
available, there are uncertainties related to the use of surrogate TRI or DMR data or, in their absence,
the use of modeling.
Use of surrogate TRI or DMR data may introduce uncertainties related to the extent to which the
surrogate OES and the OES being assessed are similar. Thus, the representativeness of the surrogate
release data towards the actual releases for the OES being assessed is uncertain.
Although no new models were developed for this release assessment, the adaptations made to and uses
of these models as part of the analysis (e.g., varying input parameters, Monte Carlo simulation) may
result in release estimates higher or lower than the actual amount. EPA used the available data to vary
input parameters in models. Where possible, EPA assigned a distribution to model input parameters
based on the data available (e.g., discrete if a full dataset was available or triangular if just a range and
mode were available,); however, the true shape of the underlying distributions is unknown in most cases,
lending uncertainty to the assessment. Additionally, for most input parameters there is uncertainty in the
extent to which the available data for the parameter distribution addresses temporal variability as well as
intra- and inter-site variability, which includes variability both within a site and between multiple sites
due to variations in process operations and conditions. The most robust input parameter dataset was
from FracFocus for the hydraulic fracturing OES, since it reflects 411 distinct sites using fracturing
fluids containing 1,4-dioxane and was taken from 2016 - 2021. However, most other input parameter
distributions were based on more limited and generic datasets from GS or ESD. Additionally, for some
parameters, sufficient data were not available to assign distributions, so EPA used a single static value.
EPA presented central tendency (50th percentile) and high-end (95th percentile) modeled release values
to capture a range of potential releases and reduce the uncertainty associated with using a single release
estimate. However, the aforementioned limitations add uncertainty in the extent to which modeled
release results capture the true distribution of potential releases from all sites that use 1,4-dioxane.
Additionally, the approaches used for estimating releases based on modeling or literature are for generic
sites, which differs from the facility-specific approach used for OES for which TRI or DMR data were
Page 319 of 570
-------
available (as described previously, the modeled estimates are for a "generic site"). This may introduce
uncertainties that differ from those of the scenarios using TRI or DMR data, described above.
For the surface cleaner OES, EPA used SHEDS-HT DTD modeling to estimate commercial use
environmental releases to surface water and land. To estimate land release, EPA used the modeled water
releases from SHEDS-HT and back-calculated a 1,4-dioxane use rate based on the expected loss fraction
to water for the OES. Then, a land release loss fraction was applied to the back-calculated use rate. The
uncertainty in this approach is due to the standard models and assumptions used to estimate loss
fractions to water and land. The main source of uncertainty from using SHEDS-HT DTD modeling is
that the modeling is for a single case study location, Liverpool, OH. It is uncertain whether the release
estimates generated from this case study are applicable to other areas of the country. EPA was unable to
estimate the number of sites in Liverpool, OH, for the OES where this modeling approach was used;
therefore, the release estimates were presented as totals for all sites as opposed to per-site estimates.
Additionally, EPA is unsure whether the use of SHEDS-HT results in a high-end or typical exposure
scenario, so the use of this data may lead to over or underestimates of releases.
Spills and Leaks
Spills and leaks may occur during multiple OES. Generally, releases and exposures from spills and leaks
are assessed within the OES where they occur, as TRI data includes releases from accidental releases
such as spills and GS/ESD typically include assessment approaches for spills where supported by data.
For example, EPA assessed releases from spills according to the Revised Hydraulic Fracturing ESD, as
discussed in Appendix E. 13. However, due to the highly variable nature of spills, there is uncertainty in
the representativeness of any data on spills towards all potential accidental releases for a given OES.
Additionally, there is uncertainty in the media of release for spills, as spill response procedures and
methods of disposal are highly depending on the nature of the spilled material.
E.8 Weight of Scientific Evidence Conclusions for Environmental Releases
Table Apx E-10 presents a summary of EPA's overall weight of scientific evidence conclusions for its
release estimates for each of the assessed OES. As discussed in Section 2.2.1.2, the weight of scientific
evidence conclusions take into account factors such as data/information quality, applicability of release
data to the OES (including considerations of temporal relevance, locational relevance), modeling
limitations such as lack of data for input parameters, and the representativeness of the release estimate
for the whole industry.
Page 320 of 570
-------
Table Apx E-10. Summary of Overall Weight of Scientific Evidence Conclusions for Environmental Release Estimates by PES
OES
Monitoring"
Modeling
Weight of Scientific Evidence
Conclusion
Notes
Air
Water
Land
Data
Quality
Rating
Air
Water
Land
Data
Quality
Rating
Air
Water
Land
Manufacturing
/
1/
M/H
X
X
X
N/A
Moderate
to Robust
Moderate to
Robust
Moderate
to Robust
Based on TRI and DMR which have
medium data quality ratings. Information
on the conditions of use of 1,4-dioxane at
facilities in TRI and DMR is limited, but
uncertainties are known.
Import and
repackaging
1/
M/H
X
X
X
N/A
Moderate
to Robust
Moderate to
Robust
Moderate
to Robust
Based on TRI and DMR which have
medium data quality ratings. Information
on the conditions of use of 1,4-dioxane at
facilities in TRI and DMR is limited, but
uncertainties are known.
Industrial uses
/
/
1/
M/H
X
X
X
N/A
Moderate
to Robust
Moderate to
Robust
Moderate
to Robust
Based on TRI and DMR which have
medium data quality ratings. Information
on the conditions of use of 1,4-dioxane at
facilities in TRI and DMR is limited, but
uncertainties are known.
Functional
fluids (open-
system)
1/
M/H
X
X
X
N/A
Moderate
to Robust
Moderate to
Robust
Moderate
to Robust
Based on TRI and DMR which have
medium data quality ratings. Information
on the conditions of use of 1,4-dioxane at
facilities in TRI and DMR is limited, but
uncertainties are known.
Functional
fluids (closed-
system)
Assessed as a part of Industrial uses OES
Slight
Slight
Slight
No data were available to estimate
releases for this OES, so it was grouped
with Industrial uses OES. There is
uncertainty in the representativeness of
the Industrial uses data for this OES.
Laboratory
chemicals
X
X
X
N/A
H
Slight to
Moderate
Slight to
Moderate
Slight to
Moderate
Assessed using Laboratory Chemicals GS
which has a high data quality rating.
Activities could vary drastically on a site-
by-site basis due to uncertainties and
limitations in the model.
Film cement
X
X
X
N/A
Not
expected
H
Slight to
Moderate
Slight to
Moderate
Slight to
Moderate
The underlying data sources for model
parameters have a high data quality rating.
Modeling may not be sufficiently
representative of all the sites for this OES.
Spray foam
application
X
X
X
N/A
M
Slight to
Moderate
Slight to
Moderate
Slight to
Moderate
Assessed using SPF GS which has a
medium data quality rating. Activities
could vary drastically on a site-by-site
basis due to uncertainties and limitations
in the model.
Page 321 of 570
-------
OES
Monitoring"
Modeling
Weight of Scientific Evidence
Conclusion
Notes
Air
Water
Land
Data
Quality
Rating
Air
Water
Land
Data
Quality
Rating
Air
Water
Land
Printing inks
(3D)
Assumed
included
in
Industria
1 uses
OES
/
Assumed
included
in
Industrial
uses OES
M/H
X
X
X
N/A
Slight
Moderate to
Robust
Slight
Based on TRI and DMR which have
medium data quality ratings. Information
on the conditions of use of 1,4-dioxane at
facilities in TRI and DMR is limited.
There is uncertainty in the
representativeness of the Industrial uses
data for this OES.
Dry film
lubricant
X
X
X
N/A
/
Not
expected
/
H
Slight to
Moderate
Slight to
Moderate
Slight to
Moderate
The underlying data sources for model
parameters have a high data quality rating.
Modeling may not be sufficiently
representative of all the sites for this OES.
Disposal
/
/
•
-------
Monitoring"
Modeling
Weight of Scientific Evidence
Conclusion
OES
Data
Data
Notes
Air
Water
Land
Quality
Rating
Air
Water
Land
Quality
Rating
Air
Water
Land
be sufficiently representative of all the
sites for this OES.
Dishwasher
X
X
X
N/A
/
/
/
H
Moderate
Moderate
Moderate
Assessed using a public comment, which
detergent
has a high data quality rating. Monte
Carlo modeling allows for parameter
variation; however, the modeling may not
be sufficiently representative of all the
sites for this OES.
Laundry
X
X
X
N/A
/
/
/
M
Moderate
Moderate
Moderate
Assessed using Laundries ESD, which has
detergent
(industrial)
a medium data quality rating. Monte
Carlo modeling allows for parameter
variation; however, the modeling may not
be sufficiently representative of all the
sites for this OES.
Laundry
X
X
X
N/A
/
/
/
M
Moderate
Moderate
Moderate
Assessed using Laundries ESD, which has
detergent
(institutional)
a medium data quality rating. Monte
Carlo modeling allows for parameter
variation; however, the modeling may not
be sufficiently representative of all the
sites for this OES.
Paints and floor
lacquer
X
X
X
N/A
/
Not
expected
/
M
Slight to
Moderate
Slight to
Moderate
Slight to
Moderate
Assessed using Automotive Spray
Painting ESD, which has a medium data
quality rating. Modeling may not be
sufficiently representative of all the sites
for this exposure scenario.
PET byproduct
/
/
/
M/H
X
X
X
N/A
Moderate
to Robust
Moderate to
Robust
Moderate
to Robust
Based on TRI and DMR which have
medium data quality ratings. Information
on the conditions of use of 1,4-dioxane at
facilities in TRI and DMR is limited, but
uncertainties are known.
Ethoxylation
/
/
/
M/H
X
X
X
N/A
Moderate
Moderate to
Moderate
Based on TRI and DMR which have
process
byproduct
to Robust
Robust
to Robust
medium data quality ratings. Information
on the conditions of use of 1,4-dioxane at
facilities in TRI and DMR is limited, but
uncertainties are known.
Hydraulic
M
M
Moderate
Moderate to
Moderate
Based on FracFocus 3.0 and the Hydraulic
fracturing
to Robust
Robust
to Robust
Fracturing ESD, which has a medium data
quality rating. Monte Carlo modeling
allows for parameter variation; however,
Page 323 of 570
-------
OES
Monitoring"
Modeling
Weight of Scientific Evidence
Conclusion
Notes
Air
Water
Land
Data
Quality
Rating
Air
Water
Land
Data
Quality
Rating
Air
Water
Land
the modeling may not be sufficiently
representative of all the sites for this OES.
E.9 TRI to CDR Crosswalk
Table Apx E-l 1 presents the TRI-CDR Crosswalk used to map facilities to the OES for each chemical. Blanks in the 2016 CDR code column
indicate there is no corresponding CDR code that matches the TRI code.
Table Apx E-ll. TRI-CDR Use Code Crosswalk
TRI
Section
TRI Description
TRI Sub-
use Code
TRI Sub-use
Code Name
2016
CDR
Code
2016 CDR Code
Name
2016 CDR Functional Use Definition
3.1.a
Manufacture:
Produce
N/A
N/A
N/A
N/A
N/A
3.1.b
Manufacture:
Import
N/A
N/A
N/A
N/A
N/A
3.1.c
Manufacture: For
on-site
use/processing
N/A
N/A
N/A
N/A
N/A
3.1.d
Manufacture: For
sale/distribution
N/A
N/A
N/A
N/A
N/A
3.1.e
Manufacture: As a
byproduct
N/A
N/A
N/A
N/A
N/A
3.1.f
Manufacture: As an
impurity
N/A
N/A
N/A
N/A
N/A
3.2.a
Processing: As a
reactant
N/A
N/A
PC
Processing as a
reactant
Chemical substance is used in chemical reactions for the
manufacturing of another chemical substance or product.
3.2.a
Processing: As a
reactant
P101
Feedstocks
N/A
N/A
N/A
3.2.a
Processing: As a
reactant
P102
Raw Materials
N/A
N/A
N/A
Page 324 of 570
-------
TRI
Section
TRI Description
TRI Sub-
use Code
TRI Sub-use
Code Name
2016
CDR
Code
2016 CDR Code
Name
2016 CDR Functional Use Definition
3.2.a
Processing: As a
reactant
P103
Intermediates
U015
Intermediates
Chemical substances consumed in a reaction to produce
other chemical substances for commercial advantage. A
residual of the intermediate chemical substance which has
no separate function may remain in the reaction product.
3.2.a
Processing: As a
reactant
P104
Initiators
U024
Process regulators
Chemical substances used to change the rate of a
chemical reaction, start or stop the reaction, or otherwise
influence the course of the reaction. Process regulators
may be consumed or become part of the reaction product.
3.2.a
Processing: As a
reactant
P199
Other
U016
Ion exchange agents
Chemical substances, usually in the form of a solid
matrix, that are used to selectively remove targeted ions
from a solution. Examples generally consist of an inert
hydrophobic matrix such as styrenedivinylbenzene or
phenol-formaldehyde, cross-linking polymer such as
divinylbenzene, and ionic functional groups including
sulfonic, carboxylic or phosphonic acids. This code also
includes aluminosilicate zeolites.
3.2.a
Processing: As a
reactant
P199
Other
U019
Oxidizing/
reducing agent
Chemical substances used to alter the valence state of
another substance by donating or accepting electrons or
by the addition or removal of hydrogen to a substance.
Examples of oxidizing agents include nitric acid,
perchlorates, hexavalent chromium compounds, and
peroxydisulfuric acid salts. Examples of reducing agents
include hydrazine, sodium thiosulfate, and coke produced
from coal.
3.2.a
Processing: As a
reactant
P199
Other
U999
Other (specify)
Chemical substances used in a way other than those
described by other codes.
3.2.b
Processing: As a
formulation
component
N/A
N/A
PF
Processing-
incorporation into
formulation, mixture,
or reaction product
Chemical substance is added to a product (or product
mixture) prior to further distribution of the product.
3.2.b
Processing: As a
formulation
component
P201
Additives
U007
Corrosion inhibitors
and anti-scaling
agents
Chemical substances used to prevent or retard corrosion
or the formation of scale. Examples include
phenylenediamine, chromates, nitrates, phosphates, and
hydrazine.
Page 325 of 570
-------
TRI
Section
TRI Description
TRI Sub-
use Code
TRI Sub-use
Code Name
2016
CDR
Code
2016 CDR Code
Name
2016 CDR Functional Use Definition
3.2.b
Processing: As a
formulation
component
P201
Additives
U009
Fillers
Chemical substances used to provide bulk, increase
strength, increase hardness, or improve resistance to
impact. Fillers incorporated in a matrix reduce production
costs by minimizing the amount of more expensive
substances used in the production of articles. Examples
include calcium carbonate, barium sulfate, silicates, clays,
zinc oxide and aluminum oxide.
3.2.b
Processing: As a
formulation
component
P201
Additives
U010
Finishing agents
Chemical substances used to impart such functions as
softening, static proofing, wrinkle resistance, and water
repellence. Substances may be applied to textiles, paper,
and leather. Examples include quaternary ammonium
compounds, ethoxylated amines, and silicone compounds.
3.2.b
Processing: As a
formulation
component
P201
Additives
U017
Lubricants and
lubricant additives
Chemical substances used to reduce friction, heat, or wear
between moving parts or adjacent solid surfaces, or that
enhance the lubricity of other substances. Examples of
lubricants include mineral oils, silicate and phosphate
esters, silicone oil, greases, and solid film lubricants such
as graphite and PTFE. Examples of lubricant additives
include molybdenum disulphide and tungsten disulphide.
3.2.b
Processing: As a
formulation
component
P201
Additives
U034
Paint additives and
coating additives not
described by other
codes
Chemical substances used in a paint or coating
formulation to enhance properties such as water
repellence, increased gloss, improved fade resistance,
ease of application, foam prevention, etc. Examples of
paint additives and coating additives include polyols,
amines, vinyl acetate ethylene emulsions, and aliphatic
polyisocyanates.
3.2.b
Processing: As a
formulation
component
P202
Dyes
U008
Dyes
Chemical substances used to impart color to other
materials or mixtures (i.e., substrates) by penetrating the
surface of the substrate. Example types include azo,
anthraquinone, amino azo, aniline, eosin, stilbene, acid,
basic or cationic, reactive, dispersive, and natural dyes.
3.2.b
Processing: As a
formulation
component
P202
Dyes
U021
Pigments
Chemical substances used to impart color to other
materials or mixtures (i.e., substrates) by attaching
themselves to the surface of the substrate through binding
Page 326 of 570
-------
TRI
Section
TRI Description
TRI Sub-
use Code
TRI Sub-use
Code Name
2016
CDR
Code
2016 CDR Code
Name
2016 CDR Functional Use Definition
or adhesion. This code includes fluorescent agents,
luminescent agents, whitening agents, pearlizing agents,
and opacifiers. Examples include metallic oxides of iron,
titanium, zinc, cobalt, and chromium; metal powder
suspensions; lead chromates; vegetable and animal
products; and synthetic organic pigments.
3.2.b
Processing: As a
formulation
component
P203
Reaction
Diluents
U030
Solvents (which
become part of
product formulation or
mixture)
Chemical substances used to dissolve another substance
(solute) to form a uniformly dispersed mixture (solution)
at the molecular level. Examples include diluents used to
reduce the concentration of an active material to achieve a
specified effect and low gravity materials added to reduce
cost.
3.2.b
Processing: As a
formulation
component
P203
Reaction
Diluents
U032
Viscosity adjustors
Chemical substances used to alter the viscosity of another
substance. Examples include viscosity index (VI)
improvers, pour point depressants, and thickeners.
3.2.b
Processing: As a
formulation
component
P204
Initiators
U024
Process regulators
Chemical substances used to change the rate of a
chemical reaction, start or stop the reaction, or otherwise
influence the course of the reaction. Process regulators
may be consumed or become part of the reaction product.
3.2.b
Processing: As a
formulation
component
P205
Solvents
U030
Solvents (which
become part of
product formulation or
mixture)
Chemical substances used to dissolve another substance
(solute) to form a uniformly dispersed mixture (solution)
at the molecular level. Examples include diluents used to
reduce the concentration of an active material to achieve a
specified effect and low gravity materials added to reduce
cost.
3.2.b
Processing: As a
formulation
component
P206
Inhibitors
U024
Process regulators
Chemical substances used to change the rate of a
chemical reaction, start or stop the reaction, or otherwise
influence the course of the reaction. Process regulators
may be consumed or become part of the reaction product.
3.2.b
Processing: As a
formulation
component
P207
Emulsifiers
U003
Adsorbents and
absorbents
Chemical substances used to retain other substances by
accumulation on their surface or by assimilation.
Examples of adsorbents include silica gel, activated
alumina, and activated carbon. Examples of absorbents
include straw oil, alkaline solutions, and kerosene.
Page 327 of 570
-------
TRI
Section
TRI Description
TRI Sub-
use Code
TRI Sub-use
Code Name
2016
CDR
Code
2016 CDR Code
Name
2016 CDR Functional Use Definition
3.2.b
Processing: As a
formulation
component
P208
Surfactants
U002
Adhesives and sealant
chemicals
Chemical substances used to promote bonding between
other substances, promote adhesion of surfaces, or
prevent seepage of moisture or air. Examples include
epoxides, isocyanates, acrylamides, phenol, urea,
melamine, and formaldehyde.
3.2.b
Processing: As a
formulation
component
P208
Surfactants
U023
Plating agents and
surface treating agents
Chemical substances applied to metal, plastic, or other
surfaces to alter physical or chemical properties of the
surface. Examples include metal surface treating agents,
strippers, etchants, rust and tarnish removers, and
descaling agents.
3.2.b
Processing: As a
formulation
component
P208
Surfactants
U031
Surface active agents
Chemical substances used to modify surface tension when
dissolved in water or water solutions or reduce interfacial
tension between two liquids or between a liquid and a
solid or between liquid and air. Examples include
carboxylates, sulfonates, phosphates, carboxylic acid,
esters, and quaternary ammonium salts.
3.2.b
Processing: As a
formulation
component
P209
Lubricants
U017
Lubricants and
lubricant additives
Chemical substances used to reduce friction, heat, or wear
between moving parts or adjacent solid surfaces, or that
enhance the lubricity of other substances. Examples of
lubricants include mineral oils, silicate and phosphate
esters, silicone oil, greases, and solid film lubricants such
as graphite and PTFE. Examples of lubricant additives
include molybdenum disulphide and tungsten disulphide.
3.2.b
Processing: As a
formulation
component
P210
Flame
Retardants
U011
Flame retardants
Chemical substances used on the surface of or
incorporated into combustible materials to reduce or
eliminate their tendency to ignite when exposed to heat or
a flame for a short period of time. Examples include
inorganic salts, chlorinated or brominated organic
compounds, and organic phosphates/phosphonates.
3.2.b
Processing: As a
formulation
component
P211
Rheological
Modifiers
U022
Plasticizers
Chemical substances used in plastics, cement, concrete,
wallboard, clay bodies, or other materials to increase their
plasticity or fluidity. Examples include phthalates,
trimellitates, adipates, maleates, and lignosulphonates.
Page 328 of 570
-------
TRI
Section
TRI Description
TRI Sub-
use Code
TRI Sub-use
Code Name
2016
CDR
Code
2016 CDR Code
Name
2016 CDR Functional Use Definition
3.2.b
Processing: As a
formulation
component
P211
Rheological
Modifiers
U032
Viscosity adjusters
Chemical substances used to alter the viscosity of another
substance. Examples include viscosity index (VI)
improvers, pour point depressants, and thickeners.
3.2.b
Processing: As a
formulation
component
P299
Other
U003
Adsorbents and
absorbents
Chemical substances used to retain other substances by
accumulation on their surface or by assimilation.
Examples of adsorbents include silica gel, activated
alumina, and activated carbon. Examples of absorbents
include straw oil, alkaline solutions, and kerosene.
3.2.b
Processing: As a
formulation
component
P299
Other
U016
Ion exchange agents
Chemical substances, usually in the form of a solid
matrix, that are used to selectively remove targeted ions
from a solution. Examples generally consist of an inert
hydrophobic matrix such as styrenedivinylbenzene or
phenol-formaldehyde, cross-linking polymer such as
divinylbenzene, and ionic functional groups including
sulfonic, carboxylic or phosphonic acids. This code also
includes aluminosilicate zeolites.
3.2.b
Processing: As a
formulation
component
P299
Other
U018
Odor agents
Chemical substances used to control odors, remove odors,
mask odors, or impart odors. Examples include
benzenoids, terpenes and terpenoids, musk chemicals,
aliphatic aldehydes, aliphatic cyanides, and mercaptans.
3.2.b
Processing: As a
formulation
component
P299
Other
U019
Oxidizing/
reducing agent
Chemical substances used to alter the valence state of
another substance by donating or accepting electrons or
by the addition or removal of hydrogen to a substance.
Examples of oxidizing agents include nitric acid,
perchlorates, hexavalent chromium compounds, and
peroxydisulfuric acid salts. Examples of reducing agents
include hydrazine, sodium thiosulfate, and coke produced
from coal.
3.2.b
Processing: As a
formulation
component
P299
Other
U020
Photosensitive
chemicals
Chemical substances used for their ability to alter their
physical or chemical structure through absorption of light,
resulting in the emission of light, dissociation,
discoloration, or other chemical reaction. Examples
include sensitizers, fluorescents, photovoltaic agents,
ultraviolet absorbers, and ultraviolet stabilizers.
Page 329 of 570
-------
TRI
Section
TRI Description
TRI Sub-
use Code
TRI Sub-use
Code Name
2016
CDR
Code
2016 CDR Code
Name
2016 CDR Functional Use Definition
3.2.b
Processing: As a
formulation
component
P299
Other
U027
Propellants and
blowing agents
Chemical substances used to dissolve or suspend other
substances and either to expel those substances from a
container in the form of an aerosol or to impart a cellular
structure to plastics, rubber, or thermo set resins.
Examples include compressed gasses and liquids and
substances which release ammonia, carbon dioxide, or
nitrogen.
3.2.b
Processing: As a
formulation
component
P299
Other
U028
Solid separation
agents
Chemical substances used to promote the separation of
suspended solids from a liquid. Examples include
flotation aids, flocculants, coagulants, dewatering aids,
and drainage aids.
3.2.b
Processing: As a
formulation
component
P299
Other
U999
Other (specify)
Chemical substances used in a way other than those
described by other codes.
3.2.c
Processing: As an
article component
N/A
N/A
PA
Processing-
incorporation into
article
Chemical substance becomes an integral component of an
article distributed for industrial, trade, or consumer use.
3.2.c
Processing: As an
article component
N/A
N/A
U008
Dyes
Chemical substances used to impart color to other
materials or mixtures (i.e., substrates) by penetrating into
the surface of the substrate. Examples types include azo,
anthraquinone, amino azo, aniline, eosin, stilbene, acid,
basic or cationic, reactive, dispersive, and natural dyes.
3.2.c
Processing: As an
article component
N/A
N/A
U009
Fillers
Chemical substances used to provide bulk, increase
strength, increase hardness, or improve resistance to
impact. Fillers incorporated in a matrix reduce production
costs by minimizing the amount of more expensive
substances used in the production of articles. Examples
include calcium carbonate, barium sulfate, silicates, clays,
zinc oxide and aluminum oxide.
3.2.c
Processing: As an
article component
N/A
N/A
U021
Pigments
Chemical substances used to impart color to other
materials or mixtures (i.e., substrates) by attaching
themselves to the surface of the substrate through binding
or adhesion. This code includes fluorescent agents,
luminescent agents, whitening agents, pearlizing agents,
Page 330 of 570
-------
TRI
Section
TRI Description
TRI Sub-
use Code
TRI Sub-use
Code Name
2016
CDR
Code
2016 CDR Code
Name
2016 CDR Functional Use Definition
and opacifiers. Examples include metallic oxides of iron,
titanium, zinc, cobalt, and chromium; metal powder
suspensions; lead chromates; vegetable and animal
products; and synthetic organic pigments.
3.2.c
Processing: As an
article component
N/A
N/A
U034
Paint additives and
coating additives not
described by other
codes
Chemical substances used in a paint or coating
formulation to enhance properties such as water
repellence, increased gloss, improved fade resistance,
ease of application, foam prevention, etc. Examples of
paint additives and coating additives include polyols,
amines, vinyl acetate ethylene emulsions, and aliphatic
polyisocyanates.
3.2.c
Processing: As an
article component
N/A
N/A
U999
Other (specify)
Chemical substances used in a way other than those
described by other codes.
3.2.d
Processing:
Repackaging
N/A
N/A
PK
Processing-
repackaging
Preparation of a chemical substance for distribution in
commerce in a different form, state, or quantity. This
includes transferring the chemical substance from a bulk
container into smaller containers. This definition does not
apply to sites that only relabel or redistribute the
reportable chemical substance without removing the
chemical substance from the container in which it is
received or purchased.
3.2.e
Processing: As an
impurity
N/A
N/A
N/A
N/A
N/A
3.2.f
Processing:
Recycling
N/A
N/A
N/A
N/A
N/A
3.3.a
Otherwise Use: As
a chemical
processing aid
N/A
N/A
U
Use-non incorporative
Activities
Chemical substance is otherwise used (e.g., as a chemical
processing or manufacturing aid).
3.3.a
Otherwise Use: As
a chemical
processing aid
Z101
Process Solvents
U029
Solvents (for cleaning
or degreasing)
Chemical substances used to dissolve oils, greases, and
similar materials from textiles, glassware, metal surfaces,
and other articles. Examples include trichloroethylene,
perchloroethylene, methylene chloride, liquid carbon
dioxide, and n-propyl bromide.
Page 331 of 570
-------
TRI
Section
TRI Description
TRI Sub-
use Code
TRI Sub-use
Code Name
2016
CDR
Code
2016 CDR Code
Name
2016 CDR Functional Use Definition
3.3.a
Otherwise Use: As
a chemical
processing aid
Z102
Catalysts
U020
Photosensitive
chemicals
Chemical substances used for their ability to alter their
physical or chemical structure through absorption of light,
resulting in the emission of light, dissociation,
discoloration, or other chemical reaction. Examples
include sensitizers, fluorescents, photovoltaic agents,
ultraviolet absorbers, and ultraviolet stabilizers.
3.3.a
Otherwise Use: As
a chemical
processing aid
Z102
Catalysts
U025
Processing aids,
specific to petroleum
production
Chemical substances added to water-, oil-, or synthetic
drilling muds or other petroleum production fluids to
control viscosity, foaming, corrosion, alkalinity and pH,
microbiological growth, hydrate formation, etc., during
the production of oil, gas, and other products from
beneath the earth's surface.
3.3.a
Otherwise Use: As
a chemical
processing aid
Z102
Catalysts
U026
Processing aids, not
otherwise listed
Chemical substances used to improve the processing
characteristics or the operation of process equipment or to
alter or buffer the pH of the substance or mixture, when
added to a process or to a substance or mixture to be
processed. Processing agents do not become a part of the
reaction product and are not intended to affect the
function of a substance or article created. Examples
include buffers, dehumidifiers, dehydrating agents,
sequestering agents, and chelators.
3.3.a
Otherwise Use: As
a chemical
processing aid
Z103
Inhibitors
U024
Process regulators
Chemical substances used to change the rate of a
chemical reaction, start or stop the reaction, or otherwise
influence the course of the reaction. Process regulators
may be consumed or become part of the reaction product.
3.3.a
Otherwise Use: As
a chemical
processing aid
Z103
Inhibitors
U025
Processing aids,
specific to petroleum
production
Chemical substances added to water-, oil-, or synthetic
drilling muds or other petroleum production fluids to
control viscosity, foaming, corrosion, alkalinity and pH,
microbiological growth, hydrate formation, etc., during
the production of oil, gas, and other products from
beneath the earth's surface.
3.3.a
Otherwise Use: As
a chemical
processing aid
Z103
Inhibitors
U026
Processing aids, not
otherwise listed
Chemical substances used to improve the processing
characteristics or the operation of process equipment or to
alter or buffer the pH of the substance or mixture, when
Page 332 of 570
-------
TRI
Section
TRI Description
TRI Sub-
use Code
TRI Sub-use
Code Name
2016
CDR
Code
2016 CDR Code
Name
2016 CDR Functional Use Definition
added to a process or to a substance or mixture to be
processed. Processing agents do not become a part of the
reaction product and are not intended to affect the
function of a substance or article created. Examples
include buffers, dehumidifiers, dehydrating agents,
sequestering agents, and chelators.
3.3.a
Otherwise Use: As
a chemical
processing aid
Z104
Initiators
U024
Process regulators
Chemical substances used to change the rate of a
chemical reaction, start or stop the reaction, or otherwise
influence the course of the reaction. Process regulators
may be consumed or become part of the reaction product.
3.3.a
Otherwise Use: As
a chemical
processing aid
Z104
Initiators
U025
Processing aids,
specific to petroleum
production
Chemical substances added to water-, oil-, or synthetic
drilling muds or other petroleum production fluids to
control viscosity, foaming, corrosion, alkalinity and pH,
microbiological growth, hydrate formation, etc., during
the production of oil, gas, and other products from
beneath the earth's surface.
3.3.a
Otherwise Use: As
a chemical
processing aid
Z104
Initiators
U026
Processing aids, not
otherwise listed
Chemical substances used to improve the processing
characteristics or the operation of process equipment or to
alter or buffer the pH of the substance or mixture, when
added to a process or to a substance or mixture to be
processed. Processing agents do not become a part of the
reaction product and are not intended to affect the
function of a substance or article created. Examples
include buffers, dehumidifiers, dehydrating agents,
sequestering agents, and chelators.
3.3.a
Otherwise Use: As
a chemical
processing aid
Z105
Reaction
Terminators
U024
Process regulators
Chemical substances used to change the rate of a
chemical reaction, start or stop the reaction, or otherwise
influence the course of the reaction. Process regulators
may be consumed or become part of the reaction product.
3.3.a
Otherwise Use: As
a chemical
processing aid
Z105
Reaction
Terminators
U025
Processing aids,
specific to petroleum
production
Chemical substances added to water-, oil-, or synthetic
drilling muds or other petroleum production fluids to
control viscosity, foaming, corrosion, alkalinity and pH,
microbiological growth, hydrate formation, etc., during
Page 333 of 570
-------
TRI
Section
TRI Description
TRI Sub-
use Code
TRI Sub-use
Code Name
2016
CDR
Code
2016 CDR Code
Name
2016 CDR Functional Use Definition
the production of oil, gas, and other products from
beneath the earth's surface.
3.3.a
Otherwise Use: As
a chemical
processing aid
Z105
Reaction
Terminators
U026
Processing aids, not
otherwise listed
Chemical substances used to improve the processing
characteristics or the operation of process equipment or to
alter or buffer the pH of the substance or mixture, when
added to a process or to a substance or mixture to be
processed. Processing agents do not become a part of the
reaction product and are not intended to affect the
function of a substance or article created. Examples
include buffers, dehumidifiers, dehydrating agents,
sequestering agents, and chelators.
3.3.a
Otherwise Use: As
a chemical
processing aid
Z106
Solution Buffers
U026
Processing aids, not
otherwise listed
Chemical substances used to improve the processing
characteristics or the operation of process equipment or to
alter or buffer the pH of the substance or mixture, when
added to a process or to a substance or mixture to be
processed. Processing agents do not become a part of the
reaction product and are not intended to affect the
function of a substance or article created. Examples
include buffers, dehumidifiers, dehydrating agents,
sequestering agents, and chelators.
3.3.a
Otherwise Use: As
a chemical
processing aid
Z199
Other
U002
Adhesives and sealant
chemicals
Chemical substances used to promote bonding between
other substances, promote adhesion of surfaces, or
prevent seepage of moisture or air. Examples include
epoxides, isocyanates, acrylamides, phenol, urea,
melamine, and formaldehyde.
3.3.a
Otherwise Use: As
a chemical
processing aid
Z199
Other
U006
Bleaching agents
Chemical substances used to lighten or whiten a substrate
through chemical reaction, usually an oxidative process
which degrades the color system. Examples generally fall
into one of two groups: chlorine containing bleaching
agents (e.g., chlorine, hypochlorites, N-chloro compounds
and chlorine dioxide); and peroxygen bleaching agents
(e.g., hydrogen peroxide, potassium permanganate, and
sodium perborate).
Page 334 of 570
-------
TRI
Section
TRI Description
TRI Sub-
use Code
TRI Sub-use
Code Name
2016
CDR
Code
2016 CDR Code
Name
2016 CDR Functional Use Definition
3.3.a
Otherwise Use: As
a chemical
processing aid
Z199
Other
U018
Odor agents
Chemical substances used to control odors, remove odors,
mask odors, or impart odors. Examples include
benzenoids, terpenes and terpenoids, musk chemicals,
aliphatic aldehydes, aliphatic cyanides, and mercaptans.
3.3.a
Otherwise Use: As
a chemical
processing aid
Z199
Other
U023
Plating agents and
surface treating agents
Chemical substances applied to metal, plastic, or other
surfaces to alter physical or chemical properties of the
surface. Examples include metal surface treating agents,
strippers, etchants, rust and tarnish removers, and
descaling agents.
3.3.a
Otherwise Use: As
a chemical
processing aid
Z199
Other
U025
Processing aids,
specific to petroleum
production
Chemical substances added to water-, oil-, or synthetic
drilling muds or other petroleum production fluids to
control viscosity, foaming, corrosion, alkalinity and pH,
microbiological growth, hydrate formation, etc., during
the production of oil, gas, and other products from
beneath the earth's surface.
3.3.a
Otherwise Use: As
a chemical
processing aid
Z199
Other
U026
Processing aids, not
otherwise listed
Chemical substances used to improve the processing
characteristics or the operation of process equipment or to
alter or buffer the pH of the substance or mixture, when
added to a process or to a substance or mixture to be
processed. Processing agents do not become a part of the
reaction product and are not intended to affect the
function of a substance or article created. Examples
include buffers, dehumidifiers, dehydrating agents,
sequestering agents, and chelators.
3.3.a
Otherwise Use: As
a chemical
processing aid
Z199
Other
U028
Solid separation
agents
Chemical substances used to promote the separation of
suspended solids from a liquid. Examples include
flotation aids, flocculants, coagulants, dewatering aids,
and drainage aids.
3.3.b
Otherwise Use: As
a manufacturing aid
N/A
N/A
U
Use-non
incorporative
Activities
Chemical substance is otherwise used (e.g., as a chemical
processing or manufacturing aid).
3.3.b
Otherwise Use: As
a manufacturing aid
Z201
Process
Lubricants
U017
Lubricants and
lubricant additives
Chemical substances used to reduce friction, heat, or wear
between moving parts or adjacent solid surfaces, or that
enhance the lubricity of other substances. Examples of
Page 335 of 570
-------
TRI
Section
TRI Description
TRI Sub-
use Code
TRI Sub-use
Code Name
2016
CDR
Code
2016 CDR Code
Name
2016 CDR Functional Use Definition
lubricants include mineral oils, silicate and phosphate
esters, silicone oil, greases, and solid film lubricants such
as graphite and PTFE. Examples of lubricant additives
include molybdenum disulphide and tungsten disulphide.
3.3.b
Otherwise Use: As
a manufacturing aid
Z202
Metalworking
Fluids
U007
Corrosion inhibitors
and antiscaling agents
Chemical substances used to prevent or retard corrosion
or the formation of scale. Examples include
phenylenediamine, chromates, nitrates, phosphates, and
hydrazine.
3.3.b
Otherwise Use: As
a manufacturing aid
Z202
Metalworking
Fluids
U014
Functional fluids
(open systems)
Liquid or gaseous chemical substances used for one or
more operational properties in an open system. Examples
include antifreezes and de-icing fluids such as ethylene
and propylene glycol, sodium formate, potassium acetate,
and sodium acetate. This code also includes substances
incorporated into metal working fluids.
3.3.b
Otherwise Use: As
a manufacturing aid
Z203
Coolants
U013
Functional fluids
(closed systems)
Liquid or gaseous chemical substances used for one or
more operational properties in a closed system. Examples
include: heat transfer agents (e.g., coolants and
refrigerants) such as polyalkylene glycols, silicone oils,
liquified propane, and carbon dioxide;
hydraulic/transmission fluids such as mineral oils,
organophosphate esters, silicone, and propylene glycol;
and dielectric fluids such as mineral insulating oil and
high flash point kerosene. This code does not include
fluids used as lubricants.
3.3.b
Otherwise Use: As
a manufacturing aid
Z204
Refrigerants
U013
Functional fluids
(closed systems)
Liquid or gaseous chemical substances used for one or
more operational properties in a closed system. Examples
include: heat transfer agents (e.g., coolants and
refrigerants) such as polyalkylene glycols, silicone oils,
liquified propane, and carbon dioxide;
hydraulic/transmission fluids such as mineral oils,
organophosphate esters, silicone, and propylene glycol;
and dielectric fluids such as mineral insulating oil and
high flash point kerosene. This code does not include
fluids used as lubricants.
Page 336 of 570
-------
TRI
Section
TRI Description
TRI Sub-
use Code
TRI Sub-use
Code Name
2016
CDR
Code
2016 CDR Code
Name
2016 CDR Functional Use Definition
3.3.b
Otherwise Use: As
a manufacturing aid
Z205
Hydraulic Fluids
U013
Functional fluids
(closed systems)
Liquid or gaseous chemical substances used for one or
more operational properties in a closed system. Examples
include: heat transfer agents (e.g., coolants and
refrigerants) such as polyalkylene glycols, silicone oils,
liquified propane, and carbon dioxide;
hydraulic/transmission fluids such as mineral oils,
organophosphate esters, silicone, and propylene glycol;
and dielectric fluids such as mineral insulating oil and
high flash point kerosene. This code does not include
fluids used as lubricants.
3.3.b
Otherwise Use: As
a manufacturing aid
Z299
Other
U013
Functional fluids
(closed systems)
Liquid or gaseous chemical substances used for one or
more operational properties in a closed system. Examples
include: heat transfer agents (e.g., coolants and
refrigerants) such as polyalkylene glycols, silicone oils,
liquified propane, and carbon dioxide;
hydraulic/transmission fluids such as mineral oils,
organophosphate esters, silicone, and propylene glycol;
and dielectric fluids such as mineral insulating oil and
high flash point kerosene. This code does not include
fluids used as lubricants.
3.3.b
Otherwise Use: As
a manufacturing aid
Z299
Other
U023
Plating agents and
surface treating agents
Chemical substances applied to metal, plastic, or other
surfaces to alter physical or chemical properties of the
surface. Examples include metal surface treating agents,
strippers, etchants, rust and tarnish removers, and
descaling agents.
3.3.c
Otherwise Use:
Ancillary or other
use
N/A
N/A
U
Use-non
incorporative
Activities
Chemical substance is otherwise used (e.g., as a chemical
processing or manufacturing aid).
3.3.c
Otherwise Use:
Ancillary or other
use
Z301
Cleaner
U007
Corrosion inhibitors
and antiscaling agents
Chemical substances used to prevent or retard corrosion
or the formation of scale. Examples include
phenylenediamine, chromates, nitrates, phosphates, and
hydrazine.
Page 337 of 570
-------
TRI
Section
TRI Description
TRI Sub-
use Code
TRI Sub-use
Code Name
2016
CDR
Code
2016 CDR Code
Name
2016 CDR Functional Use Definition
3.3.c
Otherwise Use:
Ancillary or other
use
Z301
Cleaner
U029
Solvents (for cleaning
or degreasing)
Chemical substances used to dissolve oils, greases, and
similar materials from textiles, glassware, metal surfaces,
and other articles. Examples include trichloroethylene,
perchloroethylene, methylene chloride, liquid carbon
dioxide, and n-propyl bromide.
3.3.c
Otherwise Use:
Ancillary or other
use
Z302
Degreaser
U003
Adsorbents and
Absorbents
Chemical substances used to retain other substances by
accumulation on their surface or by assimilation.
Examples of adsorbents include silica gel, activated
alumina, and activated carbon. Examples of absorbents
include straw oil, alkaline solutions, and kerosene.
3.3.c
Otherwise Use:
Ancillary or other
use
Z302
Degreaser
U029
Solvents (for cleaning
or degreasing)
Chemical substances used to dissolve oils, greases, and
similar materials from textiles, glassware, metal surfaces,
and other articles. Examples include trichloroethylene,
perchloroethylene, methylene chloride, liquid carbon
dioxide, and n-propyl bromide.
3.3.c
Otherwise Use:
Ancillary or other
use
Z303
Lubricant
U017
Lubricants and
lubricant additives
Chemical substances used to reduce friction, heat, or wear
between moving parts or adjacent solid surfaces, or that
enhance the lubricity of other substances. Examples of
lubricants include mineral oils, silicate and phosphate
esters, silicone oil, greases, and solid film lubricants such
as graphite and PTFE. Examples of lubricant additives
include molybdenum disulphide and tungsten disulphide.
3.3.c
Otherwise Use:
Ancillary or other
use
Z304
Fuel
U012
Fuels and fuel
additives
Chemical substances used to create mechanical or thermal
energy through chemical reactions, or which are added to
a fuel for the purpose of controlling the rate of reaction or
limiting the production of undesirable combustion
products, or which provide other benefits such as
corrosion inhibition, lubrication, or detergency. Examples
of fuels include coal, oil, gasoline, and various grades of
diesel fuel. Examples of fuel additives include
oxygenated compound such as ethers and alcohols,
antioxidants such as phenylenediamines and hindered
phenols, corrosion inhibitors such as carboxylic acids,
Page 338 of 570
-------
TRI
Section
TRI Description
TRI Sub-
use Code
TRI Sub-use
Code Name
2016
CDR
Code
2016 CDR Code
Name
2016 CDR Functional Use Definition
amines, and amine salts, and blending agents such as
ethanol.
3.3.c
Otherwise Use:
Ancillary or other
use
Z305
Flame Retardant
U011
Flame retardants
Chemical substances used on the surface of or
incorporated into combustible materials to reduce or
eliminate their tendency to ignite when exposed to heat or
a flame for a short period of time. Examples include
inorganic salts, chlorinated or brominated organic
compounds, and organic phosphates/phosphonates.
3.3.c
Otherwise Use:
Ancillary or other
use
Z306
Waste
Treatment
U006
Bleaching agents
Chemical substances used to lighten or whiten a substrate
through chemical reaction, usually an oxidative process
which degrades the color system. Examples generally fall
into one of two groups: chlorine containing bleaching
agents (e.g., chlorine, hypochlorites, N-chloro compounds
and chlorine dioxide); and peroxygen bleaching agents
(e.g., hydrogen peroxide, potassium permanganate, and
sodium perborate).
3.3.c
Otherwise Use:
Ancillary or other
use
Z306
Waste
Treatment
U018
Odor agents
Chemical substances used to control odors, remove odors,
mask odors, or impart odors. Examples include
benzenoids, terpenes and terpenoids, musk chemicals,
aliphatic aldehydes, aliphatic cyanides, and mercaptans.
3.3.c
Otherwise Use:
Ancillary or other
use
Z306
Waste
Treatment
U019
Oxidizing/reducing
agent
Chemical substances used to alter the valence state of
another substance by donating or accepting electrons or
by the addition or removal of hydrogen to a substance.
Examples of oxidizing agents include nitric acid,
perchlorates, hexavalent chromium compounds, and
peroxydisulfuric acid salts. Examples of reducing agents
include hydrazine, sodium thiosulfate, and coke produced
from coal.
3.3.c
Otherwise Use:
Ancillary or other
use
Z306
Waste
Treatment
U028
Solid separation
agents
Chemical substances used to promote the separation of
suspended solids from a liquid. Examples include
flotation aids, flocculants, coagulants, dewatering aids,
and drainage aids.
Page 339 of 570
-------
TRI
Section
TRI Description
TRI Sub-
use Code
TRI Sub-use
Code Name
2016
CDR
Code
2016 CDR Code
Name
2016 CDR Functional Use Definition
3.3.c
Otherwise Use:
Ancillary or other
use
Z307
Water Treatment
U006
Bleaching agents
Chemical substances used to lighten or whiten a substrate
through chemical reaction, usually an oxidative process
which degrades the color system. Examples generally fall
into one of two groups: chlorine containing bleaching
agents (e.g., chlorine, hypochlorites, N-chloro compounds
and chlorine dioxide); and, peroxygen bleaching agents
(e.g., hydrogen peroxide, potassium permanganate, and
sodium perborate).
3.3.c
Otherwise Use:
Ancillary or other
use
Z307
Water Treatment
U018
Odor agents
Chemical substances used to control odors, remove odors,
mask odors, or impart odors. Examples include
benzenoids, terpenes and terpenoids, musk chemicals,
aliphatic aldehydes, aliphatic cyanides, and mercaptans.
3.3.c
Otherwise Use:
Ancillary or other
use
Z307
Water Treatment
U019
Oxidizing/reducing
agent
Chemical substances used to alter the valence state of
another substance by donating or accepting electrons or
by the addition or removal of hydrogen to a substance.
Examples of oxidizing agents include nitric acid,
perchlorates, hexavalent chromium compounds, and
peroxydisulfuric acid salts. Examples of reducing agents
include hydrazine, sodium thiosulfate, and coke produced
from coal.
3.3.c
Otherwise Use:
Ancillary or other
use
Z307
Water Treatment
U028
Solid separation
agents
Chemical substances used to promote the separation of
suspended solids from a liquid. Examples include
flotation aids, flocculants, coagulants, dewatering aids,
and drainage aids.
3.3.c
Otherwise Use:
Ancillary or other
use
Z308
Construction
Materials
N/A
N/A
N/A
3.3.c
Otherwise Use:
Ancillary or other
use
Z399
Other
U001
Abrasives
Chemical substances used to wear down or polish
surfaces by rubbing against the surface. Examples include
sandstones, pumice, silex, quartz, silicates, aluminum
oxides, and glass.
3.3.c
Otherwise Use:
Ancillary or other
use
Z399
Other
U013
Functional fluids
(closed systems)
Liquid or gaseous chemical substances used for one or
more operational properties in a closed system. Examples
include: heat transfer agents (e.g., coolants and
Page 340 of 570
-------
TRI
Section
TRI Description
TRI Sub-
use Code
TRI Sub-use
Code Name
2016
CDR
Code
2016 CDR Code
Name
2016 CDR Functional Use Definition
refrigerants) such as polyalkylene glycols, silicone oils,
liquified propane, and carbon dioxide;
hydraulic/transmission fluids such as mineral oils,
organophosphate esters, silicone, and propylene glycol;
and dielectric fluids such as mineral insulating oil and
high flash point kerosene. This code does not include
fluids used as lubricants.
3.3.c
Otherwise Use:
Ancillary or other
use
Z399
Other
U014
Functional fluids
(open systems)
Liquid or gaseous chemical substances used for one or
more operational properties in an open system. Examples
include antifreezes and de-icing fluids such as ethylene
and propylene glycol, sodium formate, potassium acetate,
and, sodium acetate. This code also includes substances
incorporated into metal working fluids.
3.3.c
Otherwise Use:
Ancillary or other
use
Z399
Other
U018
Odor agents
Chemical substances used to control odors, remove odors,
mask odors, or impart odors. Examples include
benzenoids, terpenes and terpenoids, musk chemicals,
aliphatic aldehydes, aliphatic cyanides, and mercaptans.
3.3.c
Otherwise Use:
Ancillary or other
use
Z399
Other
U020
Photosensitive
chemicals
Chemical substances used for their ability to alter their
physical or chemical structure through absorption of light,
resulting in the emission of light, dissociation,
discoloration, or other chemical reaction. Examples
include sensitizers, fluorescents, photovoltaic agents,
ultraviolet absorbers, and ultraviolet stabilizers.
3.3.c
Otherwise Use:
Ancillary or other
use
Z399
Other
U023
Plating agents and
surface treating agents
Chemical substances applied to metal, plastic, or other
surfaces to alter physical or chemical properties of the
surface. Examples include metal surface treating agents,
strippers, etchants, rust and tarnish removers, and
descaling agents.
Page 341 of 570
-------
E.10 Developing Models that Use Monte Carlo Methods
This appendix provides background information on Monte Carlo methods, including an overview of
deterministic and stochastic processes, an overview of the implementation of Monte Carlo methods, and
a discussion of EPA's approach for building models that utilized Monte Carlo methods.
This appendix is only intended to provide general background information; information related to the
specific models for which EPA implemented Monte Carlo methods is included in Appendices E. 11
through E.13 and Appendices F.5 through F.9.
E.10.1 Background on Monte Carlo Methods
A deterministic process has a single output (or set of outputs) for a given input (or set of inputs). The
process does not involve randomness and the direction of the process is known.
In contrast, stochastic processes are non-deterministic. The output is based on random trials and can
proceed via multiple, or even infinite, directions.
Monte Carlo methods fall under the umbrella of stochastic modeling. Monte Carlo methods are a
replication technique for propagating uncertainty through a model. The model is run multiple times, and
each run uses different input values and generates different output values: each run is independent of
each other. The sample of output values is used to estimate the properties of the actual probability
distribution of the outputs.
E.10.2 Implementation of Monte Carlo Methods
The implementation of Monte Carlo methods generally follows the following steps:
1. Define probability distributions for input parameters.
2. Generate a set of input values by randomly drawing a sample from each probability distribution.
3. Execute the deterministic model calculations.
4. Save the output results.
5. Repeat steps 2 through 4 through the desired number of iterations.
6. Aggregate the saved output results and calculate statistics.
Figure Apx E-l illustrates a flowchart of a Monte Carlo method implemented in a Microsoft Excel-
based model using a Monte Carlo add-in tool, such as the Palisade @Risk software.
Page 342 of 570
-------
Results are Stored until
Desired Iterations are
Met
Iterations
Met
Results Summary
and Descriptive
Statistics
Define
Inputs
Monte Carlo
Add-In Tool
Randomly
Selected
Inputs
Deterministic Model
Outputs
FigureApx E-l. Flowchart of a Monte Carlo Method Implemented in a Microsoft Excel-Based
Model Using a Monte Carlo Add-In Tool
E.10.3 Building the Model
The steps for building a release or exposure model that incorporates Monte Carlo methods are as
follows:
1. Build the deterministic model.
2. Define probability distributions for input parameters.
3. Select model outputs for aggregation of simulation results.
4. Select simulation settings and run model.
5. Aggregate the simulation results and calculate output statistics.
Each of these steps is discussed in the subsections below.
E. 10.3.1 Build the Deterministic Model
First, the model is built as a deterministic model. EPA uses Microsoft Excel in order to use Palisade's
@Risk software that is used for probabilistic analyses in Excel. The model parameters and equations are
programmed into the spreadsheet. Model parameters are programmed in a summary table format for
transparency and to aid in the assignment of probability di stributions. Such summary tables are included
in the model-specific write-ups in Appendices E, 11 through E. 13 and Appendices F.5 through F.9.
E.10.3.2 Define Probability Distributions for Input Parameters
Defining a probability distribution for an input parameter generally involves three steps:
1. Select the model input parameters for which probability distributions will be developed.
2. Determine a probability distribution from the available data.
3. Investigate if any parameters are statistically correlated. Define a statistical correlation among
parameters if a correlation is desired.
Step 1: Select Input Parameters for Probability Distribution Development
When selecting parameters for which probability distributions will be developed, the following factors
are considered:
• The availability of data to inform a distribution.
• The dependency of the input parameters on one another.
The sensitivity of the model results to each input parameter.
Page 343 of 570
-------
Availability of Data to Inform a Distribution: Data sources to investigate for available data to inform
probability distributions of model inputs include but are not limited to the following:
• EPA Generic Scenarios,
• OECD Emission Scenario Documents,
• Peer reviewed literature,
• Published chemical assessments, and
• Other gray literature.18
Model parameters may vary greatly in their available data. There may be a single study that provides
detailed measurements or observation data. There may be multiple studies that provide limited
measurements or observations. There may be only overall statistics available for a parameter. For a
given model development, the available data goes through a systematic review process to evaluate the
data quality, integrate the data, and decide how to use the data.
Dependency of Input Parameters on One Another: The model parameters are evaluated for any
dependency on each other. When each varied parameter is sampled according to its defined probability
distribution, they are sampled independently of each other. Therefore, the value of a sampled parameter
should be independent of the other sampled parameters. An exception is if a statistical correlation is
desired among two or more parameters. Correlating sampled parameters is discussed below in Step 3.
An example of dependency is the relationship between a facility's number of operating days, annual
production volume (PV), and daily PV. These three parameters are not all independent of each other.
The annual PV may be calculated from the daily PV and the operating days. Alternatively, the daily PV
may be calculated from the annual PV and the operating days. Additionally, operating days may be
calculated from the annual PV and daily PV. It is necessary to first understand the mathematical
relationship among these parameters before selecting parameters for which probability distributions will
be developed.
Sensitivity of the Model Results to Each Input Parameter: One consideration in selecting model
parameters for probability distribution development is the sensitivity of the model outputs to each
parameter. A sensitivity analysis can inform how sensitive each model output is to each model input
parameter. EPA may choose to prioritize probability distribution development for parameters to which
model outputs are more sensitive. Since the model outputs are more sensitive to these parameters, it
would be more important to capture variability and/or uncertainty for these parameters compared to
parameters to which model outputs are less sensitive.
A sensitivity analysis is conducted by varying each desired parameter and performing a Monte Carlo
simulation. The varied range for each parameter should be consistent with the expected range in values
for the parameter. The @Risk software (Palisade. 2022b) can perform sensitivity analyses. The statistic
of the outputs for which sensitivity is measured, such as mean, mode, or a percentile, can be selected. As
the simulation is run, the software tracks how each output changes with respect to each varied input.
18 Gray literature is defined as the broad category of data/information sources not found in standard, peer-reviewed literature
databases. Gray literature includes data/information sources such as white papers, conference proceedings, technical reports,
reference books, dissertations, information on various stakeholder websites, and various databases.
Page 344 of 570
-------
Step 2: Determine a Probability Distribution
To determine a probability distribution, first, all the information known about the parameter is evaluated
(Oracle. 2017). The following considerations can help guide summarizing important information about
the parameter (Analvtica. ^ ):
• Discrete or continuous
o Consider whether the parameter is discrete or continuous. Does the parameter have a
finite or countable number of possible values? Is the parameter logical or Boolean such as
having possible values of "yes or no" or "true or false"? Can the parameter be
represented by all real numbers within a domain?
• Bounds
o Consider whether the parameter has bounds. A parameter may have a lower bound and/or
an upper bound. Alternatively, a parameter may be unbounded and can range to negative
and/or positive infinity.
• Modes
o Consider whether the parameter has one or more modes. Does the parameter have no
mode (such as represented by a uniform distribution)? If it has a mode, is it unimodal or
multimodal? If multimodal, is the parameter a combination of two or more populations?
In which case, the parameter may be best separated into its separate components and then
develop probability distributions for the individual components.
• Symmetric or skewed
o Consider whether the parameter is symmetric or skewed. If skewed, consider whether the
parameter is positively skewed (thicker upper tail) or negatively skewed (thicker lower
tail).
Second, review standard probability distributions and identify possible candidates that meet the
considerations identified in the first step (Oracle. 2017). The following are common probability
distributions:
• Uniform distribution
o A uniform distribution has finite upper and lower bounds and all values between the
bounds have equal probability.
• Triangular distribution
o A triangular distribution has finite upper and lower bounds and a modal value. The modal
value is the value that occurs most frequently. If the most frequent value is not known
another statistic, such as the mean or a percentile, could be used to define the triangular
distribution.
• Normal distribution
o The parameters of a normal distribution are its mean and standard deviation. A normal
distribution is unbounded, and values range from negative to positive infinity. If desired,
the range of values of a normal distribution may be truncated to finite bounds to prevent
unrealistic values from being sampled.
• Lognormal distribution
o If a variable is lognormally distributed, it means that the logarithm of that variable is
normally distributed. The parameters of a lognormal distribution are its mean and
standard deviation. A lognormal distribution is bounded from zero to positive infinity. A
lognormal distribution may be shifted and its upper bound truncated to fit the observed
data and prevent unrealistic values from being sampled.
Page 345 of 570
-------
Lastly, select the best suited probability distribution (Oracle. 2017). Review the available data for the
parameter to determine how to define the distribution's parameters. For example, if the only available
data are an overall range (with a minimum and a maximum), then a uniform distribution is the
appropriate distribution to use. If the only available data are an overall range and a mode, then a
triangular distribution is the appropriate distribution to use. If historical data for the parameter are
available, consider data fitting to determine the appropriate distribution and regress the distribution
parameter values.
Step 3: Check for and Define Statistical Correlations
When developing a Monte Carlo model and setting statistical distributions for parameters, EPA
evaluates possible correlations among parameters. When distributions are defined for the parameters,
each parameter is independently sampled on each iteration of the model. This may result in
combinations of parameter values that are not logical for the scenario. In the example of a model that
uses annual PV, daily PV, and operating days as parameters, there are set distributions for annual PV
and operating days, with the daily production volume calculated from the other two parameters. But
annual PV and operating days may be correlated. For example, if a site has a fixed manufacturing
capacity (as determined by the equipment size and production lines), then annual PV is a function of the
number of operating days. A facility is more likely to scale-up or scale-down their annual PV by varying
the operating days rather than varying their daily PV. Varying annual PV and operating days
independently in the model may arrive at value combinations that are not logical. For example, one
iteration may sample a high annual PV value with a low number of operating days that may result in a
high daily production rate that is not logical. In this example, a different probability distribution strategy
may be appropriate, such as defining probability distributions for daily PV and operating days since
those two parameters are likely more independent of each other than annual PV and operating days.
When developing distributions from observed data, there are statistical tests that can be performed to
indicate a statistical correlation. Two common ones are: 1) the Pearson product-moment correlation
coefficient, which measures the linear correlation between two data sets; and 2) Spearman's rank
correlation coefficient, which is a measure of rank correlation and how well a relationship between two
data sets can be described using a monotonic function. A monotonic relationship is one where the two
variables change together but not necessarily at a constant rate (Minitab. 2022). A linear correlation is
necessarily monotonic. But a monotonic correlation is not necessarily linear.
Both the Pearson and Spearman coefficients range from -1 to +1. A value close to ±1 indicates a strong
correlation (either positive or negative). A positive correlation means as one variable increases, the other
also increases. A negative correlation means as one variable increases, the other decreases. A value close
to 0 means a weak or no correlation exists between the variables. The Pearson correlation only measures
linear relationships, and the Spearman correlation only measures monotonic relationships. If two
variables are correlated by a relationship that is neither linear nor monotonic, then the Pearson and
Spearman coefficients would not be informative of the nature of the correlation (Minitab. 2022).
After testing for statistical correlations, statistical correlations can be defined for input parameters using
@Risk. @Risk only uses Spearman coefficients to define statistical correlations among input
parameters. Spearman coefficients to correlate two or more input parameters are defined through a
correlation matrix. The correlation matrix allows the Spearman coefficient to be defined for each pair of
correlated input parameters (Palisade. 2022a).
Page 346 of 570
-------
E.10.3.3 Select Model Outputs for Aggregation of Simulation Results
The last step before running the model is to select the model outputs for which statistical results are
desired. Defining these outputs in @Risk will allow the software to save the output results from each
iteration and aggregate the simulation results over all iterations together.
E.10.3.4 Select Simulation Settings and Run Model
Simulation settings must be defined before running the model. Important simulation settings include the
number of iterations, the sampling method, and the random number generator.
• Number of Iterations. Generally speaking, a larger number of iterations is desired to ensure
adequate sampling and representation of lower probability events. The number of iterations to
achieve a desired margin of error for a given confidence interval for an output can be calculated
using the Central Limit Theorem (OJ>n I'1 JO I \ <\tlisade. 2015a). The equation shows that the
margin of error is inversely proportional to the square root of the number of iterations. Therefore,
the greater the number of iterations, the smaller the margin of error. Calculating the number of
iterations can be difficult as the sample standard deviation is not known beforehand. EPA
typically uses 100,000 iterations to ensure convergence and have minimal cost to the simulation
time.
• Sampling Method. The sampling method is the method used to draw random samples from the
input parameter probability distributions. @Risk uses two methods: Latin Hypercube (the
default) and Monte Carlo. Monte Carlo sampling is a purely random sampling method. This can
lead to clustering and under-representing low probability events. Latin Hypercube sampling is a
stratified sampling method. This ensures the sampled input parameter distribution matches the
assigned probability distribution closely. EPA typically uses Latin Hypercube sampling because
it is efficient and can achieve convergence with fewer iterations than Monte Carlo sampling
(Palisade. 2018).
• Random Number Generator. The random number generator is used to generate pseudorandom
numbers that are used in an algorithm to draw random samples from the probability distributions.
The @Risk default is Mersenne Twister, which is a robust and efficient random number
generator (Palisade. 2015b).
E.10.3.5 Aggregate the Simulation Results and Produce Output Statistics
During the simulation, @Risk will save the defined model outputs for aggregation on each iteration.
After the simulation is completed, EPA can generate desired statistical results and distributions of the
defined outputs. EPA typically uses the 50th percentile and 95th percentile of the output as the central
tendency and high-end estimates, respectively.
E.ll Textile Dye Modeling Approach and Parameters for Estimating
Environmental Releases
This appendix presents the modeling approach and equations used to estimate environmental releases of
1,4-dioxane during the commercial use of textile dyes. This approach utilizes the OECD ESD on Textile
Dyes (OECD. 2017) combined with Monte Carlo simulation (a type of stochastic simulation). This ESD
includes a diagram of release and exposure points during textile dying, as shown in Figure Apx E-2.
Page 347 of 570
-------
Dve Formulation
© Dust Emissions
during Unloading
/Tn Disposal of Spent
Dyebath
Equipment Cleaning
©<§> Container
Residue
Cleaning and/or
Disposal
<Ł> Worker Exposure
During Dyeing
Operation
FigureApx E-2. Environmental Release Points (Numbered) and Occupational Exposure Points
(Lettered) During Textile Dying
Based on Figure Apx E-2, EPA identified the following release points (note that diagram points 3 and 4
were combined for ease of analysis):
• Release point 1 (RP1): Dust emissions of during unloading of solid powders to air, landfill,
POTW, or incineration;
• Release point 2 (RP2): Container residual losses to POTW, landfill, or incineration; and
• Release point 3 (RP3): Release of spent dye bath and equipment cleaning losses to POTW.
Environmental releases of textile dyes are a function of the chemical's physical properties, container
size, mass fractions, and other model parameters. Although physical properties are fixed, some model
parameters are expected to vary from one facility to another. An individual model input parameter could
either have a discrete value or a distribution of values. EPA assigned statistical distributions based on
available literature data or engineering judgment to address the variability in mass fraction of dye
formulation in the bath (Fdye dyebath). container size (Vcontainer), textile production rate (Vfabric), operating
days (OD), and container residue fractions (Fcontainer residue).
A Monte Carlo simulation was then conducted to capture variability in the model input parameters
described above. The simulation was conducted using the Latin hypercube sampling method in @Risk
(Palisade, Ithaca, NY). The Latin hypercube sampling method is a statistical method for generating a
sample of possible values from a multi-dimensional distribution. Latin hypercube sampling is a stratified
method, meaning it guarantees that its generated samples are representative of the probability density
function (variability) defined in the model. EPA performed 100,000 iterations of the model to capture
the range of possible input values, including values with low probability of occurrence.
From the distribution resulting from the Monte Carlo simulation, EPA selected the 95th and 50th
percentile values to represent a high-end release and central tendency release level, respectively. The
statistics were calculated directly in @Risk. The following subsections detail the model design equations
and parameters used for environmental release estimates.
E.ll.l Model Equations
Daily use rate of dye formulation is calculated using the following equation:
Equation Apx E-l.
Qdye_formulation_day ~ ^fabric * ^fabric * ^dye_fabric
Page 348 of 570
-------
Where:
Qdye_formuiation_day = Daily use rate of dye formulation [kg/site-day]
Vfabric = Textile production rate [kg/site-day]
Ffabric = Mass fraction of textiles treated with dye [kg/kg]
Fdye fabric = Mass fraction of dye used per mass of textile dyed [kg/kg]
Daily use rate of 1,4-dioxane formulation is calculated using the following equation:
EquationApx E-2.
Qdioxane_site_day ~ Qdye_formulation_day * Pdioxane_dye * Pdye
Where:
Qdioxane_site_day = Daily use rate of 1,4-dioxane [kg/site-day]
Qdye_formuiation_day = Daily use rate of dye formulation [kg/site-day]
Pdioxane_dye = Mass fraction of 1,4-dioxane in dye formulation [kg/kg]
Fdye = Fraction of dye containing 1,4-dioxane [kg/kg]
Containers emptied per facility is calculated using the following equation:
Equation Apx E-3.
Qdioxane site day * OD
Nr ~ ~ ~
vcontainer_unload_site_yr ~
Pdioxane_dye * ^container * 3.79 * RHOform
Where:
Ncontainer unioad_site_yr = Containers emptied per facility [containers/site-year]
Qdioxane_site_day = Daily use rate of 1,4-dioxane [kg/site-day]
OD = Operating days [days/year]
Pdioxane_dye = Mass fraction of 1,4-dioxane in dye formulation [kg/kg]
^container = Container size [gal/container]
RHOf0rm = Dye density [kg/L]
Container residual fraction is calculated using the below equations. To make the simulation more
realistic, EPA assessed container size based on the dye formulation use rate. This avoids situations
where a small container size is associated with a large use rate, such that an unrealistic number of
containers are used each year, and vice-versa.
Equation Apx E-4.
If Qdye_formulation_day ^700 kg/site-day.
Pcontainer_residue ~ Pcontainer_residue_tote
If Qdye_formulation_day '^ 200-700 kg/site-day.
Pcontainerjresidue ~ Pcontainer_residue_drum
If Qdye_formulation_day 200 kg/site-day
Pcontainer_residue ~ Pcontainer_residue_pail
Page 349 of 570
-------
Where:
Qdye_formulation_day ~
Fcontainer residue ~
Fcontainer _residue_tote ~
Fcontainer _residue_drum~
Fcontainer_residue_pail ~
Daily use rate of dye formulation [kg/site-day]
Container residual fraction [kg/kg]
Container residual fraction for totes [kg/kg]
Container residual fraction for drums [kg/kg]
Container residual fraction for pails [kg/kg]
Mass fraction of 1,4-dioxane in dye bath is calculated using the following equation:
EquationApx E-5.
Where:
Fdioxane_dyebath
Fdioxane_dye
Fdye_dyebath
Fdioxane_dyebath ~ Fdioxane ciye * Fdye_dyebath
Mass fraction of 1,4-dioxane in dye bath [kg/kg]
Mass fraction of 1,4-dioxane in dye formulation [kg/kg]
Mass fraction of the dye formulation in the dyebath [kg/kg]
Release point 2 (container residual) release per day is calculated using the following equation:
Equation Apx E-6.
Release _perDayRP2 = Qd
Where:
Release_perDayRP2 =
Qdioxane_site_day ~
Fr,
container residue
ioxane_site_day * ^containerjresidue
Container residual release from release point 2 [kg/site-day]
Daily use rate of 1,4-dioxane [kg/site-day]
Container residual fraction [kg/kg]
Release point 3 (spent dye bath and equipment cleaning) release per release day is calculated using the
following equation:
Equation Apx E-7.
Release _perD ayRP3 = Qd
ioxane_site_day
(1 ^fixation)
Where:
Release_perDayRP3 =
Qd
ioxane_site_day
Ffixation
Dye bath and equipment cleaning release from point 3 [kg/site-day]
Daily use rate of 1,4-dioxane [kg/site-day]
Fraction of dye affixed to textile during dye process [kg/kg]
E.11.2 Model Input Parameters
Table Apx E-12 summarizes the model parameters and their values for the Monte Carlo simulation.
Additional explanations of EPA's selection of the distributions for each parameter are provided after this
table. High-end and central tendency releases are estimated by selecting the 50th and 95th percentile
values from the output distribution.
Page 350 of 570
-------
Table Apx E-12. Summary of Parameter Values and Distributions Used in the Textile Release Model
Input Parameter
Symbol
Unit
Constant Model
Parameter Values
Variable Model Parameter Values
Rationale/Basis
Value
Lower
Bound
Upper
Bound
Mode
Distribution
Type
Number of Sites
Ns
sites
783
-
-
-
-
See Section E.11.3
Fraction of Dye
Containing 1,4-Dioxane
Fdye
kg/kg
1
-
-
-
-
See Section E.11.4
1,4-Dioxane Vapor
Pressure
VP
Torr
40
-
-
-
-
Physical property
1,4-Dioxane Molecular
Weight
MW
g/mol
88.1
-
-
-
-
Physical property
Operating Days
OD
days/year
312
10
312
-
Discrete
See Section E.11.5
Mass fraction of 1,4-
Dioxane in Dye
Formulation
Fdioxane dye
kg/kg
0.0000047
See Section E.11.6
Textile Production Rate
V fabric
kg/day
9,100
3,250
50,000
9,100
Triangular
See Section E.11.7
Mass Fraction of Textiles
Treated with Dye
F fabric
kg/kg
0.3
-
-
-
-
See Section E.11.8
Mass Fraction of Dye
Used Per Mass of Textile
Dyed
Fdye fabric
kg/kg
0.1
See Section E.11.9
Mass Fraction of the Dye
Formulation in the
Dyebath
Fdye dyebath
kg/kg
0.06
0.002
0.06
0.02
Triangular
See Section E.11.10
Container Size for Dye
Formulation
V container
gal
35
7
264
35
Triangular
See Section E.ll.ll
Dye density
RHOform
kg/L
1
ESD assumes a
density equal to that
of water
Container Residual
Fraction for Totes
Fcontainer residue totes
kg/kg
0.002
0.0007
0.002
0.0007
Triangular
See Section E.11.12
Container Residual
Fraction for Drums
Fcontainer residue drums
kg/kg
0.03
0.0003
0.03
0.025
Triangular
See Section E.11.13
Page 351 of 570
-------
Input Parameter
Symbol
Unit
Constant Model
Parameter Values
Variable Model Parameter Values
Rationale/Basis
Value
Lower
Bound
Upper
Bound
Mode
Distribution
Type
Container Residual
Fraction for Pails
Fcontainer residue_pails
kg/kg
0.006
0.0003
0.006
0.003
Triangular
See Section E.11.14
Fraction of Dye Product
Affixed to Textile During
Dyeing Process Substrate
F fixation
kg/kg
Multiple Triangular Distributions
See Section E.11.15
Page 352 of 570
-------
E.11.3 Number of Sites
EPA did not find data on the number of textile dye sites that specifically use dyes containing 1,4-
dioxane from systematic review. As a bounding estimate, EPA used U.S. Census and BLS data for the
NAICS code 313310, Textiles and Fabric Finishing Mills, to estimate a total of 783 sites within the
industry ( .2016V
E.11.4 Mass Fraction of Dye Containing 1,4-Dioxane
EPA did not identify chemical-specific information for this parameter systematic review; therefore, the
Agency used generic values from the ESD on the Use of Textile Dyes (OECD. 2017). The ESD
provided a single value for the mass fraction of dyes containing the chemical of interest, which is 1,4-
dioxane. The ESD assumes that 100 percent of dyes contain the chemical of interest. Therefore, EPA
could not develop a distribution of values for this parameter and used the single value of 1 kg dye with
1,4-dioxane/kg dye used from the ESD.
E.11.5 Operating Days
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
the Agency used generic values from the ESD on the Use of Textile Dyes (OECD. 2017). The ESD uses
31 data points for number of operating days from past new chemical submissions that were submitted to
EPA from 2006 through 2014. EPA modeled the number of operating days per year using a discrete
distribution comprised of the 31 data points shown in TableApx E-13 with an equal probability for
each data point.
Table Apx E-13. Discrete Data Points on the
Number of Operating Days at Textile Dye Sites
Number of Operating Days (days/yr)
10
75
111
139
200
200
312
28
79
112
166
200
222
-
33
89
115
167
200
250
-
55
93
125
167
200
261
-
72
99
125
200
200
278
-
Source: (OECD, 201.7)
E.11.6 Mass Fraction of 1,4-Dioxane in Dye Formulation
The December 2020 Final Risk Evaluation for 1,4-Dioxane (U.S. EPA. 2020c) indicates that 1,4-
dioxane is a byproduct in dye formulations and provided a single value of 0.0000047 kg 1,4-dioxane/kg
dye. EPA did not identify additional data on the mass fraction of 1,4-dioxane in textile dyes. Therefore,
EPA could not develop a distribution of values for this parameter and used the single value of 0.0000047
kg 1,4-dioxane/kg dye from the risk evaluation.
E.11.7 Textile Production Rate
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
the Agency used generic data from the ESD on the Use of Textile Dyes (( ). The ESD cites
data provided in fabric finishing new chemical submissions during 1993 and 1994. Note that the ESD
uses a "typical" value as default and does not say what the typical is based on (e.g., average, median).
EPA used the range of textile production rates and the default typical value provided in the ESD as the
Page 353 of 570
-------
lower and upper bounds and mode of the triangular distribution for this parameter, respectively.
Specifically, EPA modeled textile production rate using a triangular distribution with a lower bound of
3,520 kg/site-day, and upper bound of 50,000 kg/site-day, and a mode of 9,100 kg/site-day.
E.11.8 Mass Fraction of Textiles Treated with Dye
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
the Agency used generic data from the ESD on the Use of Textile Dyes (( ). The ESD
provided a single value for the mass fraction of all textiles treated with dyes, stating that the median
share of textiles processed per day using the primary dyestuff is 30 percent. Therefore, EPA could not
develop a distribution of values for this parameter and used the single value of 30 percent from the ESD.
E.11.9 Mass Fraction of Dye Used per Mass of Textile Dyed
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
the Agency used generic data from the ESD on the Use of Textile Dyes (OE( ). The ESD
provided a single value for the mass fraction of dye used per mass of textile dyed, stating that as a
"realistic worst case," liquid dye formulations are used in an amount of 10 percent (( ).
Therefore, EPA could not develop a distribution of values for this parameter and used the single value of
0.10 kg dye/kg textiles from the ESD.
E.11.10 Mass Fraction of the Dye Formulation in the Dyebath
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
the Agency used generic data from the ESD on the Use of Textile Dyes (( ). The ESD states
that typical dye concentrations may range from 1.5 to 2.5 percent, lighter shades may be as low as 0.2 to
0.3 percent, and heavier shades may be between 4 to 6 percent. Based on this data, EPA modeled this
parameter using a triangular distribution with the overall range of dye concentrations (0.2 to 6%) and the
mid-range of the typical concentration (2%) provided in the ESD as the lower and upper bounds and
mode of the triangular distribution, respectively.
E.ll.ll Container Size for Dye Formulation
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
the Agency used generic data from the ESD on the Use of Textile Dyes (OE( ). The ESD states
that dyes can be transported in containers ranging from 25 to 1,000 kg, but most are shipped in 35-gallon
drums (OEC ). EPA converted this range from kilograms to gallons using an assumed dye
density of 1 kg/L and a conversion factor of 3.785 L/gal. Based on this data, EPA modeled container
size using a triangular distribution with a lower bound of 7 gallons, an upper bound of 264 gallons, and a
mode of 35 gallons.
E.11.12 Container Residual Fraction for Totes
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
the Agency used generic data from standard EPA models. Specifically, EPA modeled container residual
fraction for totes using a triangular distribution with a lower bound of 0.0007 kg residual/kg dye, and
upper bound of 0.002 kg residual/kg dye, and a mode of 0.0007 kg residual/kg dye. The lower and upper
bounds of this distribution are based on the central tendency and high-end values listed in the
EPA/OPPT Bulk Transport Residual Model from the ChemSTEER User Guide ( ). EPA
used the central tendency value as the mode of the triangular distribution. Note that the underlying data
for this model comes from a 1988 study by PEI Associates Inc. that looked at literature sources and
conducted a pilot-scale experiment to determine the amount of residual material left in containers (PEI
Associates. 1988). EPA reviewed the data from this study and the underlying distribution of the
Page 354 of 570
-------
container residual loss fraction is unknown; therefore, EPA assigned a triangular distribution as
discussed above.
E.11.13 Container Residual Fraction for Drums
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
the Agency used generic data from standard EPA models. Specifically, EPA modeled container residual
fraction for drums using a triangular distribution with a lower bound of 0.0003 kg residual/kg dye, an
upper bound of 0.03 kg residual/kg dye, and a mode of 0.025 kg residual/kg dye. The lower bound is
based on the minimum value for pouring and the upper bound is based on the default high-end value in
the EPA/OPPT Drum Residual Model from the ChemSTEER User Guide (U.S. EPA. 2015a). EPA used
the central tendency value for pumping as the mode of the triangular distribution. Note that the
underlying data for this model comes from a 1988 study by PEI Associates Inc. that looked at literature
sources and conducted a pilot-scale experiment to determine the amount of residual material left in
containers (PEI Associates. 1988). EPA reviewed the data from this study and the underlying
distribution of the container residual loss fraction is unknown; therefore, the Agency assigned a
triangular distribution as discussed above.
E.11.14 Container Residual Fraction for Pails
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
the Agency used generic data from standard EPA models. Specifically, EPA modeled container residual
fraction for small containers using a triangular distribution with a lower bound of 0.0003 kg residual/kg
dye, an upper bound of 0.006 kg residual/kg dye, and a mode of 0.003 kg residual/kg dye. The lower
bound is based on the minimum value for pouring and the upper bound is based on the default high-end
value listed in the EPA/OPPT Small Container Residual Model from the ChemSTEER User Guide (
). EPA used the central tendency value for pouring as the mode of the triangular distribution.
Note that the underlying data for this model comes from a 1988 study by PEI Associates Inc. that looked
at literature sources and conducted a pilot-scale experiment to determine the amount of residual material
left in containers (PEI Associates. 1988). EPA reviewed the data from this study and the underlying
distribution of the container residual loss fraction is unknown; therefore, the Agency assigned a
triangular distribution as discussed above.
E.11.15 Fraction of Dye Product Affixed to Textile During Dyeing Process Substrate
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
the Agency used generic data from the ESD on the Use of Textile Dyes (( ). The ESD on the
Use of Textile Dyes provides a table containing ranges and averages for dye fixation percentages based
on the nine different classes of dyes ( ). Therefore, EPA modeled the fraction of dye product
affixed to textiles during dyeing process substrate using multiple triangular distributions. EPA used the
low-end of the range of dye fixation as the lower bound, the high-end of the range of dye fixation as the
upper bound, and the average dye fixation as the mode for each of the nine triangular distributions. In
the Monte Carlo simulation, each of the nine triangular distributions from the ESD has an equal
probability of being selected and used for the parameter's output. The distribution selection probabilities
and values are shown in Table Apx E-14.
Page 355 of 570
-------
Table Apx E-14.r
"riangular Distributions Ffixation
Dye Type
Dye Fixation (%) Triangular Distribution
Lower Bound
Upper Bound
Mode
Acid
85
98
93
Azoic
76
95
84
Basic
95
100
99
Direct
64
96
88
Disperse
80
100
96
Metal-Containing
82
98
94
Reactive
50
97
85
Sulfur
60
95
70
Vat
70
95
85
Source: fOECD. 2017)
E.11.16 Key Strengths, Limitations, Uncertainties, and Sensitivity Analysis
General modeling uncertainties and limitations are discussed in Section 2.2.1.3 and Appendix E.7. This
section discusses model-specific uncertainties and limitations and presents examples of sensitivity charts
that EPA developed for this model. For this model, the only 1,4-dioxane specific input parameter data is
for the concentration of 1,4-dioxane in textile dyes and only one datapoint was available. All other
parameters are based on generic data from the ESD on the Use of Textile Dyes (OH < * . ^ I ) or
standard EPA/OPPT models described in the ChemSTEER User Guide ( ) This adds
uncertainty with respect to the representativeness of the input data towards textile dying sites that use
dyes containing 1,4-dioxane.
EPA ran a sensitivity analysis in @Risk to identify the input parameters which have the largest impact
on the daily release output estimates. FigureApx E-3 shows the inputs ranked by which have the largest
effect on the mean container cleaning daily release estimate, which is release point 2 in this model.
Figure Apx E-4 similarly shows the inputs that impact the daily release from spent dyebath and
equipment cleaning, which corresponds to release point 3 in this model. The textile production rate has a
relatively large impact on both release points. As discussed in Appendix E. 11.7, EPA used a triangular
distribution based on generic data from the ESD on the Use of Textile Dyes. Similarly for the other
parameters in Figure Apx E-3 and Figure Apx E-4, EPA developed distributions based on generic data.
The chart shows nine different dye fixation parameters because EPA used data for multiple types of
dyes, as discussed in Section E. 11.15. Having a distribution for each input parameter is a strength of the
assessment; however, the representativeness of the underlying generic data used for these distributions is
a limitation, as was discussed above.
Page 356 of 570
-------
Release Point 2
- Daily Release
Inputs Ranked by Effect on Output Mean
Textile production rate
Container residual fraction for Drums
Container residual fraction for Totes
[
Container residual fraction for Pails
!
I
FigureApx E-3. Container Cleaning (Daily Release Point 2) Sensitivity Chart
Release Point 3 -
Daily Release
Inputs Ranked by Effect on Output Mean
Textile production rate
Dye fixation - formula 1
Dye fixation - formula 2
Dye fixation - formula 9
Dye fixation - formula 8
Dye fixation - formula 5
Dye fixation - formula 4
Dye fixation - formula 3
Dye fixation - formula 7
Dye fixation - formula 6
Sampled Formulation Type for Dye Fixation
FigureApx E-4. Spent Dyebath and Equipment Cleaning (Daily Release Point 3)
Sensitivity Chart
Page 357 of 570
-------
E.12 Laundry Detergent Modeling Approach and Parameters for
Estimating Environmental Releases
This appendix presents the modeling approach and equations used to estimate environmental releases of
1,4-dioxane during the industrial and institutional use of laundry detergents. This approach utilizes the
OECD ESD on the Chemicals Used in Water Based Washing Operations at Industrial and Institutional
Laundries (OECD. 2011b) combined with Monte Carlo simulation (a type of stochastic simulation).
This ESD categorizes laundry facilities into either industrial or institutional facilities. Industrial
laundries are off-site laundries that wash soiled linen such as table and bed linens, towels, diapers,
uniforms, gowns, and coats, and industrial coverings such as work uniforms, protective apparel (flame
and heat resistant), clean room apparel, mops, rugs, mats, dust tool covers, cloths, and shop or wiping
towels. Institutional laundries are on-premise laundries and the items laundered will vary by facility,
which are primarily hospitals, nursing homes, and hotels (OECD. 2011b). The ESD includes different
process parameters for industrial and institutional laundry facilities; therefore, the Agency modeled the
two types of laundry facilities separately.
In addition, laundry detergents can be in liquid or powder physical forms. Because the difference in
physical form results in different parameter distributions, EPA modeled liquid and powder detergents
separately. This ESD includes a diagram of release and exposure points during the use of laundry
detergents, as shown in Figure_Apx E-5.
©© ® Container Residue and Cleaning ©© ® Container Residue and Cleaning
©© Fugitive Air Release and Dust ©@ Fugitive An Release and Dust
Emissions During Transfers Emissions During Transfers
©@© Releases During
Operations
FigureApx E-5. Environmental Release Points (Numbered) and Occupational Exposure Points
(Letterd) During Industrial/Institutional Laundering Operations
Based on Figure Apx E-5, EPA identified the following release points:
• Release point 1 (RP1): Container residual losses to POTW, landfill, or incineration;
• Release point 2 (RP2): Fugitive air releases during container cleaning;
• Release point 3 (RP3): Fugitive air releases during container unloading;
• Release point 4 (RP4): Dust releases during container unloading;
o 4a: Uncaptured dust releases;
o 4b: Captured, uncontrolled dust releases;
o 4c: Captured and controlled dust releases;
• Release point 5 (RP5): Fugitive air releases during washing; and
Page 358 of 570
-------
• Release point 6 (RP6): Release from washing and drying operations to fugitive air, stack air, or
POTW.
Environmental releases of laundry detergent are a function of the chemical's physical properties,
container size, mass fractions, and other model parameters. Although physical properties are fixed, some
model parameters are expected to vary from one facility to another. An individual model input
parameter could either have a discrete value or a distribution of values. EPA assigned statistical
distributions based on available literature data or engineering judgment to address the variability in mass
fraction of 1,4-dioxane in the detergent (Fdioxane laundry), container size (Vcontainer), daily use rate of
detergent (Qfaciiity day), air speed (RATEair speed), duration of release (OHCOnt unload), operating days (OD),
container residue fractions (FCOntainer residue), and dust capture/control efficiency (Fdust).
A Monte Carlo simulation was conducted to capture variability in the model input parameters described
above. The simulation was conducted using the Latin hypercube sampling method in @Risk (Palisade,
Ithaca, New York). The Latin hypercube sampling method is a statistical method for generating a
sample of possible values from a multi-dimensional distribution. Latin hypercube sampling is a stratified
method, meaning it guarantees that its generated samples are representative of the probability density
function (variability) defined in the model. EPA performed 100,000 iterations of the model to capture
the range of possible input values, including values with low probability of occurrence.
From the distribution resulting from the Monte Carlo simulation, EPA selected the 95th and 50th
percentile values to represent a high-end release and central tendency release level respectively. The
statistics were calculated directly in @Risk. The following subsections detail the model design equations
and parameters used for environmental release estimates.
E.12.1 Model Equations
Daily use rate selection based on physical form of detergent is based on the following two equations, the
first being for liquid detergent and the second being for powder detergent:
EquationApx E-8.
Qfacility_day ~ Qfacility_day_liquid
or
Qfacility_day ~ Qfacility_day_powder
Where:
Qf acility_day = Daily use rate based on physical form of detergent [kg/site-day]
Qf acility dayUquid = Daily use for liquid form detergent [kg/site-day]
Qf acility_day_powder = Daily use for powder form detergent [kg/site-day]
Daily use rate of laundry detergents containing 1,4-dioxane is calculated using the equation below:
Equation Apx E-9.
Q facility _day_adjusted ~ Qf acility_day * Fformulations_dioxane
Where:
Qfacility_day_adjusted = Daily use rate of detergent containing 1,4-dioxane selected based
on the physical form of the detergent [kg/site-day]
Page 359 of 570
-------
Qfacility_day = Daily use rate based on physical form of detergent [kg/site-day]
Pformuiations_dioxane = Fraction of laundry detergents containing 1,4-dioxane [kg/kg]
Daily use rate of 1,4-dioxane is calculated using the equation below:
EquationApx E-10.
Qdioxane_day ~ Qfacility_day_adjusted * Fdioxanejaundry
Where:
Qdioxane_day = Daily usage rate of 1,4-dioxane [kg/site-day]
Qf acuity .day .adjusted = DailY use rate of detergent with 1,4-dioxane [kg/site-day]
Fdioxanejaundry = Mass fraction of 1,4-dioxane in laundry detergent [kg/kg]
Number of containers used per year is calculated using the equation below:
Equation Apx E-ll.
.. _ Qfacility_day.adjusted * OD
1V i
cont_site_yr ~
^container * 3.79 ^ * ^^^detergent
Where:
NCont_site_yr = Number of containers used per site per year [containers/site-year]
Qf acuity .day .adjusted = DailY use rate of detergent with 1,4-dioxane [kg/site-day]
OD = Operating days [days/year]
^container = Container volume [gal/container]
RHOdetergent = Detergent density [kg/L]
Vapor pressure correction factor for release points 2 and 3 is calculated using the equation below:
Equation Apx E-12.
ejaundrv/
X ~ '
A,
Fdioxane_
' MW
clean.unload ~ n i u
rdioxanejaundry , -1- rdioxanejaundry
MW + 18
Where:
Xdean unload = Vapor pressure correction factor for release points 2 and 3
[mol 1,4-dioxane/mol water]
Fdioxanejaundry = Mass fraction of 1,4-dioxane in detergent [kg/kg]
MW = 1,4-dioxane molecular weight [g/mol]
Fraction of 1,4-dioxane in wash water is calculated using the equation below:
Equation Apx E-13.
Fdioxane.wash ~ FcmUf:ion * Fcnoxanejaunciry
Where:
Fdioxane.wash = Fraction of 1,4-dioxane in wash water [kg 1,4-dioxane/kg water]
Page 360 of 570
-------
Fdilution
Fd ioxane Jaundry
Vapor pressure correction factor for release point 5 is calculated using the equation below:
EquationApx E-14.
Fdioxane_wash /
/MW
X,
washing
Fdioxane_wash _|_ ^ Fdioxane_waSh
MW
18
Where:
Xwashing
Fd ioxane _w ash
MW
Vapor pressure correction factor for release point 5
[mol 1,4-dioxane/mol water]
Fraction of 1,4-dioxane in wash water [kg 1,4-dioxane/kg water]
1,4-dioxane molecular weight [g/mol]
Container residual fraction is calculated using the following equations. To make the simulation more
realistic, EPA assessed container size based on the detergent use rate. This avoids situations where a
small container size is associated with a large use rate, such that an unrealistic number of containers are
used each year, and vice-versa:
Equation Apx E-15.
kg
If Qfacility_day > 600——
site—day
container ^residue
= F,
container _residue_tote
If Qfacility_day ~ 200 600 —
kg
site—day
Fcontainerjresidue ~ Fconf:ainer resiciuejirum
kg
If Qfacility_day < 200——
site—day
container ^residue
= F,
container _residue_pail
If physical form of detergent is powder:
container _residue
= F,
container _residue_powder
Where:
Q facility _day
Fcontainer _residue
Fcontainer _residue_tote
Fcontainer _residue_drum
Fcontainer _residue_pail
Fcontainer _residue_powder
Daily use rate based on physical form of detergent [kg/site-
day]
Container residual fraction [kg/kg]
Container residual fraction for totes [kg/kg]
Container residual fraction for drums [kg/kg]
Container residual fraction for pails [kg/kg]
Container residual fraction for solid detergents [kg/kg]
Release Point 1 site release per day is calculated using the equation below:
Page 361 of 570
-------
EquationApx E-16.
Release jperD ayRP1 = Qd
ioxane_day * Fcontainer jresidue
Where:
Release_perDayRP1 = Daily 1,4-dioxane release at release point 1 [kg/site-day]
Qdioxane_day = Daily usage rate of 1,4-dioxane [kg/site-day]
pcontainer residue = Container residual fraction [kg/kg]
Release Point 2 fugitive emissions from container cleaning for pails and drums per day is calculated using
the Penetration Model equation below (air speed <100 ft/min):
Equation Apx E-17.
Release _perDayRP2 =
5 kg
3600— *0.001 —
hr g
(8.24 x 10"8) * (MW0-835) * Xcleanun[oad * VP * jRateairspeed * (.0.25uD?ontainer opening + jjfr
To.os » ^
container _opening
Wp
Where:
Release_perDayRP2 =
MW
X clean_unload ~
VP
T
Rateair speed —
Dcontainer _opening ~
P
Release point 2 fugitive emissions from pail/drum cleaning
per day [kg/site-day]
1,4-dioxane molecular weight [g/mol]
Vapor pressure correction factor release point 2
[mol 1,4-dioxane/mol water]
1,4-dioxane vapor pressure [torr]
Ambient temperature [K]
Air speed [cm/s]
Diameter of container opening [cm]
Atmospheric pressure [atm]
Release Point 2 fugitive emissions from container cleaning per day for totes is calculated using the Mass
Transfer Coefficient Model equation below (air speed >100 ft/min):
Equation Apx E-18.
Release jper DayRP2 =
3600-— * 0.001— *
hr g
kg (1.93 x 10 7) * * Xcleanun[oad * VP * Rate° ™_speed * (id.25uD^ontainer opening
Where:
Release_perDayRP2 =
X clean_unload ~
MW
VP
T
Rateair speed —
Dcontainer _opening ~
P
fOA n011
container _opening
(VT - 5.87)2/3
Release point 2 fugitive emissions from tote cleaning per day
[kg/site-day]
Vapor pressure correction factor release point 5
[mol 1,4-dioxane/mol water]
1,4-dioxane molecular weight [g/mol]
Vapor Pressure [torr]
Ambient Temperature [K]
Air speed [cm/s]
Diameter of container opening [cm]
Atmospheric pressure [atm]
Page 362 of 570
-------
Release Point 3 fugitive emissions from unloading of pails and drums during the day is calculated using
the Penetration Model equation below (air speed <100 ft/min):
EquationApx E-19.
Release jper DayRP3 =
s kg
OHcont_unload * 3600 — * 0.001— *
(8.24 x 10"8) * (MM/0 835) * Xcleanunload * VP * * (0.25nD,
)!- + —
container _opemng J ..I 29 MW
Where:
Release_perDayRP3
X clean_unload
MW
VP
T
Rateair speed
Dcontainer _opening
P
OHcont:unload
r0'05 * Jd,
container _opening
VP
Point 3 fugitive emissions from unloading during the day
[kg/site-day]
Vapor pressure correction factor release point 5
[mol 1,4-dioxane/mol water]
1,4-dioxane molecular weight [g/mol]
Vapor pressure [torr]
Ambient temperature [K]
Air speed from EPA model [cm/s]
Diameter of the opening for containers [cm]
Atmospheric pressure [atm]
Duration of container unloading [hrs/day]
Release Point 3 fugitive emissions from unloading totes during the day is calculated using the Mass
Transfer Coefficient Model equation below (air speed >100 ft/min):
Equation Apx E-20.
Release jper DayRP3 =
OHcont unload3600—* 0.001 —*
kg (1.93 x 10 7) * (MM/0 78) * XcleanuMoad * VP * Rate°Jr%eed * (0.25?iD2container opening)^ + —^
Where:
Release_perDayRP3
X clean_unload
MW
VP
T
Rateair speed
Dcontainer _opening
P
OHcont:unload
7^0.4 nO.ii
container_opening
(Vf — 5.87) 2/s
Point 3 fugitive emissions from unloading during the day
[kg/site-day]
Vapor pressure correction factor release point 5
[mol 1,4-dioxane/mol water]
1,4-dioxane molecular weight [g/mol]
Vapor pressure [torr]
Ambient temperature [K]
Air speed from EPA model [cm/s]
Diameter of the opening for containers [cm]
Atmospheric pressure [atm]
Duration of container unloading [hours/day]
Release Point 4a dust not captured to fugitive air, water, incineration, or landfill is calculated using the
following equation:
Equation Apx E-21.
Release _perD ayRP4a = Qd
5k p1
lOXd7ie_d.Ciy dust generation
(1 " F(
dUStCapf;Ure
)
Page 363 of 570
-------
Where:
Release_perDayRP4a= Dust not captured to fugitive air, water, incineration, or landfill
[kg/site-day]
Fdustgeneration = Fraction of chemical lost during transfer of solid powders [kg/kg]
Fdustcapture = Capture efficiency for dust capture methods [kg/kg]
Release Point 4b dust captured but not controlled to stack air is calculated using the following equation:
EquationApx E-22.
Release_perDayRP4b = Qdioxane_day * Fdustgeneration * Fdustcapture * (1 — Fdustcontrol)
Where:
Release_perDayRP4b= Dust captured but not controlled to stack air [kg/site-day]
Fdustgeneration = Fraction of chemical lost during transfer of solid powders [kg/kg]
Fdustcapture = Capture efficiency for dust capture methods [kg/kg]
Fdustcontroi = Control efficiency for dust control methods [kg/kg]
Release Point 4c dust captured and controlled to incineration of landfill is calculated using the following
equation:
Equation Apx E-23.
Release_perDayRP4b = Qdioxane_day * FdUstgeneration * Fdustcavture * Fdustcontrol
Where:
Release_perDayRP4b= Dust captured but not controlled to stack air [kg/site-day]
Fdustgeneration = Fraction of chemical lost during transfer of solid powders [kg/kg]
Fdustcapture = Capture efficiency for dust capture methods [kg/kg]
Fdustcontroi = Control efficiency for dust control methods [kg/kg]
Release Point 5 fugitive emissions during washing per day is calculated when air speed <100 ft/min using
the Penetration Model in the equation shown below:
Equation Apx E-24.
Release jper DayRP5 =
S ka (8,24 X 10 8) * (MM/°'83S) * xcleanuntaad * VP55 * \^^airsveed * (0-25?lD2container opening) 1-^ + -pp
OH * 3600 — * 0.001— * N
hr q TO.os * fn ~ . Jo
u V container_opemng v 1
Where:
ReleasejperDayRPS = Point 5 fugitive emissions from washing [kg/site-day]
Xdean unload = Vapor pressure correction factor release point 5
[mol 1,4-dioxane/mol water]
MW = 1,4-dioxane molecular weight [g/mol]
VP55 = Vapor pressure of 1,4-dioxane at the laundry washwater
temperature of 55°C per the ESD [torr]
T = Ambient temperature [K]
Page 364 of 570
-------
Rateair speed = Air speed [cm/s]
Dcontainer opening = Diameter of the opening for containers [cm]
P = Atmospheric pressure [atm]
OH = Operating hours [hours/day]
Release Point 5 fugitive emissions during washing per day is calculated when air speed >100 ft/min using
the Mass Transfer Coefficient Model shown below:
EquationApx E-25.
Release jper DayRP5 =
s kq (i-93 x 10"7) * (MW0,78) * Xcleanunload * VPSS * Rate°™speed * (0.25?iD2container opentng)ljs + jm?
OHcont un,oad3600—* 0.001— * j-, *
cont unload ^ 7O.4flO.ll . _ 5.87)2/3
1 ^container_opemng\v 1 j
Where:
Release_perDayRP 5 =
X,
cleanjinload
MW
vpss
T
Rate,
D
P
OH
air_speed
container _opening
Point 5 fugitive emissions from washing [kg/site-day]
Vapor pressure correction factor release point 5
[mol 1,4-dioxane/mol water]
1,4-dioxane molecular weight [g/mol]
Vapor pressure of 1,4-dioxane at the laundry washwater
temperature of 55°C per the ESD [torr]
Ambient temperature [K]
Air speed [cm/s]
Diameter of the opening for containers [cm]
Atmospheric pressure [atm]
Operating hours [hours/day]
Release Point 6 site release per day (washing and drying) is calculated using the equations and criteria
below:
Equation Apx E-26.
If Łf=i Release_perDayRPi < Qdioxane_day:
5
Release_perDayRP6 Qdioxane_day I Release_perDayRPi
i=1
If Łf=i Release_perDayRPi > Qdioxane_day:
Liquid detergent:
Release.jperDayRP6 = Qdioxane_day ~ Release jperDayRP1
Powder detergent:
Release _perD ay RPb Qdioxane_day ReleasepgrDay RP^ Release jperDayRP 4
Where:
Release_perDayRP1
ReleasejperD ayRP2
Release_perDayRP3
ReleasejperD ayRP4
Point 1 container residual releases [kg/site-day]
Point 2 fugitive emissions from container cleaning [kg/site-
day]
Point 3 fugitive emissions from unloading [kg/site-day]
Point 4 fugitive dust emissions [kg/site-day]
Page 365 of 570
-------
Release_perDayRP 5
ReleasejperD ayRP6
Point 5 fugitive emissions from washing [kg/site-day]
Point 6 daily site releases (washing and drying) [kg/site-
day]
Daily usage rate of 1,4-dioxane [kg/site-day]
The sum of release points 1-5 emissions [kg/site-day]
ioxane_day
Łf=1 Release _perD ay RPi
E.12.2 Model Input Parameters
Table Apx E-15 summarizes the model parameters and their values for the Monte Carlo simulation.
Additional explanations of EPA's selection of the distributions for each parameter are provided after this
table. High-end and central tendency releases are estimated by selecting the 50th and 95th percentile
values from the output distribution.
Page 366 of 570
-------
Table Apx E-15. Summary of Paramel
ter Values and Distributions Used in the Industrial and Institutional Laundry E
please Model
Input Parameter
Symbol
Unit
Constant Model
Parameter Values
Variable Model Parameter Values
Rational/ Basis
Value
Lower
Bound
Upper
Bound
Mode
Distribution
Type
Operating Days
OD
days/year
260
Industrial:
19
Institutional:
249
366
260
Triangular
See Section
E.12.3
Mass Fraction of
1,4-dioxane in
Laundry Detergent
Fdioxane laundry
kg/kg
8.91E-06
5.00E-08
0.00013
Discrete
See Section
E.12.4
Daily Use Rate of
Liquid Laundry
Detergents
Qfacility day liquid
kg/day
Industrial:
35.7
Institutional:
16
Industrial:
0.116
Institutional:
0.124
Industrial:
814
Institutional:
513
Discrete
See Section
E.12.5
Daily Use Rate of
Powder Laundry
Detergents
Qfacility day_powder
kg/day
Industrial:
110.45
Institutional:
8.63
Industrial:
1.33
Institutional:
3.71
Industrial:
1,917.44
Institutional:
15
Discrete
See Section
E.12.5
Container Size
V container
gal
55
5
550
55
Triangular
See Section
E.12.6
Air Speed
RATEai, speed
cm/s
10
1.3
202.2
-
Lognormal
See Section
E.12.7
Container Residual
Fraction for Totes
Fcontainer residue totes
kg/kg
0.002
0.0002
0.002
0.0007
Triangular
See Section
E.12.8
Container Residual
Fraction for Drums
Fcontainer residue drums
kg/kg
0.03
0.017
0.03
0.025
Triangular
See Section
E.12.9
Container Residual
Fraction for Pails
Fcontainer residue_pails
kg/kg
0.006
0.0003
0.006
0.003
Triangular
See Section
E.12.10
Container Residual
Fraction for Powders
Fcontainer residue_powders
kg/kg
0.01
-
-
-
-
See Section
E.12.11
Fraction of Laundry
Detergents
Containing 1,4-
Dioxane
Fformulations dioxane
unitless
0.5
0.111
1
Industrial:
Discrete
Institutional:
Uniform
See Section
E.12.12
Page 367 of 570
-------
Input Parameter
Symbol
Unit
Constant Model
Parameter Values
Variable Model Parameter Values
Rational/ Basis
Value
Lower
Bound
Upper
Bound
Mode
Distribution
Type
Duration of Release
for Container
Unloading
OHcont unload
h/day
Industrial:
0.0043
Institutional:
0.0114
Industrial:
0.0043
Institutional:
0.0114
Industrial:
12
Institutional:
8
Uniform
See Section
E. 12.13
Fraction of
Chemical Lost
During Transfer of
Solid Powders
F dust_generation
kg/kg
0.0050
0.0010
0.03
0.005
Triangular
See Section
E. 12.14
Control Efficiency
for Dust Control
Methods
Fdust control
kg/kg
0.7900
0.0000
1
0.79
Triangular
See Section
E. 12.15
Capture Efficiency
for Dust Capture
Methods
Fdust capture
kg/kg
0.9633
0.9310
1
0.9633
Triangular
See Section
E. 12.16
Number of Sites
Ns
sites
Industrial:
2,453
Institutional:
95,533
See Section
E. 12.17
Vapor Pressure of
1,4-Dioxane at
Ambient
Temperature
VP
Torr
40
Physical
property
Vapor Pressure of
1,4-Dioxane at
Washwater
Temperature of 55°C
VP55
Torr
147
Physical
property
Molecular Weight of
1,4-Dioxane
MW
g/mol
88.1
-
-
-
-
Physical
property
Diameter of
Container Opening
Dcontainer opening
cm
5.08
-
-
-
-
See Section
E. 12.18
Diameter of Wash
Opening
Dwash opening
cm
73
-
-
-
-
See Section
E. 12.19
Page 368 of 570
-------
Input Parameter
Symbol
Unit
Constant Model
Parameter Values
Variable Model Parameter Values
Rational/ Basis
Value
Lower
Bound
Upper
Bound
Mode
Distribution
Type
Ambient
Temperature
T
K
298
-
-
-
-
Process
parameter
Ambient Pressure
P
atm
1
-
-
-
-
Process
parameter
Dilution Factor
F dilution
unitless
0.016
-
-
-
-
See Section
E. 12.20
Density of Laundry
Detergent
RHOform
kg/L
1
ESD assumes a
density equal to
that of water
Container Fill Rate
RATEfin
containers
/ hour
20
-
-
-
-
See Section
E.12.21
Page 369 of 570
-------
E.12.3 Operating Days
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
the Agency used generic data from the ESD on the Chemicals Used in Water Based Washing Operations
at Industrial and Institutional Laundries (OB ). The ESD provides the range and average of
operating days for six separate years, which EPA took the minimum, maximum, and average of the 6
years to form distributions. Specifically, EPA modeled the operating days per year using a triangular
distribution with a lower bound of 20 days per year, an upper bound of 365 days per year, and a mode of
260 days per year for industrial laundries. EPA used a triangular distribution with a lower bound of 250
days per, an upper bound of 365 days per year, and a mode of 260 days per year for institutional
laundries.
E.12.4 Mass Fraction of 1,4-Dioxane in Laundry Detergent
EPA modeled the mass fraction of 1,4-dioxane in laundry detergent using the same discrete distribution
for both industrial and institutional laundries. This is based on chemical-specific data from the
December 2020 Final Risk Evaluation for 1,4-Dioxane ( s20c) and product concentration
waiver data from the New Your State Department of Environmental Conservation (NYDEC) (NYDEC.
2023). No additional sources of data were identified from systematic review. The discrete distribution
gives equal probably to each of the 19 total data points shown in TableApx E-16 from the two
identified data sources.
Table Apx E-16. Discrete Data Points on Mass Fraction of 1,4-Dioxane in Laundry
detergent
Mass Fraction of 1,4-Dioxane in Laundry Detergents (kg 1,4-dioxane/kg detergent)
5.0E-08
4.3E-06
8.9E-06
1.3E-04
2.0E-06
4.3E-06
1.0E-05
1.3E-04
2.0E-06
5.0E-06
1.4E-05
1.3E-04
2.7E-06
5.0E-06
2.5E-05
1.3E-04
2.7E-06
8.9E-06
1.3E-04
Sources: (U.S. EPA. 2020c) and (NYDEC. 2023)
E.12.5 Daily Use Rate of Detergent
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
the Agency used generic data from the ESD on the Chemicals Used in Water Based Washing Operations
at Industrial and Institutional Laundries (OE< ). The ESD references a discrete dataset on
detergent use rates from a survey of laundry sites, comprised of 49 data points for liquid detergents and
59 data points for solid laundry detergents, as shown in Table Apx E-17 and Table Apx E-18. EPA
modeled the daily use rate of detergent using a discrete distribution comprised of these data points, with
equal probability given to each value.
Page 370 of 570
-------
TableApx E-17. Discrete Data Points on Daily
Use Rate of Liquid Detergents
Daily Use Rate of Liquid Detergents (kg/site-day)
0.12
5.61
20.5
48.5
120
0.45
5.91
26.7
62.5
124
0.72
5.94
30.1
64.1
177
1.00
6.10
31.2
66.2
180
1.46
7.08
35.7
68.1
205
1.87
9.01
36.4
86.7
207
2.43
11.1
37.3
106
290
2.63
12.7
37.3
110
376
4.07
15.3
38.3
111
814
4.10
19.1
44.6
113
-
Source: (OECD, 201.1b)
Table Apx E-18. Discrete Data Points on Daily Use
iate of Solid Detergents
Daily Use Rate of Solid Detergents (kg/site-day)
1.33
36.3
80.7
112
177
286
1.89
47.2
85.5
125
189
357
2.67
49.1
95.7
134
190
358
3.61
52.2
97.1
138
199
389
5.44
55.8
101
143
204
439
7.97
57.8
102
145
221
490
8.89
61.6
105
151
236
514
13.61
63.6
107
154
238
529
22.6
66.9
108
158
240
1,917
32.0
76.1
110
172
264
-
Source: ("OECD. 1 )
E.12.6 Container Size
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
the Agency used generic data from standard EPA models. Specifically, EPA modeled container size
using a triangular distribution with a lower bound of 5 gallons, an upper bound of 550 gallons, and a
mode of 55 gallons for industrial laundries. Because EPA expects industrial laundries to have variation
in the sizes of containers, EPA used values of 5, 55, and 550 gallons for the triangular distribution based
Page 371 of 570
-------
on the default values from the EPA/OPPT Small Container Residual Model, Drum Residual Model, and
Bulk Transport Residual Model, respectively.
EPA used a single value of 5 gallons for institutional laundries based on the ESD on the default value
for institutional laundries from the Chemicals Used in Water Based Washing Operations at Industrial
and Institutional Laundries ESD ( ).
E.12.7 Indoor Air Speed
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
the Agency used generic data from Baldwin (1998). a source known from previous EPA model
development. Baldwin (1998) measured indoor air speeds across a variety of occupational settings in the
United Kingdom. Fifty-five work areas were surveyed across a variety of workplaces. The Agency
analyzed the air speed data from Baldwin (1998) and categorized the air speed surveys into settings
representative of industrial facilities and representative of commercial facilities. EPA fit separate
distributions for these industrial and commercial settings and used the industrial distribution for laundry
facilities.
EPA fit a lognormal distribution for both data sets as consistent with the authors observations that the air
speed measurements within a surveyed location were lognormally distributed and the population of the
mean air speeds among all surveys were lognormally distributed. Since lognormal distributions are
bound by zero and positive infinity, EPA truncated the distribution at the largest observed value among
all of the survey mean air speeds from Baldwin (1998). EPA fit the air speed surveys representative of
industrial facilities to a lognormal distribution with the following parameter values: mean of 22.414
cm/s and standard deviation of 19.958 cm/s. In the model, the lognormal distribution is truncated at a
maximum allowed value of 202.2 cm/s (largest surveyed mean air speed observed in Baldwin (1998)) to
prevent the model from sampling values that approach infinity or are otherwise unrealistically large.
Baldwin (1998) only presented the mean air speed of each survey. The authors did not present the
individual measurements within each survey. Therefore, these distributions represent a distribution of
mean air speeds and not a distribution of spatially variable air speeds within a single workplace setting.
E.12.8 Container Residual Fraction for Totes
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
the Agency used generic data from standard EPA models. Specifically, EPA modeled container residual
fraction for totes using a triangular distribution with a lower bound of 0.0002 kg residual/kg detergent,
and upper bound of 0.002 kg residual/kg detergent, and a mode of 0.0007 kg residual/kg detergent. The
lower and upper bounds of this distribution are based on the central tendency and high-end values listed
in the EPA/OPPT Bulk Transport Residual Model from the ChemSTEER User Guide (U.S. EPA.
2015 a). EPA used the central tendency value as the mode of the triangular distribution. Note that the
underlying data for this model comes from a 1988 study by PEI Associates Inc. that looked at literature
sources and conducted a pilot-scale experiment to determine the amount of residual material left in
containers (PEI Associates. 1988). EPA reviewed the data from this study and the underlying
distribution of the container residual loss fraction is unknown; therefore, EPA assigned a triangular
distribution as discussed above.
E.12.9 Container Residual Fraction for Drums
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
the Agency used generic data from standard EPA models. Specifically, EPA modeled container residual
fraction for drums using a triangular distribution with a lower bound of 0.0003 kg residual/kg detergent,
Page 372 of 570
-------
an upper bound of 0.03 kg residual/kg detergent, and a mode of 0.025 kg residual/kg detergent. The
lower bound is based on the minimum value for pouring and the upper bound is based on the default
high-end value in the EPA/OPPT Drum Residual Model from the ChemSTEER User Guide (
2015a). EPA used the central tendency value for pumping as the mode of the triangular distribution.
Note that the underlying data for this model comes from a 1988 study by PEI Associates Inc. that looked
at literature sources and conducted a pilot-scale experiment to determine the amount of residual material
left in containers (PEI Associates. 1988). EPA reviewed the data from this study and the underlying
distribution of the container residual loss fraction is unknown; therefore, the Agency assigned a
triangular distribution as discussed above.
E.12.10 Container Residual Fraction for Pails
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
the Agency used generic data from standard EPA models. Specifically, EPA modeled container residual
fraction for small containers using a triangular distribution with a lower bound of 0.0003 kg residual/kg
detergent, an upper bound of 0.006 kg residual/kg detergent, and a mode of 0.003 kg residual/kg
detergent. The lower bound is based on the minimum value for pouring and the upper bound is based on
the default high-end value listed in the EPA/OPPT Small Container Residual Model from the
ChemSTEER User Guide ( ). EPA used the central tendency value for pouring as the
mode of the triangular distribution. Note that the underlying data for this model comes from a 1988
study by PEI Associates Inc. that looked at literature sources and conducted a pilot-scale experiment to
determine the amount of residual material left in containers (PEI Associates. 1988). EPA reviewed the
data from this study and the underlying distribution of the container residual loss fraction is unknown;
therefore, the Agency assigned a triangular distribution as discussed above.
E.12.11 Container Residual Fraction for Powders
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
the Agency used generic data from standard EPA models. The EPA/OPPT Solid Residuals in Transport
Containers Model provides a loss fraction 0.01 kg of solid chemicals remaining in a container per kg
transported. Therefore, EPA could not develop a distribution of values for this parameter and used the
single value 0.01 kg/kg from the model ( 2015a). Note that the underlying data for this model
comes from a 1988 study by PEI Associates Inc. that looked at literature sources and conducted a pilot-
scale experiment to determine the amount of residual material left in containers (PEI Associates. 1988).
EPA reviewed the data from this study and the underlying distribution of the container residual loss
fraction is unknown; therefore, the Agency assigned a triangular distribution as discussed above.
E.12.12 Fraction of Laundry Detergents Containing 1,4-Dioxane
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
EPA used generic data from the ESD on the Chemicals Used in Water Based Washing Operations at
Industrial and Institutional Laundries (OJ ). EPA modeled the fraction of laundry detergents
containing 1,4-dioxane using a discrete distribution comprised of survey data from laundries sites used
in the ESD. These data are on fractions of laundry detergents containing a chemical of interest, as
opposed to specifically 1,4-dioxane. Equal probability was given to each discrete survey value. Some
data points occurred multiple times in the dataset, as shown in Table Apx E-19, so each occurrence had
equal probability.
Page 373 of 570
-------
TableApx E-19. Data on the Fraction of Laundry Detergent Containing
the Chemical of Interest
Fraction of Laundry Detergent
Containing the Chemical of Interest
Number of Occurrences in
Dataset
0.111
1
0.143
2
0.167
3
0.20
14
0.25
21
0.33
60
0.50
64
1.00
57
Source: (OECD, 201.1b)
E.12.13 Duration of Release for Container Unloading
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
EPA used generic data from the ESD on the Chemicals Used in Water Based Washing Operations at
Industrial and Institutional Laundries (CM ). EPA modeled the duration of release for
container unloading using a uniform distribution. For industrial and institutional laundries, EPA
assumed the distribution had a maximum of 12 and 8 hours/day, respectively, based on the shift
durations in the ESD. The lower bound was based on the length of time to unload detergent containers
each day, calculated using the number of containers used per day and the container fill rate (see Section
0). This means that each iteration of the simulation would calculate a new lower bound based on the
parameters for that iteration.
E.12.14 Fraction of Chemical Lost During Transfer of Solid Powders
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
the Agency used generic data from standard EPA models. Specifically, EPA modeled the fraction of
chemical lost during transfer of solid powders using a triangular distribution with a lower bound of
0.001 kg dust lost/kg transferred, an upper bound of 0.03 kg dust lost/kg transferred, and a mode of
0.005 kg dust lost/kg transferred for both industrial and institutional laundries. These values were taken
from the EPA/OPPT Dust Emissions from Transferring Solids Model from the ChemSTEER User
Guide ( '015aY
E.12.15 Control Efficiency for Dust Control Methods
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
the Agency used generic data from standard EPA models. Specifically, EPA modeled the control
efficiency for dust control methods using a triangular distribution with a lower bound of 0 kg
controlled/kg transferred, an upper bound of 1 kg controlled/kg transferred, and a mode of 0.79 kg
controlled/kg transferred for both industrial and institutional laundries. These values were taken from the
EPA/OPPT Dust Emissions from Transferring Solids Model from the ChemSTEER User Guide (
).
Page 374 of 570
-------
E.12.16 Capture Efficiency for Dust Capture Methods
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
the Agency used generic data from standard EPA models. Specifically, EPA modeled the capture
efficiency for dust capture methods using a triangular distribution with a lower bound of 0.9310 kg
captured/kg transferred, an upper bound of 1 kg captured/kg transferred, and a mode of 0.9633 kg
captured/kg transferred for both industrial and institutional laundries. These values were taken from the
EPA/OPPT Dust Emissions from Transferring Solids Model from the ChemSTEER User Guide (
).
E.12.17 Number of Sites
EPA did not find data on the number of laundry sites that specifically use detergents containing 1,4-
dioxane; therefore, the Agency used generic data. As a bounding estimate for the number of industrial
laundries, EPA used U.S. Census and BLS data for the NAICS code 812330, Linen and Uniform
Supply, to estimate a total of 2,453 industrial laundry sites within the industry (\] S HI S. 2016). As a
bounding estimate for the number of institutional sites, EPA used industry information as described in
the ESD to estimate a total of 95,533 institutional laundries (OECD. 201 lb).
E.12.18 Diameter of Container Opening
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
the Agency used generic data from the ChemSTEER User Guide (U.S. EPA. ), which provides a
typical diameter of container openings as 5.08 cm. Therefore, EPA could not develop a distribution of
values for this parameter and used the single value 5.08 cm from the ChemSTEER User Guide.
E.12.19 Diameter of Wash Opening
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
the Agency used generic data from the ESD on the Chemicals Used in Water Based Washing Operations
at Industrial and Institutional Laundries (OB ). The ESD provided a single value of 73 cm for
the diameter of washer openings to estimate air releases during operation. Therefore, EPA could not
develop a distribution of values for this parameter and used the single value of 73 cm from the ESD.
E.12.20 Dilution Factor
The December 2020 Final Risk Evaluation for 1,4-Dioxane (U.S. EPA. 2020c) provided a single value
for the dilution factor of 1,4-dioxane in laundry detergents. The risk evaluation states that a dilution
factor of 0.016 was estimated assuming a high-end mass of product used (60g) in one gallon of water
( 2020c). EPA did not find any other chemical-specific data for this parameter. Therefore,
EPA could not develop a distribution of values for this parameter and used the single value of 0.016
from the 2020 RE.
E.12.21 Container Fill Rate
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
the Agency used generic data from the ChemSTEER User Guide (U.S. EPA. ), which provides a
typical fill rate of 20 containers per hour for containers with 20 to 100 gallons of liquid. Therefore, EPA
could not develop a distribution of values for this parameter and used the single value 20 containers/hour
from the ChemSTEER User Guide.
E.12.22 Key Strengths, Limitations, Uncertainties, and Sensitivity Analysis
General modeling uncertainties and limitations are discussed in Section 2.2.1.3 and Appendix E.7. This
section discusses model-specific uncertainties and limitations and presents examples of sensitivity charts
that EPA developed for this model. For this model, the only 1,4-dioxane specific input parameter data is
Page 375 of 570
-------
for the concentration of 1,4-dioxane in laundry detergent. All other parameters are based on generic data
from the ESD on the Chemicals Used in Water Based Washing Operations at Industrial and Institutional
Laundries (OECD. 2011b) or standard EPA/OPPT models described in the ChemSTEER User Guide
(U.S. EPA. 2015a). This adds uncertainty with respect to the representativeness of the input data
towards laundry sites that use detergents containing 1,4-dioxane.
EPA ran a sensitivity analysis in @Risk to identify the input parameters which have the largest impact
on the daily release output estimates. For example, FigureApx E-6 shows the inputs ranked by which
have the largest effect on the mean fugitive air release during unloading liquid laundry detergents, which
is release point 3 in this model. Figure Apx E-7 similarly shows the inputs that impact the daily release
from unloading solid laundry detergents, which corresponds to release point 4 in this model. The mass
fraction of 1,4-dioxane in laundry detergent has a relatively large impact on both release points. As
discussed in Appendix E. 12.4, EPA used a discrete dataset comprised of 19 data points for the mass
fraction of 1,4-dioxane laundry detergent. From Figure Apx E-6, the duration of release for container
unloading is actually an intermediate output of the model calculated based on the number of containers
unloaded and daily laundry detergent use rate, both of which are based on distributions from generic
data. For all other parameters in Figure Apx E-6 and Figure Apx E-7, EPA developed distributions
based on generic, not 1,4-dioxane-specific data. Having a distribution for each input parameter is a
strength of the assessment; however, the representativeness of the underlying data used for these
distributions is a limitation, as was discussed above.
Release Point 3- Daily Releases
Inputs Ranked by Effect on Output Mean
Mass Fraction of 1/l-Dioxane in Laundry Detergent
Duration of Release for Container Unloading
Air Speed
Daily Use Rate of Liquid Laundry Detergents
Operati ng Days
Fraction of Laundry Detergents Containing 1,4-Dioxane
L.
A
Figure Apx E-6. Sensitivity Chart for Fugitive Air Release During Unloading Liquid Detergents
(Daily Release Point 3) at Institutional Laundries
Page 376 of 570
-------
Release Point 4 - Daily Releases
Inputs Ranked by Effect on Output Mean
Daily Use Rate of Laundry Detergents
Mass Fraction of 1,4-Dioxane in Laundry Detergent
Fraction of chemical lost during transfer of sold
powders
Fraction of Laundry Detergents Containing 1,4-Dioxane
Capture efficiency for dust capture methods
!
|
L
I
FigureApx E-7. Sensitivity Chart for Release from Dust Generation During Unloading Solid
Detergents (Daily Release Point 4) at Industrial Laundries
E.13 Hydraulic Fracturing Modeling Approach and Parameters for
Estimating Environmental Releases
This appendix presents the modeling approach and equations used to estimate environmental releases of
1,4-dioxane during hydraulic fracturing. This approach utilizes the Revised ESD on Chemicals Used in
Hydraulic Fracturing (U.S. EPA. 2022e) combined with Monte Carlo simulation (a type of stochastic
simulation). This ESD indicates that 100 percent of hydraulic fracturing fluid chemical additives are
released and includes a diagram of release and exposure points during hydraulic fracturing, as shown in
Figure Apx E-8 (U.S. EPA. 2022eY
Page 377 of 570
-------
FigureApx E-8. Environmental Release Points (Numbered) and Occupational Exposure Points
(Lettered) During Hydraulic Fracturing
Based on Figure Apx E-8, EPA identified the following release points:
• Release point 1 (RP1): Fugitive air releases during container unloading;
• Release point 2 (RP2): Container residue losses to surface water, incineration, or landfill;
• Release point 3 (RP3): Fugitive air releases during container cleaning;
• Release point 4 (RP4): Equipment and storage tank cleaning losses to surface water, incineration,
or landfill;
• Release point 5 (RP5): Fugitive air releases during equipment and storage tank cleaning;
• Release point 6 (RP6): Release from spills to surface water (13%), land (soil) (64%), and landfill
or incineration (23%);
• Release point 7 (RP7): Release of hydraulic fracturing fluid that remains underground to deep
well injection; and
• Release point 8 (RP8): Hydraulic fracturing fluid flowback and produced wastewater to
recycle/reuse (5%), deep well injection (70%), on- or off-site treatment and discharge to surface
water (19%), or on- or off-site treatment and release to land in evaporation pits, infiltration pits,
irrigation, or road treatment (6%).
Environmental releases of hydraulic fracturing are a function of the chemical's physical properties,
container size, mass fractions, and other model parameters. Although physical properties are fixed for a
chemical, some model parameters are expected to vary from one facility to another. An individual model
input parameter could either have a discrete value or a distribution of values. EPA assigned statistical
distributions based on available literature data or engineering judgment to address the variability in
operating days (OD), mass fraction of 1,4-dioxane in fracturing fluid (Fdioxane fracturing fluid), mass fraction
of 1,4-dioxane in additive (Fdioxane additive), container container size (Vcont), annual use rate of fracturing
fluids (Qsite yr), saturation factor (Fsaturation), container cleaning losses (FCOnt cleaning), and fraction of
injected fracturing fluid that returns to the surface (Frecovered).
Page 378 of 570
-------
A Monte Carlo simulation (a type of stochastic simulation) was conducted to capture variability in the
model input parameters. The simulation was conducted using the Latin hypercube sampling method in
@Risk (Palisade, Ithaca, NY). The Latin hypercube sampling method is a statistical method for
generating a sample of possible values from a multi-dimensional distribution. Latin hypercube sampling
is a stratified method, meaning it guarantees that its generated samples are representative of the
probability density function (variability) defined in the model. EPA performed 100,000 iterations of the
model to capture the range of possible input values, including values with low probability of occurrence.
From the distribution resulting from the Monte Carlo simulation, EPA selected the 95th and 50th
percentile values to represent a high-end release and central tendency release level respectively. The
statistics were calculated directly in @Risk. The following subsections detail the model design equations
and parameters used for environmental release estimates.
E.13.1 Model Equations
Daily use rate of fracturing fluids containing 1,4-dioxane is calculated using the equation below:
EquationApx E-27.
f L RHOfracturing fluid
Qsite_day ~ Qsite_yr * gal * OD
Where:
Qsite_day = Daily use rate of fracturing fluids with 1,4-dioxane [kg/site-day]
Qsite_yr = Annual use rate of fracturing fluids with 1,4-dioxane [gal/site-year]
OD = Operating days [days/year]
RHOfracturing jiuid = Density of fracturing fluid [kg/L]
Annual use rate of 1,4-dioxane is calculated using the equation below:
Equation Apx E-28.
/ L \
Qdioxane_site_yr ~ Qsite_yr * I 3.79 I * RHOfracturing_fiuid * Fdioxane_fracturing_fluid
L \
Qdioxane_site_yr ~ Qsite_yr *
Where:
Qdioxane_site_yr = Annual use rate of 1,4-dioxane [kg/site-year]
Qsitejyr = Annual use rate of fracturing fluids with 1,4-dioxane
[gal/site-year]
RHOfracturing jiuid = Density of fracturing fluid [kg/L]
Fdioxane fracturing fluid = Mass fraction of 1,4-dioxane in hydraulic fracturing fluid
[kg/kg]
Number of containers used per year is calculated using the equation below:
Equation Apx E-29.
_ Qdioxane_site_yr
™cont_unlaod_yr ~ '
Fdioxane_additive * ^cont * ^^^/racturin5_/7uid * 3.79
Where:
Page 379 of 570
-------
NCont_uniaod_yr = Number of containers used yearly [containers/site-year]
Qdioxane_site_yr = Annual use rate of 1,4-dioxane [kg/site-year]
Fdioxane_additive = Mass fraction of 1,4-dioxane in hydraulic fracturing additive [kg/kg]
Vcont = Container size for fracturing fluids [gal]
RHOfracturing jiuid = Density of fracturing fluid [kg/L]
The vapor pressure correction factor for release point 1 (unloading) and release point 3 (container
cleaning) is calculated using the equation below:
EquationApx E-30.
Fflir,
' MW
dioxane_additive /
v
^clean_unload ~ u . i _ u
i HinYnnt? nrlrlitiiic> -L t Hi,
dioxane_additive _j_ -1- rdioxane_additive
MW 1 18
Where:
Xdean unload = Vapor pressure correction factor for RP 1 and 3 [mol 1,4-
dioxane/mol H2O]
Fdioxane_additive = Mass fraction of 1,4-dioxane in hydraulic fracturing additive [kg/kg]
MW = 1,4-dioxane molecular weight [g/mol]
The vapor pressure correction factor for release point 5 (storage tank cleaning) is calculated using the
equation below:
Equation Apx E-31.
Fdioxane_fracturing_fluid /
y _ 'MW
Atank_clean ~ n i _ n
rdioxane_fracturing_fluid , rdioxane_fracturing_fluid
MW + 18
Where:
Xtank_ciean = Vapor pressure correction factor for RP 5 [mol 1,4-
dioxane/mol H2O]
MW = 1,4-dioxane molecular weight [g/mol]
Fdioxane fracturing fluid = Mass fraction of 1,4-dioxane in hydraulic fracturing fluid
[kg/kg]
Container residual fraction is calculated using the following equations. To make the simulation more
realistic, EPA assessed container size based on the fracturing fluid use rate. This avoids situations where
a small container size is associated with a large use rate, such that an unrealistic number of containers
are used each year, and vice-versa:
Equation Apx E-32.
If Qsite-day > 1500 kg/site-day:
If Qsite-day < 1500 kg/site-day:
Where:
Fcontainer residue ~ FC0nt_Cleaning_t0te
Fcontainer_residue ~ Fconi_cieaningjirum
Page 380 of 570
-------
Fcontainer residue = Container residual fraction [kg/kg]
Fcontainer residuejote = Container residual fraction for totes [kg/kg]
pcontainer residue_drum= Container residual fraction for drums [kg/kg]
Qsite_day = Daily use rate of fracturing fluids with 1,4-dioxane [kg/site-day]
Release Point 1 daily releases per site (unloading volatile chemicals) are calculated using the AP-42
Loading Model shown in the equation below:
EquationApx E-33.
N,
cont_unload_yr j
VPhm /(qd *RATErm)
Release_perDayRP1 — Fsaturation ^actor * MW * 3785.4 * Vcont * Rate^m * Xcieanunioad * j* * R * 1000 ^
kg
Where:
Release_perDayRP1 = Release point 1 daily releases [kg/site-day]
psaturation j actor = Saturation factor [unitless]
MW = 1,4-dioxane molecular weight [g/mol]
Vcont = Container size for fracturing fluids [gal]
Xdean unload = Vapor pressure correction factor for RP 1 and 3
[mol 1,4-dioxane/mol H2O]
VP = 1,4-dioxane vapor pressure [torr]
T = Ambient temperature [K]
R = Universal gas constant [atm-cm3/gmol-K]
Ncont_uniaod_yr = Number of containers used yearly [containers/site-year]
OD = Operating days [days/year]
RATEfui = Container fill rate [containers/hour]
Release Point 2 daily releases per site (container residuals) are calculated using the following equation:
Equation Apx E-34.
Release_perDayRP2 — Qdioxane_site_day * Pcontainer residue
Where:
Release_perDayRP2 = Release point 2 daily releases [kg/site-day]
Qdioxane_site_day = Daily use rate of 1,4-dioxane [kg/site-day]
pcontainer residue = Container residual fraction [kg/kg]
Release Point 3 daily releases per site (container cleaning) are calculated using the Mass Transfer
Coefficient Model shown in the following equation:
Equation Apx E-35.
Release jper DayRP3 =
NCont unload vr * &9 ^ * 10_7) * * Xcleanuntaad * VP * RMe°a™speed * {02SnD2container ovening)\^ +
- J * 3600 — * 0.001— * 77 ^——
OD * RATEfm hr g toado.u (-JT - 5 87)/s
J 1 ^ container_opening\w 1 )
Where:
Release_perDayRP3 = Release point 3 daily releases [kg/site-day]
Xciean unload = Vapor pressure correction factor release point 3
Page 381 of 570
-------
MW
VP
T
R(Xt6airspeed
Dcontainer _opening
P
OD
Ncont_unlaod_yr
RATEfui
[mol 1,4-dioxane/mol water]
1,4-dioxane molecular weight [g/mol]
Vapor Pressure [torr]
Ambient Temperature [K]
Air speed [cm/s]
Diameter of the opening for containers [cm]
Atmospheric pressure [atm]
Operating days [days/year]
Number of containers used yearly [containers/site-year]
Container fill rate [containers/hour]
Release Point 4 daily releases per site (equipment cleaning) are calculated using the following equation:
EquationApx E-36.
Release jperD ayRP4 = Qd
Where:
Release_perDayRP4 =
Qd
ioxane_site_day
equip_cleaning
ioxane_site_day * ^equip_cleaning
Release point 4 daily releases [kg/site-day]
Daily use rate of 1,4-dioxane [kg/site-day]
Equipment cleaning loss fraction [kg/kg]
Release point 5 daily releases per site (equipment and storage tank cleaning surface losses) are
calculated using the Mass Transfer Coefficient Model shown in the equation below:
Equation Apx E-37.
Release jper DayRPS =
OHequip clean * 3600—* 0.001 — *
kg (1.93 x 10 7) * (MW0 7S) * Xtank_clean * VP * Rate°[7r8speed * (0.25?iD2container ovening)^ + —pp
7*0.4n0.il (Jf— 5 87W3
1 u container _opening\^ 1 -J-v/j
Where:
Release_perDayRP 5 =
Xtank_clean ~
MW
VP
T
Rateair speed —
Dcontainer _opening ~
P
OHeqUip_ciean ~
Release point 5 daily releases [kg/site-day]
Vapor pressure correction factor release point 5
[mol 1,4-dioxane/mol water]
1,4-dioxane molecular weight [g/mol]
Vapor Pressure [torr]
Ambient Temperature [K]
Air speed [cm/s]
Diameter of the opening for containers [cm]
Atmospheric pressure [atm]
Equipment cleaning operating hours [hours/day]
Release point 6 daily releases per site (spills) are calculated using the following equation:
Equation Apx E-38.
Release_perDayRP6 Qdioxane_site_yr * Fspill
Where:
Page 382 of 570
-------
Release_perDayRP6 = Release point 6 daily releases [kg/site-day]
Qdioxane_site_yr = Annual use rate of 1,4-dioxane [kg/site-yr]
Fspiu = Fraction of fracturing fluid spilled [kg/kg]
Release point 7 daily releases per site (deep well injection of fracturing fluid) are calculated using the
following equation:
EquationApx E-39.
Release _perD ay RP7
_ (.Q dioxane_site_yr ~ Release_pevYeaTRP^) * (l — Fcontainer_reSidUe ~ Fequip_cleaning) * (1 — Frecovered)
~ OD
Where:
Release_perDayRP7 =
Qd
ioxane_site_yr
Release_perYearRP6 =
Fcontainer residue ~
Fequip_cleaning ~
Frecovered ~
OD
Release point 7 daily releases [kg/site-day]
Annual use rate of 1,4-dioxane [kg/site-yr]
Release point 6 annual releases [kg/site-yr]
Container residual fraction [kg/kg]
Equipment cleaning loss fraction [kg/kg]
Fraction of injected fracturing fluid returning to surface [kg/kg]
Operating days [days/year]
Release point 8 daily releases per site (flowback and produced wastewater) are calculated using the
following equation:
Equation Apx E-40.
Release _perD ay RP8
0Qd
ioxane_site_yr
— Release_perYearRP6) * (l — Ft
-Fe
container jresidue r equip_cleaning
,)*?>
recovered
Where:
Release_perDayRP8 =
Qdioxane_site_yr ~
Release_perYearRP6 =
Fcontainer _residue ~
Fequip_cleaning ~
Frecovered ~
30 days/yr
Release point 8 daily releases [kg/site-day]
Annual use rate of 1,4-dioxane [kg/site-yr]
Release point 6 annual releases [kg/site-yr]
Container residual fraction [kg/kg]
Equipment cleaning loss fraction [kg/kg]
Fraction of injected fracturing fluid returning to surface [kg/kg]
E.13.2 Model Input Parameters
Table Apx E-20 summarizes the model parameters and their values for the Monte Carlo simulation.
Additional explanations of EPA's selection of the distributions for each parameter are provided after this
table. High-end and central tendency releases are estimated by selecting the 50th and 95th percentile
values from the output distribution.
Page 383 of 570
-------
Table Apx E-20. Summary of Parameter Values and Distributions Used in the Hydraulic Fracturing Release Model
Input Parameter
Symbol
Unit
Constant Model
Parameter
Values
Variable Model Parameter Values
Rationale /
Basis
Value
Lower
Bound
Upper Bound
Mode
Distribution
Type
Molecular weight of 1,4-
dioxane
MW
g/mol
88.1
-
-
-
-
Physical
property
Vapor pressure of 1,4-
dioxane
VP
torr
40
-
-
-
-
Physical
property
Gas constant
R
atm-
cm3/mol-K
82.05
-
-
-
-
Universal
constant
Ambient temperature
T
K
298
-
-
-
-
Process
parameter
Ambient pressure
P
Atm
1
-
-
-
-
Process
parameter
Number of sites
Ns
sites
411
-
-
-
-
See Section
E.13.3
Operating days
OD
days/year
16
1
72
-
Discrete
See Section
E.13.4
Container volume
(fracturing fluid)
Vcont
gal
55
20
1,000
55
Triangular
See Section
E.13.5
Density of fracturing
fluid
RHOfracturing fluid
kg/L
1
ESD assumes
a density
equal to that
of water
Diameter of container
opening
Dcontainer opening
cm
5.08
-
-
-
-
See Section
E.13.6
Diameter of equipment
opening
Dequip opening
cm
92
-
-
-
-
See Section
E.13.7
Air speed during
equipment cleaning
RATE,,, speed
ft/min
440
-
-
-
-
See Section
E.13.8
Equipment cleaning loss
fraction
Fequip cleaning
kg/kg
0.02
-
-
-
-
See Section
E.13.9
Page 384 of 570
-------
Input Parameter
Symbol
Unit
Constant Model
Parameter
Values
Variable Model Parameter Values
Rationale /
Basis
Value
Lower
Bound
Upper Bound
Mode
Distribution
Type
Container fill rate
RATEfin
containers/h
20
-
-
-
-
See Section
E.13.10
Equipment cleaning
operating hours
OHequip clean
h/day
4
-
-
-
-
See Section
E. 13.11
Spill loss fraction
Fspill
kg/kg
0.00013
4.5E-07
0.0018
0.00013
Triangular
See Section
E.13.12
Annual use rate of
fracturing fluids
containing 1,4-dioxane
Qsite_yr
gal/site-year
18013874.1
26,675.00
35,429,826.00
Discrete
See Section
E.13.13
Mass fraction of 1,4-
dioxane in hydraulic
fracturing additive
F dioxane additive
kg/kg
1.00E-04
2.3E-11
0.05
Discrete
See Section
E.13.14
Mass fraction of 1,4-
dioxane in hydraulic
fracturing fluid
F dioxane fracturing fluid
kg/kg
7.56E-08
1.00E-12
4.30E-06
Discrete
Saturation factor
F saturation factor
unitless
1
0.5
1.45
0.5
Triangular
See Section
E. 13.15
Container cleaning loss
fraction for totes
Fcont cleaning totes
kg/kg
0.002
0.0002
0.002
0.0007
Triangular
See Section
E.13.16
Container cleaning loss
fraction for drums
Fcont cleaning drums
kg/kg
0.03
0.017
0.03
0.025
Triangular
See Section
E.13.17
Fraction of injected
fracturing fluid that
returns to the surface
F recovered
kg/kg
0.75
0.02
1
0.75
Triangular
See Section
E. 13.18
Page 385 of 570
-------
E.13.3 Number of Sites
The only source of data EPA found for hydraulic fracturing sites that specifically use fracturing fluids
containing 1,4-dioxane was FracFocus. Therefore, EPA estimates 411 sites based on found the number
of hydraulic fracturing sites that reported using fracturing fluids containing 1,4-dioxane to FracFocus 3.0
(GWPC and IOGCC. 2022). EPA uses this estimate of sites that specifically use 1,4-dioxane and not an
estimate of all sites within the hydraulic fracturing industry because chemical-specific data and
assessments are preferred over assessments for generic sites that may or may not use 1,4-dioxane. That
said, these 411 sites only represent those that reported using 1,4-dioxane to FracFocus and there are
likely additional unaccounted for sites that may use 1,4-dioxane. This is an uncertainty of the
assessment.
E.13.4 Operating Days
The only source of data EPA found on the number of operating days at hydraulic fracturing sites that use
fracturing fluids containing 1,4-dioxane was FracFocus. Therefore, EPA modeled the operating days per
year using a discrete distribution of data points from FracFocus 3.0 for the 411 sites that reported using
fracturing fluids containing 1,4-dioxane (GWPC and IOGCC. 2022). The discrete distribution uses an
equal probability for each data point from FracFocus 3.0 submissions. The range of operating days and
summary statistics from the 411 FracFocus data points used in the discrete distribution are included in
TableApx E-21.
TableApx E-21. Summary Statistics on Number of Operating Days at
Hydraulic Fracturing Sites
Statistic
Operating Days (day/yr)
Maximum
72
99th Percentile
52
95th Percentile
32
50th Percentile
16
5th Percentile
4
Minimum
1
Mean
17
E.13.5 Container Size
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
the Agency used generic data from the Revised ESD on Chemicals Used in Hydraulic Fracturing (U.S.
22e). EPA modeled container size using a triangular distribution with a lower bound of 20
gallons, an upper bound of 1,000 gallons, and a mode of 55 gallons. The Revised ESD on Chemicals
Used in Hydraulic Fracturing states that hydraulic fracturing chemicals are received in drums or bulk
containers. Drums are defined as containing between 20 and 100 gallons of liquid, so EPA set the lower
bound of the triangular distribution at 20 gallons. Bulk containers (totes) are defined as containing
between 100 and 1,000 gallons of liquid, so EPA set the upper bound of the triangular distribution at
1,000 gallons. The ESD assumes 55-gallon as default for container size at wells, which EPA used as the
mode of the triangular distribution.
Page 386 of 570
-------
E.13.6 Diameter of Container Opening
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
the Agency used generic data from the ChemSTEER User Guide (U.S. EPA. ), which provides a
single diameter of container openings as 5.08 cm. Therefore, EPA could not develop a distribution of
values for this parameter and used the single value 5.08 cm from the ChemSTEER User Guide.
E.13.7 Diameter of Equipment Opening
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
the Agency used generic data from the ChemSTEER User Guide (U.S. EPA. ), which provides a
typical diameter of equipment openings as 92 cm. Therefore, EPA could not develop a distribution of
values for this parameter and used the single value 92 cm from the ChemSTEER User Guide.
E.13.8 Air Speed During Equipment Cleaning
EPA did not identify chemical-specific information for this parameter from systematic review;therefore,
the Agency used generic data from the ChemSTEER User Guide ( ), which provides a
single air speed of 440 ft/min during equipment cleaning activities. Therefore, EPA could not develop a
distribution of values for this parameter and used the single value 440 ft/min from the ChemSTEER
User Guide.
E.13.9 Equipment Cleaning Loss Fraction
EPA did not identify chemical-specific information for this parameter from systematic review used
generic data from standard EPA models. The EPA/OPPT Multiple Process Vessel Residual Model
provides a single loss fraction 0.02 kg of material remaining as equipment residual per kg of material
processed. Therefore, EPA could not develop a distribution of values for this parameter and used the
single value 0.02 kg/kg from the model ( 2015a).
E.13.10 Container Fill Rate
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
the Agency used generic data from the ChemSTEER User Guide (U.S. EPA. ), which provides a
typical fill rate of 20 containers per hour for drums and totes. Therefore, EPA could not develop a
distribution of values for this parameter and used the single value 20 containers/hour from the
ChemSTEER User Guide.
E.13.11 Equipment Cleaning Operating Hours
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
the Agency used generic data from the ChemSTEER User Guide ( ), which provides a
single duration of 4 hours/day for equipment cleaning of multiple vessels. Therefore, EPA could not
develop a distribution of values for this parameter and used the single value 4 hours/day from the
ChemSTEER User Guide.
E.13.12 Spill Loss Fraction
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
the Agency used generic data from the Revised ESD on Chemicals Used in Hydraulic Fracturing (U.S.
22e). The ESD recommends a default loss fraction of 0.00013 kg of fracturing fluid spilled per
kg of fracturing fluid handled. The ESD also indicates that the minimum loss fraction is 4.5 x 10~7 and a
maximum loss fraction is 0.0018. Therefore, EPA assessed a triangular distribution with a lower bound
of 4.5x 10~7 kg fracturing fluid spilled per kg fracturing fluid handled, and upper bound of 0.0018 kg
fracturing fluid spilled per kg fracturing fluid handled, and a mode of 0.00013 kg fracturing fluid spilled
per kg fracturing fluid handled.
Page 387 of 570
-------
E.13.13 Annual Use Rate of Fracturing Fluids Containing 1,4-Dioxane
The only source of data EPA found for hydraulic fracturing sites that specifically use fracturing fluids
containing 1,4-dioxane was FracFocus. Therefore, EPA modeled the annual use rate of fracturing fluids
containing 1,4-dioxane using a discrete distribution based on data obtained from FracFocus 3.0 for the
411 sites that reported using fracturing fluids containing 1,4-dioxane (GWPC and IOGCC. 2022). The
distribution uses an equal probability for each of the discrete data points from FracFocus 3.0. The range
of operating days and summary statistics from the 411 FracFocus data points used in the discrete
distribution are included in TableApx E-22. This range of annual use rate falls within the values
provided in the Revised ESD on Chemicals Used in Hydraulic Fracturing.
Table Apx E-22. Summary Statistics on the Annual Use Rate of
Fracturing Fluids at Hydraulic Fracturing Sites
Statistic
Annual Use Rate of Fracturing Fluids (gal/site-yr)
Maximum
35,429,826
99th Percentile
29,427,500
95th Percentile
25,644,872
50th Percentile
18,013,874
5th Percentile
6,136,351
Minimum
26,675
Mean
16,930,474
E.13.14 Mass Fraction of 1,4-Dioxane in Hydraulic Fracturing Additive/Fluid
The only source of data EPA found for hydraulic fracturing sites that specifically use fracturing fluids
containing 1,4-dioxane was FracFocus. Therefore, EPA modeled the mass fraction of 1,4-dioxane in the
hydraulic fracturing additive using a discrete distribution based on data from FracFocus 3.0 for the 411
sites that reported using fracturing fluids containing 1,4-dioxane (GWPC and IOGCC. 2022). The range
of mass fractions and summary statistics from the 411 FracFocus data points used in the discrete
distribution are included in Table Apx E-23.
Because hydraulic fracturing sites typically receive hydraulic fracturing additives, which are then
blending on-site into the fracturing fluid to be injected into the ground, a separate parameter for the mass
fraction of 1,4-dioxane in the hydraulic fracturing fluid was developed. EPA modeled this parameter
with discrete data from FracFocus 3.0 for the 411 sites that reported using fracturing fluids containing
1,4-dioxane (GWPC and IOGCC. 2022). The range of mass fractions and summary statistics from the
411 FracFocus data points used in the discrete distribution are included in Table Apx E-23.
EPA suspected that there may be a correlation between the mass fraction of 1,4-dioxane in the fracturing
fluid additive received at the sites and in the final hydraulic fracturing fluid that is injected into the
ground because the additive, and therefore 1,4-dioxane concentration, is essentially just diluted from the
mixing of various additives and water as a carrier fluid. Initial analysis in @Risk of the mass fraction of
1,4-dioxane in hydraulic fracturing additive and the mass fraction of 1,4-dioxane in hydraulic fracturing
fluid using a Pearson correlation resulted in a coefficient of 0.6, which indicates a moderately strong
correlation between the two sets of data. Due to the correlation between these two parameters, EPA
calculated the distributions for these parameters using equal probability of submitted pairs of the mass
Page 388 of 570
-------
fraction of 1,4-dioxane in hydraulic fracturing additive and the mass fraction of 1,4-dioxane in hydraulic
fracturing fluid from FracFocus 3.0 submissions (GWPC and IOGCC. 2022).
TableApx E-23. Summary Statistics on the Mass Fractions of 1,4-Dioxane in Hydraulic
Fracturing Additives and Fluids
Statistic
Mass Fraction of 1,4-Dioxane in
Hydraulic Fracturing Additive
Mass Fraction of 1,4-Dioxane in
Hydraulic Fracturing Fluid
Maximum
0.05
4.3E-06
99th Percentile
0.05
2.8E-06
95th Percentile
0.05
1.0E-06
50th Percentile
1.0E-04
7.6E-08
5th Percentile
1.0E-04
9.2E-09
Minimum
2.8E-11
1.0E-12
Mean
2.8E-03
2.7E-07
E.13.15
Saturation Factor
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
the Agency used generic data from the Chemical Engineering Branch Manual for the Preparation of
Engineering Assessments, Volume 1 [CEB Manual] ( ). The CEB manual indicates that
the saturation concentration was reached or exceeded by misting with a maximum saturation factor of
1.45 during splash filling. The CEB manual indicates that the saturation factor for bottom filling was
expected to be about 0.5 ( ). The underlying distribution of this parameter is not known;
therefore, EPA assigned triangular distributions, since triangular distribution is completely defined by
range and mode of a parameter. Because a mode was not provided for this parameter, EPA assigned a
mode value of 0.5 for bottom filling as bottom filling minimizes volatilization ( ). This
value also corresponds to the typical value provided in the ChemSTEER User Guide 0 v ( I \ j'i a)
for the EPA/OAQPS AP-42 Loading Model for drums.
E.13.16 Container Residual Fraction for Totes
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
the Agency used generic data from standard EPA models. Specifically, EPA modeled container residual
fraction for totes using a triangular distribution with a lower bound of 0.0002 kg residual/kg fracturing
fluid additive, and upper bound of 0.002 kg residual/kg fracturing fluid additive, and a mode of 0.0007
kg residual/kg fracturing fluid additive. The lower and upper bounds of this distribution are based on the
central tendency and high-end values listed in the EPA/OPPT Bulk Transport Residual Model from the
ChemSTEER User Guide ( ). EPA used the central tendency value as the mode of the
triangular distribution. Note that the underlying data for this model comes from a 1988 study by PEI
Associates Inc. that looked at literature sources and conducted a pilot-scale experiment to determine the
amount of residual material left in containers (PEI Associates. 1988). EPA reviewed the data from this
study and the underlying distribution of the container residual loss fraction is unknown; therefore, EPA
assigned a triangular distribution as discussed above.
E.13.17 Container Residual Fraction for Drums
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
the Agency used generic data from standard EPA models. Specifically, EPA modeled container residual
Page 389 of 570
-------
fraction for drums using a triangular distribution with a lower bound of 0.017 kg residual/kg fracturing
fluid additive, an upper bound of 0.03 kg residual/kg fracturing fluid additive, and a mode of 0.025 kg
residual/kg fracturing fluid additive. The lower bound is based on the minimum value for pumping and
the upper bound is based on the default high-end value in the EPA/OPPT Drum Residual Model from
the ChemSTEER User Guide ( ). EPA used the central tendency value for pumping as
the mode of the triangular distribution. Note that the underlying data for this model comes from a 1988
study by PEI Associates Inc. that looked at literature sources and conducted a pilot-scale experiment to
determine the amount of residual material left in containers (PEI Associates. 1988). EPA reviewed the
data from this study and the underlying distribution of the container residual loss fraction is unknown;
therefore, EPA assigned a triangular distribution as discussed above.
E.13.18 Fraction of Injected Fracturing Fluid that Returns to the Surface
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
the Agency used generic data from the Revised ESD on Chemicals Used in Hydraulic Fracturing (U.S.
22e). The Revised ESD o provides a range of fractions of injected fracturing fluid that returns to
the surface from three separate data sources, with a total range of 2 to 100 percent of fracturing fluid that
is injected into the ground being recovered at the surface ( Ze). The ESD uses 75 percent
as the default value. Based on this data, EPA modeled the fraction of injected fracturing fluid that
returns to the surface using a triangular distribution with a lower bound of 0.02 kg returned/kg injected,
an upper bound of 1 kg returned/kg injected, and a mode of 0.75 kg returned/kg injected. The remaining
amount is assumed to remain underground as a source of release (release point 6).
E.13.19 Key Strengths, Limitations, Uncertainties, and Sensitivity Analysis
General modeling uncertainties and limitations are discussed in Section 2.2.1.3 and Appendix E.7. This
section discusses model-specific uncertainties and limitations and presents examples of sensitivity charts
that EPA developed for this model. For multiple input parameters to this model, EPA used data from
FracFocus 3.0 for 41 1 sites that reported using fracturing fluids containing 1,4-dioxane (GWPC and
IOGCC. 2022). This is a strength of the assessment because these data are specific to sites that use 1,4-
dioxane in the United States. However, a limitation is that reporting to FracFocus is voluntary, so there
is uncertainty in the extent to which the data from these 411 sites are representative of all hydraulic
fracturing sites in the United States that use fracturing fluids containing 1,4-dioxane. All other input
parameters to the model are based on generic data from the Revised ESD on Chemicals Used in
Hydraulic Fracturing (U.S. EPA. 2022e) or standard EPA/OPPT models described in the ChemSTEER
User Guide (U.S. EPA. 2015a). This adds uncertainty with respect to the representativeness of the
generic input data towards hydraulic fracturing sites that use fracturing fluids containing 1,4-dioxane.
EPA ran a sensitivity analysis in @Risk to identify the input parameters which have the largest impact
on the daily release output estimates. For example, FigureApx E-9 shows the key inputs ranked by
decreasing impact on the mean fugitive air release during unloading hydraulic fracturing fluid additives,
which is release point 1 in this model. Figure Apx E-10 similarly shows the inputs that impact the daily
release from flowback and produced water, which corresponds to release point 8 in this model.
Figure Apx E-10 shows a dependency of the flowback and produced water release on loss fractions
from other release points like container cleaning and spills because this release point is in part based on
a mass balance approach, assuming 100 percent release of 1,4-dioxane by subtracting upstream releases.
The mass fraction of 1,4-dioxane in fracturing fluid additives received at sites and in the final fracturing
fluid formulation that is injected into the ground have the largest impact on both release point 1 and 8.
These two mass fraction parameters are based on 411 datapoints from FracFocus 3.0 and are paired,
meaning that there is a correlation between the two parameters. The annual use rate of fracturing fluids
containing 1,4-dioxane, which also impacts both release points, is similarly based on 411 datapoints
Page 390 of 570
-------
from FracFocus 3.0. For all other parameters in FigureApx E-9 and FigureApx E-10, EPA developed
distributions based on generic, not 1,4-dioxane-specific data. Having a distribution for each input
parameter is a strength of the assessment; however, the representativeness of the underlying data used
for these distributions is a limitation, as was discussed above.
Release Point 1- Daily Releases
Inputs Ranked by Effect on Output Mean
Mass Fraction of 1,4-Dioxane in Fracturing Additive and Fluid
(paired)
Operating Days
Saturation Factor
Annual use rate of fracturing fluids containing 1,4-dioxane
1
LI
Container size for fracturing fluids
1
Figure Apx E-9. Sensitivity Chart for Fugitive Air Release During Unloading (Daily Release Point
1) at Hydraulic Fracturing Sites
Page 391 of 570
-------
Release Point 8- Daily Releases
Inputs Ranked by Effect on Output Mean
Mass Fraction of lr4-Dioxane in Fracturing Additive and
Fluid (paired)
Annual use rate of fracturing fluids containing 1,4-
dioxane
Fraction of injected fracturing fluid that returns to the
surface
Operating Days
Container cleaning loss fraction for totes
II
Container deaning loss fraction for drums
Spil loss fraction
FigureApx E-10. Sensitivity Chart for Release from Flowback and Produced Water (Daily
Release Point 8) at Hydraulic Fracturing Sites
E.14 Dish Soap and Dishwasher Detergent Modeling Approach and
Parameters for Estimating Environmental Releases
This appendix presents the modeling approach and equations used to estimate environmental releases of
1,4-dioxane during the industrial and commercial use of dish soaps and dishwasher detergents. This
approach utilizes data from a public comment (P&G. 20231 concentration data from New York State
Department of Environmental Conservation (NYDEC) approved waivers for 1,4-dioxane in consumer
products (NYDEC. 20231 and standard EPA models combined with Monte Carlo simulation (a type of
stochastic simulation). Figure Apx E-l 1 is a diagram of the release and exposure points during the use
of dish soap and dishwasher detergent.
Page 392 of 570
-------
B
Air Emissions to
Unloading Dish Soap Worker Breathing
from Containers zone from Washing
Soap Formulation
Dish Soap/
Detergent Arrives in
Containers
Container Disposal
Hand Washing or
Automated Dish
Washing
4
¦ Dirty Water
FigureApx E-ll. Environmental Release Points (Numbered) and Occupational Exposure Points
(Lettered) During Industrial and Commercial Use of Dish Soap and Dishwasher Detergent
Based on Figure Apx E-ll, EPA identified the following release points:
• Release point 1 (RP1): Fugitive air releases during container unloading;
• Release point 2 (RP2): Container disposal losses to landfill;
• Release point 3 (RP3): Fugitive air releases during washing; and
• Release point 4 (RP4): Dirty water down the sink to POTW.
Environmental releases of dish soap and dishwasher detergent are a function of the chemical's physical
properties, daily throughput of soap/detergent, container size, mass fractions, and other model
parameters. Although physical properties are fixed, some model parameters are expected to vary from
one facility to another. An individual model input parameter could either have a discrete value or a
distribution of values. EPA assigned statistical distributions based on available literature data or
engineering judgment to address the variability in mass fraction of 1,4-dioxane in the soap or detergent
(Fdioxane soap/detergent), container size (Vcont), daily use rate of soap or detergent (QSOap/detergent day), air speed
(RATEair), duration of release (OHSOap/diswasher), saturation factor (fsat), container residue fractions (LFCOnt),
and diameter of sink opening (Dsink).
A Monte Carlo simulation was conducted to capture variability in the model input parameters described
above. The simulation was conducted using the Latin hypercube sampling method in @Risk (Palisade,
Ithaca, New York). The Latin hypercube sampling method is a statistical method for generating a
sample of possible values from a multi-dimensional distribution. Latin hypercube sampling is a stratified
method, meaning it guarantees that its generated samples are representative of the probability density
function (variability) defined in the model. EPA performed 100,000 iterations of the model to capture
the range of possible input values, including values with low probability of occurrence.
From the distribution resulting from the Monte Carlo simulation, EPA selected the 95th and 50th
percentile values to represent a high-end release and central tendency release level respectively. The
statistics were calculated directly in @Risk. The following subsections detail the model design equations
and parameters used for environmental release estimates.
Page 393 of 570
-------
E.14.1 Model Equations
Facility annual throughput is calculated using the equations below, the first being for dish soap and the
second being for dishwasher detergent:
EquationApx E-41.
Qsoap_yr ~ Qsoap_day * OD
or
Qdetergent_yr ~ Qdetergent_day * OD
Where:
Qsoap_yr
Qdetergent_yr
Q soap_day
Q detergent_day
OD
Annual use rate of dish soap [kg/site-yr]
Annual use rate of dishwasher detergent [kg/site-yr]
Daily use rate of dish soap [kg/site-day]
Daily use rate of dishwasher detergent [kg/site-day]
Operating days [days/yr]
Daily use rate of 1,4-dioxane is calculated using the equations below, the first being for dish soap and
the second being for dishwasher detergent:
Equation Apx E-42.
Qd
ioxane_day
= Q
soap_day * Pdioxane_soap
or
Qd
ioxane_day
Qdetergent_day * ^dioxane _deter gent
Where:
Qdioxane_day
Q soap_day
Q detergent_day
Fdioxane_soap
Fdioxane_detergent
Daily use rate of 1,4-dioxane [kg/site-day]
Daily use rate of dish soap [kg/site-day]
Daily use rate of dishwasher detergent [kg/site-day]
Mass fraction of 1,4-dioxane in dish soap [kg/kg]
Mass fraction of 1,4-dioxane in dishwasher detergent [kg/kg]
Annual use rate of 1,4-dioxane is calculated using the equation below:
Equation Apx E-43.
Qd
ioxane_yr
Qd
ioxane_day
* OD
Where:
Qdioxane_yr
Qdioxane_day
OD
Annual use rate of 1,4-dioxane [kg/site-yr]
Daily use rate of 1,4-dioxane [kg/site-day]
Operating days [days/yr]
Number of containers unloaded per year is calculated using the equation below:
Page 394 of 570
-------
EquationApx E-44.
JV,
Q soap/deter gent.
yr
Where:
Ncont_unload _yr
Qsoap_yr
Qdetergent_yr
Vcont
RHO
cont_unload_yr ~ i
Vcont* 3.79 RHO
Number of containers unloaded per site per year [containers/site-
year]
Annual use rate of dish soap [kg/site-yr]
Annual use rate of dishwasher detergent [kg/site-yr]
Container volume [gal/container]
Dish soap/detergent density [kg/L]
Number of containers unloaded per day is calculated using the equation below:
Equation Apx E-45.
JV,
JV,
cont_unload_day
cont_unload_yr
~OD
Where:
JV,
cont_unload _day
JV,
cont_unload _yr
OD
Number of containers unloaded per site per day [containers/site-
day]
Number of containers unloaded per site per year [containers/site-
year]
Operating days [days/yr]
Daily operating hours for unloading containers is calculated using the equation below:
Equation Apx E-46.
OH.
JV,
cont_unload_yr
unload_cont
OD* RATE,
unload
Where:
OH.
JV,
unload _cont
cont_unload _yr
OD
RATEunload
Daily operating hours for unloading containers [hours/day]
Number of containers unloaded per site per year [containers/site-
year]
Operating days [days/yr]
Container unloading rate [containers/hour]
Release Point 1 daily release per site (fugitive emissions during unloading) is calculated using the
EPA/OAQPS AP-42 Loading Model equation below:
Equation Apx E-47.
Release_perDayRP1=
s kg
0 HUnload_cont * * fsat * MW * (3785.4 * Vcont) * RATEunioaci * Fdioxane_soap/deter gent *
VP
760
3600 *T * R
Page 395 of 570
-------
Where:
Release_perDayRP1 =
OHunioaci cont —
fsat ~
MW
Vcont ~
RATEunload —
Fdioxane_soap ~
Fdioxane_detergent ~
VP
T
R
Daily 1,4-dioxane release at release point 1 [kg/site-day]
Daily operating hours for unloading containers [hours/day]
Saturation factor [dimensionless]
1,4-dioxane molecular weight [g/mol]
Container volume [gal/container]
Container unloading rate [containers/hour]
Mass fraction of 1,4-dioxane in dish soap [kg/kg]
Mass fraction of 1,4-dioxane in dishwasher detergent [kg/kg]
Vapor pressure of l,4-dioxane[torr]
Ambient temperature [K]
Universal gas constant [atm-cm3/gmol-L]
Release Point 2 daily release per site (container disposal) is calculated using the equations and criteria
below:
EquationApx E-48.
If Ncont_unload_yr
-------
0Hsoap
OHdiShwaSher
MW
Fdioxane_soap
Fdioxane_detergent
VP316
^355
RATEair
Dsink
Twash_soap
Twash_auto
P
Daily operating hours for hand washing [hours/day]
Daily operating hours for dishwasher operation [hours/day]
1,4-dioxane molecular weight [g/mol]
Mass fraction of 1,4-dioxane in dish soap [kg/kg]
Mass fraction of 1,4-dioxane in dishwasher detergent [kg/kg]
Vapor pressure of 1,4-dioxane at a hand washing temperature of
316 K [torr]
Vapor pressure of 1,4-dioxane at a dishwasher temperature of 355
K [torr]
Air speed [cm/s]
Diameter of sink opening or dishwasher vent [cm]
Dish soap wash water temperature [K]
Dishwasher water temperature [K]
Atmospheric pressure [atm]
Release Point 4 daily release per site (dirty water) is calculated using the equation below:
EquationApx E-50.
3
Release_perDayRP4 Qdioxane_day I Release_perDayRPi
i=1
Where:
Release_perDayRP4 = Daily 1,4-dioxane release at release point 4[kg/site-day]
Qdioxane_day = Daily use rate of 1,4-dioxane [kg/site-day]
Łf=1 Release_perDayRPi = The sum of release points 1-3 emissions [kg/site-day]
E.14.2 Model Input Parameters
Table Apx E-24 summarizes the model parameters and their values for the Monte Carlo simulation.
Additional explanations of EPA's selection of the distributions for each parameter are provided after this
table. High-end and central tendency releases are estimated by selecting the 50th and 95th percentile
values from the output distribution.
Page 397 of 570
-------
TableApx E-24. Summary of Parameter Values and Distributions Used in the Industrial and Commercial Use of Dish Soap and
Dishwasher Detergent Release Model
Input Parameter
Symbol
Unit
Constant Model
Parameter Values
Variable Model Parameter Values
Rationale/
Basis
Value
Lower
Bound
Upper
Bound
Mode
Distribution
Type
Facility Daily Throughput -
Dish Soap
Qsoap day
kg/site-day
7.2
3
7.2
-
Uniform
See Section
E.14.3
Facility Daily Throughput -
Dishwasher Detergent
Qdetergent day
kg/site-day
6.4
3.2
6.4
-
Uniform
See Section
E.14.4
Concentration of 1,4-Dioxane in
Dish Soap
Fdioxane soap
kg/kg
8.4E-06
3.00E-08
2.04E-04
-
Discrete
See Section
E.14.5
Concentration of 1,4-Dioxane in
Dishwasher Detergent
Fdioxane detergent
kg/kg
8.4E-06
4.00E-07
5.76E-05
-
Discrete
See Section
E.14.6
Saturation Factor
fiat
dimensionless
0.5
0.5
1.45
0.5
Triangular
See Section
E.14.7
Container Size
Vcont
gal
1
1
20
1
Triangular
See Section
E.14.8
Container Residual Loss
Fraction
LF cont
kg/kg
0.003
0.0003
0.006
0.003
Triangular
See Section
E.14.9
Diameter of Sink Opening
Dsink
cm
51.3
51.3
76.9
51.3
Triangular
See Section
E.14.10
Release Duration for
Dishwashers
Otldishwasher
hrs/day
2.5
0
2.5
-
Uniform
See Section
E. 14.11
Release Duration for Dish Soap
Otlsoap
hours/day
8
-
-
-
-
Assumed Full-
Shift
Number of Sites
Nsites
sites
773,851
-
-
-
-
See Section
E.14.12
Operating Days
OD
days/yr
350
-
-
-
-
See Section
E.14.13
Container Unloading Rate
RATEunload
containers/hr
60
-
-
-
-
See Section
E.14.14
Page 398 of 570
-------
Input Parameter
Symbol
Unit
Constant Model
Parameter Values
Variable Model Parameter Values
Rationale/
Basis
Value
Lower
Bound
Upper
Bound
Mode
Distribution
Type
Density of Soap & Detergent
RHO
kg/L
1
EPA assumes a
density equal to
that of water
1,4-Dioxane Ambient Vapor
Pressure
VP
torr
40
-
-
-
-
Physical
property
1,4-Dioxane Molecular Weight
MW
g/mol
88.1
-
-
-
-
Physical
property
Ambient Temperature
T
K
298
-
-
-
-
Process
Parameter
Ambient Pressure
P
atm
1
-
-
-
-
Process
Parameter
Universal Gas Constant
R
atm-
cm3/gmol-L
82.05
-
-
-
-
Universal
constant
Dish Soap Wash Water
Temperature
Twash soap
K
316
-
-
-
-
See Section
E.14.15
Dishwasher Water Temperature
Twash auto
K
355
—
—
—
—
See Section
E.14.16
1,4-Dioxane Vapor Pressure at
316 K
VP316
torr
79.35
-
-
-
-
Physical
property
1,4-Dioxane Vapor Pressure at
355 K
VP355
torr
161
-
-
-
-
Physical
property
Air Speed
RATEair
ft/min
100
-
-
-
-
See Section
E.14.17
Page 399 of 570
-------
E.14.3 Facility Daily Throughput - Dish Soap
EPA did not identify chemical-specific or industry-specific information for this parameter from
systematic review; therefore, EPA used generic data from the Consumer Exposure Model (CEM). For
dish soap, the CEM cites a use rate of 75 g to 125 g per use of dish soap, with a use duration of 5 to 20
minutes. EPA scaled up these consumer use rates from the CEM by assuming an 8-hour shift duration
for occupational settings. Based on this, EPA modeled facility daily throughput using a uniform
distribution with a lower bound of 3 kg/site-day and an upper bound of 7.2 kg/site-day.
E.14.4 Facility Daily Throughput - Dishwasher Detergent
EPA did not identify chemical-specific or industry-specific information for this parameter from
systematic review; therefore, the Agency used generic data from the Consumer Exposure Model (CEM).
For dishwasher detergent, the CEM cites a use rate of 20 to 40 grams of detergent per cycle. The public
comment states that there are up to 160 cycles run per day at commercial dishwashing locations (P&G.
2023). Therefore, EPA scaled up the consumer values from the CEM for an occupational setting by
multiplying 20 to 40 grams by 160 cycles/day. Based on this, EPA modeled facility daily throughput
using a uniform distribution with a lower bound of 3.2 kg/site-day and an upper bound of 6.4 kg/site-
day.
E.14.5 Concentration of 1,4-Dioxane in Dish Soap
EPA found data on the concentration of 1,4-dioxane in dish soap from literature sources (Lin et at..
2017; Saraii and Shirvani. . Oavarani et at.. 2012; Makino et at.. 2006; Wala-Jerzykiewicz and
Szymanowsl |), the December 2020 Final Risk Evaluation for 1,4-Dioxane ( 020c).
and product concentration waiver data from the NY DEC CNYDEC. 2023). EPA modeled the
concentration of 1,4-dioxane in dish soap using a discrete distribution based on the 42 data points from
the aforementioned sources, as shown in Table Apx E-25, with equal probability given to each discrete
data point.
Table Apx E-25. Discrete Data Points on Concentration of 1,4-Dioxane in Dish Soap
Concentration of 1,4-Dioxane in Dish Soap (kg 1,4-dioxane/kg soap)
3.0E-08
2.9E-06
8.4E-06
1.0E-05
5.8E-05
4.0E-07
2.9E-06
8.4E-06
1.0E-05
2.0E-04
7.0E-07
3.7E-06
8.4E-06
1.2E-05
-
7.5E-07
4.5E-06
8.4E-06
1.2E-05
-
1.2E-06
4.8E-06
8.4E-06
1.2E-05
-
2.0E-06
4.8E-06
1.0E-05
1.2E-05
-
2.4E-06
5.0E-06
1.0E-05
1.4E-05
-
2.5E-06
7.9E-06
1.0E-05
1.4E-05
-
2.5E-06
7.9E-06
1.0E-05
2.0E-05
-
2.5E-06
8.4E-06
1.0E-05
5.1E-05
-
Sources: fNYDEC. 2023; U.S. EPA. 2020c; Lin et al.. 2017; Saraii and Shirvani. 2C . 'arani
et al., 2012; Makino et al., 2006; Wala-Jerzvkiewicz and Szvmanowski, 1998)
E.14.6 Concentration of 1,4-Dioxane in Dishwasher Detergent
EPA found data on the concentration of 1,4-dioxane in dishwasher detergent from literature sources (Lin
et al.. 2017; Saraii and Shirvani. 2017; Davarani et al.. 2012; Makino et al.. 2006; Wala-Jerzykiewicz
Page 400 of 570
-------
and Szymanowski. 1998). the December 2020 Final Risk Evaluation for 1,4-Dioxane (
2020c). and product concentration waiver data from the NY DEC ( C. 2023). EPA modeled the
concentration of 1,4-dioxane in dishwasher detergents using a discrete distribution based on the 42 data
points from the aforementioned sources, as shown in TableApx E-26, with equal probability given to
each discrete data point.
Table Apx E-26. Discrete Data Points on Concentration of 1,4-
Dioxane in Dishwasher Detergent
Concentration of 1,4-Dioxane in Dishwasher Detergent
(kg 1,4-dioxane/kg detergent)
4.0E-07
3.0E-06
8.4E-06
1.0E-05
5.1E-05
8.6E-07
4.5E-06
8.4E-06
1.0E-05
5.8E-05
8.6E-07
4.8E-06
8.4E-06
1.0E-05
-
2.0E-06
4.8E-06
8.4E-06
1.2E-05
-
2.4E-06
5.0E-06
9.7E-06
1.2E-05
-
2.5E-06
6.5E-06
9.7E-06
1.2E-05
-
2.5E-06
7.9E-06
1.0E-05
1.2E-05
-
2.5E-06
7.9E-06
1.0E-05
1.4E-05
-
2.9E-06
8.4E-06
1.0E-05
1.4E-05
-
2.9E-06
8.4E-06
1.0E-05
2.0E-05
-
Sources: (NYDEC. 2023; U.S. EPA. 2020c; Lin et ah. 2017; Saraii and
Sliirvani, 2017; Davarani et ah, 2012; Makino et ah, 2006; Wala-Jerzvkiewicz
and Szvmanowst » (nnS)
E.14.7 Saturation Factor
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
the Agency used generic data from the Chemical Engineering Branch Manual for the Preparation of
Engineering Assessments, Volume 1 [CEB Manual] (U.S. EPA. 1991). The CEB manual indicates that
the saturation concentration was reached or exceeded by misting with a maximum saturation factor of
1.45 during splash filling. The CEB manual indicates that the saturation factor for bottom filling was
expected to be about 0.5 ( ). The underlying distribution of this parameter is not known;
therefore, EPA assigned triangular distributions, since triangular distribution is completely defined by
range and mode of a parameter. Because a mode was not provided for this parameter, EPA assigned a
mode value of 0.5 for bottom filling as bottom filling minimizes volatilization ( ). This
value also corresponds to the typical value provided in the ChemSTEER User Guide ( a)
for the EPA/OAQPS AP-42 Loading Model for small containers.
E.14.8 Container Size
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
the Agency used generic data provided in a public comment and the ChemSTEER User Guide (U.S.
). The public comment indicated that liquid dish soap and detergent are commonly packaged
in 1- and 5-galIon containers, with 1 -gallon containers the most common size (P&G. 2023). EPA
expects sites to have variation in the sizes of soap/detergent containers, so EPA also used information
from the ChemSTEER User Guide ( ), which defines small containers as containing
Page 401 of 570
-------
between 5 and 20 gallons of liquid. Based on these data, EPA modeled container size using a triangular
distribution with a lower bound of 1 gallon, an upper bound of 20 gallons, and a mode of 1 gallon.
E.14.9 Container Residual Loss Fraction
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
the Agency used generic data from standard EPA models. Specifically, EPA modeled container residual
loss fraction for drums using a triangular distribution with a lower bound of 0.0003 kg residual/kg
dishwashing product, an upper bound of 0.006 kg residual/kg dishwashing product, and a mode of 0.003
kg residual/kg dishwashing product. The mode and upper bound of the distribution are based on the
central tendency and high-end values listed in the EPA/OPPT Small Container Residual Model from the
ChemSTEER User Guide ( ). Note that the underlying data for this model comes from a
1988 study by PEI Associates Inc. that looked at literature sources and conducted a pilot-scale
experiment to determine the amount of residual material left in containers (PEI Associates. 1988). EPA
reviewed the data from this study and the underlying distribution of the container residual loss fraction is
unknown; therefore, EPA assigned a triangular distribution as discussed above.
E.14.10 Diameter of Sink Opening
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
the Agency used generic data provided in a public comment. The public comment states that the most
common 3-compartment sink size is 16" x 20", though they can be up to 30" x 24" (P&G. 2023). The
model requires a diameter of a circular opening, so EPA converted the surface area of the rectangles
from the public comment to circles with 51.3 and 76.9 cm diameters. Based on this, EPA modeled the
diameter of the sink used for dishwashing using a triangular distribution with a lower bound and mode
of 51.3 cm and an upper bound of 76.9 cm.
E.14.11 Release Duration for Dishwashers
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
the Agency used generic data provided in a public comment. The public comment indicates that there
are approximately 160 dishwashing cycles per 8-hour shift, meaning each cycle is approximately 3
minutes in length (P&G. 2023). Additionally, the comment explains that each dishwashing cycle is
comprised of loading dirty dishes into the dishwasher, washing, and emptying dishwashers. Since
potential vapor releases are only expected when the cycle is completed and the dishwasher is open, EPA
approximated this as one third of the cycle time, or up to 2.5 hours/day. Based on this, EPA modeled the
duration of release for dishwasher cycles using a uniform distribution with a lower bound of 0 hours/day
and an upper bound of 2.5 hours/day. The uniform distribution uses 0 hours/day at the low end to
account for machines that drain before being opened and thus have a smaller potential for vapor
releases.
E.14.12 Number of Sites
EPA did not find data on the number of industrial and commercial sites that specifically use dish soaps
and detergents containing 1,4-dioxane. As a bounding estimate for the number of use sites, EPA used
U.S. Census and BLS data for the NAICS codes 623300 (Continuing Care Retirement Communities and
Assisted Living Facilities for the Elderly), 713900 (Other Amusement and Recreation Industries),
721100 (Traveler Accommodation), 721300 (Rooming and Boarding Houses, Dormitories, and
Workers' Camps), 722300 (Special Food Services), 722400 (Drinking Places (Alcoholic Beverages)),
and 722500 (Restaurants and Other Eating Places) to estimate a total of 773,851 sites within the industry
(r S HIS ). This is the same estimate described in Section E.l.
Page 402 of 570
-------
E.14.13 Operating Days
EPA did not identify chemical-specific information for this parameter from systematic review. The
Agency could not develop a distribution of values for this parameter and assumed operation occurs 7
days/week, 50 weeks/year, for a total of 350 days/year.
E.14.14 Container Unloading Rate
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
the Agency used generic data from the ChemSTEER User Guide ( ). The ChemSTEER
User Guide provides a typical fill rate of 60 containers per hour for containers smaller than 20 gallons of
liquid. Therefore, EPA could not develop a distribution of values for this parameter and used the single
value 60 containers/hour from the ChemSTEER User Guide.
E.14.15 Dish Soap Wash Water Temperature
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
the Agency used generic data provided in a public comment. The public comment provided information
on the temperature of wash water in handwashing, indicating that dish sink water is kept at or above 110
°F (316 K) according to food code (P&G. 2023). This was the only data point available for this
parameter. Therefore, EPA could not develop a distribution of values for this parameter and used the
single value of 316 K.
E.14.16 Dishwasher Water Temperature
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
the Agency used generic data provided in a public comment. The public comment provided information
on the temperature of wash water in automated dishwashers, indicating that a high temperature
dishwashing machine operates at up to 180°F (355 K) (P&G. 2023). This was the only data point
available for this parameter. Therefore, EPA could not develop a distribution of values for this
parameter and used the single value of 355 K.
E.14.17 Indoor Air Speed
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
the Agency used generic data from Baldwin (1998). a source known from previous EPA model
development. Baldwin (1998) measured indoor air speeds across a variety of occupational settings in the
United Kingdom. Fifty-five work areas were surveyed across a variety of workplaces. EPA analyzed the
air speed data from Baldwin (1998) and categorized the air speed surveys into settings representative of
industrial facilities and representative of commercial facilities.
EPA fit a lognormal distribution for both data sets as consistent with the authors observations that the air
speed measurements within a surveyed location were lognormally distributed and the population of the
mean air speeds among all surveys were lognormally distributed. Since lognormal distributions are
bound by zero and positive infinity, EPA truncated the distribution at the largest observed value among
all of the survey mean air speeds from Baldwin (1998). The Agency fit the air speed surveys
representative of industrial facilities to a lognormal distribution with the following parameter values:
mean of 22.414 cm/s and standard deviation of 19.958 cm/s. In the model, the lognormal distribution is
truncated at a maximum allowed value of 202.2 cm/s (largest surveyed mean air speed observed in
Baldwin (1998)) to prevent the model from sampling values that approach infinity or are otherwise
unrealistically large.
Page 403 of 570
-------
Baldwin (1998) only presented the mean air speed of each survey. The authors did not present the
individual measurements within each survey. Therefore, these distributions represent a distribution of
mean air speeds and not a distribution of spatially variable air speeds within a single workplace setting.
E.14.18 Key Strengths, Limitations, Uncertainties, and Sensitivity Analysis
General modeling uncertainties and limitations are discussed in Section 2.2.1.3 and Appendix E.7. This
section discusses model-specific uncertainties and limitations and presents examples of sensitivity charts
that EPA developed for this model. For this model, the only 1,4-dioxane specific input parameter data is
for the concentration of 1,4-dioxane in dish soaps and detergents. All other parameters are based on
generic data from a variety of sources. For some parameters, EPA used information from a public
comment; this information is not 1,4-dioxane-specific but is industry-specific as the information comes
directly from a manufacturer of soaps and detergents (P&G. 2023). For other parameters, EPA used
generic data from standard EPA/OPPT models described in the ChemSTEER User Guide (U.S. EPA.
2015a) and from the Consumer Exposure Model. While EPA did scale values from the Consumer
Exposure Model for application in this commercial use model, the consumer data and scaling approach
add uncertainty to the model. Further, the use of generic data adds uncertainty with respect to the
representativeness of the generic input data towards dishwashing sites that use soaps and detergent
containing 1,4-dioxane.
EPA ran a sensitivity analysis in @Risk to identify the input parameters which have the largest impact
on the daily release output estimates. For example, FigureApx E-12 shows the inputs ranked by which
have the largest effect on the mean release from disposal of empty soap containers, which is release
point 2 in this model. Figure Apx E-13 similarly shows the inputs that impact the daily release from
dishwashing (e.g., fugitive releases and dirty water disposal), which corresponds to release point 4 in
this model. The mass fraction of 1,4-dioxane in soaps has the largest impact on both releases. This mass
fraction is based on 42 datapoints from literature sources, the December 2020 Final Risk Evaluation for
1,4-Dioxane, and product concentration waiver data from the NYDEC, as discussed in Appendix E. 14.5.
The use of this 1,4-dioxane-specific data from multiple different sources is a strength of the assessment.
For all other parameters in Figure Apx E-12 and Figure Apx E-13, EPA developed distributions based
on generic, not 1,4-dioxane-specific data. Having a distribution for each input parameter is a strength of
the assessment; however, the representativeness of the underlying data used for these distributions is a
limitation, as was discussed above.
Page 404 of 570
-------
Release Point 2 - Daily
Inputs Ranked by Effect on Output Mean
Mass Fraction of Dioxane in Soap
Container Size
Container Residual Loss Fraction
Daily Throughput of Dish Soap
1
FigureApx E-12. Sensitivity Chart for Container Disposal (Daily Release Point 2) at Dishwashing
Sites
Mass Fraction of Dioxane in Soap
Diameter of Sink Opening
Daily Throughput of Dish Soap
Container Residual Loss Fraction
Container Size
Saturation Facter
Release Point 4- Daily
Inputs Ranked by Effect on Output Mean
1
Figure Apx E-13. Sensitivity Chart for Releases from Dishwashing (Daily Release Point 4) at
Dishwashing Sites
Page 405 of 570
-------
Appendix F OCCUPATIONAL EXPOSURES
F,1 Calculating Acute and Chronic Inhalation Exposures and Dermal
Doses
For inhalation exposures, this risk evaluation assessed 1,4-dioxane exposures to workers in occupational
settings, presented as 8-hour TWA. The 8-hour TWA exposures were used to calculate average daily
concentration (ADC) for chronic, non-cancer risks, and lifetime average daily concentration (LADC) for
chronic, cancer risks. Refer to Appendix G.2 of the December 2020 Final Risk Evaluation for 1,4-
Dioxane ( )20c) for the equations EPA used for these inhalation exposure calculations. Refer
to Appendix G.3 of the 2020 RE for sample calculations.
For dermal exposures, this risk evaluation assessed 1,4-dioxane exposures to worker in occupational
settings, presented as daily dermal potential dose rates (mg/day). The potential dose rates were then used
to calculate acute retained doses (ARD), and chronic retained doses (CRD) for non-cancer and cancer
risks. Refer to Appendix G.7.6 of the December 2020 Final Risk Evaluation for 1,4-Dioxane (U.S. EPA.
2020c) for the equations EPA used for these dermal dose calculations. Refer to Appendix G.3 of the
December 2020 Risk Evaluation for sample calculations.
F.2 Approach for Estimating Number of Workers and Occupational Non-
users
EPA used the same approach for estimating the number of workers and occupational non-users (ONUs)
potentially exposed to the OES (listed in Section 3.1.1) as presented in the December 2020 Final Risk
Evaluation for 1,4-Dioxane (U.S. EPA. 2020c). Refer to Appendix G.5 of the December 2020 Risk
Evaluation for explanation of this approach.
TableApx F-l contains a summary of the total number of workers and ONUs for each supplemental
OES corresponding to estimated exposures for this supplemental risk evaluation.
Page 406 of 570
-------
Table Apx F-l. Summary of Total Number of Workers and ONUs Pol
entially Exposed
to 1,4-Dioxane for Each Supplemental OES"
OES
Total Exposed
Workers
Total Exposed
ONUs
Total Exposed
Number of
Facilities
Notes
Textile dye
5,353
2,634
7,987
783
Bounding estimate based on U.S. Census Bureau
data for NAICS code 313310, Textiles and Fabric
Finishing Mills.
Antifreeze
182,615
18,096
200,711
84,383
Bounding estimate based on U.S. Census Bureau
data for NAICS codes 811111, General
Automotive Repair, and 811198, All Other
Automotive Repair and Maintenance.
Surface cleaner
552,300
32,133
584,433
55,998
Bounding estimate for the industry is based on U.S.
Census Bureau data for NAICS code 561720,
Janitorial Services.
Dish Soap
465,270
881,870
1,347,140
773,851
Bounding estimate for the industry is based on U.S.
Census Bureau data for NAICS codes 623300,
713900,721100,721300,722300,722400,and
722500.
Dishwasher
detergent
465,270
881,870
1,347,140
773,851
Bounding estimate for the industry is based on U.S.
Census Bureau data for NAICS codes 623300,
713900,721100,721300,722300,722400,and
722500.
Laundry detergent
(industrial)
66,231
7,359
73,590
2,453
Bounding estimate based on U.S. Census Bureau
data for NAICS code 812330, Linen and Uniform
Supply.
Laundry detergent
(institutional)
573,198
Unknown
Unknown
95,533
Bounding estimate based on industry information
as described in the ESD on Water Based Washing
operations at Industrial and Institutional Laundries
COECD.: ).
Paint and floor
lacquer
111,511
11,050
122,561
33,648
Bounding estimate based on U.S. Census Bureau
data for NAICS code 811121, Automotive Body,
Paint, and Interior Repair and Maintenance.
PET byproduct
43,528
17,195
60,723
1,695
Bounding estimate based on U.S. Census Bureau
data for NAICS codes 325211 and 326113.
Page 407 of 570
-------
OES
Total Exposed
Workers
Total Exposed
ONUs
Total Exposed
Number of
Facilities
Notes
Ethoxylation
process byproduct
64,926
24,835
89,761
2,730
Bounding estimate based on U.S. Census Bureau
data for NAICS codes 325110, 325199, 325611,
325613, and 325998.
Hydraulic
fracturing
46,315
26,007
72,322
411
Estimate for the number of facilities is based on the
number of fracking sites that reported using 1,4-
dioxane to FracFocus 3.0 (GWPC and IOGCC.
2022). Estimates for number of workers and ONUs
are based on per site estimates from U.S. Census
Bureau data for NAICS codes 213111 and 213112,
multiplied by the number of fracking sites from
FracFocus 3.0.
a EPA's approach and methodology for using U.S. Census Bureau data to estimate the number of facilities using 1,4-dioxane and the number of workers and
ON Us potentially exposed to 1.4-dioxane can be found in the December 2020 Final Risk Evaluation for 1,4-Dioxane (U.S. EPA, 2020c).
Page 408 of 570
-------
F.3 Occupational Dermal Exposure Assessment Method
To assess dermal exposure, EPA used the same modeling approach as that described in Appendix G.7 of
the December 2020 Final Risk Evaluation for 1,4-Dioxane ( )20c). Specifically, EPA used
the EPA Dermal Exposure to Volatile Liquids Model to calculate the dermal retained dose for each
COU included in this supplemental risk evaluation. The equation modifies the EPA 2-Hand Dermal
Exposure to Liquids Model by incorporating a "fraction absorbed (fabs)" parameter to account for the
evaporation of volatile chemicals and a "protection factor (PF)" to account for glove use. The ECETOC
TRA v3 model represents the protection factor of gloves as a fixed, assigned protection factor equal to 5,
10, or 20 (Marquart et al..: ). Given the limited state of knowledge about the protection afforded by
gloves in the workplace, EPA utilize the PF values of the ECETOC TRA v3 model (Marquart et al..
2017) as shown in TableApx F-2 rather than attempt to derive new values.
The fraction absorbed (fabs) for 1,4-dioxane is estimated to be 0.86 in commercial settings with lower
indoor wind speeds and 0.78 in industrial settings with higher indoor wind flows based on a theoretical
framework provided by Kasting and Miller (2006), indicating that 86 or 78 percent of the applied dose is
retained by the stratum corneum, the outermost layer of the epidermis, and absorbed systemically.
Additional details on this approach can be found in the December 2020 Final Risk Evaluation for 1,4-
Dioxane ( )20c).
Table Apx F-2. Glove Protection Factors for Different Dermal Protection Strategies from
ECETOC TRA v3
Dermal Protection Characteristics
Setting
Protection
Factor (PF)
a. No gloves used, or any glove/gauntlet without permeation data and
without employee training
1
b. Gloves with available permeation data indicating that the material of
construction offers good protection for the substance
Industrial and
Commercial
Uses
5
c. Chemically resistant gloves (i.e., as b above) with "basic" employee
training
10
d. Chemically resistant gloves in combination with specific activity training
(e.g., procedure for glove removal and disposal) for tasks where dermal
exposure can be expected to occur
Industrial Uses
Only
20
Source: (Marauart et al.. 2017)
Occupational Dermal Exposure Assessment Bins
The December 2020 Final Risk Evaluation for 1,4-Dioxane included six "bins" of OES (Bins 1 through
6) for the occupational dermal analysis ( 2020c). This supplemental risk evaluation builds off
that analysis with the inclusion of nine additional "bins" of OES, described below.
Bin 7: covers the use of 1,4-dioxane present in textile dyes, which EPA expects may involve both
commercial and industrial facilities. Workers may be exposed to 1,4-dioxane during unloading and
transferring of dye products, transport container cleaning, and textile dyeing machine operation (OECD.
2017).
Page 409 of 570
-------
No Gloves Used: Actual use of gloves at textile dyeing facilities in the United States is uncertain. EPA
assumes workers may not wear gloves or may only wear gloves for abrasion protection or gripping that
are not chemical resistant during routine operations.
Gloves Used with a Protection Factor of 5. 10. and 20: According to the GS on the Use of Textile Dyes,
workers may wear proper chemical-specific personal protective equipment ( ). EPA
assumes gloves may offer a range of protection, depending on the type of glove and employee training
provided.
Bin 8: covers the use of 1,4-dioxane present in antifreeze. Workers may be exposed to 1,4-dioxane
during container unloading and transferring, container cleaning, and filling of antifreeze into mechanical
equipment (Stefl and George. 2014).
No Gloves Used: Actual use of gloves at facilities using antifreeze is uncertain. EPA assumes workers
may not wear gloves or may only wear gloves for abrasion protection or gripping that are not chemical
resistant during routine operations.
Gloves Used with a Protection Factor of 5 and 10: Workers may wear chemical-resistant gloves in
accordance with the associated safety data sheets. Gloves may offer a range of protection, depending on
the type of glove and employee training provided. A glove protection factor of 20 is not applied to this
bin because the use of antifreeze is expected to be commercial and a protection factor of 20 is only
applicable to industrial settings, per TableApx F-2.
Bin 9: covers the use of 1,4-dioxane in surface cleaner. Workers may be exposed to 1,4-dioxane during
dilution of cleaner (if needed), transferring the formulations into application equipment, applying the
formulation to a surface, and wiping the cleaner off the surface (OECD. 2015).
No Gloves Used: Actual use of gloves at facilities using surface cleaner is uncertain. EPA assumes
workers may not wear gloves or may only wear gloves for abrasion protection or gripping that are not
chemical resistant during routine operations.
Gloves Used with a Protection Factor of 5 and 10: Workers may wear chemical-resistant gloves in
accordance with the associated safety data sheets. Gloves may offer a range of protection, depending on
the type of glove and employee training provided. A glove protection factor of 20 is not applied to this
bin because the use of surface cleaners is expected to be commercial and a protection factor of 20 is only
applicable to industrial settings, per Table Apx F-2.
Bin 10: covers the use of 1,4-dioxane in dish soap. EPA expects workers may be exposed to 1,4-dioxane
during the use of dish soap from unloading the dish soap, rinsing empty dish soap containers (if
performed), and dish washing operations.
No Gloves Used: Actual use of gloves at facilities using dish soap is uncertain. EPA assumes workers
may not wear gloves or may only wear gloves for abrasion protection or gripping that are not chemical
resistant during routine operations.
Gloves Used with a Protection Factor of 5 and 10: Workers may wear chemical-resistant gloves in
accordance with the associated safety data sheets. Gloves may offer a range of protection, depending on
the type of glove and employee training provided. A glove protection factor of 20 is not applied to this
Page 410 of 570
-------
bin because the use of dish soap is expected to be commercial and a protection factor of 20 is only
applicable to industrial settings, per TableApx F-2.
Bin 11: covers the use of 1,4-dioxane in dishwasher detergent. EPA expects workers to be exposed to
1,4-dioxane during use of dishwasher detergent from unloading and transferring formulation into
machine and rinsing empty dish detergent containers (if performed).
No Gloves Used: Actual use of gloves at facilities using dishwasher detergent is uncertain. EPA assumes
workers may not wear gloves or may only wear gloves for abrasion protection or gripping that are not
chemical resistant during routine operations.
Gloves Used with a Protection Factor of 5 and 10: Workers may wear chemical-resistant gloves in
accordance with the associated safety data sheets. Gloves may offer a range of protection, depending on
the type of glove and employee training provided. A glove protection factor of 20 is not applied to this
bin because the use of dishwasher detergent is expected to be commercial and a protection factor of 20 is
only applicable to industrial settings, per Table Apx F-2.
Bin 12: covers the use of 1,4-dioxane in laundry detergent, which EPA expects may involve both
commercial and industrial facilities. Workers may be exposed to 1,4-dioxane during use of laundry
detergent from transfer operations, container cleaning, handling damp laundry, and other operational
activities (OECD. 2 ).
No Gloves Used: Actual use of gloves at facilities using laundry detergent is uncertain. EPA assumes
workers may not wear gloves or may only wear gloves for abrasion protection or gripping that are not
chemical resistant during routine operations.
Gloves Used with a Protection Factor of 5. 10. and 20: According to the ESD on Chemicals Used in
Water-Based Washing Operations at Industrial and Institutional Laundries, workers may wear proper
chemical-specific personal protective equipment (OECD. 2 ). Gloves may offer a range of
protection, depending on the type of glove and employee training provided.
Bin 13: covers the use of 1,4-dioxane in paint and floor lacquer, which EPA expects may involve both
commercial and industrial facilities. Workers may be exposed to 1,4-dioxane during use of paint and
floor lacquer from quality testing of formulations, transferring the formulations into application
equipment (if used), applying the formulation to a substrate, and maintenance and cleaning activities
(OECD. 20091
No Gloves Used: Actual use of gloves at facilities using paint and floor lacquer is uncertain. EPA
assumes workers may not wear gloves or may only wear gloves for abrasion protection or gripping that
are not chemical resistant during routine operations.
Gloves Used with a Protection Factor of 5. 10. and 20: NIOSH recommends that workers wear gloves
impervious to paints and floor lacquer to prevent skin contact and avoid possible dermal exposure route
(Hills et at.. 1989). Gloves may offer a range of protection, depending on the type of glove and
employee training provided.
Bin 14: covers the presence of 1,4-dioxane as a byproduct in industrial facilities performing PET
manufacturing. Workers may be exposed to 1,4-dioxane during PET manufacture from transferring of
produced PET containing 1,4-dioxane as a byproduct and equipment cleaning ( ).
Page 411 of 570
-------
No Gloves Used: Actual use of gloves at facilities conducting PET manufacture processes is uncertain.
EPA assumes workers may not wear gloves or may only wear gloves for abrasion protection or gripping
that are not chemical resistant during routine operations.
Gloves Used with a Protection Factor of 5. 10. and 20: According to the GS on Use of Additives in
Plastic Compounding, workers typically wear suitable gloves ( ). Gloves may offer a
range of protection, depending on the type of glove and employee training provided.
Bin 15: covers the presence of 1,4-dioxane as a byproduct in industrial facilities performing ethoxylation
processes. EPA expects workers to may be exposed to 1,4-dioxane during ethoxylation processes from
transferring ethoxylated products containing 1,4-dioxane as a byproduct and equipment cleaning.
No Gloves Used: Actual use of gloves at facilities conducting ethoxylation processes is uncertain. EPA
assumes workers may not wear gloves or may only wear gloves for abrasion protection or gripping that
are not chemical resistant during routine operations.
Gloves Used with a Protection Factor of 5. 10. and 20: Workers may wear chemical-resistant gloves in
accordance with the associated safety data sheets. Gloves may offer a range of protection, depending on
the type of glove and employee training provided. A glove protection factor of 20 is applied to this bin
because ethoxylation processes occur in industrial settings.
Bin 16: covers the use of 1,4-dioxane in hydraulic fracturing, which EPA expects may involve both
commercial and industrial settings because workers may be part of a larger company with multiple
industrial facilities or from commercial contractor companies hired to support the fracturing operations.
Workers may be exposed to 1,4-dioxane during multiple activities involved in hydraulic fracturing
operations, including container unloading and transferring, container cleaning, and equipment cleaning
( 2022e).
No Gloves Used: Actual use of gloves at hydraulic fracturing facilities is uncertain. EPA assumes
workers may not wear gloves or may only wear gloves for abrasion protection or gripping that are not
chemical resistant during routine operations.
Gloves Used with a Protection Factor of 5. 10. and 20: The ESD on Chemicals Used in Hydraulic
Fracturing indicates that workers may wear proper chemical-specific personal protective equipment
( 2022e). Gloves may offer a range of protection, depending on the type of glove and
employee training provided.
F.4 Occupational Exposure Scenarios
This appendix includes a process description, worker activities, estimates of the number of potentially
exposed workers and ONUs, worker inhalation exposure assessment details, and key uncertainties in the
exposure assessment for each OES. The process descriptions included in this appendix are applicable to
the OES as a whole, including general information that is applicable to both the environmental release
and occupational exposure assessments.
F.4.1 Textile Dye
Process Description
1,4-Dioxane is present in textile dyes as an unintentional byproduct in ethoxylated substances that may
be used as a formulation component in textile dyes ( 20c). EPA has identified 1,4-dioxane
in a textile dye formulation at a concentration of 4.7 ppm ( 2020c). According to the ESD on
Page 412 of 570
-------
the Use of Textile Dyes, liquid dye formulations arrive at facilities in containers ranging from 25 kg to
1,000 kg, with 35-gallon drums being the most common container size (OECD, 2017). Dyes are
typically unloaded manually into equipment but may also be supplied to equipment via automated feed
lines. Textile substrates are immersed in a bath in which the dye is dispersed, heated, and agitated in a
batch process. Fibers in the textile substrates absorb a portion of the textile dye solution to produce the
final desired product. The remaining spent dye bath is disposed of, typically to a POTW for treatment
(OECD. 2017V
The volume of 1,4-dioxane present in textile dyes is unknown. Additionally, the number and location of
sites that use textile dyes containing 1,4-dioxane are unknown. According to the ESD on the Use of
Textile Dyes, textile dye facilities operate over a range of 31 to 295 days per year (OECD. 2017). EPA
modeled the 1,4-dioxane use rate for a generic site using the ESD on the Use of Textile Dyes to estimate
releases, resulting in a 50th and 95th percentile 1,4-dioxane use rate of 0.0027 and 0.0057 kg/site-day,
respectively. The flow diagram with release and exposure points from the ESD on the Use of Textile
Dyes is shown in Figure_Apx F-l (OECD. 2017) below. For additional information on the modeling and
associated input parameters used to estimate the daily use rate, refer to Appendix E. l 1.
(T) Dust Emissions
dunns Unloading
Dve Formulation
Disposal of Spent
Dyebath
Equipment Cleaning
©<3> Container
Residue
Cleaning and/or
Disposal
<Ł> Worker Exposure
During Dye in s
Operation
O = Environmental Releases:
1. Transfer operation losses of dust emissions (release to POTW, air, incineration, or landfill).
2. Container residues from dye transport container (release to POTW, incineration or landfill).
3. Disposal of spent dyebath (release to POTW).
4. Equipment cleaning (release to POTW)
= Occupational Exposures:
A. Inhalation (solid particulate dyes only) and dermal exposure during equipment loading/container unloading.
B. Inhalation (solid particulate dyes only) and dermal exposure during container cleaning.
C. Dermal exposure during dyeing operation.
FigureApx F-l. Environmental Release and Occupational Exposure Points During Textile Dying
Worker Activities
Workers are potentially exposed to 1,4-dioxane during the use of textile dyes from unloading and
transferring dye product, transport container cleaning, and machine operation (OECD. 2017). These
activities are all potential sources of worker exposure through dermal contact and inhalation of 1,4-
dioxane in liquid dye.
The ESD on the Use of Textile Dyes indicates that workers may connect transfer lines or manually
unload chemicals from transport containers into dyeing equipment or storage (OECD. 2017). Dermal
exposure is expected for both automated and manual unloading activities. Workers may experience
Page 413 of 570
-------
inhalation and dermal exposure to 1,4-dioxane while rinsing containers used to transport textile dyes.
Workers may be exposed to 1,4-dioxane in the liquid dyebath during removal of dyed goods after batch
processes or during handling of dyed rolls of material (« ).
According to the ESD on the Use of Textile Dyes, workers at sites that use textile dyes may wear proper
chemical-specific personal protective equipment (OECD. 2017). Workers may wear safety glasses,
goggles, aprons, respirators, and/or masks (OECD.: ). EPA did not find information that indicates
the extent that engineering controls and worker PPE are used at facilities that use textiles dyes in the
United States.
ONUs include employees that work at the sites where textile dyes are used, but they do not directly
handle the chemical and are therefore expected to have lower inhalation exposures and are not expected
to have dermal exposures through contact with liquids or solids. ONUs for this scenario include
supervisors, managers, and other employees that may be in the dyeing area but do not perform tasks that
result in the same level of exposure as those workers that engage in tasks related to the use of textile
dyes.
Number of Potentially Exposed Workers and ONUs
EPA used U.S. Census and BLS data for the NAICS code 313310, Textiles and Fabric Finishing Mills,
to estimate a total of 783 sites, 5,353 workers, and 2,634 ONUs, which corresponds to an estimated
average of seven workers and three ONUs per site ( 2016). For additional information on the
steps used to estimate the number of potentially exposed workers and ONUs, refer to Appendix G.5 of
the 2020 Risk Evaluation for 1,4-Dioxane (U.S. EPA. 2020c).
Worker Inhalation Exposure Assessment
1,4-Dioxane is present in textile dyes as an unintentional byproduct in ethoxylated substances that may
be used as a formulation component in textile dyes ( 20c). The information and data quality
evaluation to assess occupational exposures during use of textile dye is listed in Table Apx F-3 and
described in detail below.
Table Apx F-3. Textile Dye Worker Exposure Data Evaluation
Worker Activity
or Sampling
Location
Data Type
Number of
Samples
Overall Data
Quality
Determination
Source Reference
Unknown
PBZ Monitoring
14
High
(OSHA. 2020)
EPA assessed occupational inhalation exposures for this OES using OSHA's Chemical Exposure Health
Data (CEHD) (OSHA. 2020). EPA obtained CEHD for 1,4-dioxane from the OSHA webpaee. including
sampling data from 1984 to the present (data were pulled in mid-2022). EPA then edited the resulting
data download by excluding all sample types except for personal and area samples (e.g., excluding wipe
samples, bulk samples) and excluding blank samples. EPA converted the CEHD from parts per million
(ppm) to mg/m3 by multiplying the values by the molecular weight of 1,4-dioxane and dividing by the
molar volume. EPA then mapped the CEHD to 1,4-dioxane OES. To map the CEHD, EPA used the SIC
codes reported in the CEHD and corresponding SIC descriptions to identify the most likely OES for the
establishment at which the inhalation monitoring data was taken. In some cases, EPA searched the
internet for the establishment name to identify the types of products manufactured at the facility to aid
the OES mapping process. Due to the subjectivity of OES mapping and broadness of SIC codes, OES
mapping is an uncertainty of the assessment.
Page 414 of 570
-------
For this OES, monitoring data were available in CEHD from four sites with SIC codes 2399 (All Other
Misc. Textile Product Mills), 3111 (Leather and Hide Tanning and Finishing), 5136 (Men and Boy's
Clothing and Furnishings), and 2326 (Men and Boy's Work Clothing). EPA determined these four sites
to be clothing manufacturers, which may use textile dyes. Note that data were also available in CEHD
for one site that reported the SIC code 3143 (Footwear Manufacturing); however, EPA determined that
1,4-dioxane may be used in a variety of ways within footwear manufacturing such that the potential for
use in textile dyes was low. 1,4-Dioxane may be used as a functional fluid for automated footwear
production machines, a detergent in washing footwear before distribution, or as a polymerization
catalyst to make plastic components of shoes. In addition, footwear is often composed of leather or
plastic, which would not use textile dyes. Therefore, EPA excluded the data for this one site from that
used for assessing occupational inhalation exposures for this OES.
TableApx F-5 shows the discrete inhalation monitoring points from the CEHD set that EPA mapped to
the textile dyes OES. The majority of data are from 1991 and 1992, with a smaller portion from 2010.
The data include 14 inhalation monitoring data points, 12 of which are PBZ samples and two are area
samples, from four different sites. For two of the sites, all air concentrations were non-detect for 1,4-
dioxane. EPA included the data from one of these sites because bulk sampling at the site indicated the
presence of 1,4-dioxane. However, the Agency EPA excluded the data from the second site because all
PBZ, area, and bulk samples at the site were non-detect for 1,4-dioxane, so it is questionable if the site
handles 1,4-dioxane. The excluded data is denoted in Table Apx F-5. CEHD does not include
information on worker activities for PBZ samples or sampling locations for area samples, therefore
EPA's assessment assumes that that the remaining samples are relevant to this assessment. However, it
is uncertain the extent to which all potential worker activities are represented in these data.
The CEHD includes an inspection number, which corresponds to the OSHA visit at the facility, and a
sampling number, which corresponds to the worker sampling event at the facility. EPA combined
samples with the same inspection and sampling numbers into the same 8-hour TWA because these
correspond to the same worker and the same day. Therefore, combining these exposure results is more
reflective of full-shift exposures for the worker than the individual short-term samples. For samples with
detected values, the Agency translated the sample results into 8-hour TWA concentrations by assuming
that exposure concentration is zero for the time remaining in the 8-hour durations. EPA made this
assumption because the data include multiple samples for the same worker, thus increasing the
likelihood that the data reflect all tasks with potential 1,4-dioxane exposures.
Where non-detect values were included in the dataset, EPA first calculated the LOD for the sample. The
Agency assumed the use of NIOSH method 1602, which has an estimated LOD of 0.01 mg/sample. To
calculate LOD in terms of an air concentration, The Agency divided the limit of 0.01 mg/sample by the
sampled air volume provided in the CEHD, which converted from L to m3. For the non-detect values,
EPA then used the LOD divided by two in subsequent central tendency (50th percentile) and high-end
(95th percentile) calculations. The Agency used the LOD/V2 for approximating an air concentration for
non-detect samples because the geometric standard deviation of the underlying datasets are less than
three ( |). Because greater than 50 percent of the monitoring data results are non-detect
for 1,4-dioxane, this method for the calculation of statistics will result in potentially biased estimates.
EPA then used the air concentrations and LOD/V2 as shown in Table Apx F-5 to calculate full shift (8-
hour TWA) central tendency (50th percentile) and high-end (95th percentile) inhalation exposures for
workers. EPA used these central tendency and high-end values to calculate the ADC and LADC. The
calculated values are summarized in Table Apx F-4. Equations for calculating ADC and LADC are
Page 415 of 570
-------
presented in Appendix G of the December 2020 Final Risk Evaluation for 1,4-Dioxane (U.S. EPA.
2020c).
Exposure data for ONUs were not available. The ONU exposures are anticipated to be lower than
worker exposures since ONUs do not typically directly handle the chemical. Only inhalation exposures
to vapors or incidental dermal exposures may be expected to ONUs.
TableApx F-4. Inhalation Exposures of Workers for the Use of Textile Dye Based on Monitoring
Data
Exposure Type
Central Tendency
(50th Percentile)
(mg/m3)"
High-End
(95th Percentile) (mg/m3)"
Drafl Rl- eslimales
S-hour TWA Exposure Concentrations
0.07
74
Average Daily Concentration (ADC)
0.040
71.15
T.ifetime Average Daily Concentration ([.ADO
I pchil
n m 6
ed eslimales
36 49
8-hour TWA Exposure Concentrations
0.81
15
Average Daily Concentration (ADC)
0.49
14
Lifetime Average Daily Concentration (LADC)
0.19
7.4
a See Table_Apx F-3 for corresponding references.
b For select OESs, updates to exposure estimates were made via information provided by the SACC and public
comments.
Page 416 of 570
-------
Table Apx F-5. Occupational Inhalation Monitoring Data
or Textile Dyes
Row
#
Type
of
Sample
Worker Activity or
Sample Location
No. of
Samples
Sample
Date
Sample
Time
(min)
1,4-Dioxane
Exposure
Concentration
(mg/m3)
EPA
Determined
8-hour TWA
(mg/m3)"
Source
Overall Data
Quality
Determination
1
PBZ
Unknown
1
1/17/1991
23
32.9
8.4 (8-hour TWA
for the same
worker from rows
1-7)
COSHA. 2020)
High
2
PBZ
Unknown
1
1/17/1991
39
21.3
COSHA. 2020)
High
3
PBZ
Unknown
1
1/17/1991
32
26.2
COSHA. 2020)
High
4
PBZ
Unknown
1
1/17/1991
12
13.7
(OSUA. 2020)
High
5
PBZ
Unknown
1
1/17/1991
21
4.6
(OSUA. 2020)
High
6
PBZ
Unknown
1
1/17/1991
18
26.5
COSHA. 2020)
High
7
PBZ
Unknown
1
1/17/1991
30
28.1
COSHA. 2020)
High
8
PBZ
Unknown
1
1/17/1991
25
41.1
9.8 (8-hour TWA
for the same
worker from rows
8-15)
COSHA. 2020)
High
9
PBZ
Unknown
1
1/17/1991
35
33.5
(OSUA. 2020)
High
10
PBZ
Unknown
1
1/17/1991
23
8.1
(OSUA. 2020)
High
11
PBZ
Unknown
1
1/17/1991
24
33.2
COSHA. 2020)
High
12
PBZ
Unknown
1
1/17/1991
23
15.4
COSHA. 2020)
High
13
PBZ
Unknown
1
1/17/1991
26
18.4
COSHA. 2020)
High
14
PBZ
Unknown
1
1/17/1991
22
17.0
(OSUA. 2020)
High
15
PBZ
Unknown
1
1/17/1991
10
31.4
COSHA. 2020)
High
16
PBZ
Unknown
1
12/10/1992
5
ND (LOD =
0.53)
0.059 (8-hour
TWA for the
same worker
from rows 9-22)
(OSUA. 2020)
High
17
PBZ
Unknown
1
12/10/1992
5
ND (LOD =
0.67)
COSHA. 2020)
High
18
PBZ
Unknown
1
12/10/1992
5
ND (LOD =
0.67)
COSHA. 2020)
High
19
PBZ
Unknown
1
12/10/1992
7
ND (LOD =
2.94)
(OSUA. 2020)
High
Page 417 of 570
-------
Row
#
Type
of
Sample
Worker Activity or
Sample Location
No. of
Samples
Sample
Date
Sample
Time
(min)
1,4-Dioxane
Exposure
Concentration
(mg/m3)
EPA
Determined
8-hour TWA
(mg/m3)"
Source
Overall Data
Quality
Determination
20
PBZ
Unknown
1
12/10/1992
5
ND (LOD =
0.67)
COSHA. 2020)
High
21
PBZ
Unknown
1
12/10/1992
5
ND (LOD =
0.67)
COSHA. 2020)
High
22
PBZ
Unknown
1
12/10/1992
5
ND (LOD =
0.67)
(OSUA. 2020)
High
23
PBZ
Unknown
1
12/9/1992
5
ND (LOD =
0.67)
(OSUA. 2020)
High
24
PBZ
Unknown
1
12/9/1992
5
ND (LOD =
0.67)
(OSUA. 2020)
High
25
PBZ
Unknown
1
12/9/1992
5
ND (LOD =
0.67)
0.054 (8-hour
TWA for the
same worker
COSHA. 2020)
High
26
PBZ
Unknown
1
12/9/1992
5
ND (LOD =
0.67)
COSHA. 2020)
High
27
PBZ
Unknown
1
12/9/1992
5
ND (LOD =
0.67)
from rows 23-29)
(OSUA. 2020)
High
28
PBZ
Unknown
1
12/9/1992
5
ND (LOD =
0.67)
(OSUA. 2020)
High
29
PBZ
Unknown
1
12/9/1992
5
ND (LOD =
3.33)
COSHA. 2020)
High
30
PBZ
Unknown
1
6/3/1992
59
134
COSHA. 2020)
High
31
PBZ
Unknown
1
6/3/1992
48
ND (LOD =
2.83)
17.2 (8-hour
TWA for the
same worker
from rows 30-34)
(OSUA. 2020)
High
32
PBZ
Unknown
1
6/3/1992
55
ND (LOD =
2.47)
COSHA. 2020)
High
33
PBZ
Unknown
1
6/3/1992
53
ND (LOD =
2.56)
COSHA. 2020)
High
Page 418 of 570
-------
Row
#
Type
of
Sample
Worker Activity or
Sample Location
No. of
Samples
Sample
Date
Sample
Time
(min)
1,4-Dioxane
Exposure
Concentration
(mg/m3)
EPA
Determined
8-hour TWA
(mg/m3)"
Source
Overall Data
Quality
Determination
34
PBZ
Unknown
1
6/3/1992
44
ND (LOD =
3.09)
COSHA. 2020)
High
35
PBZ
Unknown
1
6/3/1992
53
ND (LOD =
2.58)
1.0 (8-hour TWA
for the same
worker from rows
35-39)
COSHA. 2020)
High
36
PBZ
Unknown
1
6/3/1992
60
ND (LOD =
2.28)
(OSUA. 2020)
High
37
PBZ
Unknown
1
6/3/1992
55
ND (LOD =
2.48)
(OSUA. 2020)
High
38
PBZ
Unknown
1
6/3/1992
46
ND (LOD =
2.97)
(OSUA. 2020)
High
39
PBZ
Unknown
1
6/3/1992
58
ND (LOD =
2.35)
COSHA. 2020)
High
40
PBZ
Unknown
1
6/3/1992
60
ND (LOD =
2.28)
0.60 (8-hour
TWA for the
same worker
from rows 40-42)
COSHA. 2020)
High
41
PBZ
Unknown
1
6/3/1992
60
ND (LOD =
2.28)
(OSUA. 2020)
High
42
PBZ
Unknown
1
6/3/1992
56
ND (LOD =
2.44)
(OSUA. 2020)
High
43
PBZ
Unknown
1
6/3/1992
54
ND (LOD =
2.52)
0.60 (8-hour
TWA for the
same worker
from rows 43-45)
COSHA. 2020)
High
44
PBZ
Unknown
1
6/3/1992
61
ND (LOD =
2.23)
COSHA. 2020)
High
45
PBZ
Unknown
1
6/3/1992
58
ND (LOD =
2.34)
(OSUA. 2020)
High
46
Area
Unknown
1
7/15/2010
69
ND (LOD =
0.86)
0.09
(OSUA. 2020)
Excluded6
Page 419 of 570
-------
Row
#
Type
of
Sample
Worker Activity or
Sample Location
No. of
Samples
Sample
Date
Sample
Time
(min)
1,4-Dioxane
Exposure
Concentration
(mg/m3)
EPA
Determined
8-hour TWA
(mg/m3)"
Source
Overall Data
Quality
Determination
47
Area
Unknown
1
7/15/2010
270
ND (LOD =
0.25)
0.10
COSHA. 2020)
Excluded6
48
PBZ
Unknown
1
7/15/2010
244
ND (LOD =
0.22)
0.08
COSHA. 2020)
Excluded6
49
PBZ
Unknown
1
7/15/2010
150
ND (LOD =
0.39)
0.09
(OSUA. 2020)
Excluded6
50
PBZ
Unknown
1
7/15/2010
155
ND (LOD =
0.39)
0.09
(OSUA. 2020)
Excluded6
51
PBZ
Unknown
1
7/15/2010
294
ND (LOD =
0.2)
0.09
(OSUA. 2020)
Excluded6
PBZ = Personal breathing zone; ND = Non-detect for 1,4-dioxane; LOD = limit of detection; TWA = time-weighted average
a The 8-hour TWA calculations use LOD/V2 for non-detect values because the geometric standard deviations of the underlying datasets are all <3.
6 As explained prior to this table, these data points were excluded from the analysis of central tendency and high-end worker exposures because all PBZ, area,
and bulk sampling at this site was non-detect for 1,4-dioxane; therefore, it is questionable if the site handles 1,4-dioxane.
Page 420 of 570
-------
Key Uncertainties
The OSHA CEHD monitoring data does not include process information or worker activities; therefore,
there is uncertainty as to which worker activities these data cover and whether all potential workers
activities are represented in this data. Additionally, these data are from four facilities, and it is unclear
how representative the data are for all sites and all workers across the United States. Approximately half
of OSHA CEHD used for this assessment are from the 1990s and the other half are from 2010.
Therefore, the age of the monitoring data can also introduce uncertainty.
As discussed above, EPA used half the detection limit for the non-detect values in the central tendency
and high-end exposure calculations. Due to the high number of non-detects (11 of the 14 TWAs were
non-detect), this method may result in bias ( 4a). Additional uncertainties are listed in
Section 3.1.2.4.
F.4.2 Antifreeze
Process Description
1,4-Dioxane is present in antifreeze as an unintentional byproduct of certain ethoxylated substances that
may be used as a formulation component in antifreeze ( EOc). One public comment
indicates that 1,4-dioxane is produced as a byproduct from the production of polyester polyols, with 1,4-
dioxane distilled from the polyol mixture and condensed with glycol (Huntsman. 2023). This glycol
mixture, containing 3 percent 1,4-dioxane, is sold to glycol manufacturers who purify and blend the
glycol into antifreeze (Huntsman. 2023). However, this OES only reflects the use of antifreeze
containing 1,4-dioxane, the processing/blending of antifreeze is covered in the "Industrial uses" OES.
EPA has identified 1,4-dioxane concentrations in antifreeze ranging from 0.01 to 86 ppm (U.S. EPA.
2020c). Antifreeze is formulated for use in motor vehicles and other mechanical equipment to prevent
freezing of engine fluids (Stefl and George. , ). EPA did not find any container specific information
on 1,4-dioxane in antifreeze; however, EPA expects the antifreeze formulation to ship to automotive
maintenance facilities as a liquid in drums or smaller containers. Antifreeze is manually added to
engines and is typically replaced every 2 to 3 years. Upon completion of use, the spent antifreeze may be
recycled or disposed to municipal waste treatment facilities (Stefl and George. 2014).
The volume of 1,4-dioxane present in antifreeze is unknown. Additionally, the number and location of
sites that use antifreeze containing 1,4-dioxane are unknown. EPA modeled the 1,4-dioxane use rate
using the consumer exposure model, which indicates a use rate of 0.15 kg of antifreeze/job. The 0.15
kg/job represents a "top-up" amount and recommended a use rate of 2 kg/job to represent a full
replacement of antifreeze in a car. EPA assumes facilities use antifreeze 5 days/week, 50 weeks/year or
250 days/year. For additional information on the modeling and associated input parameters used to
estimate the daily use rate, refer to Appendix F.5.
Worker Activities
Workers are potentially exposed to 1,4-dioxane during multiple activities involved in use of antifreeze,
including container unloading and transferring, container cleaning, and filling of antifreeze into
mechanical equipment (Stefl and George. 2014). These activities are all potential sources of worker
exposure through dermal contact to liquid and inhalation of volatile chemical vapors.
Workers may don personal protective equipment (PPE) during the use of antifreeze in accordance with
the associated safety data sheets. EPA did not find information that indicates the extent to which
engineering controls are present or worker PPE are worn at U.S. facilities that use antifreeze.
Page 421 of 570
-------
ONUs include employees that work at the sites where antifreeze is used, but they do not directly handle
the chemical and are therefore expected to have lower inhalation exposures and are not expected to have
dermal exposures through contact with liquids. ONUs for this scenario include supervisors, managers,
and other employees that may be in the filling area but do not perform tasks that result in the same level
of exposures as those workers that engage in tasks related to the use of antifreeze.
Number of Potentially Exposed Workers and ONUs
EPA estimated the number of workers and occupational non-users potentially exposed to 1,4-dioxane in
antifreeze using 2016 BLS data for NAICS codes 811111, General Automotive Repair, and 811198, All
Other Automotive Repair and Maintenance. Using BLS data, EPA estimated a total of 84,383 sites, two
workers per site, and 0.2 ONUs per site (U.S. BLS. 2016). For additional information on the steps used
to estimate the number of potentially exposed workers and ONUs, refer to Appendix G.5 of the 2020
Risk Evaluation for 1,4-Dioxane (U.S. EPA. 2020c).
Worker Inhalation Exposure Assessment
1,4-Dioxane is present in antifreeze as an unintentional byproduct of certain ethoxylated substances that
may be used as formulation components in antifreeze ( 2020c). The information and data
quality evaluation to assess occupational exposures during use of antifreeze is listed in TableApx F-6
and described below.
Table Apx F-6. Antifreeze Data Source Evaluation
Worker Activity or
Sampling Location
Data Type
Number of
Samples
Overall Data
Quality
Determination
Source Reference
Unloading/transferring
antifreeze from
containers
Input parameters for
Monte Carlo modeling
N/A
Higha
( 2022b)
Unloading/transferring
antifreeze from
containers
Input parameters for
Monte Carlo modeling
N/A
Higha
COECD, 2020)
a This is the rating for the underlying data used in the model, and not the Monte Carlo model itself.
EPA did not find relevant inhalation monitoring data for the use of antifreeze. Therefore, EPA modeled
1,4-dioxane air concentrations using a Monte Carlo modeling approach, which is described in Appendix
F.7. This modeling approach utilizes the EPA AP-42 Loading Model and the EPA Mass Balance
Inhalation Model, with variation in input parameters for container size, jobs per day, concentration of
1,4-dioxane in antifreeze, ventilation rate, mixing factor, and saturation factor based on available data.
Table Apx F-7 provides a summary of the modeled inhalation exposures.
Page 422 of 570
-------
Table Apx F-7. Modeled Occupational Inhalation Exposures for Antifreeze
Statistic
1,4-Dioxane Exposure Concentration, 8-Hour-TWA
(mg/m3)
Maximum
1.8E-05
99th Percentile
2.1E-06
95th Percentile
9.8E-07
50th Percentile
1.3E-07
5th Percentile
7.3E-09
Minimum
3.2E-12
Mean
2.7E-07
EPA used the 50th and 95th percentile modeled 8-hour TWA exposures values presented in TableApx
F-7 to calculate the central tendency and high-end ADC and LADC for workers, respectively. The
calculated values are summarized in Table Apx F-8. Equations for calculating ADC and LADC are
presented in Appendix G of the December 2020 Final Risk Evaluation for 1,4-Dioxane (U.S. EPA.
2020c).
Exposure data for ONUs were not available. The ONU exposures are anticipated to be lower than
worker exposures since ONUs do not typically directly handle the chemical. Only inhalation exposures
to vapors or incidental dermal exposures may be expected to ONUs.
Table Apx F-8. Inhalation Exposures of Workers for the Use of Antifreeze Based on Modeling
Exposure Type
Central Tendency
(50th Percentile) (mg/m3)"
High-End
(95th Percentile)
(mg/m3)"
Drafl Rl- estimates
8-hour TWA Exposure Concentrations
2.18E-08
1.10E-07
Average Daily Concentration (ADC)
2.10E-08
1.06E-07
Lifetime Average Daily Concentration (LADC)
8.34E-09
5.44E-08
I pchilccl estimates
8-hour TWA Exposure Concentrations
1.3E-07
9.8E-07
Average Daily Concentration (ADC)
1.2E-07
9.4E-07
Lifetime Average Daily Concentration (LADC)
4.8E-08
4.8E-07
a See Table_Apx F-6 for corresponding references.
b For select OESs, updates to exposure estimates were made via information provided by the SACC and public
comments.
Key Uncertainties
Due to a lack of data specific to 1,4-dioxane for this use, EPA used assumptions and values from the
Automotive Detailing MRD, Automotive Lubricant ESD, EPA AP-42 Loading Model, EPA Mass
Page 423 of 570
-------
Balance Inhalation Model, and Near-Field/Far-Field Brake Model. The uncertainties associated with this
modeling approach are described in Section 3.1.2.4.
In addition, the commercial use rate of antifreeze was scaled up from the consumer use rate provided by
the SHEDS-HT model, using the number of jobs per day from the Automotive Detailing MRD,
Automotive Lubricant GS. These scaling factors may overestimate exposure if the actual number of jobs
at commercial sites is lower or may underestimate exposure if the actual number of jobs at commercial
sites is higher.
F.4.3 Surface Cleaner
Process Description
1,4-Dioxane is present an unintentional byproduct in ethoxylated substances that may be used as a
formulation component in surface cleaners ( 2020c). EPA has identified concentrations of 1,4-
dioxane in surface cleaners ranging from 0.36 to 9.0 ppm ( 2015b). In addition, EPA has
reviewed the New York Department of Environmental Conservation (NYDEC) database of waivers for
cleaning, personal care, and cosmetic products not meeting the proposed maximum concentrations of
1,4-dioxane in these products (2 ppm by the end of 2022 and 1 ppm by the end of 2023) (NYDEC.
2023). Using the product names/descriptions in the database, EPA determine which products in were
likely relevant to commercial surface cleaners. EPA found that the concentration of 1,4-dioxane in
commercial surface cleaners in this waiver database ranged from 2.2 ppm to 75.7 ppm ( ,023).
EPA used this maximum concentration of 75.7 ppm in the occupational dermal exposure assessment in
Section 3.1.2.2.
Surface cleaners are used to disinfect and remove unwanted foreign matter from various types of
surfaces (Naev and Theiner. 2020). EPA did not find any container specific information on 1,4-dioxane
in surface cleaners; however, EPA expects formulation to arrive as a liquid in small containers of
various sizes. Surface cleaners may be aqueous, semi-aqueous, or non-aqueous. Aqueous and semi-
aqueous cleaners may be diluted with water prior to use. The cleaner is typically spray applied to the
surface and wiped off (OE ).
The volume of 1,4-dioxane present in surface cleaners is unknown. Additionally, the number and
location of sites that use surface cleaners containing 1,4-dioxane are unknown. EPA modeled the 1,4-
dioxane use rate the SHEDS-HT case study from Liverpool, OH, resulting in a central tendency and
high-end 1,4-dioxane use rate of 79 and 85 g/site-day, respectively. EPA assumes facilities use surface
cleaners 5 days/week, 50 weeks/year or 250 days/year.
Worker Activities
During the use of surface cleaners, workers are potentially exposed during the dilution of cleaner (if
needed), transferring the formulations into application equipment, applying the formulation to a surface,
and wiping the cleaner off the surface. These activities are all potential sources of worker exposure
through dermal contact to liquid and inhalation of vapors (( ).
EPA did not find information that indicates the extent that engineering controls and worker PPE are used
at facilities that use surface cleaners in the United States.
ONUs include employees that work at the site where surface cleaners are used, but they do not directly
handle the chemical and are therefore expected to have lower inhalation exposures and vapor-through-
skin uptake. Additionally, dermal exposures through contact with liquids are not expected. ONUs
include supervisors, managers, and other employees that may be in the cleaning area but do not perform
Page 424 of 570
-------
tasks that result in the same level of exposures as workers that engage in tasks related to the use of
surface cleaner.
Number of Potentially Exposed Workers and ONUs
EPA used U.S. Census and BLS data for the NAICS code 561720, Janitorial Services, to estimate a total
of 55,998 sites, 552,300 workers, and 32,144 ONUs, which corresponds to an estimated average of 9.9
workers and 0.6 ONUs per site ( S. 2016). For additional information on the steps used to
estimate the number of potentially exposed workers and ONUs, refer to Appendix G.5 of the 2020 Risk
Evaluation for 1,4-Dioxane (U.S. EPA. 2020c).
Worker Inhalation Exposure Assessment
1,4-Dioxane is present an unintentional byproduct in ethoxylated substances that may be used as a
formulation component in surface cleaners ( 2020c). The information and data quality
evaluation to assess occupational exposures during use of surface cleaner is listed in TableApx F-9 and
described below.
Table Apx F-9. Surface Cleaner Worker Exposure Data Evaluation
Worker Activity
or Sampling
Location
Data Type
Number of
Samples
Overall Data
Quality
Determination
Source Reference
Use of surface
cleaning products
PBZ Monitoring
49
High
(Harlev et al..
2021)
Table Apx F-10 shows the 1,4-dioxane inhalation monitoring data available in published literature
related to the use of surface cleaners (Harley et at.. 2021). This data is from a study in which 49 PBZ
samples were taken in 2019 during the use of surface cleaners in domestic kitchens and bathrooms. The
study does not provide the discrete values for the 49 samples but does provide the geometric mean and
maximum of the 49 samples, which are 0.57 |ig/m3 and 7.38 |ig/m3, respectively. In this study, personal
air monitoring was conducted on 50 consumers while they cleaned their homes with standard cleaning
products for 30 minutes. The volunteers were asked to clean their own kitchen and bathroom using their
regular cleaning products while wearing a small backpack containing personal air monitoring
equipment. For this OES, EPA did not find air monitoring of workers or other occupational non-users;
therefore, EPA uses the data from Harley (2021). which is for consumer use, as surrogate for
occupational exposures. EPA expects that both consumers and workers utilize similar practices for
surface cleaning such that the inhalation exposure potential is similar between the two. EPA recognizes,
however, that workers are more likely to conduct surface cleaning at a higher frequency or for longer
durations than consumers. Therefore, EPA used available information to determine the appropriate
exposure durations for workers, which is described further below.
EPA converted the geometric mean and maximum 30-minute air concentration values into 8-hour
TWAs by assuming that commercial workers may perform cleaning activities over their entire 8-hour
shift. Therefore, to convert the 30-minute geometric mean and maximum air concentrations from Harley
(2021) to 8-hour TWAs, EPA assumed the air concentrations were representative of the entire 8-hour
shift. EPA then used these values to calculate the ADC and LADC. The calculated values are
summarized in Table Apx F-l 1. Equations for calculating ADC and LADC are presented in Appendix
G of the December 2020 Final Risk Evaluation for 1,4-Dioxane ( 3c).
Page 425 of 570
-------
Exposure data for ONUs were not available. The ONU exposures are anticipated to be lower than
worker exposures since ONUs do not typically directly handle the chemical. Only inhalation exposures
to vapors or incidental dermal exposures may be expected to ONUs.
TableApx F-10. Inhalation Exposures of Workers for the Use of Surface Cleaner Based on
Monitoring Data
Exposure Type
Central Tendency
(Geometric Mean)
(mg/m3)"
High-End
(Maximum)
(mg/m3)"
Draft RE estimates6
8-hour TWA Exposure Concentrations
2.9E-04
3.70E-03
Average Daily Concentration (ADC)
2.79E-04
3.56E-03
Lifetime Average Daily Concentration (LADC)
1.11E-04
1.82E-03
I pdated estimates
S-hour TWA Exposure Concentrations
5.7E U4
7.4E U3
Average Daily Concentration (ADC)
5.5E-04
7.1E-03
Lifetime Average Daily Concentration (LADC)
2.2E-04
3.7E-03
a See Table_Apx F-9 for corresponding references.
b For select OESs, updates to exposure estimates were made via information provided by the SACC and public
comments.
Page 426 of 570
-------
Table Apx F-ll. Occupational Inhalation Monitoring Data for Surface Cleaner
Row
#
Type of
Sample
Worker Activity or
Sample Location
Number
of
Samples
Sample
Date
Sample
Time
1,4-Dioxane Exposure
Concentration
(mg/m3)
EPA
Determined
8-hour TWA
(mg/m3)
Source
Overall Data
Quality
Determination
1
Personal
Use of surface
cleaning products
49a
2019
30 min
0.00057
(Geometric mean)
0.00057
(Geometric
mean)
(Harlev et
ah. 2021)
High
0.0074
(Maximum)
0.0074
(Maximum)
TWA = Time-weighted average
a Source did not include discrete values for each of the 49 samples but provided the geometric mean and maximum.
Page 427 of 570
-------
Key Uncertainties
The literature source EPA used for this COU did not present discrete sampling values, so EPA used the
geometric mean and maximum of the sample results as they were provided in the source (Harlev et at..
2021). The representativeness of these values towards the central tendency and high-end exposures is
uncertain due to the lack of discrete data provided and inability to verify summary statistics.
Additionally, there is uncertainty in how the literature source accounted for non-detect values in the
geometric mean calculation.
It is unknown whether the activities performed in this study accurately reflect all surface cleaning
scenarios or the cleaning industry as whole. Also, EPA assumed that cleaning activities occur over four
hours per day per the Draft Furnishing Cleaning GS ( 322a). Besides the Furnishing
Cleaning GS, ERG did nott identify any other sources to estimate frequency and duration of cleaners.
This assumption may result in an underestimate or overestimate of exposures if cleaning occurs over a
different timeframe. Additional uncertainties are listed in Section 3.1.2.4.
F.4.4 Dish Soap
Process Description
1,4-Dioxane has been identified as an unintentional component in dish soaps ( )20c).
Sources indicate 1,4-dioxane content in dish soaps ranges from 0.03 to 204 ppm ( s20c; Lin
et at.. 2017; Saraii and Shirvani. 2017; Makino et at.. 2006; Wala-Jerzykiewicz and Szymanowski.
1998). Note that some sources identify "dishwashing liquids"; EPA assumed these products may be
either dish soaps or dishwashing detergents. Additionally, some of these data are for 1,4-dioxane
concentrations in consumer dish soaps; however, EPA expects similar formulations may be used
commercially. In addition, the Agency reviewed the NYDEC database of waivers for cleaning, personal
care, and cosmetic products not meeting the proposed maximum concentrations of 1,4-dioxane in these
products (2 ppm by the end of 2022 and 1 ppm by the end of 2023) (NYDEC. 2023). Using the product
names/descriptions in the database, EPA determine which products in were likely relevant to
commercial dish soaps and detergents; the Agency could not generally distinguish between dish soaps
and detergents. EPA found that the concentration of 1,4-dioxane in commercial dish soaps and
detergents in this waiver database ranged from 2.4 to 57.6 ppm (NYDEC. 2023). Given all the available
data, EPA used this maximum concentration of 204 ppm in the occupational dermal exposure
assessment in Section 3.1.2.2.
EPA expects formulations containing 1,4-dioxane contaminant to arrive as a liquid in small containers
of various sizes, such as one-gallon containers (P&G. 2023). Dish soap may be dispensed directly into
sinks using a pump affixed to the top of the soap bottle, with an automatic dosing system, or by free
pouring from the bottle. The standard method for commercial dishwashing is to use a three-compartment
sink, with the first compartment used to wash dishes in soapy water, the second used for rinsing with
water, and the third used to rinse the dishes in a sanitizing solution. Dish sink water is kept at or above
1 10 "F. Workers scrub dishes with sponges, clothes, or brushes in the soapy water (P&G. 2023). Dirty
water containing the used dish soap is rinsed down sink drains to POTWs (ATSDK JO I
The volume of 1,4-dioxane present in dish soaps is unknown. Additionally, the number and location of
sites that use dish soaps containing 1,4-dioxane are unknown. EPA assumes facilities use dish soaps 5
days/week, 50 weeks/year or 250 days/year. The Agency modeled the 1,4-dioxane use rate using the
SHEDS-HT case study from Liverpool, OH to estimate releases, resulting in a central tendency and
high-end 1,4-dioxane use rate of 64.6 and 64.8 g/site-day, respectively.
Page 428 of 570
-------
Worker Activities
EPA expects workers to be potentially exposed to 1,4-dioxane during the use of dish soap from
dispensing the dish soap and washing operations. While the dish soap is diluted during use, workers may
handle the undiluted dish soap when dispensing it into sinks, depending on the dispending method used
{i.e., automated vs. manual). These activities are all potential sources of worker exposure through
dermal contact to liquid and inhalation of vapors. Dishwashing workers may wash dishes over their
entire 8-hour shift; however, workers are likely to perform other jobs throughout their shift. It is likely
that dermal exposure only occurs when workers have their hands in soapy sink water, which has been
estimated to be 40 minutes per shift ( 3. 2023). Note that the dermal exposure model discussed in
Section 3.1.1.3 does not have a term for dermal exposure duration, as it is based on a single dermal
contact event leaving a specific quantity on the skin.
Additionally, dishwashers may wear dishwashing gloves to mitigate potential dermal exposures (P&G.
2023). EPA did not find information that indicates the extent that engineering controls and worker PPE
are used at facilities that use dish soap in the United States.
ONUs include employees that work at the sites where dish soaps are used, but they do not directly
handle the chemical and are therefore expected to have lower inhalation exposures and are not expected
to have dermal exposures by contact with liquids. ONUs for this scenario include supervisors, managers,
and other employees that may be in the washing area are but do not perform tasks that result in the same
level of exposure as those workers that engage in tasks related to the use of dish soaps.
Number of Potentially Exposed Workers and ONUs
To estimate the number of workers, EPA used U.S. Census and BLS data for the following NAICS
codes: 623300, 713900, 721100, 721300, 722300, 722400, and 722500. EPA estimated a total of
773,851 sites, 0.6 workers per site, and 1.1 ONUs per site ( ) For additional information
on the steps used to estimate the number of potentially exposed workers and ONUs, refer to Appendix
G.5 of the 2020 Risk Evaluation for 1,4-Dioxane (U.S. EPA. 2020c).
Worker Inhalation Exposure Assessment
1,4-Dioxane has been identified as an unintentional component in dish soaps ( !20c). The
information and data quality evaluation to assess occupational exposures during use of dish soap is listed
in Table Apx F-12 and described below.
Table Apx F-12. Dish Soap Worker Exposure Data Evaluation
Worker Activity or Sampling
Location
Data Type
Number of
Samples
Overall Data
Quality
Determination
Source
Reference
Unloading detergent into
sinks/machines and cleaning
dishes
Input parameters
for Monte Carlo
modeling
N/A
Higha
fP&G. 2023)
a This is the rating for the underlying data used in the model, and not the Monte Carlo model itself.
EPA did not find relevant inhalation monitoring data for the use of dish soaps and detergents. Therefore,
EPA modeled 1,4-dioxane air concentrations using a Monte Carlo modeling approach, which is
described in Appendix F. 10. This modeling approach utilizes standard EPA models with industry-
specific information for many of the model input parameters {e.g., sink size, wash temperature). For
other parameters like ventilation rate and mixing factor, EPA used generic data from standard sources.
Page 429 of 570
-------
Appendix F. 10 has an explanation of each input parameter to the model. TableApx F-13 provides a
summary of the modeled inhalation exposures.
Table Apx F-13. Modeled Occupational Inhalation Exposures for Dish Soap
Statistic
1,4-Dioxane Exposure Concentration, 8 Hour-TWA
(mg/m3)
Maximum
0.61
99th Percentile
4.4E-02
95th Percentile
1.0E-02
50th Percentile
1.1E-03
5th Percentile
7.5E-05
Minimum
9.6E-07
Mean
3.2E-03
EPA used the 50th and 95th percentile modeled 8-hour TWA exposures values presented in Table Apx
F-13 to calculate the central tendency and high-end ADC and LADC for workers, respectively. The
calculated values are summarized in Table Apx F-14. Equations for calculating ADC and LADC are
presented in Appendix G of the December 2020 Final Risk Evaluation for 1,4-Dioxane (U.S. EPA.
2020c).
Exposure data for ONUs were not available. The ONU exposures are anticipated to be lower than
worker exposures since ONUs do not typically directly handle the chemical. Only inhalation exposures
to vapors or incidental dermal exposures may be expected to ONUs.
Table Apx F-14. Inhalation Exposures of Workers for the Use of Dish Soaps Based on Modeling
Exposure Type
Central Tendency
(50th Percentile) (mg/m3)"
High-End
(95th Percentile)
(mg/m3)"
Draft Rl- estimates
8-hour TWA Exposure Concentrations
1.0C
2.\d
Average Daily Concentration (ADC)
1.0
2.0
T.ifetime Average Daily Concentration (T.ADC)
I pdal
n 3QR
jd esli males
1 ^3
8-hour TWA Exposure Concentrations
1.1E-03
1.0E-02
Average Daily Concentration (ADC)
1.1E-03
1.0E-02
Lifetime Average Daily Concentration (LADC)
4.4E-04
5.1E-03
a See Table_Apx F-12 for corresponding references.
b For select OESs, updates to exposure estimates were made via information provided by the SACC and public
comments.
c All data were non-detect; EPA presented the LOD/2 for the central tendency value.
d All data were non-detect; EPA presented the LOD for the high-end value.
Page 430 of 570
-------
Key Uncertainties
There are no directly relevant GSs or ESDs for the use of dish soaps; therefore, EPA developed this
model using standard EPA models for the expected release and exposure points. Due to a lack of data
specific to 1,4-dioxane for this use, EPA used industry-specific data from a public comment along with
standard default values from sources like the ChemSTEER User Guide for the model input parameters.
In addition, the use rate of dish soaps in the model is based on a value from the Consumer Exposure
Model that was scaled up for commercial use. This scaling approach adds uncertainty to the assessment.
Additional uncertainties are listed in Section 3.1.2.4.
F.4.5 Dishwasher Detergent
Process Description
1,4-Dioxane has been identified as an unintentional component in dishwasher detergent containing
ethoxylated surfactants ( 320c). Sources indicate 1,4-dioxane content in dishwasher
detergents ranges from 0.86 to 51 ppm ( Z020c; Lin et at.. 2017; Saraii and Shirvani. 2017;
Davarani et ai. 2012; Makino et at.. 2006; Wala-lerzykiewicz and Szymanowski. 1998). Note that some
sources identify "dishwashing liquids"; EPA assumed these products may be either dish soaps or
dishwashing detergents. Additionally, some of these data are for 1,4-dioxane concentrations in consumer
dishwashing detergents; however, EPA expects similar formulations may be used commercially. In
addition, EPA reviewed the NYDEC database of waivers for cleaning, personal care, and cosmetic
products not meeting the proposed maximum concentrations of 1,4-dioxane in these products (2 ppm by
the end of 2022 and 1 ppm by the end of 2023) ( C. 2023). Using the product names/descriptions
in the database, EPA determined which products in were likely relevant to commercial dish soaps and
detergents; EPA could not generally distinguish between dish soaps and detergents. EPA found that the
concentration of 1,4-dioxane in commercial dish soaps and detergents in this waiver database ranged
from 2.4 to 57.6 ppm ( 023). Given all the available data, EPA used this maximum
concentration of 57.6 ppm in the occupational dermal exposure assessment in Section 3.1.2.2.
Professional dish detergent products are sold in 1- to 5-gallon containers designed to prevent spilling
when the container is overturned and to be compatible with dispensing equipment (P&G. 2023). Some
dishwashing establishments use dispensing systems to automatically dispense the amount of detergent
needed into the dishwashing machine (HCPA. 2023). Workers load dirty dishes into a dish rack, open
the door to the machine, slide the dish rack into the machine, then close the door, with dish detergent
dispensed into the machine once the door is closed (P&G. 2023). Once the washing cycle is complete,
workers remove the rack of clean dishes and insert a rack of dirty dishes. Dishwasher machine
temperatures range between 120 and 180 °F (P&G. 2023). Dirty water containing the used dishwasher
detergent and 1,4-dioxane are rinsed down machine drains to POTWs (ATSDR. ).
The volume of 1,4-dioxane present in dishwasher detergents is unknown. Additionally, the number and
location of sites that use dishwasher detergents containing 1,4- are unknown. EPA did not identify data
on facility operating schedules. EPA assumes facilities use 1,4-dioxane 5 days/week, 50 weeks/year or
250 days/year. EPA modeled the 1,4-dioxane use rate using the SHEDS-HT case study from Liverpool,
OH to estimate releases, resulting in a 50th and 95th percentile 1,4-dioxane use rate of 1.44 g/site-day.
Worker Activities
EPA expects workers to be potentially exposed to 1,4-dioxane when handling dish detergent and when
dishwashing machines are opened, as detergents may be present if the machine has a wash solution
reservoir. While the dish detergent is diluted during use, workers may come into contact with the
undiluted dish detergent if a manual dispensing method is used or when attaching an automated
dispensing system to the container. These activities are all potential sources of worker exposure through
Page 431 of 570
-------
dermal contact to liquid and inhalation of vapors. Dishwashing workers may operate dishwashing
machines over their entire 8-hour shift; however, inhalation exposures are expected only when the
dishwasher machine door is opened ( 3. 2023). Note that the dermal exposure model discussed in
Section 3.1.1.3 does not have a term for dermal exposure duration, as it is based on a single dermal
contact event leaving a specific quantity on the skin.
ONUs include employees that work at the sites where dishwasher detergents are used, but they do not
directly handle the chemical and are therefore expected to have lower inhalation exposures and are not
expected to have dermal exposures through contact with liquids. ONUs for this scenario include
supervisors, managers, and other employees that may be in the dishwashing area are but do not perform
tasks that result in the same level of exposure as those workers that engage in tasks related to the use of
dishwasher detergent.
Number of Potentially Exposed Workers and ONUs
To estimate the number of workers, EPA used U.S. Census and BLS data for the following NAICS
codes: 623300, 713900, 721100, 721300, 722300, 722400, and 722500. EPA estimated a total of
773,851 sites, 0.6 workers per site, and 1.1 ONUs per site (U.S. BLS. ). For additional information
on the steps used to estimate the number of potentially exposed workers and ONUs, refer to Appendix
G.5 of the 2020 Risk Evaluation for 1,4-Dioxane (U.S. EPA. 2020c).
Worker Inhalation Exposure Assessment
EPA used the same modeling approach as discussed for dish soap in Appendix F.4.4 to estimate
inhalation exposures to 1,4-dioxane during the use of dishwasher detergent. EPA modified the input
parameters to the model to account for the differences between using dish soap versus detergent,
particularly for the cleaning stage. For dish soap, workers continuously wash dishes over an open sink
whereas, for dishwasher detergents, workers load dishes into the dishwasher, run the dishwasher, and
unload the dishes from the dishwasher. This model accounts for the reduced time during which workers
are potentially exposed during automatic dishwashing {i.e., just the time when the dishwasher is open).
The model also accounts for differences in wash temperature between hand washing and using
automated dishwashers. See Appendix F.10 for detailed explanations of each input parameter.
TableApx F-15 provides a summary of the modeled inhalation exposures for use of dishwasher
detergents.
Table Apx F-15. Modeled Occupational Inhalation Exposures for Dishwasher
)etergent
Statistic
1,4-Dioxane Exposure Concentration, 8 Hour-TWA
(mg/m3)
Maximum
0.15
99th Percentile
1.1E-02
95th Percentile
4.5E-03
50th Percentile
5.9E-04
5th Percentile
3.4E-05
Minimum
1.3E-08
Mean
1.3E-03
EPA used the 50th and 95th percentile modeled 8-hour TWA exposures values presented in Table Apx
F-15 to calculate the central tendency and high-end ADC and LADC for workers, respectively. The
Page 432 of 570
-------
calculated values are summarized in TableApx F-16. Equations for calculating ADC and LADC are
presented in Appendix G of the December 2020 Final Risk Evaluation for 1,4-Dioxane (U.S. EPA.
2020c).
Exposure data for ONUs were not available. The ONU exposures are anticipated to be lower than
worker exposures since ONUs do not typically directly handle the chemical. Only inhalation exposures
to vapors or incidental dermal exposures may be expected to ONUs.
Table Apx F-16. Inhalation Exposures of Workers for the Use of Dishwasher Detergents Based on
Modeling
Exposure Type
Central Tendency
(50th Percentile) (mg/m3)"
High-End
(95th Percentile)
(mg/m3)"
Draft \U:. estimates
8-hour TWA Exposure Concentrations
1.0C
2.\d
Average Daily Concentration (ADC)
1.0
2.0
Lifetime Average Daily Concentration (LADC)
0.398
1.03
I pdated estimates
8-hour TWA Exposure Concentrations
5.9E-04
4.5E-03
Average Daily Concentration (ADC)
5.7E-04
4.3E-03
Lifetime Average Daily Concentration (LADC)
2.3E-04
2.2E-03
a See Table_Apx F-12 for corresponding references.
b For select OESs, updates to exposure estimates were made via information provided by the SACC and public
comments.
c All data were non-detect; EPA presented the LOD/2 for the central tendency value.
d All data were non-detect; EPA presented the LOD for the high-end value.
Key Uncertainties
Since EPA used the same approach as discussed for dish soap in Appendix F.4.4, the same key
uncertainties in that appendix apply.
F.4.6 Laundry Detergent (Industrial and Institutional)
Process Description
1,4-Dioxane is found in laundry detergents due to its presence as an unintentional byproduct in certain
ethoxylated substances that may be used as formulation components (U.S. EPA. 2020c). Laundries can
be classified into two main categories in the United States: industrial and institutional ( ).
For both categories, the laundered items are loaded into the mechanical washers and the laundry is
washed using water and a detergent appropriate for the item type and soil loading. Washing is completed
in a continuous process composed of a series of cycles. The wash cycle is typically followed by a rinse
cycle to remove the of the detergent chemicals. Although many facilities may have on-site wastewater
treatment, most of these treatment technologies are designed to remove dirt and oil, not detergent
chemicals. Subsequently, the wastewater is transferred down drains to a POTW. A flow diagram
including release and exposure points from the ESD on Water Based Washing Operations at Industrial
and Institutional Laundries is presented in FigureApx F-2 (01 ).
Page 433 of 570
-------
GXiD(©/'Container Residue and Cleaning ©O©Container Residue md Cleaning
(3)(4) Fugitive Air Release and Dust
Emission's During Transfers
(J)(4) Fugitive An*Release and Dust
Emissions During Transfers
k
Liquid Laundry
Cleaning Products
Received 111 Drums.
Totes, 01 Tank Tracks
rr\
iAj Connecting
Transfer Lines
Automatic
Loading of
Chemicals
Manual
Loading of
Chemicals
Scooping
or Pouring
Solid or Liquid Laundry
Cleaning Products
Received m Pails.
Drums, or Totes
Laundry
Received 111
Tracks or —
Generated
On-site
Sorting
Weighing
Washing Machine
(25 -95 111111)
Laige Object-, ipem
papet etc 1 to Landfill
yC) Handling Damp
Laundrv
(D®© Releases During
Operations
and
Str amine
Hanging
and
Folding
Laundry
Returned to
Customers
Exuu'ui es:
(A 1 Dermal and inhalation exposure 60111 connecting transfei lines or from ^cooping and pouring.
(JB) Dermal and inhalation exposure dunne container cleaning uf containers cleaned on-sitet
(O Dermal exposure from handling damp laundry and inhalation exposure to vaporized chemicals dining operations.
Releases:
0 Transport container residue released to water, incineration, or landfill.
(2) Open surface losses to air during transport container cleaning (if containers cleaned on-site; volatile chemicals only).
CD Transfer operation losses to air from unloading and transferring laundry cleaning product (volatile chemicals only).
/•'"V
V4J Dust losses dating unloading and transferring solids (powdered laundry products only).
© Releases to air within the workers' breathing zone from operations.
(gj Washing water discharge to POTW (non-volatile and volatile chemicals) and evaporation losses to air during
washing and drying operations I volatile chemicals only).
FigureApx F-2. Environmental Release and Occupational Exposure Points During Industrial/Institutional Laundering Operation
Page 434 of 570
-------
Industrial Laundries: Industrial laundries wash soiled garments, linens, etc., received from hospitals,
repair shops, doctor's offices, industrial sites, as well as other customers (QEt O 1 I h). EPA did not
find specific container information for 1,4-dioxane in industrial laundry detergents; however, the ESD
on Water Based Washing Operations at Industrial and Institutional Laundries indicates that industrial
laundry detergents are typically transported as a liquid or powder in drums, totes, or bulk tanker trucks
(( ) 1,4-Dioxane can be present in institutional laundry detergents at concentrations from
0.05 to 14 ppm (U.S. EPA. 2020c). In addition, EPA reviewed the NY DEC database of waivers for
cleaning, personal care, and cosmetic products not meeting the proposed maximum concentrations of
1,4-dioxane in these products (2 ppm by the end of 2022 and 1 ppm by the end of 2023) (NYDEC.
2023). Using the product names/descriptions in the database, EPA determine which products in were
likely relevant to laundry detergents; EPA could not generally distinguish between institutional and
industrial laundry detergents. EPA found that the concentration of 1,4-dioxane in laundry detergents in
this waiver database ranged from 2.0 to 129 ppm (NYDEC. 2023). Given all the available data, EPA
used this maximum concentration of 129 ppm in the occupational dermal exposure assessment in
Section 3.1.2.2.
The volume of 1,4-dioxane present in industrial laundry detergents is unknown. Additionally, the
number and location of sites that use industrial laundry detergents containing 1,4-dioxane as a are
unknown. According to the ESD on Water Based Washing Operations at Industrial and Institutional
Laundries, industrial laundry facilities operate over a range of 20 to 365 days per year (I ).
EPA modeled the 1,4-dioxane use rate for a generic site using the ESD on Water Based Washing
Operations at Industrial and Institutional Laundries to estimate releases, resulting in 50th and 95th
percentile 1,4-dioxane use rates of approximately 7/10 5 and 0.0013 kg/site-day in both industrial
power and liquid laundry detergents, respectively ( ). For additional information on the
modeling and associated input parameters used to estimate the daily use rate, refer to Appendix E. 11.16.
Institutional Laundries: Institutional laundries are typically located within a hospital, nursing home,
hotel, or other institutional facility ( ). EPA did not find specific container information for
1,4-dioxane in institutional laundry detergents; however, the ESD on Water Based Washing Operations
at Industrial and Institutional Laundries indicates that institutional laundry detergents are typically
transported as a liquid or powder in 5-galIon pails (QEt ). EPA used the same concentrations of
1,4-dioxane in laundry detergents as discussed above, as these data do not distinguish between industrial
and institutional laundry detergents.
The volume of 1,4-dioxane present in institutional laundry detergents is unknown. Additionally, the
number and location of sites that use institutional laundry detergents containing 1,4-dioxane as a
contaminant are unknown. According to the ESD on Water Based Washing Operations at Industrial and
Institutional Laundries, institutional laundry facilities operate over a range of 250 to 365 days per year
(OE( ). EPA modeled the 1,4-dioxane use rate for a generic site using the ESD on Water Based
Washing Operations at Industrial and Institutional Laundries to estimate releases, resulting in 50th and
95th percentile 1,4-dioxane use rates of approximately 2,2/ 10 5 and 1 /10 4 kg/site-day in power
detergents and 3.4x 10~5 and 0.0014 kg/site-day in liquid detergents, respectively ( ). For
additional information on the modeling and associated input parameters used to estimate the daily use
rate, refer to Appendix E. 11.16.
Worker Activities
Workers are potentially exposed to 1,4-dioxane in laundry detergents during transfer operations,
container cleaning, handling damp laundry, and other operational activities, which are expected for both
Page 435 of 570
-------
industrial and institutional laundries (OB ). These activities are all potential sources of worker
exposure through dermal contact or inhalation exposure to solid or liquid chemicals.
During the use of laundry detergents, workers may be exposed during manual loading of solid or liquid
detergent chemicals into the washing machine ( ). Automatic liquid injection systems may
be employed which reduce worker exposure; however, workers may still be exposed when connecting
transfer lines or transferring the liquid chemicals from the transport container to storage tanks. Solid
detergents are less frequently used than liquid detergents due to their increased risk of exposure from
dusts and inability to be automatically loaded into machines (OECD. i ).
The 2011 ESD on The Chemicals Used in Water-Based Washing Operations at Industrial and
Institutional Laundries indicates that PPE may be required in both industrial and institutional laundry
settings in the case of handling substances that may be corrosive or produce dust or vapors that can be
inhaled, or if workers' hands are constantly immersed in water or wash solutions containing detergents
(OE( ). However, these situations are not typical for most activities at industrial and
institutional laundries.
ONUs include employees that work at the sites where laundry detergent is used, but they do not directly
handle the chemical and are therefore expected to have lower inhalation exposures and are not expected
to have dermal exposures through contact with liquids. ONUs for this scenario include supervisors,
managers, and other employees that may be in the laundry areas but do not perform tasks that result in
the same level of exposures as those workers that engage in tasks related to the use of laundry
detergents.
Number of Potentially Exposed Workers and ONUs
For industrial laundries, EPA used U.S. Census and BLS data for the NAICS code 812330, Linen and
Uniform Supply, to estimate a total of 2,453 sites, 27 workers per site, and 3 ONUs per site ( ^
2016). EPA estimated the number of institutional laundries based on industry information as described
in the ESD on Water Based Washing operations at Industrial and Institutional Laundries, resulting in a
total of 95,533 sites and 6 workers per site. The number of ONUs per institutional laundry site is
unknown (OECD. ^ ). For additional information regarding the steps used to estimate the number of
potentially exposed workers and ONUs, refer to Appendix G.5 of the 2020 Risk Evaluation for 1,4-
Dioxane (U.S. EPA. 2020c).
Worker Inhalation Exposure Assessment
1,4-Dioxane is found in laundry detergents due to its presence as an unintentional byproduct in certain
ethoxylated substances that may be used as formulation components (U.S. EPA. 2020c). The
information and data quality evaluation to assess occupational exposures during use of laundry detergent
is listed in Table Apx F-17 and described below.
Page 436 of 570
-------
Table Apx F-17. Laundry Detergent Worker Exposure Data Evaluation
Worker Activity or
Sampling Location
Data Type
Number of
Samples
Overall Data
Quality
Determination
Source/
Reference
Unloading detergent into
machines, cleaning empty
detergent containers, laundry
operations
Input parameters for
Monte Carlo modeling
N/A
Medium0
(OECD.
2 )
a This is the rating for the underlying data used in the model, and not the Monte Carlo model itself.
EPA did not find relevant inhalation monitoring data for the use of laundry detergent. Therefore, EPA
modeled 1,4-dioxane air concentrations using a Monte Carlo modeling approach, which is described in
Appendix F.8. This modeling approach utilizes the EPA/OPPT Penetration Model, EPA/OPPT Mass
Transfer Coefficient Model, EPA Mass Balance Inhalation Model, and Generic Model for Central
Tendency and High-End Inhalation Exposure to Total and Respirable Particulates Not Otherwise
Regulated (PNOR), with variation in input parameters for mass fraction of 1,4-dioxane in detergent,
ventilation rate, mixing factor, and total/respirable PNOR concentrations based on available data. To
compile a full-shift estimate, EPA combined exposure estimates for all activities, ensuring that the total
exposure duration for all activities combined did not exceed the shift length, which could be 8, 10, or 12
hours per the OECD ESD on the Chemicals Used in Water Based Washing Operations at Industrial and
Institutional Laundries (Of ). Container unloading and cleaning duration was calculated by
taking the number of containers unloaded and dividing by fill rate and operating days. Laundry
operation duration was calculated by taking the total work shift duration and subtracting the duration of
container unloading and cleaning. TableApx F-18 and TableApx F-19 present the modeled 8-hour,
10-hour, and 12-hour TWA exposures for industrial and institutional laundries, respectively.
Table Apx
7-18. Modeled Occupational
nhalation Exposures for Industrial Laundries
Statistic
1,4-Dioxane
Exposure, 8h-
TWA Vapor
(mg/m3)
1,4-Dioxane
Exposure,
lOh-TWA
Vapor (mg/m3)
1,4-Dioxane
Exposure,
12h-TWA
Vapor (mg/m3)
1,4-Dioxane
Exposure, 8h-TWA
Total Particulate
(mg/m3)
1,4-Dioxane Exposure,
8h-TWA Respirable
Particulate
(mg/m3)
Maximum
3.9E-02
4.8E-02
5.8E-02
1.9E-03
6.4E-04
99th
Percentile
2.9E-02
3.5E-02
4.2E-02
1.7E-03
5.4E-04
95th
Percentile
2.1E-02
2.5E-02
2.9E-02
1.4E-03
4.0E-04
50th
Percentile
8.6E-04
9.9E-04
1.1E-03
5.6E-05
1.4E-05
5th
Percentile
1.1E-05
1.3E-05
1.5E-05
6.5E-07
2.0E-07
Minimum
1.2E-06
1.1E-06
1.3E-06
8.0E-09
2.9E-09
Mean
4.8E-03
5.6E-03
6.4E-03
3.2E-04
8.2E-05
Page 437 of 570
-------
Table Apx F-19. Modeled
Occupational Inhalation Exposures for Institutional
sundries
Statistic
1,4-Dioxane
Exposure, 8h-
TWA Vapor
(mg/m3)
1,4-Dioxane
Exposure, lOh-
TWA Vapor
(mg/m3)
1,4-Dioxane
Exposure, 12h-
TWA Vapor
(mg/m3)
1,4-Dioxane
Exposure, 8h-
TWA Total
Particulate
(mg/m3)
1,4-Dioxane
Exposure, 8h-
TWA Respirable
Particulate
(mg/m3)
Maximum
3.9E-02
4.7E-02
5.6E-02
1.9E-03
6.4E-04
99th
Percentile
2.2E-02
2.7E-02
3.2E-02
1.7E-03
5.4E-04
95th
Percentile
1.6E-02
1.9E-02
2.3E-02
1.4E-03
4.0E-04
50th
Percentile
6.5E-04
7.6E-04
8.7E-04
5.6E-05
1.4E-05
5th
Percentile
8.4E-06
1.0E-05
1.2E-05
6.5E-07
2.0E-07
Minimum
1.0E-06
1.1E-06
1.2E-06
4.2E-09
2.9E-09
Mean
3.7E-03
4.3E-03
4.9E-03
3.2E-04
8.2E-05
EPA used the 50th and 95th percentile modeled 8-hour TWA exposures from TableApx F-18 and
TableApx F-19 to calculate the central tendency and high-end ADC and LADC for laundry detergents,
based on the timeframe for the available health hazard data. The calculated values are summarized in
Table Apx F-20 and Table Apx F-20, respectively. Equations for calculating ADC and LADC are
presented in Appendix G of the December 2020 Final Risk Evaluation for 1,4-Dioxane (U.S. EPA.
2020c).
Exposure data for ONUs were not available. The ONU exposures are anticipated to be lower than
worker exposures since ONUs do not typically directly handle the chemical. Only inhalation exposures
to vapors or incidental dermal exposures may be expected to ONUs.
Table Apx F-20. Inhalation Exposures of Workers for the Use of Laundry Detergent in Industrial
Laundries Based on Modeling
Exposure Type
Physical Form
Central Tendency
(50th Percentile)
(mg/m3)"
High-End
(95th Percentile)
(mg/m3)"
Draft RE estimates''
8-hour TWA Exposure
Concentrations
Liquid detergents: vapor
5.2E-04
1.9E-03
Solid detergents: total particulate
1.1E-04
2.0E-04
Solid detergents: respirable
particulate
3.5E-05
6.7E-05
Average Daily Concentration
(ADC)
Liquid detergents: vapor
4.96E-04
1.80E-03
Solid detergents: total particulate
1.01E-04
1.92E-04
Solid detergents: respirable
particulate
3.38E-05
6.40E-05
Lifetime Average Daily
Concentration (LADC)
Liquid detergents: vapor
1.97E-04
9.22E-04
Solid detergents: total particulate
4.03E-05
9.84E-05
Solid detergents: respirable
particulate
1.34E-05
3.28E-05
Page 438 of 570
-------
Exposure Type
Physical Form
Central Tendency
(50th Percentile)
(mg/m3)"
High-End
(95th Percentile)
(mg/m3)"
Updated estimates''
8-hour TWA Exposure
Concentrations
Liquid detergents: vapor
8.6E-04
2.1E-02
Solid detergents: total particulate
5.6E-05
1.4E-03
Solid detergents: respirable
particulate
1.4E-05
4.0E-04
Average Daily Concentration
(ADC)
Liquid detergents: vapor
8.3E-04
2.0E-02
Solid detergents: total particulate
5.4E-05
1.4E-03
Solid detergents: respirable
particulate
1.4E-05
3.9E-04
Lifetime Average Daily
Concentration (LADC)
Liquid detergents: vapor
3.3E-04
1.0E-02
Solid detergents: total particulate
2.2E-05
7.0E-04
Solid detergents: respirable
particulate
5.5E-06
2.0E-04
a See Table Apx F-17 for corresponding references.
b For select OESs, updates to exposure estimates were made via information provided by the SACC and public
comments.
TableApx F-21. Acute and Chronic Inhalation Exposures of Workers for the Use of Laundry
Detergent in Institutional Laundries Based on Modeling
Exposure Type
Physical Form
Central Tendency
(50th Percentile)
(mg/m3)"
High-End
(95th
Percentile)
(mg/m3)"
Dial'l RL cslimak-s
8-hour TWA Exposure
Concentrations
Liquid detergents: vapor
4.10E-04
1.45E-03
Solid detergents: total particulate
1.05E-04
2.00E-04
Solid detergents: respirable particulate
3.51E-05
6.65E-05
Average Daily Concentration
(ADC)
Liquid detergents: vapor
3.94E-04
1.39E-03
Solid detergents: total particulate
1.01E-04
1.92E-04
Solid detergents: respirable particulate
3.38E-05
6.40E-05
Lifetime Average Daily
Concentration (LADC)
Liquid detergents: vapor
1.57E-04
7.14E-04
Solid detergents: total particulate
4.03E-05
9.84E-04
Solid detergents: respirable particulate
1.34E-05
3.28E-05
I pdalcd cslimak-s
8-hour TWA Exposure
Concentrations
Liquid detergents: vapor
6.5E-04
1.6E-02
Solid detergents: total particulate
5.6E-04
1.4E-03
Solid detergents: respirable particulate
1.4E-05
4.0E-04
Liquid detergents: vapor
6.3E-04
1.5E-05
Page 439 of 570
-------
Exposure Type
Physical Form
Central Tendency
(50th Percentile)
(mg/m3)"
High-End
(95th
Percentile)
(mg/m3)"
Average Daily Concentration
(ADC)
Solid detergents: total particulate
5.4E-05
1.4E-03
Solid detergents: respirable particulate
1.4E-05
3.9E-04
Lifetime Average Daily
Concentration (LADC)
Liquid detergents: vapor
2.5E-04
7.9E-03
Solid detergents: total particulate
2.2E-05
7.0E-04
Solid detergents: respirable particulate
5.5E-06
2.0E-04
a See Table_Apx F-17 for corresponding references.
b For select OESs, updates to exposure estimates were made via information provided by the SACC and public
comments.
Key Uncertainties
Due to a lack of data specific to 1,4-dioxane for this use, EPA used assumptions and values from the
ESD on Water Based Washing Operations at Industrial and Institutional Laundries and EPA models to
estimate inhalation exposures during container transfers, container cleaning, and laundry operations (see
Appendix F.8). The uncertainties associated with this modeling approach are described in Section
3.1.2.4.
F.4.7 Paint and Floor Lacquer
Process Description
EPA identified 1,4-dioxane present in consumer paints and floor lacquer as an unintentional byproduct
in formulation components (U.S. EPA. 2020c). Concentrations of 1,4-dioxane in consumer paints and
floor lacquer range from 0.02 to 30 ppm ( 20c). These consumer products could potentially
be used commercially. Additionally, 1,4-dioxane is present as an unintentional component of
commercial automotive refinishing coatings and architectural paints/coatings (Franz et at.. , ). Based
on this information, EPA assesses this OES as the commercial use of paints, coatings, and lacquers.
Based on the products identified, of the available GS and ESD, the ESD on Coating Application via
Spray-Painting in the Automotive Refinishing Industry (OECD. 2009) and the ESD on the Coating
Industry (Paints, Lacquers and Varnishes) ( ) are the most applicable; however, the latter
ESD contains relatively limited information, mostly focused on general process information.
Paint and coating formulations are typically transported as a liquid in drums and are loaded into the
reservoir of application equipment (OECD. 2009). The application procedure depends on the type of
paint or floor lacquer and the type of substrate. The paint or lacquer may be applied to the substrate via
spray, brush, or roller application. Following application, the paint or lacquer is allowed to cure or dry.
The curing process may involve air drying, baking, or radiation curing, depending on the substrate being
painted or coated (OECD. 2009).
The volume of 1,4-dioxane present in paints and floor lacquer is unknown. Information from the CDR
indicate that 1,4-dioxane is imported and present in paint and coatings as a formulation component (
20a). Additionally, the number and location sites that use paints and floor lacquer containing
1,4-dioxane are unknown. The ESD on Coating Application via Spray Painting in the Automotive
Refinishing Industry (referenced due to identification of 1,4-dioxane in automotive refinishing coatings)
indicates a default of 250 days/year of operation (OECD. 201 la). Using the default values from the ESD
and the concentration of 1,4-dioxane above (0.02 to 30 ppm), EPA calculates a daily use rate of 1,4-
dioxane at an automotive refinishing site of 3.2><10~8 to 4,8/10 5 kg/site-day.
Page 440 of 570
-------
Worker Activities
Workers are potentially exposed to 1,4-dioxane in paint and floor lacquer formulations during multiple
activities, including quality testing of formulations, transferring the formulations into application
equipment (if used), applying the formulation to a substrate, and maintenance and cleaning activities
(OECD. 2009). These activities are all potential sources of worker exposure through dermal contact to
liquid and inhalation of 1,4-dioxane vapors.
During application of paint or floor lacquer, workers may manually apply the formulation with a variety
of application techniques, including spray application, brush application, dipping, or rolling (OECD.
2009). All application methods have potential exposure points for workers. Some application methods
may be automated, which reduces the potential for worker exposures. For example, if the dip coating
apparatus has an enclosed reservoir, this reduces the potential for 1,4-dioxane vapors to escape and
become available for worker inhalation and vapor-through-skin exposure (OECD. 2009). The extent of
automated application processes and use of open versus closed systems in the various industries that
conduct paint or floor lacquer applications is unknown.
A NIOSH evaluation of a small parts and vehicle painting facility revealed that half-face respirators with
organic vapor cartridges were available to workers at the identified site (Hills et at.. 1989). The workers
mainly used brushes for paint application but occasionally used spray gun applicators for brief periods
of time. NIOSH suggests implementing a respiratory protection program for the painters; details of
which can be found in the NIOSH publication, Guide to Industrial Respiratory Protection, DHHS
(NIOSH) publication number 87-1 16 (NIOS. 7). NIOSH also recommends wearing gloves
impervious to the paints and solvents to prevent skin contact and avoid possible dermal exposure route
(Hills et at.. 1989). EPA did not find any additional information regarding PPE used at facilities that
apply paints and floor lacquer.
ONUs include employees that work at the sites where paint and floor lacquer is used, but they do not
directly handle the chemical and are therefore expected to have lower inhalation exposures and vapor-
through-skin uptake and are not expected to have dermal exposures through contact with liquids. ONUs
for this scenario include supervisors, managers, and other employees that may be in the application areas
but do not perform tasks that result in the same level of exposures as those workers that engage in tasks
related to the use of paint and floor lacquer.
Number of Potentially Exposed Workers and ONUs
EPA used U.S. Census and BLS data for the NAICS code 811121, Automotive Body, Paint, and Interior
Repair and Maintenance, to estimate a total of 33,648 sites, 111,511 workers, and 11,050 ONUs, which
corresponds to an estimated average of three workers and 0.3 ONUs per site ( 16). For
additional information on the steps used to estimate the number of potentially exposed workers and
ONUs, refer to Appendix G.5 of the 2020 Risk Evaluation for 1,4-Dioxane (U.S. EPA. 2020c).
Worker Inhalation Exposure Assessment
EPA identified 1,4-dioxane present in commercial paints and floor lacquer as an unintentional byproduct
in formulation components (U.S. EPA. 2020c). The information and data quality evaluation to assess
occupational exposures during use of paints and floor lacquer is listed in Table Apx F-22 and described
below.
Page 441 of 570
-------
TableApx F-22. Paint and Floor Lacquer Worker Exposure Data Evaluation
Worker Activity or
Sampling Location
Data Type
Number of
Samples
Overall Data
Quality
Determination
Source Reference
Applying tape to parts that
are not to be painted,
painting, touch-ups
PBZ Monitoring
17
High
(Hills et al. 1989)
TableApx F-24 shows the 17 discrete inhalation monitoring data points available in published literature
for the use of paint and floor lacquer containing 1,4-dioxane (Hills et at.. 1989). This data is from a
NIOSH study in which PBZ samples were taken at a military vehicle manufacturing site in 1987. The
study was conducted in the final processing where approximately 47 workers touch-up vehicles and
perform quality checks. The worker activities captured in this sampling include taping vehicles prior to
painting, painting vehicles, and performing paint touch-ups. The study does not identify where 1,4-
dioxane is present at the site, which is a limitation of this data.
The NIOSH report provided 17 PBZ sample results, three of which are 8-hour TWAs and the remaining
14 of which were taken over a shorter period of time. Many of these, however, are still close to a full
shift duration of 8 hours. EPA converted these 14 samples into 8-hour TWAs by assuming no exposure
for the remainder of the eight hours. EPA made this assumption because the site analyzed in the study
was not strictly a vehicle painting site. As such, workers may spend time doing other jobs that did not
involve formulations containing 1,4-dioxane. Therefore, EPA assumed that sampling occurred for the
duration of the employee's painting tasks where there was potential exposure to 1,4-dioxane.
Four of the 17 samples were non-detect for 1,4-dioxane. The study indicated that the LOD for all
samples was 0.1 mg/m3 of 1,4-dioxane. For the non-detect samples, EPA used the LOD divided by two
for subsequent central tendency and high-end calculations. EPA used this method for approximating a
concentration for non-detect samples because the geometric standard deviation of the dataset is greater
than three ( 4a).
EPA used the 8-hour TWA air concentration measurements and LOD/2 (for the non-detects) to calculate
central tendency (50th percentile) and high-end exposures (95th percentile). EPA used these values to
calculate the ADC and LADC. The calculated values are summarized in Table Apx F-23. Equations for
calculating ADC and LADC are presented in Appendix G of the December 2020 Final Risk Evaluation
for 1,4-Dioxane (U.S. EPA. 2020cY
Exposure data for ONUs were not available. The ONU exposures are anticipated to be lower than
worker exposures since ONUs do not typically directly handle the chemical. Only inhalation exposures
to vapors or incidental dermal exposures may be expected to ONUs.
Page 442 of 570
-------
TableApx F-23. Inhalation Exposures of Workers for the Use of Paint and Floor Lacquer Based
on Monitoring Data
Exposure Type
Central Tendency
(50th Percentile) (mg/m3)"
High-End
(95th Percentile)
(mg/m3)"
8-hour TWA Exposure Concentrations
0.210
1.20
Average Daily Concentration (ADC)
0.202
1.15
Lifetime Average Daily Concentration
(LADC)
0.080
0.592
a See Table_Apx F-22 for corresponding references.
Page 443 of 570
-------
Table Apx F-24,
Occupational Inhalation JV
onitoring Data for Paint and Floor Lacquer
Row
#
Type of
Sample
Worker
Activity or
Sample
Location
Number
of
Samples
Sample
Date
Sample
Time
1,4-Dioxane
Exposure
Concentration
(mg/m3)
EPA
Determined
8-hour TWA
(mg/m3)1'
Source
Overall Data
Quality
Determination
1
Personal
N/A
1
10/27/1987
480
ND (LOD = 0.1
mg/m3)
0.05
(Hills et al„
1989}
High
2
Personal
N/A
1
10/27/1987
480
ND (LOD = 0.1
mg/m3)
0.05
(Hills et al..
1989")
High
3
Personal
N/A
1
10/27/1987
480
ND (LOD = 0.1
mg/m3)
0.05
(Hills et al..
1989")
High
4
Personal
N/A
1
10/27/1987
463
0.1
0.10
(Hills et al..
1989")
High
5
Personal
N/A
1
10/27/1987
457
0.2
0.19
(Hills et al..
1989")
High
6
Personal
N/A
1
10/27/1987
456
0.5
0.48
(Hills et al..
1989")
High
7
Personal
N/A
1
10/27/1987
439
0.1
0.09
(Hills et al..
1989")
High
8
Personal
N/A
1
10/27/1987
441
0.7
0.64
(Hills et al..
1989")
High
9
Personal
N/A
1
10/27/1987
428
1.3
1.7
(Hills et al..
1989")
High
10
Personal
N/A
1
10/27/1987
251
1.7
0.89
(Hills et al..
1989")
High
11
Personal
N/A
1
10/27/1987
148
0.7
0.22
(Hills et al..
1989")
High
12
Personal
N/A
1
10/27/1987
456
1.3
1.24
(Hills et al..
1989")
High
Page 444 of 570
-------
Row
#
Type of
Sample
Worker
Activity or
Sample
Location
Number
of
Samples
Sample
Date
Sample
Time
1,4-Dioxane
Exposure
Concentration
(mg/m3)
EPA
Determined
8-hour TWA
(mg/m3)1'
Source
Overall Data
Quality
Determination
13
Personal
N/A
1
10/27/1987
229
0.4
0.19
(Hills et al„
1989}
High
14
Personal
N/A
1
10/27/1987
145
0.7
0.21
(Hills et al..
1989")
High
15
Personal
N/A
1
10/27/1987
347
1.0
0.72
(Hills et al..
1989")
High
16
Personal
N/A
1
10/27/1987
410
1.4
1.2
(Hills et al..
1989")
High
17
Personal
N/A
1
10/27/1987
400
ND (LOD = 0.1
mg/m3)
0.040
(Hills et al..
1989")
High
ND = non-detect for 1,4-dioxane; LOD = limit of detection; TWA = time-weighted average
a The 8-hour TWA calculations use LOD/2 for non-detect values because the geometric standard deviations of the underlying datasets are all >3.
Page 445 of 570
-------
Key Uncertainties
As discussed above, EPA translated short-term samples from the NIOSH HHE report (Hills et at.. 1989)
into 8-hour TWAs by assuming no exposure for the remainder of the eight hours after the sampling
duration. This assumption may result in underestimation of exposures if workers perform additional
activities that may result in exposures to 1,4-dioxane that were not captured in the monitoring performed
in the NIOSH HHE report. However, the data set did include full-shift monitoring, which EPA included
in this analysis. Additionally, these data are from one facility, and it is unclear how representative the
data are for all sites and all workers across the United States. The monitoring performed for the NIOSH
HHE was completed in the 1980s; therefore, the age of the monitoring data can also introduce
uncertainty.
As discussed above, EPA used half the detection limit for the non-detect values in the central tendency
and high-end exposure calculations. Due to the high number of non-detects (13 of the 17 TWAs were
non-detect), this method may result in bias ( 4a). Additional uncertainties are listed in
Section 3.1.2.4.
F.4.8 Spray Foam Application
Process Description
There are three main types of spray polyurethane foam (SPF): two-component high-pressure, two-
component low-pressure, and one-component foam (OCF) ( ). The low-pressure and
OCF types are available for DIY-use, but the high-pressure type is only available for professional use. A
safety data sheet (SDS) identified in the Preliminary Information on Manufacturing, Processing,
Distribution, Use, and Disposal: 1,4-Dioxane indicate that 1,4-dioxane is present in open- and closed-
cell SPFs, which are subsets of two-component high-pressure SPFs ( 1017a. b). Although one
SDS has been identified where 1,4-dioxane was listed as an ingredient, it could also be a byproduct and
the concentration could vary by the type of SPF.
High-pressure SPF is used for larger insulation applications, as an air sealant in hybrid insulations, and
in roofing applications. The components are typically stored in 55-gallon drums. The operator pumps
both components (sides A and B) through heated tubes from the supply tanks into a nozzle. 1,4-Dioxane
is a component in side B with concentrations typically around 0.1 percent U.S. EPA ( I,
2017a). Sides A and B begin to react in the nozzle and are sprayed at elevated pressures and
temperatures (>150 °F and 1,200 psi). The formulation may be applied via hand-held spray gun or
automated spray system. Closed-cell foam could be applied in layers. As the foam cures, it expands up
to 120 times its original size. After curing, the foam may be trimmed or cut. Trimmings and waste foam
are collected and disposed.
The volume of 1,4-dioxane present in spray polyurethane foams is unknown. In 2008, U.S. production
of two-component spray foams reached 365 million in 2008 ( >a). The GS on Application
of Spray Foam Insulation indicates a default of 260 days/year of operation ( 1018a).
Worker Activities
Workers are potentially exposed to 1,4-dioxane during the application of spray polyurethane foam while
unloading SPF chemicals into spray rig equipment, transport container cleaning, SPF application, and
trimming of the applied and hardened SPF insulation ( |a). These activities are all
potential sources of worker exposure through dermal contact to liquid and the inhalation of mist or
vapors. Exposure during equipment cleaning is not expected, as the spray equipment is a closed system
that is flushed with solvent; workers do not come into contact with the inside of the equipment.
Page 446 of 570
-------
During application of spray foam insulation, workers may manually apply the formulation via hand-held
spray gun or employ an automated spray system ( a). Both types of application are
potential exposure points for workers. Typically, the main engineering controls used by SPF applicators
are containment and ventilation. A containment system is often made up of plastic sheeting or cardboard
secured to walls to isolate the work zone, thus reducing the potential for airborne chemicals to enter the
building ventilation systems. Ventilation systems, including active exhaust and air supply systems, are
typically used to avoid accumulation of chemical vapors and particulate emissions near the application
area (U.S. EPA. 2018a).
According to the GS on Application of Spray Polyurethane Foam Insulation, workers at sites that apply
SPF insulation are expected to wear proper chemical-specific personal protective equipment (
2018a). Workers may wear chemical-resistant gloves, protective clothing (e.g., long sleeves, body suit,
coveralls), eye and face protection (e.g., safety glasses, chemical goggles), and respiratory protection.
Additionally, an SPF sprayer may wear a full-face, air-supplied respirator with chemical protective
coveralls and chemical protective gloves (U.S. EPA. 2018a). The appropriate PPE may vary for the
specific application.
ONUs include employees that work at the sites where spray polyurethane foam is applied, but they do
not directly handle the chemical and are therefore expected to have lower inhalation exposures and
vapor-through-skin uptake and are not expected to have dermal exposures through contact with liquids.
ONUs for this scenario include supervisors, managers, and other employees that may be in the
application areas but do not perform tasks that result in the same level of exposures as those workers that
engage in the tasks related to the use of spray polyurethane foam.
Number of Potentially Exposed Workers and ONUs
EPA estimated the number of potentially exposed workers and ONUs in Appendix G.6.7 of the
December 2020 Final Risk Evaluation for 1,4-Dioxane ( >20c).
Worker Inhalation Exposure Assessment
EPA estimated occupational inhalation exposures during the use of spray polyurethane foam containing
1,4-dioxane in Section 2.4.1.1.9 of the December 2020 Final Risk Evaluation for 1,4-Dioxane (U.S.
20c). EPA did not conduct additional analyses of occupational inhalation exposures for spray
polyurethane foam for this supplemental risk evaluation. Refer to the December 2020 Final Risk
Evaluation for 1,4-Dioxane for additional details.
Key Uncertainties
Key uncertainties are listed in Section 2.4.1.1.9 of the December 2020 Final Risk Evaluation for 1,4-
Dioxane (U.S. EPA. 2020c).
F.4.9 Polyethylene Terephthalate Byproduct
Process Description
1,4-Dioxane has been identified as a byproduct in the manufacture of PET plastics (1 c. i V \ JO I . ).
PET is produced by the esterification of terephthalic acid to form bishydroxyethyl terephthalate (BHET)
(Forkner et at.. 2004). BHET polymerizes in a transesterification reaction catalyzed by antimony oxide
to form PET (Forkner et at.. 2004). 1,4-Dioxane is produced as a byproduct in polyol reactors and is
distilled from the product and condensed along with water and/or glycol, with 1,4-dioxane present at 3
percent in the condensed off-take material (Huntsman. 2023). EPA used this concentration of 3 percent
in the occupational dermal exposure assessment in Section 3.1.2.2. Off-take material comprised of water
and 1,4-dioxane is loaded into trucks for off-site disposal, such as through Class I underground injection,
Page 447 of 570
-------
and off-take mixtures comprised of condensed glycol and 1,4-dioxane is sent to glycol manufacturers for
glycol recovery and antifreeze blending (Huntsman. 2023).
In 2014, 20.6 million metric tons of PET were used in the United States (McDaniel and DesLauriers.
2015). The volume of 1,4-dioxane produced as a byproduct of PET manufacturing is unknown. Due to
lack of information, EPA does not present annual or daily site throughputs. EPA assumes facilities that
produce 1,4-dioxane as a byproduct during PET manufacturing operate 5 days/week, 50 weeks/year or
250 days/year.
Worker Activities
Workers are potentially exposed to 1,4-dioxane during activities such as loading of waste containing
1,4-dioxane into trucks, equipment cleaning, and maintenance activities ( ) ( untsman.
2023). These activities are potential sources of worker exposure through dermal contact to liquid and
inhalation of volatile chemical vapors.
According to the GS on Use of Additives in Plastic Compounding, workers may wear suitable gloves,
hearing protection, and eye protection (U.S. EPA. 2 ). Facilities may use forced ventilation
techniques to reduce worker exposure to vapors. Local exhaust ventilation may be used in areas where
there is potential for the formation of particulates or vapors (\ v < < \ .0 -1 ^)- Workers wear
appropriate PPE during plant operations, which may include supplied air respirators (Huntsman. 2023).
EPA did not find information that indicates the extent that and worker PPE is used at facilities that
manufacture PET in the United States.
ONUs include employees that work at the sites where PET is manufactured, but they do not directly
handle the chemical and are therefore expected to have lower inhalation exposures and are not expected
to have dermal exposures through contact with liquids or solids. ONUs for this scenario include
supervisors, managers, and other employees that may be in the manufacturing areas but do not perform
tasks that result in the same level of exposures as those workers that engage in tasks related to the
manufacture of PET.
Number of Potentially Exposed Workers and ONUs
To estimate the number of workers, EPA used U.S. Census and BLS data for NAICS codes 325211 and
3261 13. EPA estimated a total of 1,695 sites, 43,528 workers, and 17,195 ONUs ( >. 2016). For
additional information on the steps used to estimate the number of potentially exposed workers and
ONUs, refer to Appendix G.5 of the December 2020 Final Risk Evaluation for 1,4-Dioxane (
2020c).
Worker Inhalation Exposure Assessment
1,4-Dioxane has been identified as a byproduct in the manufacture of polyethylene terephthalate (PET)
(\ v << \ \ ). Occupational exposure to 1,4-dioxane in PET was determined using monitoring data
provided by Chemical Exposure Health Data (OSHA. 2020). The information and data quality
evaluation to assess occupational exposures during manufacture of PET is listed in Table Apx F-25 and
summarized below.
Page 448 of 570
-------
Table Apx F-25. Polyethylene Terephthalate (PET) Byproduct Worker Exposure Data Evaluation
Worker Activity
or Sampling
Location
Data Type
Number of
Samples
Overall Data
Quality
Determination
Source
Reference
Unknown
PBZ Monitoring
11
High
(oc.n \ :020)
Various
PBZ Monitoring
63
High
(DAK Americas.
2023)
Around tanks
during loading
and maintenance
Area Monitoring
4
High
(Huntsman. 2023)
EPA assessed occupational inhalation exposures for this OES using OSHA's CEHD (OSHA. 2020) and
data from two public comments (DAK Americas. 2023; Huntsman. 2023).
For detailed information on where/how CEHD was obtained and mapped to OES, see Appendix F.4.1.
For this OES, monitoring data were available in CEHD from five sites with SIC codes 3089 (All Other
Plastics Product Manufacturing), 2653 (Corrugated and Solid Fiber Box Manufacturing), 3052 (Rubber
and Plastics Hoses and Belting Manufacturing), and 3069 (All Other Rubber Product Manufacturing).
All sites were determined to be manufacturers of plastic products (foams, packaging, etc.). TableApx
F-27 shows the 11 discrete worker inhalation monitoring data points from CEHD for this OES, all of
which are PBZ samples, from five different sites. The data are from 1985 to 1994. For one of these sites,
all air concentrations were non-detect for 1,4-dioxane. EPA excluded the data from this site when
calculating central tendency and high-end exposures for this OES because all samples at the site were
non-detect for 1,4-dioxane, meaning it is questionable if the site handles 1,4-dioxane. CEHD does not
include information on the worker activities included in the PBZ sampling, therefore EPA's assessment
assumed that all remaining samples are relevant to this assessment. Furthermore, it is uncertain the
extent to which all potential worker activities are represented in these data. As discussed in Appendix
F.4.1, EPA combined CEHD sample results with the same inspection number and sampling number to
attempt to construct a full-shift exposure concentration.
For the CEHD samples with detected values, EPA translated the sample results into 8-hour TWA
concentrations by assuming that the exposure concentration is zero for the time remaining in the 8-hour
durations. EPA made this assumption because the data include multiple samples for the same worker,
thus increasing the likelihood that the data reflect all tasks with potential 1,4-dioxane exposures. Where
non-detect values were included in the dataset, EPA calculated the LOD for each sample and used the
LOD/2 or LOD/V2 for subsequent central tendency (50th percentile) and high-end (95th percentile)
calculations, depending on the geometric standard deviation of the datasets ( 4a).
One public comment provides 63 personal breathing zone samples taken from 5 different sites over 1998
to 2023, as shown in Table Apx F-27 ( lericas. 2023). The sample results include worker
activity information and sample durations. In summary, 41 of these samples had sample durations
shorter than 92 minutes and the rest were for durations exceeding 520 hours. Based on this, EPA
excluded the results for the short-term samples because they appear to be task-based and not
representative of full shift exposure. Because all but two samples were non-detect for 1,4-dioxane, this
exclusion is not expected to have a significant impact on the results. Because these data include detailed
worker activities, for ONU exposure, EPA used samples taken during worker activities that do not
appear to directly involve 1,4-dioxane, such as patrolling process areas. The exclusions and data used for
Page 449 of 570
-------
ONU exposures are denoted as such in TableApx F-27. For central tendency and high-end exposure
calculations, EPA assessed exposures as the LOD/V2 for non-detect samples because the geometric
standard deviation of each site's dataset is less than three ( 4a).
Another public comment provides four area samples taken in 2019 at one facility, as shown in
Table Apx F-27 (Huntsman. 2023). Samples were taken near tanks during routine activities and loading,
as well as from a third-floor tower during maintenance activities. Because the public comment does not
provide context to determine if the location of the area sampling is representative of where workers
perform activities, EPA assesses these data as representing general area conditions, which are more
representative of ONU than worker exposures. All data were non-detect for 1,4-dioxane; therefore, EPA
assessed ONU exposures as the LOD/V2 because the geometric standard deviation of the dataset is less
than three ( la).
EPA then used the 8-hour TWAs as shown in Table Apx F-27 to calculate full shift (8-hour TWA)
central tendency (50th percentile) and high-end (95th percentile) inhalation exposures for workers and
ONUs. EPA used these values to calculate the ADC and LADC. The calculated values are summarized
in Table Apx F-26. Equations for calculating ADC and LADC are presented in Appendix G of the
December 2020 Final Risk Evaluation for 1,4-Dioxane ( s20c).
Table Apx F-26. Inhalation Exposures of Workers for PET Byproduct Based on Monitoring Data
Exposure Type
Central Tendency
(50th Percentile)
(mg/m3)"
High-End
(95th Percentile)
(mg/m3)"
W orker (Draft Rl- estimates)
8-hour TWA Exposure Concentrations
4.7
47
Average Daily Concentration (ADC)
4.52
45.2
Lifetime Average Daily Concentration
(LADC)
1.80
23.2
Worker (updated estimates)
S-hour TWA Exposure Concentrations
0.74
5.9
Average Daily Concentration (ADC)
0.71
5.7
Lifetime Average Daily Concentration
(LADC)
0.28
2.9
OM (updated estimates)
8-hour TWA Exposure Concentrations
0.21
0.23
Average Daily Concentration (ADC)
0.20
0.22
Lifetime Average Daily Concentration
(LADC)
8.0E-02
0.11
a See Table_Apx F-25 for corresponding references.
b For select OESs, updates to exposure estimates were made via information provided by the SACC and public
comments.
Page 450 of 570
-------
Table Apx F-
27. Occupational Inhalation Monitoring
Data for Polyethylene Terephthalate (PET) Byproduct
Row
#
Type of
Sample
Worker Activity or Sample
Location
Number
of
Samples
Sample
Date
Sample
Time
(min)
1,4-Dioxane
Exposure
Concentration
(mg/m3)
EPA
Determined
8-hour TWA
(mg/m3)"
Source
Overall Data
Quality
Determination
1
PBZ
Unknown
1
4/23/1985
270
ND (LOD=0.74)
0.21
(OSHA. 2020)
High
2
PBZ
Unknown
1
4/23/1985
270
10.5
5.9
(OSHA. 2020)
High
3
PBZ
Unknown
1
4/23/1985
270
11.2
6.3
(OSHA. 2020)
High
4
PBZ
Unknown
1
1/10/1994
112
ND (LOD=1.37)
0.90 (8-hr
(OSHA. 2020)
Excluded6
5
PBZ
Unknown
1
1/10/1994
169
ND (LOD=0.9)
TWA for the
same worker
from rows 4-
7)
(OSHA. 2020)
Excluded6
6
PBZ
Unknown
1
1/10/1994
79
ND (LOD=1.93)
(OSHA. 2020)
Excluded6
7
PBZ
Unknown
1
1/10/1994
95
ND (LOD=1.61)
(OSHA. 2020)
Excluded6
8
PBZ
Unknown
1
1/10/1994
130
ND (LOD=2.16)
1.7 (8-hr
(OSHA. 2020)
Excluded6
9
PBZ
Unknown
1
1/10/1994
83
ND (LOD=3.38)
TWA for the
same worker
from rows 8-
11)
(OSHA. 2020)
Excluded6
10
PBZ
Unknown
1
1/10/1994
90
ND (LOD=3 .12)
(OSHA. 2020)
Excluded6
11
PBZ
Unknown
1
1/10/1994
174
ND (LOD=1.61)
(OSHA. 2020)
Excluded6
12
PBZ
Unknown
1
1/10/1994
162
ND (LOD=l .l)
1.1 (8-hr
(OSHA. 2020)
Excluded6
13
PBZ
Unknown
1
1/10/1994
79
ND (LOD=2.27)
TWA for the
same worker
from rows 12-
15)
(OSHA. 2020)
Excluded6
14
PBZ
Unknown
1
1/10/1994
127
ND (LOD=1.41)
(OSHA. 2020)
Excluded6
15
PBZ
Unknown
1
1/10/1994
93
ND (LOD=1.92)
(OSHA. 2020)
Excluded6
16
PBZ
Unknown
1
3/15/1991
20
8.6
2.4 (8-hr
(OSHA. 2020)
High
17
PBZ
Unknown
1
3/15/1991
30
9.9
(OSHA. 2020)
High
18
PBZ
Unknown
1
3/15/1991
30
9.6
TWA for the
same worker
from rows 16-
21)
(OSHA. 2020)
High
19
PBZ
Unknown
1
3/15/1991
40
6.0
(OSHA. 2020)
High
20
PBZ
Unknown
1
3/15/1991
25
ND (LOD=4)
(OSHA. 2020)
High
21
PBZ
Unknown
1
3/15/1991
30
ND (LOD=3.33)
(OSHA. 2020)
High
Page 451 of 570
-------
Row
#
Type of
Sample
Worker Activity or Sample
Location
Number
of
Samples
Sample
Date
Sample
Time
(min)
1,4-Dioxane
Exposure
Concentration
(mg/m3)
EPA
Determined
8-hour TWA
(mg/m3)"
Source
Overall Data
Quality
Determination
22
PBZ
Unknown
1
3/15/1991
30
7.4
3.4 (8-hour
TWA for the
same worker
from rows 22-
26)
(OSHA. 2020)
High
23
PBZ
Unknown
1
3/15/1991
30
28.2
(OSHA. 2020)
High
24
PBZ
Unknown
1
3/15/1991
30
8.3
(OSHA. 2020)
High
25
PBZ
Unknown
1
3/15/1991
30
7.6
(OSHA. 2020)
High
26
PBZ
Unknown
1
3/15/1991
30
ND (LOD=3.33)
(OSHA. 2020)
High
27
PBZ
Unknown
1
8/12/1993
75
4.1
4.3 (8-hour
TWA for the
same worker
from rows 27-
33)
(OSHA. 2020)
High
28
PBZ
Unknown
1
8/12/1993
75
6.2
(OSHA. 2020)
High
29
PBZ
Unknown
1
8/12/1993
10
6.2
(OSHA. 2020)
High
30
PBZ
Unknown
1
8/12/1993
75
4.7
(OSHA. 2020)
High
31
PBZ
Unknown
1
8/12/1993
75
5.6
(OSHA. 2020)
High
32
PBZ
Unknown
1
8/12/1993
75
2.4
(OSHA. 2020)
High
33
PBZ
Unknown
1
8/12/1993
60
4.5
(OSHA. 2020)
High
34
PBZ
Unknown
1
8/12/1993
457
4.9
4.7
(OSHA. 2020)
High
35
PBZ
Unknown
1
11/20/1990
74
0.9
0.14
(OSHA. 2020)
High
36
PBZ
Collecting process samples in
Polymer 1* building
1
1/21/2003
20
ND (LOD=21.7)
0.64
(DAK
Americas.
2023)
Excluded17
37
PBZ
Collecting process samples in
Polymer 1* building
1
1/22/2003
15
ND (LOD=28.9)
0.64
(DAK
Americas,
2023)
Excluded17
38
PBZ
Collecting process samples in
Polymer 1* building
1
1/28/2003
27
ND (LOD=16.1)
0.64
(DAK
Americas,
2023)
Excluded17
39
PBZ
While operators obtain process
samples
1
7/8/1998
22
ND (LOD=1.3)
0.041
(DAK
Americas.
2023)
Excluded17
40
PBZ
While operators obtain process
samples
1
7/8/1998
33
ND (LOD=0.8)
0.041
(DAK
Americas.
2023)
Excluded17
Page 452 of 570
-------
Row
#
Type of
Sample
Worker Activity or Sample
Location
Number
of
Samples
Sample
Date
Sample
Time
(min)
1,4-Dioxane
Exposure
Concentration
(mg/m3)
EPA
Determined
8-hour TWA
(mg/m3)"
Source
Overall Data
Quality
Determination
41
PBZ
While operators obtain process
samples
1
7/8/1998
20
ND (LOD=1.5)
0.044
(DAK
Americas.
2023)
Excluded17
42
PBZ
While operators obtain process
samples
1
5/11/2001
15
ND (LOD=6.6)
0.15
(DAK
Americas.
2023)
Excluded17
43
PBZ
While operators obtain process
samples
1
5/11/2001
15
ND (LOD=6.5)
0.14
(DAK
Americas,
2023)
Excluded17
44
PBZ
While operators obtain process
samples
1
5/12/2001
15
ND (LOD=6.7)
0.15
(DAK
Americas,
2023)
Excluded17
45
PBZ
L1&L2 Stripper, L1&L2 MY33,
1-6 Crystallizer, L1&L2 paste,
L1&L2 water column, C01, C02,
E01, L1&L2 seal pot, L1&L2
water column
1
5/3/2007
671
ND (LOD=4.3)
4.3
(DAK
Americas.
2023)
High
46
PBZ
Cleaned immersion vessels 1st
floor, opened PTA feeder 2nd
floor restart cutter 3rd floor,
swapped L2 paste pump &
flushed with EG, put heads on L2
heat exchanger
1
5/3/2007
675
ND (LOD=5)
5.0
(DAK
Americas,
2023)
High
47
PBZ
Inspecting sprays at immersion
vessels (pre-polymer & final
polymer) and clean and swap
baskets - total of (4) baskets of
waste dumped into waste buggy
1
5/22/2008
34
ND (LOD=4.7)
0.23
(DAK
Americas.
2023)
Excluded17
48
PBZ
Inspecting sprays at immersion
vessels (pre-polymer & final
polymer) and clean and swap
baskets - total of (4) baskets of
waste dumped into waste buggy
1
6/5/2008
25
ND (LOD=6.5)
0.24
(DAK
Americas,
2023)
Excluded17
Page 453 of 570
-------
Row
#
Type of
Sample
Worker Activity or Sample
Location
Number
of
Samples
Sample
Date
Sample
Time
(min)
1,4-Dioxane
Exposure
Concentration
(mg/m3)
EPA
Determined
8-hour TWA
(mg/m3)"
Source
Overall Data
Quality
Determination
49
PBZ
Inspecting sprays at immersion
vessels (L2 pre-polymer & final
polymer - Line 1 upset)
1
7/24/2008
52
9.7
1.1
(DAK
Americas.
Excluded17
2023)
50
PBZ
Changing pot filters - Pot filters -
Finisher
1
12/9/2005
48
ND (LOD= 0.2)
0.016
(DAK
Americas.
Excluded17
2023)
51
PBZ
Changing pot filters - Pot filters -
Finisher
1
12/9/2005
43
ND (LOD=0.3)
0.017
(DAK
Americas,
Excluded17
2023)
52
PBZ
Changing pot filters - Pot filters -
Finisher
1
12/9/2005
43
ND (LOD=0.2)
0.015
(DAK
Americas,
Excluded17
2023)
53
PBZ
Changing pot filters - Pot filters -
Up Flow
1
12/9/2005
46
ND (LOD=0.3)
0.017
(DAK
Americas.
Excluded17
2023)
54
PBZ
Changing pot filters - Pot filters -
Up Flow
1
12/9/2005
43
ND (LOD=0.3)
0.017
(DAK
Americas.
Excluded17
2023)
55
PBZ
Changing pot filters - Pot filters -
Up Flow
1
12/9/2005
42
ND (LOD=0.3)
0.016
(DAK
Americas,
Excluded17
2023)
56
PBZ
Cleaning/raking finisher hot well
1
4/13/2005
37
ND (LOD=0.7)
0.039
(DAK
Americas.
Excluded17
2023)
57
PBZ
Cleaning/raking finisher hot well
1
4/14/2005
28
ND (LOD=0.7)
0.030
(DAK
Americas.
Excluded17
2023)
58
PBZ
Cleaning/raking finisher hot well
1
4/15/2005
32
ND (LOD=0.7)
0.034
(DAK
Americas,
Excluded17
2023)
Page 454 of 570
-------
Row
#
Type of
Sample
Worker Activity or Sample
Location
Number
of
Samples
Sample
Date
Sample
Time
(min)
1,4-Dioxane
Exposure
Concentration
(mg/m3)
EPA
Determined
8-hour TWA
(mg/m3)"
Source
Overall Data
Quality
Determination
59
PBZ
Cleaning/raking finisher hot well
1
4/18/2005
25
ND (LOD=0.7)
0.027
(DAK
Americas.
2023)
Excluded17
60
PBZ
Cleaning/raking finisher hot well
1
4/28/2005
30
ND (LOD=0.7)
0.032
(DAK
Americas.
2023)
Excluded17
61
PBZ
Cleaning/raking finisher hot well
1
4/29/2005
28
ND (LOD=0.7)
0.030
(DAK
Americas,
2023)
Excluded17
62
PBZ
Cleaned dryer screens; pulled
finisher hot well screens twice
and sprayed down; drainer recirc.
Pump for the finisher hot well
twice; pulled 1 o'clock liquid
samples; took band filter paper
out, etc....
1
9/28/2022
681
ND (LOD=0.7)
0.72
(DAK
Americas,
2023)
High
63
PBZ
Worked around finisher hot well;
pulled samples; pulled water
sample on band filter; pulled up
flow heat exchanger sample
(glycol is in this system); worked
on recrystallizer to free up clumps
(some AA possible); rodded mix
tank chute, etc..
1
9/29/2022
658
ND (LOD=0.7)
0.70
(DAK
Americas,
2023)
High
64
PBZ
Hot well screen raking, pot filter
finisher & up flow filter cleaning,
collection of liquid samples,
cleaned out mix tank chute 4-5
times, valved out Nash pump,
assisted w/pot filter (finisher)
swap over, took out band filter
paper, around hot wells cleaning
up waste around them, etc...
1
9/28/2022
675
ND (LOD=0.7)
0.72
(DAK
Americas,
2023)
High
Page 455 of 570
-------
Row
#
Type of
Sample
Worker Activity or Sample
Location
Number
of
Samples
Sample
Date
Sample
Time
(min)
1,4-Dioxane
Exposure
Concentration
(mg/m3)
EPA
Determined
8-hour TWA
(mg/m3)"
Source
Overall Data
Quality
Determination
65
PBZ
Performed pot filter work and hot
well inspections and raking,
collected liquid samples (4 total),
in admin bldg. for a bit, outside,
walkthrough laser bldg., helped
w/crystallizer beds, walked up to
silos, collected chip samples,
swept up chip, looked at chippers
1
9/29/2022
618
ND (LOD=0.7)
0.66
(DAK
Americas.
2023)
High
66
PBZ
Hot well raking and pot filter
(finisher and up flow) cleaning
1
9/28/2022
68
ND (LOD=l .l)
0.11
(DAK
Americas,
2023)
Excluded17
67
PBZ
Hot well raking and pot filter
(finisher and up flow) cleaning
1
9/29/2022
86
ND (LOD=l .l)
0.14
(DAK
Americas.
2023)
Excluded17
68
PBZ
Process sample collection
(collected 4 liquid samples) - hot
wells, CP recycle sample, water
sample on reflux system
1
9/28/2022
20
ND (LOD=3 .6)
0.11
(DAK
Americas.
2023)
Excluded17
69
PBZ
Process sample collection
(collected 4 liquid samples) - hot
wells, CP recycle sample, water
sample on reflux system
1
9/29/2022
19
ND (LOD=3 .6)
0.10
(DAK
Americas.
2023)
Excluded17
70
PBZ
Completed routine operations in
L-l: cleaned PP1, PP2, and DRR
glycol immersion vessels, locked
out cutter, preventative
maintenance in L building,
assisted moving chemicals to 4th
floor, unlocked pressure test on
Hx 65, locked out HTM pump 08,
unlocked red toner pump on H-2,
*pump stopped
1
1/31/2023
520
ND (LOD=0.1)
0.091
(DAK
Americas.
2023)
High
Page 456 of 570
-------
Row
#
Type of
Sample
Worker Activity or Sample
Location
Number
of
Samples
Sample
Date
Sample
Time
(min)
1,4-Dioxane
Exposure
Concentration
(mg/m3)
EPA
Determined
8-hour TWA
(mg/m3)"
Source
Overall Data
Quality
Determination
71
PBZ
Completed routine operations in
H-2: cleaned PP1, PP2, and DRR
glycol immersion vessels
(samples 07 and 08), went offsite
for physical- approx. 3 1/2 hrs,
unlocked heat exchanger 39 and
filled w/glycol in H-2, locked out
cutter, PM in L-building, locked
out HTM pumpO 08B (was
leaking HTM) in H-2, unlocked
HX55 for 24hr pressure test
1
1/31/2023
636
ND (LOD=0.8)
0.74
(DAK
Americas.
2023)
High
72
PBZ
Cleaned H-l PP1 Glycol
Immersion Vessel: turned
vent/fan on, opened vessel,
removed/placed half
screen/cleaned basket, rinsed
w/water hose *typically closes
vessel after task (took approx. 5
mins), but left open for sample
purposes *task done every 12
hours
1
1/31/2023
30
ND (LOD=1.7)
0.074
(DAK
Americas,
2023)
Excluded17
73
PBZ
Cleaned H-l PP2 Glycol
Immersion Vessel: turned
vent/fan on, opened vessel,
removed/placed half
screen/cleaned basket, rinsed
w/water hose *typically closes
vessel after task (took approx. 5
mins), but left open for sample
purposes *task done every 12
hours
1
1/31/2023
30
ND (LOD=1.6)
0.070
(DAK
Americas.
2023)
Excluded17
74
PBZ
Cleaned H-2 PP1 Glycol
Immersion Vessel: turned fan on,
opened vessel, removed/placed
1
1/31/2023
30
ND (LOD=1.7)
0.074
(DAK
Americas.
2023)
Excluded17
Page 457 of 570
-------
Row
#
Type of
Sample
Worker Activity or Sample
Location
Number
of
Samples
Sample
Date
Sample
Time
(min)
1,4-Dioxane
Exposure
Concentration
(mg/m3)
EPA
Determined
8-hour TWA
(mg/m3)"
Source
Overall Data
Quality
Determination
screen w/pulley, cleaned basket,
rinsed w/water hose. Took
approx. 5 min *internal vent left
off during sample by accident
*task done every 12 hours
75
PBZ
Cleaned H-2 PP2 Glycol
Immersion Vessel: turned fan on,
opened vessel, removed/placed
screen w/pulley, cleaned basket,
rinsed w/water hose. Took
approx. 5 min *internal vent left
off during sample by accident
1
1/31/2023
30
ND (LOD=1.7)
0.074
(DAK
Americas,
2023)
Excluded17
76
PBZ
Completed routine operations in
H-2: cleaned PP1, PP2, and DRR
Glycol Immersion Vessels
(sample 06); collected/ran
samples of PP1 and DRR and
composite samples, 6073 ring
main samples, pressure
test/LOTO heat exchangers,
locked out cutters, cleaned 25
column bottom pump,
housekeeping
1
2/7/2023
673
ND (LOD=0.7)
0.74
(DAK
Americas,
2023)
High
77
PBZ
Cleaned L Building PP1 glycol
immersion vessel: Turned
vent/fan on, opened vessel,
removed/placed screen
w/mechanic pulley, cleaned
basket w/scraper/hand tools and
rinsed w/water hose. * Typically
closes vessel after task (15 mins)
but left open for sample purposes,
task completed every 12 hours
1
2/7/2023
30
ND (LOD=1.7)
0.074
(DAK
Americas,
2023)
Excluded17
Page 458 of 570
-------
Row
#
Type of
Sample
Worker Activity or Sample
Location
Number
of
Samples
Sample
Date
Sample
Time
(min)
1,4-Dioxane
Exposure
Concentration
(mg/m3)
EPA
Determined
8-hour TWA
(mg/m3)"
Source
Overall Data
Quality
Determination
78
PBZ
Cleaned L Building PP2 glycol
immersion vessel: Turned
vent/fan on, opened vessel,
removed/placed screen
w/mechanic pulley, cleaned
basket w/scraper/hand tools and
rinsed w/water hose. * Typically
closes vessel after task (15 mins)
but left open for sample purposes,
task completed every 12 hours
1
2/7/2023
30
ND (LOD=1.7)
0.074
(DAK
Americas.
2023)
Excluded17
79
PBZ
Cleaned H-2 DRR glycol
immersion vessel: Turned
vent/fan on, opened vessel,
removed/placed screen
w/mechanic pulley, cleaned
basket w/scraper/hand tools and
rinsed w/water hose. * Typically
closes vessel after task (15 mins)
but left open for sample purposes,
task completed every 12 hours
1
2/7/2023
35
ND (LOD=1.4)
0.072
(DAK
Americas.
2023)
Excluded17
80
PBZ
Cleaned H-l DRR glycol
immersion vessel, opened vessel,
removed/placed screen
w/mechanic pulley, cleaned
basket w/scraper/hand tools and
rinsed w/water hose. *typically
closes vessel after task (15 mins)
but left open for sample purposes,
task completed every 12 hours
1
2/7/2023
36
ND (LOD=1.4)
0.072
(DAK
Americas.
2023)
Excluded17
81
PBZ
Completed routine operations in
L building: clean PP1, PP2, and
DRR (samples 04 and 05);
collected/ran sample in L
building, collected composite
1
2/7/2023
562
ND (LOD=0.9)
0.74
(DAK
Americas,
2023)
High
Page 459 of 570
-------
Row
#
Type of
Sample
Worker Activity or Sample
Location
Number
of
Samples
Sample
Date
Sample
Time
(min)
1,4-Dioxane
Exposure
Concentration
(mg/m3)
EPA
Determined
8-hour TWA
(mg/m3)"
Source
Overall Data
Quality
Determination
samples in L building, unlocked
59 pump, housekeeping, unlocked
HDM pump, locked/brought up
cutters, swapped/brought up
compressors and fans, collected
field readings
82
PBZ
Final Polymer Immersion Vessel
- Operator cleaning vessel
1
2/9/2018
40
ND (LOD=0.7)
0.042
(DAK
Americas.
Excluded17
2023)
83
PBZ
Cleaning PP2 and Final Polymer
Immersion Vessels
1
2/23/2018
70
ND (LOD=0.4)
0.037
(DAK
Americas,
Excluded17
2023)
84
PBZ
Cleaning PP1 and PP2 Immersion
Vessels
1
3/2/2018
70
1.1
0.16
(DAK
Americas.
Excluded17
2023)
85
PBZ
Cleaning PP2 and Final
Immersion Vessels
1
3/9/2018
92
ND (LOD=0.3)
0.039
(DAK
Americas.
Excluded17
2023)
86
PBZ
Cleaning PP1 , PP2, and Final
Immersion Vessels
1
3/30/2018
70
ND (LOD=0.4)
0.037
(DAK
Americas,
Excluded17
2023)
87
PBZ
Resin CP
1
2/26/2013
560
ND (LOD=0.3)
0.24
(DAK
Americas,
High
2023)
88
PBZ
Resin CP - Hot wells, Patrols,
Filter Change - Post Finisher
1
7/9/2013
600
ND (LOD=0.3)
0.22
(DAK
Americas.
High
2023)
89
PBZ -
ONU
CP 4th Floor Patrol
1
5/23/2006
390
ND (LOD=0.3)
0.19
(DAK
Americas.
High
2023)
Page 460 of 570
-------
Row
#
Type of
Sample
Worker Activity or Sample
Location
Number
of
Samples
Sample
Date
Sample
Time
(min)
1,4-Dioxane
Exposure
Concentration
(mg/m3)
EPA
Determined
8-hour TWA
(mg/m3)"
Source
Overall Data
Quality
Determination
90
PBZ -
ONU
CP 4th Floor Patrol
1
5/24/2006
420
ND (LOD=0.3)
0.19
(DAK
Americas.
High
2023)
91
PBZ -
ONU
CP Field Patrol 4th Floor
1
5/25/2006
405
ND (LOD=0.3)
0.19
(DAK
Americas.
High
2023)
92
PBZ -
ONU
CP Field Patrol 4th Floor
1
5/31/2006
405
ND (LOD=0.3)
0.19
(DAK
Americas,
High
2023)
93
PBZ -
ONU
CP Field Patrol 4th Floor
1
6/1/2006
405
ND (LOD=0.3)
0.19
(DAK
Americas,
High
2023)
94
PBZ -
ONU
CP Field Patrol
1
6/6/2006
405
ND (LOD=0.3)
0.19
(DAK
Americas.
High
2023)
95
PBZ -
ONU
Routine Patrols
1
2/26/2013
555
ND (LOD=0.3)
0.24
(DAK
Americas.
High
2023)
96
PBZ -
ONU
Staple CP - Routine Patrols
1
7/9/2013
600
ND (LOD=0.3)
0.22
(DAK
Americas,
High
2023)
97
PBZ -
ONU
Routine Patrols
1
7/10/2013
580
ND (LOD=0.3)
0.22
(DAK
Americas.
High
2023)
98
PBZ -
ONU
Routine Patrols
1
7/10/2013
580
ND (LOD=0.3)
0.22
(DAK
Americas.
High
2023)
99
Area-
ONU
Area sample near vacuum pump
for Tank 19 during loading of
tanker trucks
1
6/11/2019
467
ND (LOD=0.31)
0.21
(Huntsman,
High
2023)
Page 461 of 570
-------
Row
#
Type of
Sample
Worker Activity or Sample
Location
Number
of
Samples
Sample
Date
Sample
Time
(min)
1,4-Dioxane
Exposure
Concentration
(mg/m3)
EPA
Determined
8-hour TWA
(mg/m3)"
Source
Overall Data
Quality
Determination
100
Area-
ONU
Area sample near pump during
routine maintenance tasks, 3rd
floor tower
1
4/22/2019
356
ND (LOD=0.40)
0.21
("Huntsman,
2023)
High
101
Area-
ONU
Area sample near Tank 19 during
routine duties around tank
1
4/22/2019
368
ND (LOD=0.40)
0.21
(Huntsman,
2023)
High
102
Area-
ONU
Area sample near pump during
routine maintenance tasks, 3rd
floor tower
1
4/22/2019
353
ND (LOD=0.40)
0.21
(Huntsman.
2023)
High
PBZ = Personal breathing zone; ND = non-detect for 1,4-dioxane; LOD = limit of detection; TWA = time-weighted average
a The 8-hour TWA calculations use the LOD/2 or the LOD/V2 for non-detect values, depending on the geometric standard deviation of the dataset.
b As explained prior to this table, these data points were excluded from the analysis of central tendency and high-end worker exposures because all PBZ, area,
and bulk sampling at this site was non-detect for 1,4-dioxane; therefore, it is questionable if the site handles 1,4-dioxane.
c As discussed prior to this table, EPA excluded these samples due to the short sample durations.
Page 462 of 570
-------
Key Uncertainties
The OSHA CEHD monitoring data does not include process information or worker activities; therefore,
there is uncertainty as to which worker activities these data cover and whether all potential workers
activities are represented in this data. Additionally, these data are from five facilities, and it is unclear
how representative the data are for all sites and all workers across the United States. The OSHA CEHD
used for this assessment is from the 1980s and 1990s. Therefore, the age of the monitoring data may also
introduce uncertainty. EPA used half the detection limit for the non-detect values in the central tendency
and high-end exposure calculations. This introduces uncertainty into the assessment because the true
value of 1,4-dioxane is unknown (although expected to be between zero and the level of detection).
Additional uncertainties are listed in Section 3.1.2.4.
F.4.10 Ethoxylation Process Byproduct
Process Description
1,4-Dioxane may be formed as a byproduct of ethoxylation reactions and sulfonation processes during
manufacture of ingredients that are then used for a variety of applications, such as personal care
products, cleaning products, coatings, and certain pharmaceuticals (HHS. 2016) (Dow Chemical. 2023).
Polyethoxylated raw materials are widely used in cosmetic products as emulsifiers, foaming agents, and
dispersants (Black et at.. 2001). They are produced by polymerizing ethylene oxide, usually with a fatty
alcohol, to form polyethoxylated alcohols which may be used to synthesize other products such as
sulfated surface-active agent. During the ethoxylation process, 1,4-dioxane can be formed as a
byproduct by the dimerization of ethylene oxide (Black et at.. 2001).
It should be noted that there are existing technologies in operation which may mitigate the formation of
1,4-dioxane, including carefully controlling reactant ratios and the rapid neutralization of sulfonation
products with sodium hydroxide to prevent 1,4-dioxane formation in the ethoxylation process (HCPA.
2023). Additionally, there may be post-processing steps can remove any 1,4-dioxane that is formed from
ethoxylation (ACC. 2023). These steps include vacuum or steam stripping.
In cosmetic ethoxylated raw materials and ethoxylated alkyl sulfates, 1,4-dioxane has been detected at
concentrations of 0.48 to 1,410 ppm (U.S. EPA. 2020c; Saraii and Shirvani. 2017; Davarani et at.. 2012;
Black et at.. 2001). Information submitted through public comments indicates that 1,4-dioxane may be
produced as a byproduct at 1 to 30 ppm during the ethoxylation process inside reactors (Dow Chemical.
2023). The surfactants with residual 1,4-dioxane are then pumped into containers such as rail cars and
sent to downstream formulators. Subsequently, releases to onsite wastewater treatment may occur when
process equipment and rail cars are rinsed (Dow Chemical. 2023).
The volume of 1,4-dioxane produced as a byproduct of ethoxylation reactions is unknown. Due to lack
of information, EPA does not present annual or daily site throughputs. EPA assumes facilities that
produce 1,4-dioxane as a byproduct during ethoxylation reactions operate 5 days/week, 50 weeks/year,
or 250 days/year.
Worker Activities
Workers are potentially exposed to 1,4-dioxane during ethoxylation processes through loading
ethoxylation products into transport containers, taking quality control samples, and equipment cleaning
(Dow Chemical. 2023). All of these activities are potential sources of worker exposure through dermal
contact to liquid and inhalation of volatile chemical vapors.
Suitable PPE may be worn in accordance with safety data sheets. In addition to engineering and
administrative controls, workers may use PPE such as hard hats, safety glasses or goggles, chemical
Page 463 of 570
-------
resistant gloves, and chemical resistant suits, depending on the worker task (Dow Chemical. 2023). EPA
did not find information that indicates the extent that and worker PPE is used at facilities that conduct
ethoxylation processes in the United States.
ONUs include employees that work at the sites where ethoxylation processes occur, but they do not
directly handle the chemicals and are therefore expected to have lower inhalation exposures and are not
expected to have dermal exposures through contact with liquids. ONUs for this scenario include
supervisors, managers, and other employees that may be in the process areas but do not perform tasks
that result in the same level of exposures as those workers that engage in tasks related to ethoxylation.
Number of Potentially Exposed Workers and ONUs
To estimate the number of workers, EPA used U.S. Census and BLS data for the following NAICS
codes: 325110, 325199, 325611, 325613, and 325998. EPA estimated a total of 2,730 sites, 64,926
workers, and 24,835 ONUs (\] S HI S. 2016). For additional information on the steps used to estimate
the number of potentially exposed workers and ONUs, refer to Appendix G.5 of the 2020 Risk
Evaluation for 1,4-Dioxane ( 1020c).
Worker Inhalation Exposure Assessment
1,4-Dioxane may be formed as a byproduct of reactions based on condensing ethylene oxide or ethylene
glycol during manufacture of detergents, shampoos, surfactants, some food additives, and certain
pharmaceuticals (HHS. 2016). Occupational exposure to 1,4-dioxane in ethoxylation process byproduct
was determined using monitoring data provided by Chemical Exposure Health Data (OJ 320). The
information and data quality evaluation to assess occupational exposures during the ethoxylation process
is listed in TableApx F-28 and described below.
Table Apx F-28. Ethoxylation Process
byproduct Worker Exposure Dal
ta Evaluation
Worker Activity
or Sampling
Location
Data Type
Number of
Samples
Overall Data
Quality
Determination
Source Reference
Unknown
PBZ Monitoring
1
High
( HA.. 2020)
See Table Apx
F-30
PBZ Monitoring
8
High
(Dow Chemical. 2023)
EPA assessed occupational inhalation exposures for this OES using OSHA's CEHD (OSHA. 2020) and
data provided in a public comment (Dow Chemical. 2023). Table Apx F-30 shows the one 8-hour TWA
from CEHD and the eight data points from the public comment.
For detailed information on where/how CEHD was obtained and mapped to OES, see Appendix F.4.1.
For this OES, monitoring data were available in CEHD from one site with SIC code 2841 (Soap and
Other Detergent Manufacturing). This site was determined to be a detergent manufacturer.
The one CEHD 8-hour TWA is comprised of multiple short-term samples with the same inspection
number and sampling number, as shown in Table Apx F-30. EPA's rationale and process for combining
samples with the same inspection and sampling numbers is described in Appendix F.4.1. The combined
sample duration was 381 minutes, which is close to the full-shift duration of 8 hours (480 minutes). EPA
translated this into 8-hour TWA concentration by assuming that the exposure concentration is zero for
the time remaining in the 8-hour shift.
Page 464 of 570
-------
A public comment provides eight 8-hour TWA personal breathing zone sample points taken in 2017-
2021 (Dow Chemical. 2023). Samples were taken for workers performing unloading and laboratory
activities. All data were non-detect for 1,4-dioxane; therefore, EPA assessed exposures as the LOD/V2
because the geometric standard deviation of the dataset is less than three ( 94a).
EPA then used the 8-hour TWAs as shown in TableApx F-30 to calculate full shift (8-hour TWA)
central tendency (50th percentile) and high-end (95th percentile) inhalation exposures for workers. EPA
used these values to calculate the ADC and LADC. The calculated values are summarized in Table Apx
F-29. Equations for calculating ADC and LADC are presented in Appendix G of the December 2020
Final Risk Evaluation for 1,4-Dioxane ( )c).
The ONU exposures are anticipated to be lower than worker exposures since ONUs do not typically
directly handle the chemical. Only inhalation exposures to vapors or incidental dermal exposures may be
expected to ONUs.
Table Apx F-29. Inhalation Exposures of Workers for the Ethoxylation Process
Byproduct Based on Monitoring Data
Exposure Type
Central Tendency
(50th Percentile) (mg/m3)"
High-End
(95th Percentile) (mg/m3)"
Draft \U:. estimates6
8-hour TWA Exposure
Concentrations
1.2 (single value)
Average Daily Concentration
(ADC)
1.15 (single value)
Lifetime Average Daily
Concentration (LADC)
0.459 (single value)
I pdatcd estimates
8-hour TWA Exposure
Concentrations
0.56
1.1
Average Daily Concentration
(ADC)
0.54
1.1
Lifetime Average Daily
Concentration (LADC)
0.21
0.54
a See Table_Apx F-28 for corresponding references.
b For select OESs, updates to exposure estimates were made via information provided by the SACC and public
comments.
Page 465 of 570
-------
Table Apx F<
50. Occupational Inhalation IV
onitoring Data for Ethoxylation Process
Jyproduct
Row
#
Type of
Sample
Worker Activity or
Sample Location
Number
of
Samples
Sample
Date
Sample
Time (min)
1,4-Dioxane
Exposure
Concentration
(mg/mJ)
EPA
Determined
8-hour TWA
(mg/m3)"
Source
Overall Data
Quality
Determination
1
PBZ
Unknown
1
6/16/2000
84
0.76
0.76 (8-hour
(OSUA. 2020)
High
2
PBZ
Unknown
1
6/16/2000
50
0.81
(OSUA. 2020)
High
3
PBZ
Unknown
1
6/16/2000
32
ND (LOD = 2.3)
(OSUA. 2020)
High
4
PBZ
Unknown
1
6/16/2000
65
ND (LOD = 1.1)
TWA for the
same worker
from rows 1-7)
(OSHA. 2020)
High
5
PBZ
Unknown
1
6/16/2000
60
ND (LOD = 1.2)
(OSHA. 2020)
High
6
PBZ
Unknown
1
6/16/2000
60
ND (LOD = 1.2)
(OSHA. 2020)
High
7
PBZ
Unknown
1
6/16/2000
30
ND (LOD = 2.4)
(OSHA. 2020)
High
8
PBZ
SCO EXP Laboratory
Technician
1
9/15/2017
Unknown -
8-hour TWA
ND (LOD = 0.79)
0.56
(Dow
Chemical
2023)
High
9
PBZ
SCO EXP
Loader/Unloader Rail
Car
1
9/15/2017
Unknown -
8-hour TWA
ND (LOD = 1.5)
1.1
(Dow
Chemical
2023)
High
10
PBZ
SCO EXP Laboratory
Technician
1
9/16/2017
Unknown -
8-hour TWA
ND (LOD = 0.76)
0.54
(Dow
Chemical
2023)
High
11
PBZ
SCO EXP Laboratory
Technician
1
9/17/2017
Unknown -
8-hour TWA
ND (LOD = 0.79)
0.56
(Dow
Chemical
2023)
High
12
PBZ
SCO EXP Laboratory
Technician
1
6/11/2018
Unknown -
8-hour TWA
ND (LOD = 0.79)
0.56
(Dow
Chemical
2023)
High
13
PBZ
SCO EXP
Loader/Unloader Rail
Car
1
6/12/2018
Unknown -
8-hour TWA
ND (LOD = 1.6)
1.1
(Dow
Chemical
2023)
High
14
PBZ
SCO EXP Laboratory
Technician
1
6/24/2021
Unknown -
8-hour TWA
ND (LOD = 0.18)
0.13
(Dow
Chemical
2023)
High
Page 466 of 570
-------
Row
#
Type of
Sample
Worker Activity or
Sample Location
Number
of
Samples
Sample
Date
Sample
Time (min)
1,4-Dioxane
Exposure
Concentration
(mg/mJ)
EPA
Determined
8-hour TWA
(mg/m3)"
Source
Overall Data
Quality
Determination
15
PBZ
SCO EXP
Maintenance
Loader/Unloader Rail
Car
1
6/24/2021
Unknown -
8-hour TWA
ND (LOD = 0.18)
0.13
(Dow
Chemical.
2023)
High
PBZ = Personal breathing zone; ND = non-detect for 1,4-dioxane; LOD = limit of detection; TWA = time-weighted average
a The 8-hour TWA calculations use the LOD/2 or the LOD/V2 for non-detect values, depending on the geometric standard deviation of the dataset.
Page 467 of 570
-------
Key Uncertainties
The OSHA CEHD monitoring data does not include process information or worker activities; therefore,
there is uncertainty as to which worker activities these data cover and whether all potential workers
activities are represented in this data. Additionally, the OSHA CEHD only include one 8-hour TWA
from one facility. Therefore, EPA cannot determine the statistical representativeness of this data point
(e.g., high-end, central tendency) towards potential exposures from this COU. Further, it is unclear how
representative the data are for all sites and all workers across the United States. The OSHA CEHD point
used for this assessment is from the year 2000. Therefore, the age of the monitoring data can also
introduce uncertainty. Additional uncertainties are listed in Section 3.1.2.4.
F.4.11 Hydraulic Fracturing
Process Description
Facilities have self-reported to FracFocus 3.0 that 1,4-dioxane is present in hydraulic fracturing fluid
additives, as scale inhibitors, additives, biocides, friction reducers, and surfactants (GWPC and lOGCC.
2022). EPA also expects that 1,4-dioxane is present as an unintentional component in hydraulic
fracturing fluids, due to its presence as a byproduct in ethoxylated substances. According to the
FracFocus 3.0 database, 1,4 dioxane is present in weight fractions ranging from 2,3/10 " to 0.05 within
hydraulic fracturing additives and 1.00xl0~12 to 4.30xl0~6 in hydraulic fracturing fluids (GWPC and
IOGCC. 20221
Hydraulic fracturing stimulates an existing oil or gas well by injecting a pressurized fluid containing
chemical additives into the well ( 2022e). Hydraulic fracturing differs from conventional
drilling, which involves the use of a mechanical drilling rig to drill vertically down. Hydraulic fracturing
is often used where conventional drilling cannot reach because hydraulic fracturing can be done both
vertically and horizontally, allowing for greater access to oil- and natural gas-bearing rock.
EPA did not find specific container information for 1,4-dioxane in hydraulic fracturing; however, the
Revised ESD on Hydraulic Fracturing indicates that hydraulic fracturing fluids are typically transported
as a liquid in totes, drums, or bulk containers. Hydraulic fracturing fluid formulations are charged to a
temporary storage tank, or they may be charged to a mixing tank with other additives to formulate the
final fracturing fluid that is injected into the well ( 022e).
Multiple types of wastewaters are created by hydraulic fracturing: flowback water, produced water, and
naturally occurring wastewater. The ESD indicates that 100 percent of chemical additives such as 1,4-
dioxane are released during the hydraulic fracturing process, with a portion entrapped in the shale
formation and the remaining returning to the surface in the various types of wastewaters, as described
below ( I022e).
After formulation, the hydraulic fracturing fluid is pumped into a wellbore where it cracks and
permeates the surrounding rock ( 2022e). A portion of the fracturing fluid, including any
chemical additives such as 1,4-dioxane, may remain in the underground shale formation. The remaining
fluid will return to the surface as flowback water that flows back to the surface from the well. Flowback
water is the first wastewater to return to the surface after hydraulic fracturing ( )22e).
Naturally Occurring Water: exists in the rock formation prior to hydraulic fracturing. Initially flowback
water is mostly fracturing fluid, which includes 1,4-dioxane. However, over time, it becomes primarily
composed of naturally existing water from the rock formation ( 22e).
Page 468 of 570
-------
Produced Water: is defined as "water trapped in underground formations that is brought to the surface
along with oil or gas" (U.S. EPA. 2022e). Produced water returns to the surface of the well after
flowback water. Produced water may contain many constituents, water and compounds from the rock
formation, oil or gas from the rock formation, and smaller portions of hydraulic fracturing fluid
including 1,4-dioxane (U.S. EPA. 2022e).
Wastewater containing chemical additives such as 1,4-dioxane is stored and accumulated at the surface
for eventual reuse or disposal ( 22e). Typical storage facilities include open air
impoundments and closed containers. This wastewater is collected and may be taken to disposal wells,
recyclers, wastewater treatment plants (on- or off-site), or in some cases the water may be left in pits to
evaporate or infiltrate or be used for irrigation or road treatment ( I022e).
FracFocus 3.0 reports 411 sites that utilize 1,4-dioxane in hydraulic fracturing fluid. These sites are
located throughout the United States (GWPC an X. 2022). FracFocus 3.0 also reports that a
typical number of operating days per year is 1 to 72 days/year (GWPC and IOGCC. 2022). EPA
modeled the 1,4-dioxane use rate for a generic site using data from FracFocus 3.0 and the Revised ESD
on Hydraulic Fracturing to estimate releases, resulting in a 50th and 95th percentile 1,4-dioxane use rate
of 0.3 and 5.18 kg/site-day, respectively. A flow diagram including release and exposure points from the
Draft ESD on Hydraulic Fracturing is presented in FigureApx F-3 ( I022e).
For additional information on the modeling and associated input parameters used to estimate the daily
use rate, refer to Appendix E.13.
Page 469 of 570
-------
Chemical
Additives
Fracturing Fluid Mixing
Temporary
Storage
Tank
© s
®®^>
0®<3>
Container
Unloading
Container
Cleaning
Equipment
Cleaning
On* or off-site
treatment and
discharge to
surface water
Deep well
injection
Recyc led reused
to oil well
Land (evaporation
ponds, percolation
ponds, irrigation)
®
_ OilGas to
Refinery
= Occupational Exposures:
A. Container unloading
B. Equipment cleaning
C. Container cleaning
O = Environmental Releases:
1. Unloading volatile chemicals
2. Container residuals
3. Container cleaning
4. Equipment cleaning residuals
5. Equipment cleaning
6. Release to surface water, land (soil), landfill, or incineration from spills
7. Release to deep well injection from fracturing fluid that remains underground
8. Flowback and produced wastewater release
FigureApx F-3. Environmental Release and Occupational Exposure Points During Hydraulic
Fracturing
Worker Activities
Workers are potentially exposed to 1,4-dioxane during multiple activities involved in hydraulic
fracturing operations, including container unloading and transferring, container cleaning, and equipment
cleaning (U.S. EPA. 2022e). These activities are all potential sources of worker exposure through
dermal contact to liquid and inhalation of volatile chemical vapors and are included in the exposure
modeling described in Appendix F.9. Depending on how sites manage flowback and produced
wastewater, workers may also potentially be exposed to chemical additives such as 1,4-dioxane in this
wastewater during handling or treatment. However, this exposure point is not included in the ESD on the
Use of Chemicals in Hydraulic Fracturing, so is not included in the modeling in Appendix F.9.
The ESD on the Use of Chemicals in Hydraulic Fracturing indicates that workers may connect transfer
lines to pump chemical additives directly from transport containers, or manually unload chemicals from
transport containers into mixing tanks or injection system (U.S. EPA 2022e). Dermal exposure may
occur during both automated and manual unloading activities. Container cleaning and equipment
cleaning are typically manual activities. (U.S. EPA 2022e).
The ESD on Chemicals Used in Hydraulic Fracturing suggests that workers consult the Safety Data
Sheet (SDS) which may identify specific hazards and recommend the appropriate personal protective
Page 470 of 570
-------
equipment (PPE) (U.S. EPA. 2022e). EPA did not find information that indicates the extent that
engineering controls and worker PPE are used at facilities that use in the United States.
ONUs include employees that work at the sites where hydraulic fracturing chemicals are used, but they
do not directly handle the chemicals and are therefore expected to have lower inhalation exposures and
are not expected to have dermal exposures through contact with liquids. ONUs for this scenario include
supervisors, managers, and other employees that may be in the oil/gas well area but do not perform tasks
that result in the same level of exposures as those workers that engage in tasks related to the use of
fracturing chemicals.
Number of Potentially Exposed Workers and ONUs
Use of hydraulic fracturing chemicals are expected to fall within NAICS codes 213111, Drilling Oil and
Gas Wells, and 213112, Support Activities for Oil and Gas Operations. EPA estimated a total of 14,193
sites, 46,315 workers, and 26,007 ONUs (\] S HI S. 2016). The number of sites conducting hydraulic
fracturing using 1,4-dioxane is provided by FracFocus 3.0 data, with a total of 41 1 sites (GWPC and
IOGCC. 2022). For additional information on the steps used to estimate the number of potentially
exposed workers and ONUs, refer to Appendix G.5 of the 2020 Risk Evaluation for 1,4-Dioxane (U.S.
20c).
Worker Inhalation Exposure Assessment
Facilities have self-reported to FracFocus 3.0 that 1,4-dioxane is present in hydraulic fracturing fluid
additives, such as scale inhibitors, additives, biocides, friction reducers, and surfactants (GWPC and
IOGCC. 2022). The information and data quality evaluation to assess occupational exposures during
hydraulic fracturing is listed in Table Apx F-31 and described below.
Table Apx F-31. Hydraulic Fracturing Worker Exposure Data Evaluation
Worker Activity or Sampling
Location
Data Type
Number of
Samples
Overall Data
Quality
Determination
Source
Reference
Unloading hydraulic fracturing
fluid additives, cleaning empty
additive containers, equipment
cleaning
Input parameters for
Monte Carlo modeling
N/A
Mediuma
(U.S. EPA.
2022e)
N/A
Mediuma
(GWPC and
CC.
2022)
a This is the rating for the underlying data used in the model, and not the Monte Carlo model itself.
EPA did not find relevant inhalation monitoring data for the use of hydraulic fracturing fluids.
Therefore, the Agency modeled 1,4-dioxane air concentrations using a Monte Carlo modeling approach,
which is described in Appendix 0. This modeling approach utilizes the EPA AP-42 Loading Model,
EPA/OPPT Mass Transfer Coefficient Model, and EPA Mass Balance Inhalation Model, with variation
in input parameters for mass fraction of 1,4-dioxane in hydraulic fracturing additive and fluid, saturation
factor, container size, use rate of fracturing fluid, ventilation rate, and mixing factor based on available
data. During modeling, EPA noted that if the durations for all individual hydraulic fracturing activities
were summed, the total exposure time can exceed a full shift duration of eight hours. To avoid this, the
time spent unloading containers and cleaning containers was capped at two hours each, since the other
activity for equipment cleaning occurs over four hours {i.e., 2 hours for container unloading + 2 hours
for container cleaning + 4 hours for equipment cleaning = 8 hours). This is a limitation of the assessment
Page 471 of 570
-------
because EPA is unsure the extent to which the assessed activity durations are representative of real-
world conditions. A summary of the modeled exposures is presented in TableApx F-32.
TableApx F-32. Modeled Occupational Inhalation Exposures
for Hydraulic Fracturing
Statistic
1,4-Dioxane Exposure Concentration, 8-Hour
TWA (mg/m3)
Maximum
298
99th Percentile
6.9
95th Percentile
1.8
50th Percentile
9.1E-02
5th Percentile
3.7E-03
Minimum
6.6E-09
Mean
0.50
EPA used the 50th and 95th percentile modeled 8-hour TWA exposures values presented in Table Apx
F-33 to calculate the central tendency and high-end ADC and LADC, respectively. The calculated
values are summarized in Table Apx F-32. Equations for calculating ACD and LADC are presented in
Appendix G of the December 2020 Final Risk Evaluation for 1,4-Dioxane 0 1 S 1 T \ 2020c).
Exposure data for ONUs were not available. The ONU exposures are anticipated to be lower than
worker exposures since ONUs do not typically directly handle the chemical. Only inhalation exposures
to vapors or incidental dermal exposures may be expected to ONUs.
Table Apx F-33. Inhalation Exposures of Wor
ters for Hydraulic Fracturin
g Based on Modeling
Exposure Type
Central Tendency
(50th Percentile) (mg/m3)"
High-End
(95th Percentile)
(mg/m3)"
Draft Rl- estimates
8-hour TWA Exposure Concentrations
2.87
66.8
Average Daily Concentration (ADC)
0.177
18.5
Lifetime A\ crime Daily Concentration (T.ADC )
I pchile
o 070
d cslimules
0 40
8-hour TWA Exposure Concentrations
9.1E-02
1.8
Average Daily Concentration (ADC)
5.6E-03
0.49
Lifetime Average Daily Concentration (LADC)
2.2E-03
0.25
a See Table_Apx F-31 for corresponding references.
b For select OESs, updates to exposure estimates were made via information provided by the SACC and public
comments.
Key Uncertainties
EPA used assumptions and values from the ESD on Hydraulic Fracturing and various EPA models to
estimate inhalation exposures during container transfers, container cleaning, and equipment cleaning
Page 472 of 570
-------
within the hydraulic fracturing COU (see Appendix F.9). The uncertainties associated with this
modeling approach are described in Section 3.1.2.4.
EPA also used data from FracFocus 3.0 ("GWPC and IOGCC. 2022) to inform input parameters for the
exposure calculations. FracFocus contains self-reported data; therefore, the extent to which these data
represent operations across multiple sites throughout the United States is unclear.
F.5 Summary of Occupational Inhalation Exposures
A summary of the inhalation exposure estimates previously discussed is included in Table Apx F-34.
The table presents high-end and central tendency inhalation exposures by condition of use. The table
also indicates whether the source data are monitoring values or modeled estimates. For more details on
how each inhalation exposure was estimated, see Appendix F.4.
Page 473 of 570
-------
Table Apx F-34. Estimated Inhalation Exposure (mg/m3) for Workers During Various Conditions of Use
OES
Category
Exposure
ime-t'rame
Exposure
Frequency
(day/vear)
8-hour TWA
Exposures
Chronic, Non-cancer
Exposures
Chronic, Cancer
Exposures
8-Hour
Data
Points
Total
Samples
Sources & Notes
Data Type
Cs-hTWA (mj»/mJ)
ADCs-n iwa (mg/m3)
LADCs-h iwa (mg/m3)
HE
CT
HE
CT
HE
CT
HE
CT
Textile dye
(draft RE
estimates)"
Worker
8-hour
250
31
74
6.6E-02
84
7.9E-03
43
3.1E-03
14
51
OSHA CEHD from
1991-2010
(OSHA. 2020). 51
PBZ samples, from
which 14 8-h
TWAs were
derived.
Monitoring
Data
Textile dye
(updated
estimates)"
Worker
8-hour
250
157
15
0.81
14
0.49
7.4
0.19
14
51
OSHA CEHD from
1991-2010
(OSHA. 2020). 51
PBZ samples, from
which 14 8-h
TWAs were
derived.
Monitoring
Data
Antifreeze
(draft RE
estimates)"
Worker
8-hour
250
250
1.1E-07
2.2E-08
1.1E-07
2.1E-08
5.4E-08
8.3E-09
N/A
N/A
Monte Carlo
Simulation results
Monte Carlo
Modeling
Antifreeze
(updated
estimates)"
Worker
8-hour
250
250
9.8E-07
1.3E-07
9.4E-07
1.2E-07
4.8E-07
4.8E-08
N/A
N/A
Monte Carlo
Simulation results
Monte Carlo
Modeling
Surface
cleaner (draft
RE
estimates)"
Worker
8-hour
250
250
3.7E-03
2.9E-04
3.6E-03
2.8E-04
1.8E-03
1.1E-04
49
49
(Barley et al.
Monitoring
Data
2021)
Surface
cleaner
(updated
estimates)"
Worker
8-hour
250
250
7.4E-03
5.7E-04
7.1E-03
5.5E-04
3.7E-03
2.2E-04
49
49
(Bailey et al.
Monitoring
Data
2021)
Dish Soap
(draft RE
estimates)"
Worker
8-hour
250
250
2.1
1.0
2.0
1.0
1.0
4.0E-01
29
29
(Belanger et al.,
1980)
Monitoring
Data
Dish Soap
(updated
estimates)"
Worker
8-hour
250
250
1.0E-02
1.1E-03
1.0E-02
1.1E-03
5.1E-03
4.4E-04
N/A
N/A
Monte Carlo
Simulation
results
Monte Carlo
Modeling
Page 474 of 570
-------
OES
Category
Exposure
ime-frame
Exposure
Frequency
(day/vear)
8-hour TWA
Exposures
Chronic, Non-cancer
Exposures
Chronic, Cancer
Exposures
8-Hour
Data
Points
Total
Samples
Sources & Notes
Data Type
Cs-h i wa (mjj/mJ)
ADCs-n iwa (mjj/mJ)
LADCs-n iwa (mjj/mJ)
HE
CT
HE
CT
HE
CT
HE
CT
Dishwasher
detergent
(draft RE
estimates)"
Worker
8-hour
250
250
2.1
1.0
2.0
1.0
1.0
4.0E-01
29
29
(Belanger et al.,
1980)
Monitoring
Data
Dishwasher
detergent
(updated
estimates)"
Worker
8-hour
250
250
4.5E-03
5.9E-04
4.3E-03
5.7E-04
2.2E-03
2.3E-04
N/A
N/A
Monte Carlo
Simulation
results
Monte Carlo
Modeling
Laundry
detergent
(industrial)
(draft RE
estimates)"
Worker
(vapor)
8-hour
250
250
1.9E-03
5.2E-04
1.8E-03
5.0E-04
9.2E-04
2.0E-04
N/A
N/A
Monte Carlo
Simulation results
Monte Carlo
Modeling
Laundry
detergent
(industrial)
(updated
estimates)"
Worker
(vapor)
8-hour
250
250
2.1E-02
8.6E-04
2.0E-02
8.3E-04
1.0E-02
3.3E-04
N/A
N/A
Monte Carlo
Simulation results
Monte Carlo
Modeling
Laundry
detergent
(industrial)
(draft RE
estimates)"
Worker
(Total
Particulates)
8-hour
250
250
2.0E-04
1.1E-04
1.9E-04
1.0E-04
9.8E-05
4.0E-05
N/A
N/A
Monte Carlo
Simulation results
Monte Carlo
Modeling
Laundry
detergent
(industrial)
(updated
estimates)"
Worker
(Total
Particulates)
8-hour
250
250
1.4E-03
5.6E-05
1.4E-03
5.4E-05
7.0E-04
2.2E-05
N/A
N/A
Monte Carlo
Simulation results
Monte Carlo
Modeling
Laundry
detergent
(industrial)
(draft RE
estimates)"
Worker
(Respirable
Particulates)
8-hour
250
250
6.7E-05
3.5E-05
6.4E-05
3.4E-05
3.3E-05
1.3E-05
N/A
N/A
Monte Carlo
Simulation results
Monte Carlo
Modeling
Page 475 of 570
-------
OES
Category
Exposure
ime-frame
Exposure
Frequency
(day/vear)
8-hour TWA
Exposures
Chronic, Non-cancer
Exposures
Chronic, Cancer
Exposures
8-Hour
Data
Points
Total
Samples
Sources & Notes
Data Type
Cs-h i wa (mjj/mJ)
ADCs-n iwa (mjj/mJ)
LADCs-n iwa (mjj/mJ)
HE
CT
HE
CT
HE
CT
HE
CT
Laundry
detergent
(industrial)
(updated
estimates)"
Worker
(Respirable
Particulates)
8-hour
250
250
4.0E-04
1.4E-05
3.9E-04
1.4E-05
2.0E-04
5.5E-06
N/A
N/A
Monte Carlo
Simulation results
Monte Carlo
Modeling
Laundry
detergent
(institutional)
(draft RE
estimates)"
Worker
(vapor)
8-hour
250
250
1.4E-03
4.1E-04
1.4E-03
3.9E-04
7.1E-04
1.6E-04
N/A
N/A
Monte Carlo
Simulation results
Monte Carlo
Modeling
Laundry
detergent
(institutional)
(updated
estimates)"
Worker
(vapor)
8-hour
250
250
1.6E-02
6.5E-04
1.5E-02
6.3E-04
7.9E-03
2.5E-04
N/A
N/A
Monte Carlo
Simulation results
Monte Carlo
Modeling
Laundry
detergent
(institutional)
(draft RE
estimates)"
Worker
(Total
Particulates)
8-hour
250
250
2.0E-04
1.1E-04
1.9E-04
1.0E-04
9.8E-05
4.0E-05
N/A
N/A
Monte Carlo
Simulation results
Monte Carlo
Modeling
Laundry
detergent
(institutional)
(updated
estimates)"
Worker
(Total
Particulates)
8-hour
250
250
1.4E-03
5.6E-05
1.4E-03
5.4E-05
7.0E-04
2.2E-05
N/A
N/A
Monte Carlo
Simulation results
Monte Carlo
Modeling
Laundry
detergent
(institutional)
(draft RE
estimates)"
Worker
(Respirable
Particulates)
8-hour
250
250
6.7E-05
3.5E-05
6.4E-05
3.4E-04
3.3E-05
1.3E-05
N/A
N/A
Monte Carlo
Simulation results
Monte Carlo
Modeling
Laundry
detergent
(institutional)
(updated
estimates)"
Worker
(Respirable
Particulates)
8-hour
250
250
4.0E-04
1.4E-05
3.9E-04
1.4E-05
2.0E-04
5.5E-06
N/A
N/A
Monte Carlo
Simulation results
Monte Carlo
Modeling
Page 476 of 570
-------
OES
Category
Exposure
ime-frame
Exposure
Frequency
(day/vear)
8-hour TWA
Exposures
Chronic, Non-cancer
Exposures
Chronic, Cancer
Exposures
8-Hour
Data
Points
Total
Samples
Sources & Notes
Data Type
Cs-h i wa (mjj/mJ)
ADCs-n iwa (mjj/mJ)
LADCs-n iwa (mjj/mJ)
HE
CT
HE
CT
HE
CT
HE
CT
Paint and
floor lacquer
Worker
8-hour
250
250
1.2
0.21
1.2
0.20
0.59
8.0E-02
17
17
("Hills et aL 1989)
Monitoring
Data
Polyethylene
terephthalate
(PET)
byproduct
(draft RE
estimates)"
Worker
8-hour
250
250
47
4.7
45
4.5
23
1.8
11
35
OSHA CEHD from
1985-1994
("OSHA. 2020)
Monitoring
Data
Polyethylene
terephthalate
(PET)
byproduct
(updated
estimates)"
Worker
8-hour
250
250
5.9
0.74
5.7
0.71
2.9
0.28
62
62
O
1'
(<
Pi
(I
2(
2(
SHA CEHD from
585-1994
)SHA. 2020) and
iblic comments
)AK Americas,
)23; Huntsman,
)23)
Monitoring
Data
Polyethylene
terephthalate
(PET)
byproduct
(updated
estimates)"
ONU
8-hour
250
250
0.23
0.21
0.22
0.20
0.11
8.0E-02
14
14
Pi
(I
2(
2(
iblic comments
)AK Americas,
)23; Huntsman,
)23)
Monitoring
Data
Ethoxylation
process
byproduct
(draft RE
estimates)"
Worker
8-hour
250
250
1.2
1.2
1.2
1.2
5.9E-01
4.6E-01
1
7
OSHA CEHD from
1985-1994
("OSHA. 2020)
Monitoring
Data
Ethoxylation
process
byproduct
(updated
estimates)"
Worker
8-hour
250
250
1.1
0.56
1.1
0.54
0.54
0.21
9
15
OSHA CEHD from
1985-1994
("OSHA. 2020) and
public comment
("Dow Chemical.
2023)
Monitoring
Data
Hydraulic
fracturing
(draft RE
estimates)"
Worker
8-hour
72
1
67
2.9
19
1.1E-02
9.5
4.4E-03
N/A
N/A
Monte Carlo
Simulation results
Monte Carlo
Modeling
Page 477 of 570
-------
OES
Category
Exposure
ime-frame
Exposure
Frequency
(day/vear)
8-hour TWA
Exposures
Chronic, Non-cancer
Exposures
Chronic, Cancer
Exposures
8-Hour
Data
Points
Total
Samples
Sources & Notes
Data Type
Cs-h i wa (mjj/mJ)
ADCs-n iwa (mjj/mJ)
LADCs-n iwa (mjj/mJ)
HE
CT
HE
CT
HE
CT
HE
CT
Hydraulic
fracturing
(updated
estimates)"
Worker
8-hour
72
1
1.78
9.1E-02
0.49
5.6E-03
0.25
2.2E-03
N/A
N/A
Monte Carlo
Simulation results
Monte Carlo
Modeling
CT = central tendency; HE = high-end
a For select OESs, updates to exposure estimates were made via information provided by the SACC and public comments.
Page 478 of 570
-------
F.6 Summary of Weight of Scientific Evidence Conclusions in Inhalation Exposure Estimates
TableApx F-35 provides a summary of EPA's weight of scientific evidence conclusions in its inhalation exposure estimates for each of the
Occupational Exposure Scenarios assessed.
Table Apx F-35. Summary of Weight of Scientific Evidence Conclusions in Inhalation Exposure Estimates by PES
OES
Weight of Scientific Evidence Conclusion in Inhalation Exposure Estimates
Textile dye
8-hour TWA inhalation exposure estimates are assessed using OSHA's CEHD. Factors that increase the strength of evidence for
this OES are that the exposure data are directly relevant to the OES (as opposed to surrogate), that OSHA CEHD has a high overall
data quality determination, and consistency within the dataset (all measurements are taken by OSHA through NIOSH method
1602). The data includes personal and area samples from multiple sites, which increases the variability of the dataset. Factors that
decrease the strength of the evidence for this OES include the low number of data points, uncertainty in the representativeness of the
monitoring data for all sites in this OES, and uncertainty in the representativeness of the older monitoring data towards more current
operations (some data were from 1991-1992). Additionally, worker activity descriptions are not provided in the dataset and there
was a high number of non-detects present. Based on this information, EPA has concluded that the weight of scientific evidence for
this assessment is moderate and provides a plausible estimate of exposures in consideration of the strengths and limitations of
reasonably available data.
Antifreeze
8-hour TWA inhalation exposure estimates are assessed using Monte Carlo modeling with information from the OECD ESD on
Chemical Additives used in Automotive Lubricants, the EPA MRD on Commercial Use of Automotive Detailing Products, and
EPA/OPPT models. Factors that increase the strength of evidence for this OES are that the ESD and MRD used have high overall
data quality determinations, high number of data points (simulation runs), and full distributions of input parameters. The Monte
Carlo modeling accounts for the entire distribution of input parameters, calculating a distribution of potential exposure values that
represents a larger proportion of sites than a discrete value. Factors that decrease the strength of the evidence for this OES include
that the ESD and MRD are not directly applicable to antifreeze uses (used as surrogate), uncertainty in the representativeness of
evidence to all sites, and uncertainty in the use of generic default values from the ESD and MRD for sites that specifically use 1,4-
dioxane. Additionally, EPA scaled up a consumer antifreeze use rate to a commercial use rate based on information in the ESD and
MRD, which increases uncertainty. Based on this information, EPA has concluded that the weight of scientific evidence for this
assessment is moderate and provides a plausible estimate of exposures in consideration of the strengths and limitations of
reasonably available data.
Surface cleaner
8-hour TWA inhalation exposure estimates are assessed using monitoring data from published literature. Factors that increase the
strength of evidence for this OES are that the exposure data are directly relevant to the OES (as opposed to surrogate), that the
literature has a medium overall data quality determination, and consistency within the dataset (all measurements are taken via the
same method). Additionally, the literature includes information on worker activities during sampling. Factors that decrease the
strength of the evidence for this OES include the lack of variability (only one study), uncertainty in the representativeness of the
monitoring data for all sites in this OES, uncertainty from using summary statistics from the study (discrete sample results not
provided), and uncertainty in whether the activities performed in this study accurately reflect all surface cleaning scenarios or the
Page 479 of 570
-------
OES
Weight of Scientific Evidence Conclusion in Inhalation Exposure Estimates
cleaning industry as whole. Based on this information, EPA has concluded that the weight of scientific evidence for this assessment
is moderate to robust and provides a plausible estimate of exposures in consideration of the strengths and limitations of reasonably
available data.
Dish soap
8-hour TWA inhalation exposure estimates are assessed using Monte Carlo modeling with EPA/OPPT models, using input data
from the NYDEC waiver database (NYDEC, 2023). a public comment (P&G. 2023). and standard EPA/OPPT default values.
Factors that increase the strength of evidence for this OES are that the exposure estimates are directly relevant to the OES (as
opposed to surrogate), that the public comment contains directly relevant data and has a high overall data quality determination,
high number of data points (simulation runs), and full distributions of input parameters. The Monte Carlo modeling accounts for the
entire distribution of input parameters, calculating a distribution of potential exposure values that represents a larger proportion of
sites than a discrete value. Factors that decrease the strength of the evidence for this OES include uncertainty in the
representativeness of evidence to all sites and uncertainty in the representativeness of some standard EPA/OPPT default values
towards real-world sites that use 1,4-dioxane. Based on this information, EPA has concluded that the weight of scientific evidence
for this assessment is moderate and provides a plausible estimate of exposures in consideration of the strengths and limitations of
reasonably available data.
Dishwasher
detergent
Since EPA used the same approach as discussed for dish soap, the same information and weight of scientific evidence conclusion
apply.
Laundry detergent
8-hour TWA inhalation exposure estimates are assessed using Monte Carlo modeling with information from the ESD on Water
Based Washing Operations at Industrial and Institutional Laundries and EPA/OPPT models. Factors that increase the strength of
evidence for this OES are that the exposure estimates are directly relevant to the OES (as opposed to surrogate), that the ESD on
Industrial and Institutional Laundries has a medium overall data quality determination and was peer reviewed, high number of data
points (simulation runs), and full distributions of input parameters. The Monte Carlo modeling accounts for the entire distribution of
input parameters, calculating a distribution of potential exposure values that represents a larger proportion of sites than a discrete
value. Also, EPA was able to separately estimate exposures for industrial and institutional laundry settings. Factors that decrease the
strength of the evidence for this OES include uncertainty in the representativeness of evidence to all sites and uncertainty in the
representativeness of generic values in the ESD towards real-world sites that use 1,4-dioxane. Based on this information, EPA has
concluded that the weight of scientific evidence for this assessment is moderate and provides a plausible estimate of exposures in
consideration of the strengths and limitations of reasonably available data.
Paint and floor
lacquer
8-hour TWA inhalation exposure estimates are assessed using monitoring data from a NIOSH HHE. Factors that increase the
strength of evidence for this OES are that the exposure data are directly relevant to the OES (as opposed to surrogate), that the
literature has a high overall data quality determination, and consistency within the dataset (all measurements are taken via the same
NIOSH method). Factors that decrease the strength of the evidence for this OES include the low number of data points, lack of
variability (one study), uncertainty in the representativeness of the monitoring data for all sites and worker activities in this OES,
and uncertainty in the representativeness of the older monitoring data towards more current operations (data were from 1989).
Additionally, some of these data were short-term samples that EPA converted to 8-hour TWAs by assuming there was no exposure
for the remainder of the 8 hours after the sampling duration, which adds uncertainty. This assumption may result in underestimation
Page 480 of 570
-------
OES
Weight of Scientific Evidence Conclusion in Inhalation Exposure Estimates
of exposures if workers perform additional activities that may result in exposures to 1,4-dioxane that were not captured in the
monitoring performed. Based on this information, EPA has concluded that the weight of scientific evidence for this assessment is
moderate and provides a plausible estimate of exposures in consideration of the strengths and limitations of reasonably available
data.
Polyethylene
terephthalate (PET)
byproduct
8-hour TWA inhalation exposure estimates are assessed using OSHA's CEHD and data from public comments (DAK Americas,
2023; Huntsman. 2023). Factors that increase the strength of evidence for this OES are that the exposure data are directiv relevant to
the OES (as opposed to surrogate), that all exposure data sources have a high overall data quality determination, the exposure data
from the public comments is from 1998 through 2023 and includes detailed worker activity descriptions, and that the exposure data
represents multiple sites. Factors that decrease the strength of the evidence for this OES include uncertainty in the
representativeness of the monitoring data for all sites and worker activities in this OES and uncertainty in the representativeness of
the older monitoring data from OSHA towards more current operations (data were from 1985-1994). Additionally, worker activity
descriptions are not provided in the OSHA CEHD dataset. Based on this information, EPA has concluded that the weight of
scientific evidence for this assessment is moderate to robust and provides a plausible estimate of exposures in consideration of the
strengths and limitations of reasonably available data.
Ethoxylation
process byproduct
8-hour TWA inhalation exposure estimates are assessed using OSHA's CEHD and data from a public comment (Dow Chemical,
2023). Factors that increase the strength of evidence for this OES are that the exposure data are directlv relevant to the OES (as
opposed to surrogate), both data sources have a high overall data quality determination, the monitoring data from the public
comment is dated 2023 and includes worker activity descriptions, and the data are from multiple sites. Factors that decrease the
strength of the evidence for this OES include the low number of data points, age of the OSHA data (data are from 2000), and
uncertainty in the representativeness of the monitoring data for all sites and worker activities in this OES. Additionally, worker
activity descriptions are not provided in the OSHA CEHD dataset. Based on this information, EPA has concluded that the weight of
scientific evidence for this assessment is moderate and provides a plausible estimate of exposures in consideration of the strengths
and limitations of reasonably available data.
Hydraulic
fracturing
8-hour TWA inhalation exposure estimates are assessed using Monte Carlo modeling with information from the Revised ESD on
Hydraulic Fracturing and FracFocus 3.0. Factors that increase the strength of evidence for this OES are that the exposure estimates
are directly relevant to the OES (as opposed to surrogate), that the Revised ESD on Hydraulic Fracturing and FracFocus 3.0 have
medium overall data quality determinations, that the Revised ESD has undergone peer review by OECD, the high number of data
points (simulation runs), and the full distributions of input parameters. The Monte Carlo modeling accounts for the entire
distribution of input parameters, calculating a distribution of potential exposure values that represents a larger proportion of sites
than a discrete value. Factors that decrease the strength of the evidence for this OES include the uncertainties and limitations in the
representativeness of the estimates for sites that specifically use 1,4-dioxane because the default values from the Revised ESD on
Hydraulic Fracturing. Additionally, the duration of exposure for container unloading and cleaning activities is uncertain. To avoid
unrealistic output parameters, exposure duration was capped at 2 hours for each activity. This is a limitation of the assessment
because there is uncertainty in the extent to which the assessed activity durations are representative of real-world conditions. Based
on this information, EPA has concluded that the weight of scientific evidence for this assessment is moderate to robust and provides
a plausible estimate of exposures in consideration of the strengths and limitations of reasonably available data.
Page 481 of 570
-------
F.7 Antifreeze Modeling Approach and Parameters for Estimating
Occupational Inhalation Exposures
This appendix presents the modeling approach used to estimate occupational inhalation exposures to
1,4-dioxane during the commercial use of antifreeze. EPA expects that the main source of occupational
inhalation exposure during the use of antifreeze is from the unloading of antifreeze from containers into
vehicles. Therefore, this approach applies a stochastic modeling approach to the EPA/OAQPS AP-42
Loading Model, which estimates air releases during liquid transfer operations, and the EPA/OPPT Mass
Balance Model, which estimates the corresponding inhalation exposures resulting from these air
releases.
Inhalation exposure to chemical vapors is a function of the chemical's physical properties, ventilation
rate of the container loading area, type of loading method, and other model parameters. While physical
properties are fixed for a chemical, some model parameters are expected to vary from one facility to
another. An individual model input parameter could either have a discrete value or a distribution of
values. EPA assigned statistical distributions based on available literature data or engineering judgment
to address the variability in parameters such as ventilation rate (RATEventiiation), mixing factor (Fmixing),
saturation factor (Fsaturation), concentration of 1,4-dioxane in antifreeze (Fdioxane), container size (Vcont),
and number of jobs per day (Njobs).
A Monte Carlo simulation was conducted to capture variability in the model input parameters described
above. The simulation was conducted using the Latin hypercube sampling method in @Risk (Palisade,
Ithaca, NY). The Latin hypercube sampling method is a statistical method for generating a sample of
possible values from a multi-dimensional distribution. Latin hypercube sampling is a stratified method,
meaning it guarantees that its generated samples are representative of the probability density function
(variability) defined in the model. EPA performed 100,000 iterations of the model to capture the range
of possible input values, including values with low probability of occurrence.
From the distribution resulting from the Monte Carlo simulation, EPA selected the 95th and 50th
percentile values to represent a high-end exposure and central tendency exposure level respectively. The
statistics were calculated directly in @Risk. The following subsections detail the model design equations
and parameters used for inhalation exposure estimates.
F.7.1 Model Equations
Daily use rate of antifreeze at commercial sites is calculated using the following equation:
Equation Apx F-l.
Qantifreeze_day ~ Qconsumer * ^jobs
Where:
Qantifreeze_day = Commercial daily use rate of antifreeze [kg/site-day]
Qconsumer = Consumer use rate of antifreeze [kg/job]
Njobs = Commercial antifreeze jobs per day [jobs/day]
Annual use rate of antifreeze at commercial sites is calculated using the following equation:
Page 482 of 570
-------
EquationApx F-2.
Where:
Q anti freeze _yr
OD
Q anti freeze_day
Qantifreeze_yr ~ OD * Qantifreeze_day
Commercial annual use rate of antifreeze [kg/site-year]
Operating days [days/site-year]
Commercial daily use rate of antifreeze [kg/site-day]
The number of antifreeze container used per year is calculated using the following equation:
Equation Apx F-3.
N,
Q antifreeze^
yr
Where:
Ncont_site_yr
Q anti freeze _yr
Vcont
cont_site_yr ~ I kg
3-79^m*1~*Vcont
Number of antifreeze containers used per year [containers/site-year]
Commercial annual use rate of antifreeze [kg/site-year]
Antifreeze container size [gal]
Duration of release for container activities is calculated using the following equation:
Equation Apx F-4.
OH,
cont unload ~
Ncont_site_yr
OD * RATE fin
Where:
OHcont:unload
Ncont_site_yr
OD
RATE fin
= Duration of release for container activities [hours]
= Number of antifreeze containers used per year [containers/site-year]
= Operating days [days/site-year]
= Container fill/unloading rate [containers/hour]
Vapor pressure correction factor is calculated using the following equation:
Equation Apx F-5.
X =
Fdioxane j
MW
Fdioxane _|_ 1 Pdioxane
MW
18
Where:
MW
Fdioxane
Vapor pressure correction factor [mol dioxane/mol water]
Molecular weight of 1,4-dioxane [g/mol]
1,4-dioxane concentration in antifreeze [kg/kg]
Vapor generation rate of 1,4-dioane during container unloading is calculated using the following equation:
Page 483 of 570
-------
EquationApx F-6.
tvapor_generation
= F<
saturation
* MW * 3785.4 * V,
cont
RATEfm VP
* x * * ¦
3600
760 T*R
Where:
Qvapor generation
Fsaturation
MW
Vcont
RATE fin
VP
T
R
Vapor generation rate of l,4-dioxane[g/s]
Saturation factor [unitless]
Molecular weight of 1,4-dioxane [g/mol]
Antifreeze container size [gal]
Container fill/unloading rate [containers/hour]
Vapor pressure of 1,4-dioxane [torr]
Ambient temperature [K]
Universal gas constant [atm-cm3/gmol-K]
Volumetric concentration of 1,4-dioxane in air during unloading is calculated using the following
equation:
Equation Apx F-7.
^ 170000 * T * QVaporgeneration
MW * RATEventnation * Fm/w
mixing
Where:
r
Qvapor generation
MW
RATEventiiation
F ¦ ¦
1 mixing
Volumetric concentration of 1,4-dioxane in air [ppm]
Ambient temperature [K]
Vapor generation rate of 1,4-dioxane [g/s]
Molecular weight of 1,4-dioxane [g/mol]
Ventilation rate [ftVmin]
Mixing factor [unitless]
8-hour TWA mass concentration of 1,4-dioxane in air is calculated using the following equation. Note
that this equation assumes that no exposure occurs for the remainder of the 8-hour shift after container
unloading takes place:
Equation Apx F-8.
ConcentrationEP1 =
Cv * MW OHcont unioad
Vm
8
Where:
Concentration
Cv
MW
0HCOnt_unlOad
Vm
EP1
8-hour TWA mass concentration of 1,4-dioxane in air [mg/m3]
Volumetric concentration of 1,4-dioxane in air [ppm]
Molecular weight of 1,4-dioxane [g/mol]
Duration of release for container activities [hours]
Molar volume [L/mol]
F.7.2 Modeling Input Parameters
Table Apx F-36 summarizes the model parameters and their values for the Monte Carlo simulation.
Additional explanations of EPA's selection of the distributions for each parameter are provided after this
Page 484 of 570
-------
table. High-end and central tendency exposures are estimated by selecting the 50th and 95th percentile
values from the output distribution.
Page 485 of 570
-------
Table Apx F-36. Summary of Parameter Values and Distributions Used in the Antifreeze Exposure Modeling
Input Parameter
Symbol
Unit
Constant Model
Parameter Values
Variable Model Parameter Values
Rationale/Basis
Value
Lower
Bound
Upper
Bound
Mode
Distribution
Type
Container Size
Vcont
gal
0.125
0.03125
20
0.125
Triangular
See Section F.7.3
Jobs per Day
Njobs
jobs/day
9
1
9
-
Discrete
See Section F.7.4
Concentration of
1,4-dioxane in
Antifreeze
F dioxane
kg/kg
0.000086
0.00000001
0.000086
Uniform
See Section F.7.5
Ventilation Rate
RATEventilation
ft3/min
500
500
10,000
3,000
Triangular
See Section F.7.6
Mixing Factor
F mixing
Dimensionless
0.1
0.1
1
0.5
Triangular
See Section F.7.7
Saturation Factor
F saturation
Dimensionless
1
0.5
1
0.5
Triangular
See Section F.7.8
Vapor Pressure of
1,4-dioxane
VP
torr
40
-
-
-
-
Physical property
Molecular Weight
of 1,4-dioxane
MW
g/mol
88.1
-
-
-
-
Physical property
Ambient
Temperature
T
K
298
-
-
-
-
Process parameter
Universal Gas
Constant
R
atm-cm3/gmol-K
82.05
-
-
-
-
Universal constant
Molar Volume
Vm
L/mol
24.45
-
-
-
-
Physical property
Use Rate of
Antifreeze
Qconsumer
Kg/job
2
0.15
2
—
Uniform
See Section F.7.9
Operating Days
OD
days/year
250
-
-
-
-
See Section F.4.2
Fill Rate of
Containers
RATEfin
containers/hour
60
-
-
-
-
See Section F.7.10
Operating Hours
OHccnt unload
hours/day
8
-
-
-
-
Process parameter
Page 486 of 570
-------
F.7.3 Container Size
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
the Agency used generic data from the EPA MRD on Commercial Use of Automotive Detailing
Products (U.S. EPA. 2022b) and the OECD ESD on Chemical Additives used in Automotive Lubricants
(OECD. 2020). The MRD identifies a minimum container size of 4 ounces (0.03125 gal) and a default
container size of 16 ounces (0.125 gallons) ( 2022b). The ESD identifies a maximum
container size of 20 gallons (OECD. 2020). Based on these data, EPA modeled container size using a
triangular distribution with a lower bound of 0.03125 gallons, an upper bound of 20 gallons, and a mode
of 0.125 gallons.
F.7.4 Jobs per Day
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
the Agency used generic data from standard EPA models and the EPA MRD on Commercial Use of
Automotive Detailing Products ( 2b). The EPA Brake Servicing Near-Field/Far-Field
Inhalation Exposure Model indicates one to four cars are serviced per day and the MRD indicates up to
nine cars are serviced per day ( 22b). Based on this, EPA modeled this parameter with a
uniform distribution that assigns equal probability for each whole number of jobs from one to nine
jobs/day.
F.7.5 Concentration of 1,4-Dioxane in Antifreeze
EPA modeled concentration of 1,4-dioxane in antifreeze using a uniform distribution from a lower
bound of 1,00x 10~8 kg 1,4-dioxane/kg antifreeze to an upper bound of 8,60/10 5 kg 1,4-dioxane/kg
antifreeze. This is based on the December 2020 Final Risk Evaluation for 1,4-Dioxane (U.S. EPA.
2020c). which indicates that 1,4-dioxane is a byproduct in antifreeze at concentrations ranging from 0.01
to 86 ppm. EPA did not identify additional data on the concentration of 1,4-dioxane in antifreeze.
F.7.6 Ventilation Rate
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
the Agency used generic data from the CEB Manual ( ), which indicates general
ventilation rates in industry range from 500 to 10,000 ftVmin, with a typical value of 3,000 ftVmin. The
underlying distribution of this parameter is not known; therefore, EPA assigned a triangular distribution,
which is completely defined by the range and mode of a parameter. EPA assumed the mode is equal to
the typical value provided by the CEB Manual ( ).
F.7.7 Mixing Factor
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
the Agency used generic data from the CEB Manual ( ), which indicates mixing factors
may range from 0.1 to 1, with 1 representing ideal mixing. The CEB Manual references the 1988
ACGIH Ventilation Handbook, which suggests the following factors and descriptions: 0.67 to 1 for best
mixing; 0.5 to 0.67 for good mixing; 0.2 to 0.5 for fair mixing; and 0.1 to 0.2 for poor mixing. The
underlying distribution of this parameter is not known; therefore, EPA assigned a triangular distribution,
which is completely defined by the range and mode of a parameter. The mode for this distribution was
not provided; therefore, EPA assigned a mode value of 0.5 based on the typical value provided in the
ChemSTEER User Guide 0 v ^ \ -<' * «) for the EPA/OPPT Mass Balance Inhalation Model.
F.7.8 Saturation Factor
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
the Agency used generic data from the CEB Manual ( ), which indicates that the
Page 487 of 570
-------
saturation concentration was reached or exceeded by misting with a maximum saturation factor of 1.45
during splash filling. The CEB manual indicates that the saturation factor for bottom filling was
expected to be about 0.5 ( ). The underlying distribution of this parameter is not known;
therefore, EPA assigned a triangular distribution, which is completely defined by range and mode of a
parameter. Because a mode was not provided for this parameter, EPA assigned a mode value of 0.5 for
bottom filling as bottom filling minimizes volatilization ( ). This value also corresponds
to the typical value provided in the ChemSTEER User Guide (U.S. EPA. I \t) for the EPA/OAQPS
AP-42 Loading Model for small containers.
F.7.9 Use Rate of Antifreeze per Job
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
the Agency used generic data from the Consumer Exposure Model (CEM). The December 2020 Final
Risk Evaluation for 1,4-Dioxane provided a single value of 0.15 kg/job for the consumer use rate of
antifreeze from the CEM (U.S. EPA. 2020c). The 0.15 kg/job represents a "top-up" amount and a use
rate of 2 kg/job represents a full replacement of antifreeze in a car. Therefore, EPA modeled the use rate
for antifreeze to be a uniform distribution with a lower-bound of 0.15 kg/job and an upper bound of 2
kg/job.
F.7.10 Container Fill Rate
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
the Agency used generic data from the ChemSTEER User Guide ( ). The ChemSTEER
User Guide provides a typical fill rate of 60 containers per hour for small containers and bottles, which
are anything less than 20 gallons in capacity. Therefore, EPA could not develop a distribution of values
for this parameter and used the single value 60 containers/hour from the ChemSTEER User Guide.
F.7.11 Key Strengths, Limitations, Uncertainties, and Sensitivity Analysis
General modeling uncertainties and limitations are discussed in Section 2.2.1.3 and Appendix E.7
regarding assigning distributions to input parameters, using generic data for some input parameter
distributions, and using static values for other input parameters also apply to the exposure modeling.
EPA ran a sensitivity analysis in @Risk to identify the input parameters which have the largest impact
on the 8-hour TWA occupational inhalation exposure estimate. FigureApx F-4 shows the inputs ranked
by which have the largest effect on the 8-hour TWA occupational inhalation exposure. The ventilation
rate and concentration of 1,4-dioxane in antifreeze have the relatively largest impacts on the exposure
estimate. As discussed in Appendix F.7.5, the concentration of 1,4-dioxane in antifreeze is based on a
range from the December 2020 Final Risk Evaluation for 1,4-Dioxane (U.S. EPA. 2020c). EPA did not
find any additional data on concentration of 1,4-dioxane in antifreeze. The ventilation rate, as well as all
the other input parameters in Figure Apx F-4 are based on generic, not 1,4-dioxane specific data.
Having a distribution for each input parameter is a strength of the assessment; however, the
representativeness of the underlying data used for these distributions towards is a limitation.
Page 488 of 570
-------
8-hr TWA of 1,4-Dioxane
Inputs Ranked by Effect on Output Mean
Ventilation Rate
Concentration of 1,4-dioxane in Antifreeze
Mixing Factor
Jobs per Day
Saturation Factor
Container Size
FigureApx F-4. Sensitivity Chart for 8-Hour TWA Occupational Inhalation Exposure to 1,4-
Dioxane at Antifreeze Use Sites
F.8 Laundry Detergent Modeling Approach and Parameters for
Estimating Occupational Inhalation Exposures
This appendix presents the modeling approach used to estimate occupational inhalation exposures to
1,4-dioxane during the industrial and institutional use of laundry detergents. This approach utilizes the
OECD ESD on the Chemicals Used in Water Based Washing Operations at Industrial and Institutional
Laundries (OECD. 2011b) combined with Monte Carlo simulation (a type of stochastic simulation).
This ESD categorized laundry facilities into either industrial or institutional facilities, as described in
Appendix E.12. Because the ESD includes different process parameters for industrial and institutional
laundries, EPA modeled the two types of laundry facilities separately. In addition, laundry detergents
can be in liquid or powder physical forms. The difference in physical form results in different parameter
distributions. Therefore, EPA modeled liquid and powder detergents separately. This ESD includes a
diagram of release and exposure points during the use of laundry detergents, as shown in Figure Apx
F-5.
Page 489 of 570
-------
©©©Container Residue and Cleaning ©©©Container Residue and Cleaning
©© Fugitive Air Release and Dust @@ Fugitive Air Release and Dust
Emissions During Transfers Emissions During Transfers
©@© Releases During
Operations
FigureApx F-5. Environmental Release Points (Numbered) and Occupational Exposure Points
(Lettered) During Industrial/Institutional Laundering Operations
Based on Figure Apx F-5, EPA identified the following exposure points (note that exposure point 1
corresponds to diagram point A, point 2 to diagram point B, and points 3 - 4 to diagram point C):
• Exposure point 1 (EP1): Exposure to vapors during container transfers;
• Exposure point 2 (EP2): Exposure to vapors during container cleaning;
• Exposure point 3 (EP3): Exposure to vapors during laundry operations;
• Exposure point 4 (EP4): Exposure to total particulates over all activities; and
• Exposure point 5 (EP5): Exposure to respirable particulates over all activities.
To estimate inhalation exposures to vapors, this model utilizes the previously modeled vapor releases for
each corresponding release point, as explained in Appendix E. 11.16. To calculate a full-shift TWA, the
1,4-dioxane concentrations calculated for each exposure point above are multiplied by their respective
exposure durations, then summed and divided by the total workday duration (8, 10, or 12 hours per the
ESD).
Inhalation exposure to chemical vapors is a function of the chemical's physical properties, ventilation
rate of the container loading area, type of loading method, and other model parameters. Although
physical properties are fixed for a chemical, some model parameters are expected to vary from one
facility to another. An individual model input parameter could either have a discrete value or a
distribution of values. EPA assigned statistical distributions based on available literature data or
engineering judgment to address the variability in parameters such as ventilation rate (RATEventiiation),
mixing factor (Fmixing), total and respirable particulate concentration (Cparticuiate) and mass fraction of 1,4-
dioxane (F dioxane laundry) •
A Monte Carlo simulation was then conducted to capture variability in the model input parameters
described above. The simulation was conducted using the Latin hypercube sampling method in @Risk
(Palisade, Ithaca, NY). The Latin hypercube sampling method is a statistical method for generating a
sample of possible values from a multi-dimensional distribution. Latin hypercube sampling is a stratified
method, meaning it guarantees that its generated samples are representative of the probability density
function (variability) defined in the model. EPA performed 100,000 iterations of the model to capture
the range of possible input values, including values with low probability of occurrence.
Page 490 of 570
-------
From the distribution resulting from the Monte Carlo simulation, EPA selected the 95th and 50th
percentile values to represent a high-end exposure and central tendency exposure level respectively. The
statistics were calculated directly in @Risk. The following subsections detail the model design equations
and parameters used for inhalation exposure estimates.
F.8.1 Model Equations
Exposure point 1 (container transfers) volumetric concentration of 1,4-dioxane is calculated using the
following equation:
EquationApx F-9.
Release_perDayRP3
Cvl = (1.75 x 105) * T
Where:
\OHcont unload * (3600^) * (o-OOl^
MW * RATEventuation
* Fmixing
Cvl = Volumetric concentration of 1,4-dioxane in air for exposure
point 1 [ppm]
T = Ambient temperature [K]
Release_perDayRP2 = Daily vapor release for release point 3, Appendix E. 11.16
[kg/site-day]
OHcont unload = Duration of release for container unloading, Appendix E. 11.16
[hours/day]
MW = Molecular weight of 1,4-dioxane [g/mol]
RATEventnation = Ventilation rate [ftVmin]
Pmixing = Mixing factor [unitless]
Exposure point 1 (container transfers) mass concentration of 1,4-dioxane is calculated using the
following equation:
Equation Apx F-10.
Cvl * MW
Concentration_VaporEP1 = —
nn
Where:
Concentration_VaporEP1= Mass concentration of 1,4-dioxane in air for exposure
point 1 [mg/m3]
Cvl = Volumetric concentration of 1,4-dioxane in air for exposure
point 1 [ppm]
MW = Molecular weight of 1,4-dioxane [g/mol]
Vm = Molar volume [L/mol]
Exposure point 2 (container cleaning) volumetric concentration of 1,4-dioxane is calculated using the
following equation:
Page 491 of 570
-------
EquationApx F-ll.
Release_perDayRp2
Ncont_site_yr
* (3600-^pJ * (0.001^^
\OD * RATE fin v in / x
Cv2 = (1.75 x 105) *T*
MW * RATEventnation
* Fmixing
Where:
Cv2 = Volumetric concentration of 1,4-dioxane in air for exposure
point 2 [ppm]
T = Ambient temperature [K]
Release_perDayRP2 = Daily vapor release for release point 2, see Appendix E. 11.16
[kg/site-day]
NCont_site_yr = Number of detergent containers used per year, see Appendix
E.11.16
[containers/site-year]
OD = Operating days, see Appendix E. 11.16 [days/year]
RATEfui = Container fill/unload rate, see Appendix E.11.16 [containers/hour]
MW = Molecular weight of 1,4-dioxane [g/mol]
RATEventiiation = Ventilation rate [ftVmin]
Pmixing = Mixing factor [unitless]
Exposure point 2 (container cleaning) mass concentration of 1,4-dioxane is calculated using the
following equation:
Equation Apx F-12.
C oncentrationV ap orEP2
Cv2 * MW
Vm
Where:
Concentration_VaporEP2= Mass concentration of 1,4-dioxane in air for exposure
point 2 [mg/m3]
Cv2 = Volumetric concentration of 1,4-dioxane in air for exposure
point 2 [ppm]
MW = Molecular weight of 1,4-dioxane [g/mol]
Vm = Molar volume [L/mol]
Exposure point 3 (laundry washing operations) volumetric concentration of 1,4-dioxane is calculated
using the following equation:
Equation Apx F-13.
Release_perDayRPS
i OH *(3600 4~) * 10-001 —]
C„3 = (1-75 x 10=) . T . A ^ hlL\ LiL
MW * RATEventuation
* Fmixing
Where:
Page 492 of 570
-------
Cv3 = Volumetric concentration of 1,4-dioxane in air for exposure
point 3 [ppm]
T = Ambient temperature [K]
Release_perDayRP5 = Daily vapor release for release point 5, see Appendix E.l 1.16
[kg/site-day]
OH = Operating hours, see Appendix E. 11.16 [hours/day]
MW = Molecular weight of 1,4-dioxane [g/mol]
RATEventnation = Ventilation rate [ftVmin]
Fmixina = Mixing factor [unitless]
Exposure point 3 (laundry washing operations) mass concentration of 1,4-dioxane is calculated using the
following equation:
EquationApx F-14.
Cv3 * MW
Concentration_VaporEP3 = —
nn
Where:
Concentration_VaporEP3= Mass concentration of 1,4-dioxane in air for exposure
point 3 [mg/m3]
Cv3 = Volumetric concentration of 1,4-dioxane in air for exposure
point 3 [ppm]
MW = Molecular weight of 1,4-dioxane [g/mol]
Vm = Molar volume [L/mol]
The total full-shift vapor exposure (8-, 10-, and 12-hour TWAs) accounting for EP1 through EP3 is
calculated using the following equation:
Equation Apx F-15.
Vapor_Exposure_TWA
(concentration_VaporEP1 * OHcont unload + Concentration_VaporEP2 * + Concentration-VaPorEP3 * (OH - OHcont unload -
~ OH
Where:
Vapor_Exposure_TWA =
Concentration_VaporEP1 =
Concentration_VaporEP2 =
Concentration_VaporEP3 =
OHcont:unload ~
OD
Ncont_site_yr ~
RATEfui =
Full-shift TWA of 1,4-dioxane vapor exposure [mg/m3]
Mass concentration of 1,4-dioxane in air for exposure
point 1 [mg/m3]
Mass concentration of 1,4-dioxane in air for exposure
point 2 [mg/m3]
Mass concentration of 1,4-dioxane in air for exposure
point 3 [mg/m3]
Duration of release for container unloading, see Appendix
E.l 1.16 [hours/day]
Operating days, see Appendix E. 11.16 [days/year]
Number of detergent containers used per year, see
Appendix E. 11.16 [containers/site-year]
Container fill/unload rate, see Appendix E. 11.16
Page 493 of 570
-------
containers/hour]
OH = Operating hours, see Appendix E. 11.16 [hours/day]
Exposure point 4 (total particulate exposure) mass concentration of 1,4-dioxane is calculated using the
following equation:
EquationApx F-16.
Concentration_particulate_tOtal ^particulate_total * ^dioxane_laundry
Where:
Concentration_particulate_total = Air concentration of total 1,4-dioxane particles in the
worker's breathing zone [mg/m3]
^particulate_totai = Air concentration of all particles in the worker's breathing zone
[mg/m3]
Fdioxanejaundry = Mass fraction of 1,4-dioxane in laundry detergent, see Appendix
E.11.16 [kg/kg]
Exposure point 5 (respirable particulate exposure) mass concentration of 1,4-dioxane is calculated using
the following equation:
Equation Apx F-17.
Concentration_particulate_respirable CparticuiatereSpjra2,;e * Fdioxanejaundry
Where:
Concentration_particulate_respirable = Air concentration of respirable 1,4-dioxane
particles in the worker's breathing zone [mg/m3]
^particulate respirable = Air concentration of all respirable particles in the worker's breathing
zone [mg/m3]
Fdioxanejaundry = Mass fraction of 1,4-dioxane in laundry detergent, see Appendix
E.11.16 [kg/kg]
F.8.2 Model Input Parameters
Table Apx F-37 summarizes the model parameters and their values for the Monte Carlo simulation.
Additional explanations of EPA's selection of the distributions for each parameter are provided after this
table. High-end and central tendency exposures are estimated by selecting the 50th and 95th percentile
values from the output distribution.
Page 494 of 570
-------
Table Apx F-37. Summary of Parameter Values and Distributions Used in the Laundry Detergent Exposure Modeling
Input Parameter
Symbol
Unit
Constant Model
Parameter
Values
Variable Model Parameter Values
Rationale/
Basis
Value
Lower
Bound
Upper Bound
Mode
Distribution
Type
Ventilation Rate
RATEventilation
ft3/min
500
500
10,000
3,000
Triangular
See Section
F.8.3
Mixing Factor
F mixing
dimensionless
0.1
0.1
1
0.5
Triangular
See Section
F.8.4
Total Particulate
Concentration
Cparticulate total
mg/m3
15
0.01
15
9.5
Triangular
See Section
F.8.5
Respirable Particulate
Concentration
Cparticulate respirable
mg/m3
5
0.018
Institutional:
5
Industrial:
5
Institutional:
0.21
Industrial:
1.3
Triangular
See Section
F.8.6
Molecular Weight of 1,4-
Dioxane
MW
g/mol
88.1
-
-
-
-
Physical
property
Ambient Temperature
T
K
298
-
-
-
-
Process
parameter
Molar Volume
Vm
L/mol
24.45
-
-
-
-
Physical
property
Page 495 of 570
-------
F.8.3 Ventilation Rate
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
the Agency used generic data from the CEB Manual ( ), which indicates general
ventilation rates in industry range from 500 to 10,000 ftVmin, with a typical value of 3,000 ftVmin. The
underlying distribution of this parameter is not known; therefore, the Agency assigned a triangular
distribution, which is completely defined by the range and mode of a parameter. EPA assumed the mode
is equal to the typical value provided by the CEB Manual (U.S. EPA. 1991).
F.8.4 Mixing Factor
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
the Agency used generic data from the CEB Manual ( ), which indicates mixing factors
may range from 0.1 to 1, with 1 representing ideal mixing. The CEB Manual references the 1988
ACGIH Ventilation Handbook that suggests the following factors and descriptions: 0.67 to 1 for best
mixing; 0.5 to 0.67 for good mixing; 0.2 to 0.5 for fair mixing; and 0.1 to 0.2 for poor mixing. The
underlying distribution of this parameter is not known; therefore, EPA assigned a triangular distribution,
which is completely defined by the range and mode of a parameter. The mode for this distribution was
not provided; therefore, EPA assigned a mode value of 0.5 based on the typical value provided in the
ChemSTEER User Guide ( v N \ .) for the EPA/OPPT Mass Balance Inhalation Model.
F.8.5 Total Particulate Concentration
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
the Agency used generic data from a standard EPA model. Specifically, EPA modeled the total
particulate concentration using a triangular distribution with a lower bound of 0.01 mg/m3, an upper
bound of 15 mg/m3, and a mode of 9.5 mg/m3 for both industrial and institutional laundries. These
values were taken from EPA's Generic Model for Central Tendency and High-End Inhalation Exposure
to Total and Respirable Particulates Not Otherwise Regulated. This model utilizes inhalation monitoring
data from OSHA, which are analyzed by industry type (at the 2-digit or 3-digit NAICS code level). EPA
specifically used the data for NAICS industry group 81 (Other Services, Except Public Administration)
because this includes the NAICS code relevant to this OES, which is 812330, Linen and Uniform
Supply. For this industry group, the Generic Model for Central Tendency and High-End Inhalation
Exposure to Total and Respirable Particulates Not Otherwise Regulated indicates a total PNOR
concentration ranging from 0.01 to 699 mg/m3, with a mean of 9.5 mg/m3. EPA used the low-end of this
range and the mean as the lower bound and mode of the triangular distribution for this model. EPA used
the OSHA permissible exposure limit (PEL) for total particulates of 15 mg/m3 as the upper bound of the
distribution per the Generic Model for Central Tendency and High-End Inhalation Exposure to Total and
Respirable Particulates Not Otherwise Regulated, which indicates assessments should not assume that
the PEL is exceeded without case-specific data.
F.8.6 Respirable Particulate Concentration
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
the Agency used generic data from a standard EPA model. Specifically, EPA modeled the respirable
particulate concentration using a triangular distribution with lower bound of 0.018 mg/m3, an upper
bound of 5 mg/m3, and a mode of 1.3 mg/m3 for industrial laundries and 0.21 mg/m3 for institutional
laundries. These values were taken from EPA's Generic Model for Central Tendency and High-End
Inhalation Exposure to Total and Respirable Particulates Not Otherwise Regulated for NAICS industry
group 81 (Other Services, Except Public Administration) as described above in Section F.8.5. For this
industry group, the Generic Model for Central Tendency and High-End Inhalation Exposure to Total and
Respirable Particulates Not Otherwise Regulated indicates a respirable PNOR concentration ranging
Page 496 of 570
-------
from 0.018 to 19 mg/m3, with a mean of 1.3 mg/m3. EPA used the low-end of this range and the mean as
the lower bound and mode of the triangular distribution for this model. EPA used the OSHA permissible
exposure limit (PEL) for respirable particulates of 5 mg/m3 as the upper bound of the distribution per the
Generic Model for Central Tendency and High-End Inhalation Exposure to Total and Respirable
Particulates Not Otherwise Regulated, which indicates assessments should not assume that the PEL is
exceeded without case-specific data.
F.8.7 Key Strengths, Limitations, Uncertainties, and Sensitivity Analysis
General modeling uncertainties and limitations are discussed in Section 2.2.1.3 and Appendix E.7
regarding assigning distributions to input parameters, using generic data for some input parameter
distributions, and using static values for other input parameters also apply to the exposure modeling.
EPA ran a sensitivity analysis in @Risk to identify the input parameters which have the largest impact
on the 8-hourr TWA vapor and total particulate occupational inhalation exposure estimates. FigureApx
F-6 shows the inputs ranked by which have the largest effect on the 8-hour TWA occupational
inhalation exposure to 1,4-dioxane vapors at institutional laundries. Figure Apx F-7 similarly shows the
inputs that impact the 8-hour TWA occupational inhalation exposure to 1,4-dioxane particulates at
industrial laundries. The mass fraction of 1,4-dioxane in laundry detergent has the largest impact on both
forms of inhalation exposure. As discussed in Appendix E. 12.4, EPA used a discrete dataset comprised
of 19 data points for the mass fraction of 1,4-dioxane laundry detergent. For all other parameters in
Figure Apx F-6 and Figure Apx F-7, EPA developed distributions based on generic—not 1,4-dioxane-
specific data. Having a distribution for each input parameter is a strength of the assessment; however,
the representativeness of the underlying data used for these distributions is a limitation.
Total Vapor 8h-TWA of 1,4-Dioxane
Inputs Ranked by Effect on Output Mean
Mass Fraction of 1,4-Dioxane in Laundry Detergent
Air Speed
Duration of Release for Container Unloading
Ventilation Rate
Operating Days
Mixing Factor
Daily Use Rate of Liquid Laundry Detergents
Fraction of Laundry Detergents Containing lr4-Dioxane
Figure Apx F-6. Sensitivity Chart for 8-Hour TWA Occupational Inhalation Exposure to 1,4-
Dioxane Vapor at Institutional Laundries
m
~z
i
Page 497 of 570
-------
Total Particulate 8h-TWA of 1,4-Dioxane
Inputs Ranked by Effect on Output Mean
Mass Fraction of 1,4-Dioxane in Laundry
Detergent
Total PNOR Concentration
jL
FigureApx F-7. Sensitivity Chart for 8-Hour TWA Occupational Inhalation Exposures to 1,4-
Dioxane Total Particulates at Industrial Laundries
F.9 Hydraulic Fracturing Modeling Approach and Parameters for
Estimating Occupational Inhalation Exposures
This appendix presents the modeling approach used to estimate occupational inhalation exposures to
1,4-dioxane during hydraulic fracturing. This approach utilizes the Revised ESD on Chemicals Used in
Hydraulic Fracturing (U.S. EPA. 2022e) combined with Monte Carlo simulation (a type of stochastic
simulation). This ESD includes a diagram of release and exposure points during hydraulic fracturing, as
shown in Figure Apx F-8.
Page 498 of 570
-------
FigureApx F-8. Environmental Release Points (Numbered) and Occupational Exposure Points
(Lettered) During Hydraulic Fracturing
Based on Figure Apx F-8, EPA identified the following release points (note that exposure point 1
corresponds to diagram point A, exposure point 2 to diagram point B, and exposure point 3 to diagram
point C):
• Exposure point 1 (EP1): Exposure to vapors during container unloading and/or transferring;
• Exposure point 2 (EP2): Exposure to vapors during container cleaning; and
• Exposure point 3 (EP3): Exposure to vapors during equipment cleaning.
To calculate a full-shift TWA, the 1,4-dioxane concentrations calculated for each exposure point above
are multiplied by their respective exposure durations, then summed and divided by the total workday
duration (8 hours per the ESD).
Inhalation exposure to chemical vapors is a function of the chemical's physical properties, ventilation
rate of the container loading area, type of loading method, and other model parameters. Although
physical properties are fixed for a chemical, some model parameters are expected to vary from one
facility to another. An individual model input parameter could either have a discrete value or a
distribution of values. EPA assigned statistical distributions based on available literature data or
engineering judgment to address the variability in parameters such as ventilation rate (RATEventiiation)
and mixing factor (Fmixing).
A Monte Carlo simulation was then conducted to capture variability in the model input parameters
described above. The simulation was conducted using the Latin hypercube sampling method in @Risk
(Palisade, Ithaca, New York). The Latin hypercube sampling method is a statistical method for
generating a sample of possible values from a multi-dimensional distribution. Latin hypercube sampling
is a stratified method, meaning it guarantees that its generated samples are representative of the
probability density function (variability) defined in the model. EPA performed 100,000 iterations of the
model to capture the range of possible input values, including values with low probability of occurrence.
Page 499 of 570
-------
From the distribution resulting from the Monte Carlo simulation, EPA selected the 95th and 50th
percentile values to represent a high-end exposure and central tendency exposure level respectively. The
statistics were calculated directly in @Risk. The following subsections detail the model design equations
and parameters used for inhalation exposure estimates.
F.9.1 Model Equations
Duration of exposure for container unloading is calculated using the following equation:
EquationApx F-18.
OH,
N,
cont_unload_yr
cont_exposures
OD * RATEfill adjUSteci
Where:
OH,
cont_exposures
N,
cont_unlaod_yr
OD
RATE finad justed
Duration of exposure for container unloading [hours/day]
Number of containers unloaded annually, see Appendix E.13
[containers/site-year]
Operating days in a year, see Appendix E.13 [days/year]
Container fill rate that is adjusted so that the release duration does
not exceed 24 hours [containers/hour]
To make the simulation more realistic and account for subsequent exposure points 2 and 3, EPA set a
maximum exposure duration for container unloading (exposure point 1) of 2 hours per day, assuming
workers would not be unloading containers for a full shift. Therefore, the duration of exposure for
container unloading is adjusted with the following equation:
Equation Apx F-19.
If OHcont exposures ^ 2
OHcont eXp0Sures_adjusted ~ 2
If OH,
contexposures
<2
OHcont exp0sures_adjusted ~ OHcont exposures
Where:
OH,
cont_exposures_adjusted= Duration of exposure for container unloading adjusted so that it is
capped at 2 hours/day [hours/day]
OH,
cont_exposures
Duration of exposure for container unloading [hours/day]
Exposure point 1 (container unloading) volumetric concentration in air for 1,4-dioxane is calculated
using the EPA Mass Balance Inhalation Model shown in the following equation:
Page 500 of 570
-------
EquationApx F-20.
The lesser of:
Release,,
1000 g
Cv i —
3600 sec
In *
(1.7 x 105) * T * jj , " yRn ka
cont_unload_yr j
/ (OP * RATEfill) * hr
MW * RATEventuation
* Fmixing
Or
Cvi = (1 x 106) * x
clean_unload
VP
760
Where:
a
Volumetric concentration of 1,4-dioxane in air for exposure point 1
[PPm]
Ambient temperature [K]
Release point 1 daily releases, see Appendix E.13 [kg/site-day]
Number of containers used yearly, see Appendix E. 13
[containers/site-year]
Operating days, see Appendix E.13 [days/year]
Container fill/unloading rate, see Appendix E.13 [containers/hour]
1,4-dioxane molecular weight [g/mol]
Ventilation rate [ftVmin]
Mixing factor [unitless]
Vapor pressure correction factor for container unloading and
Cleaning, see Appendix E.13 [mol dioxane/mol water]
= Vapor pressure of 1,4-dioxane [torr]
Exposure point 1 (container unloading) mass concentration of 1,4-dioxane in air is calculated using the
following equation:
vl
Release_perDayRP1
Ncont_unlaod_yr
OD
RATEfui
MW
RATEventiiation
F ¦ ¦
1 mixing
X clean_unload
VP
Equation Apx F-21.
ConcentrationEP1 = C.
•vl
MW
Where:
ConcentrationEP1 = Mass concentration of 1,4-dioxane in air for exposure point 1
r /
c
Vl
MW
Vm
[mg/m3]
Volumetric concentration of 1,4-dioxane in air for exposure point 1
[ppm]
Molecular weight of 1,4-dioxane [g/mol]
Molar volume [L/mol]
Exposure point 2 (container cleaning) volumetric concentration in air for 1,4-dioxane is calculated using
the EPA Mass Balance Inhalation Model shown in the following equation:
Page 501 of 570
-------
EquationApx F-22.
The lesser of:
Release,,
1000 g
CV2 —
3600 sec
In *
(1.7 x 105) * T * jj , " ^ ka
cont_unload_yr j
/ (OP * RATEfill) * hr
MW * RATEventuation
* Fmixing
Or
Cv2 = (1 X 106) * X
clean_unload
VP
760
Where:
a
v2
Release_perDayRP3
Ncont_unlaod_yr
OD
RATEfui
MW
RATEventiiation
F ¦ ¦
1 mixing
X clean_unload
VP
Volumetric concentration of 1,4-dioxane in air for exposure point 2
[PPm]
Ambient temperature [K]
Release point 3 daily releases [kg/site-day]
Number of containers used yearly, see Appendix E. 13
[containers/site-year]
Operating days, see Appendix E.13 [days/year]
Container fill/unloading rate, see Appendix E.13 [containers/hour]
Molecular weight of 1,4-dioxane [g/mol]
Ventilation rate [ftVmin]
Mixing factor [unitless]
Vapor pressure correction factor for container unloading and
Cleaning, see Appendix E.13 [mol dioxane/mol water]
Vapor pressure of 1,4-dioxane [torr]
Exposure point 2 (container cleaning) mass concentration of 1,4-dioxane in air is calculated using the
following equation:
Equation Apx F-23.
ConcentrationEP2 = C.
v2
MW
Where:
ConcentrationEp2 = Mass concentration of 1,4-dioxane in air for exposure point 2
r /
c
v2
MW
Vm
[mg/m3]
Volumetric concentration of 1,4-dioxane in air for exposure point 2
[ppm]
Molecular weight of 1,4-dioxane [g/mol]
Molar volume [L/mol]
Exposure point 3 (equipment cleaning) volumetric concentration in air for 1,4-dioxane is calculated
using the EPA Mass Balance Inhalation Model shown in the following equation:
Page 502 of 570
-------
EquationApx F-24.
The lesser of:
(1.7 x io5)*r
5^ rr, ^e^eClSeperDayRp5
1000 g
kg
OH.
CV3 —
equip_clean
3600 sec
hr
MW * RATEventilation
* F'mixing
Or
Cv3 = (1 x 106) *xt
tank_clean
VP
760
Where:
C-,
v3
Release_perDayRPS =
OHeqU(p ciean —
MW
RATEventiiation —
p . . =
1 mixing
Xtank_clean ~
VP
Volumetric concentration of 1,4-dioxane in air for exposure point 3
[PPm]
Ambient temperature [K]
Release point 5 daily releases [kg/site-day]
Equipment cleaning operating hours [hours/day]
Molecular weight of 1,4-dioxane [g/mol]
Ventilation rate [ftVmin]
Mixing factor [unitless]
Vapor pressure correction factor for equipment and storage tank
cleaning [mol dioxane/mol water]
1,4-dioxane vapor pressure [torr]
Exposure point 3 (equipment cleaning) mass concentration of 1,4-dioxane in air is calculated using the
following equation:
Equation Apx F-25.
ConcentrationEP3 = *
Jv3
MW
Where:
ConcentrationEP3
MW
Vm
Mass concentration of 1,4-dioxane in air for exposure point 3
[mg/m3]
Volumetric concentration of 1,4-dioxane in air for exposure point 3
[ppm]
Molecular weight of 1,4-dioxane [g/mol]
Molar volume [L/mol]
The total vapor 8-hour TWA based on the mass concentrations of 1,4-dioxane for exposure points 1
through 3 is calculated using the following equation:
Equation Apx F-26.
T otal_V apor_TWA
{Concentration^ * 0Hcon^ eXpOSUres acijUS^eci + Concentvcition^p2 * 0Hcon^ eXpOSUres acijUS^eci + Coneentvcition* 0HeqUip Ciean))
Page 503 of 570
-------
Where:
TotalVapor _TWA =
ConcentrationEP1 =
ConcentrationEP2 =
ConcentrationEP3 =
OH,
cont_exposures_ad justed
OH,
equip_clean
Full-shift 8-hour TWA of 1,4-dioxane vapor exposure [mg/m3]
Mass concentration of 1,4-dioxane in air for exposure point 1
[mg/m3]
Mass concentration of 1,4-dioxane in air for exposure point 2
[mg/m3]
Mass concentration of 1,4-dioxane in air for exposure point 3
[mg/m3]
Duration of exposure for container unloading adjusted so that it is
capped at 2 hours/day [hours/day]
Duration equipment cleaning releases, see Appendix E.13
[hours/day]
F.9.2 Model Input Parameters
Table Apx F-38 summarizes the model parameters and their values for the Monte Carlo simulation.
Additional explanations of EPA's selection of the distributions for each parameter are provided after this
table. High-end and central tendency exposures are estimated by selecting the 50th and 95th percentile
values from the output distribution.
Page 504 of 570
-------
Table Apx F-38. Summary of Parameter Values and Distributions Used in the Hydraulic Fracturing Exposure Modeling
Input Parameter
Symbol
Unit
Constant Model
Parameter Values
Variable Model Parameter Values
Rationale/ Basis
Value
Lower
Bound
Upper
Bound
Mode
Distribution
Type
Ventilation Rate
RATEventilation
ft3/min
132,000
132,000
237,600
—
Uniform
See Section F.9.3
Mixing Factor
F mixing
none
0.1
0.1
1
0.5
Triangular
See Section F.9.4
Vapor Pressure of 1,4-dioxane
VP
Torr
40
—
—
—
—
Physical property
Molecular Weight of 1,4-
dioxane
MW
g/mol
88.1
—
—
—
—
Physical property
Ambient Temperature
T
K
298
—
—
—
—
Process
parameter
Universal Gas Constant
R
atm-
cm3/gmol-K
82.05
—
—
—
—
Universal
constant
Molar Volume
Vm
L/mol
24.45
—
—
—
—
Physical property
Page 505 of 570
-------
F.9.3 Ventilation Rate
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
the Agency used generic data from the CEB Manual ( ), which indicates general outdoor
ventilation rates in industry range from 132,000 to 237,600 ftVmin in outdoor conditions. The
underlying distribution of this parameter is not known; therefore, EPA assigned a uniform distribution,
since a uniform distribution is completely defined by range of a parameter.
F.9.4 Mixing Factor
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
the Agency used generic data from the CEB Manual ( ), which indicates mixing factors
may range from 0.1 to 1, with 1 representing ideal mixing. The CEB Manual references the 1988
ACGIH Ventilation Handbook which suggests the following factors and descriptions: 0.67 to 1 for best
mixing; 0.5 to 0.67 for good mixing; 0.2 to 0.5 for fair mixing; and 0.1 to 0.2 for poor mixing. The
underlying distribution of this parameter is not known; therefore, EPA assigned a triangular distribution,
which is completely defined by the range and mode of a parameter. The mode for this distribution was
not provided; therefore, EPA assigned a mode value of 0.5 based on the typical value provided in the
ChemSTEER User Guide ( ) for the EPA/OPPT Mass Balance Inhalation Model.
F.9.5 Key Strengths, Limitations, Uncertainties, and Sensitivity Analysis
General modeling uncertainties and limitations are discussed in Section 2.2.1.3 and Appendix E.7
regarding assigning distributions to input parameters, using generic data for some input parameter
distributions, and using static values for other input parameters also apply to the exposure modeling.
EPA ran a sensitivity analysis in @Risk to identify the input parameters which have the largest impact
on the 8-hour TWA occupational inhalation exposure estimate. FigureApx F-9 shows the inputs ranked
by which have the largest effect on the 8-hour TWA occupational inhalation exposure. Similar to the
sensitivity analysis for the daily release estimates in Appendix E.13.19, the mass fraction of 1,4-dioxane
in fracturing fluid additives received at sites and in the final fracturing fluid formulation that is injected
into the ground have the largest impact on the exposure estimate. These two mass fraction parameters
are based on 411 datapoints from FracFocus 3.0 and are paired, meaning that there is a correlation
between the two parameters. The annual use rate of fracturing fluids containing 1,4-dioxane, which also
impacts the exposure estimate, is similarly based on 411 datapoints from FracFocus 3.0. For all other
parameters in Figure Apx F-9, EPA developed distributions based on generic, not 1,4-dioxane-specific
data. Having a distribution for each input parameter is a strength of the assessment; however, the
representativeness of the underlying data used for these distributions towards is a limitation.
Page 506 of 570
-------
8-hr TWA of lf4-Dioxane
Inputs Ranked by Effect on Output Mean
Mass Fraction of 1,4-Dioxane in Fracturing Additive and
Fluid (paired)
Container size for fracturing fluids
Operating Days
Mixing Factor
Annual use rate of fracturing fluids containing 1,4-
dioxane
1
Saturation Factor
1
i
FigureApx F-9. Sensitivity Chart for 8-Hour TWA Occupational Inhalation Exposure to 1,4-
Dioxane at Hydraulic Fracturing Sites
F.10 Dish Soap and Dishwasher Detergent Modeling Approach and
Parameters for Estimating Occupational Inhalation Exposures
This appendix presents the modeling approach used to estimate occupational inhalation exposures to
1,4-dioxane during the industrial and commercial use of dish soaps and dishwasher detergents. This
approach utilizes standard EPA models combined with Monte Carlo simulation (a type of stochastic
simulation). Figure Apx F-10 is a diagram of the release and exposure points during the use of dish soap
and dishwasher detergent.
Page 507 of 570
-------
Unloading Dish Soap Worker Breathing
from Containers Zone from Washing
Container Disposal
FigureApx F-10. Environmental Release Points (Numbered) and Occupational Exposure Points
(Lettered) During Industrial and Commercial Use of Dish Soap & Dishwasher Detergent
Based on Figure Apx F-10, EPA identified the following exposure points (note that exposure point 1
corresponds to diagram point A and exposure point 2 to diagram point B):
• Exposure point 1 (EP1): Exposure to vapors during container unloading; and
• Exposure point 2 (EP2): Exposure to vapors during washing.
To estimate inhalation exposures to vapors, this model utilizes the previously modeled vapor releases for
each corresponding release point, as explained in Appendix E.14. To calculate a full-shift TWA, the 1,4-
dioxane concentrations calculated for each exposure point above are multiplied by their respective
exposure durations, then summed and divided by the total workday duration of 8 hours.
Inhalation exposure to chemical vapors is a function of the chemical's physical properties, ventilation
rate of the container loading area, type of loading method, and other model parameters. Although
physical properties are fixed for a chemical, some model parameters are expected to vary from one
facility to another. An individual model input parameter could either have a discrete value or a
distribution of values. EPA assigned statistical distributions based on available literature data or
engineering judgment to address the variability in parameters such as ventilation rate (RATEventiiation),
mixing factor (Fmixing), and mass fraction of 1,4-dioxane in the soap or detergent (Fdioxane soap/detergent),
A Monte Carlo simulation was then conducted to capture variability in the model input parameters
described above. The simulation was conducted using the Latin hypercube sampling method in @Risk
(Palisade, Ithaca, New York). The Latin hypercube sampling method is a statistical method for
generating a sample of possible values from a multi-dimensional distribution. Latin hypercube sampling
is a stratified method, meaning it guarantees that its generated samples are representative of the
probability density function (variability) defined in the model. EPA performed 100,000 iterations of the
model to capture the range of possible input values, including values with low probability of occurrence.
From the distribution resulting from the Monte Carlo simulation, EPA selected the 95th and 50th
percentile values to represent a high-end exposure and central tendency exposure level respectively. The
Page 508 of 570
-------
statistics were calculated directly in @Risk. The following subsections detail the model design equations
and parameters used for inhalation exposure estimates.
F.10.1 Model Equations
Vapor pressure correction factor for exposure points 1 and 2 is calculated using the equation below:
EquationApx F-27.
Fdioxane_
X
e_soap /detergentj
soap/detergent
MW
Fdioxane_soap / deter gent , 1 Fdioxane_soap / deter gent
MW + 18
Where:
X,
soap
X,
detergent
Fdioxane_soap
Fdioxane_detergent
MW
Vapor pressure correction factor for dish soap
[mol 1,4-dioxane/mol water]
Vapor pressure correction factor for dishwasher detergent
[mol 1,4-dioxane/mol water]
Mass fraction of 1,4-dioxane in dish soap [kg/kg]
Mass fraction of 1,4-dioxane in dishwasher detergent [kg/kg]
1,4-dioxane molecular weight [g/mol]
Exposure point 1 (container unloading) volumetric concentration of 1,4-dioxane is calculated using the
following equation:
Equation Apx F-28.
Exposure point 1 (container unloading) volumetric concentration in air for 1,4-dioxane is calculated
using the EPA Mass Balance Inhalation Model shown in the following equation:
The lesser of:
(1.7 x 105) *T
Release_perDay *
1000 g
kg
OH
Cv i —
unload_cont
3600 sec
hr
MW * RATEventilation
* F'mixing
Or
Cvl = (1 x 106) * x
soap/detergent
VP
760
Where:
Cv i
T
Release_perDayRP1 =
OHunioaci cont —
MW
RATE.
F ¦ ¦
1 mixing
ventilation
Volumetric concentration of 1,4-dioxane in air for exposure point 1
[ppm]
Ambient temperature [K]
Release point 1 daily releases, see Appendix E.14 [kg/site-day]
Daily operating hours for unloading containers [hours/day]
1,4-dioxane molecular weight [g/mol]
Ventilation rate [ftVmin]
Mixing factor [unitless]
Page 509 of 570
-------
Xsoap = Vapor pressure correction factor for dish soap
[mol 1,4-dioxane/mol water]
Xdetergent = Vapor pressure correction factor for dishwasher detergent
[mol 1,4-dioxane/mol water]
VP = Vapor pressure of 1,4-dioxane [torr]
Exposure point 1 (container unloading) mass concentration of 1,4-dioxane is calculated using the
following equation:
EquationApx F-29.
Cvl * MW
Concentration_VaporEP1 = —
nn
Where:
Concentration_VaporEP1= Mass concentration of 1,4-dioxane in air for exposure
point 1 [mg/m3]
Cvl = Volumetric concentration of 1,4-dioxane in air for exposure
point 1 [ppm]
MW = Molecular weight of 1,4-dioxane [g/mol]
Vm = Molar volume [L/mol]
Exposure point 2 (washing) volumetric concentration of 1,4-dioxane is calculated using the following
equation:
Equation Apx F-30.
The lesser of:
(1.7 x 105) *T
ReleasejperDay *
1000 g
kg
OH.
CV2 —
soap/dishwasher
3600 sec
hr
MW * RATE,
ventilation rmixing
* Fn
Or
Cv2 = ( 1X106)*X
soap/detergent
VP
760
Where:
CV2
T
Release_perDayRP3 =
OHsoap —
MW
RATE.
ventilation
mixing
Volumetric concentration of 1,4-dioxane in air for exposure point 2
[ppm]
Ambient temperature [K]
Release point 3 daily releases, see Appendix E.14 [kg/site-day]
Daily operating hours for hand washing [hours/day]
Daily operating hours for dishwasher operation [hours/day]
1,4-dioxane molecular weight [g/mol]
Ventilation rate [ftVmin]
Mixing factor [unitless]
Page 510 of 570
-------
Xsoap = Vapor pressure correction factor for dish soap
[mol 1,4-dioxane/mol water]
Xdetergent = Vapor pressure correction factor for dishwasher detergent
[mol 1,4-dioxane/mol water]
VP = Vapor pressure of 1,4-dioxane [torr]
Exposure point 2 (washing) mass concentration of 1,4-dioxane is calculated using the following
equation:
EquationApx F-31.
Cv2 * MW
Concentration_VaporEp2 = —
nn
Where:
Concentration_VaporEp2= Mass concentration of 1,4-dioxane in air for exposure
point 2 [mg/m3]
Cv2 = Volumetric concentration of 1,4-dioxane in air for exposure
point 2 [ppm]
MW = Molecular weight of 1,4-dioxane [g/mol]
Vm = Molar volume [L/mol]
The total full-shift vapor exposure (8-hour TWA) accounting for EP1 and EP2 is calculated using the
following equation:
Equation Apx F-32.
(Concentration_VaporEP1 * 0Hunload cont + Concentration_VaporEP2 * 0Hsoap/detergent)
Vapor_Exposure_TWA = - -—-
Where:
Vapor_Exposure_TWA = Full-shift TWA of 1,4-dioxane vapor exposure [mg/m3]
Concentration_VaporEP1= Mass concentration of 1,4-dioxane in air for exposure
point 1 [mg/m3]
Concentration_VaporEP2= Mass concentration of 1,4-dioxane in air for exposure
point 2 [mg/m3]
OHunioad cont = Daily operating hours for unloading containers [hours/day]
0Hsoap = Daily operating hours for hand washing [hours/day]
0Hdishwasher = Daily operating hours for dishwasher operation [hours/day]
F.10.2 Model Input Parameters
Table Apx F-39 summarizes the model parameters and their values for the Monte Carlo simulation.
Additional explanations of EPA's selection of the distributions for each parameter are provided after this
table. High-end and central tendency exposures are estimated by selecting the 50th and 95th percentile
values from the output distribution.
Page 511 of 570
-------
TableApx F-39. Summary of Parameter Values and Distributions Used in the Industrial and Commercial Use of Dish Soap and
Dishwasher Detergent Exposure Modeling
Input Parameter
Symbol
Unit
Constant Model
Parameter
Values
Variable Model Parameter Values
Rationale/Basis
Value
Lower
Bound
Upper Bound
Mode
Distribution
Type
Ventilation Rate
RATEventilation
ft3/min
3,000
500
10,000
3,000
Triangular
See Section F.10.3
Mixing Factor
F mixing
dimensionless
0.5
0.1
1
0.5
Triangular
See Section F.10.4
Vapor Pressure of 1,4-
Dioxane
VP
Torr
40
—
—
—
—
Physical property
Molecular Weight of 1,4-
Dioxane
MW
g/mol
88.1
—
—
—
—
Physical property
Ambient Temperature
T
K
298
—
—
—
—
Process parameter
Molar Volume
Vm
L/mol
24.45
—
—
—
—
Physical property
Page 512 of 570
-------
F.10.3 Ventilation Rate
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
the Agency used generic data from the CEB Manual ( ), which indicates general
ventilation rates in industry range from 500 to 10,000 ftVmin, with a typical value of 3,000 ftVmin. The
underlying distribution of this parameter is not known; therefore, EPA assigned a triangular distribution,
which is completely defined by the range and mode of a parameter. EPA assumed the mode is equal to
the typical value provided by the CEB Manual ( ).
F.10.4 Mixing Factor
EPA did not identify chemical-specific information for this parameter from systematic review; therefore,
the Agency used generic data from the CEB Manual ( ), which indicates mixing factors
may range from 0.1 to 1, with 1 representing ideal mixing. The CEB Manual references the 1988
ACGIH Ventilation Handbook that suggests the following factors and descriptions: 0.67 to 1 for best
mixing; 0.5 to 0.67 for good mixing; 0.2 to 0.5 for fair mixing; and 0.1 to 0.2 for poor mixing. The
underlying distribution of this parameter is not known; therefore, EPA assigned a triangular distribution,
which is completely defined by the range and mode of a parameter. The mode for this distribution was
not provided; therefore, the Agency assigned a mode value of 0.5 based on the typical value provided in
the ChemSTEER User Guide ( ) for the EPA/OPPT Mass Balance Inhalation Model.
F.10.5 Key Strengths, Limitations, Uncertainties, and Sensitivity Analysis
General modeling uncertainties and limitations are discussed in Section 2.2.1.3 and Appendix E.7
regarding assigning distributions to input parameters, using generic data for some input parameter
distributions, and using static values for other input parameters also apply to the exposure modeling.
EPA ran a sensitivity analysis in @Risk to identify the input parameters which have the largest impact
on the 8-hour TWA occupational inhalation exposure estimate. FigureApx F-l 1 shows the inputs
ranked by which have the largest effect on the 8-hour TWA occupational inhalation exposure for the use
of dish soaps. Figure Apx F-12 shows the same for the use of dishwasher detergents. The model uses a
mass balance approach, which is why the sensitivity charts show that the exposures are dependent on the
release estimates. Both figures show similar input parameter dependency; however, the exposure
associated with the use of dishwasher detergents is also dependent on the duration of time that the
dishwasher is open, as shown in Figure Apx F-12. This is different than the use of dish soaps, for which
exposure may occur during the entire manual dish washing process.
Similar to the sensitivity analysis for the daily release estimates in Appendix E. 14.18, the mass fraction
of 1,4-dioxane in soaps and detergents have the largest impact on the exposure estimates. This mass
fraction is based on 42 datapoints from literature sources, the December 2020 Final Risk Evaluation for
1,4-Dioxane, and product concentration waiver data from the NYDEC, as discussed in Appendix E. 14.5.
The use of this 1,4-dioxane-specific data from multiple different sources is a strength of the assessment.
For all other parameters in Figure Apx F-l 1 and Figure Apx F-12, EPA developed distributions based
on generic, not 1,4-dioxane-specific data. Having a distribution for each input parameter is a strength of
the assessment; however, the representativeness of the underlying data used for these distributions
towards is a limitation.
Page 513 of 570
-------
8-hr TWA
Inputs Ranked by Effect on Output Mean
Mass Fraction of Dioxane in Soap
Ventilation Rate
Mixing Factor
Daily Throughput of Dish Soap
Saturation Factor
Container Residual Loss Fraction
Container Size
Diameter of Sink Opening
II
1
[
1
I
1
FigureApx F-ll. Sensitivity Chart for 8-Hour TWA Occupational Inhalation Exposure to 1,4-
Dioxane at Sites Using Dish Soap
8-hr TWA
Inputs Ranked by Effect on Output Mean
Mass Fraction of Dioxane in Soap
Ventilation Rate
Mixing Factor
Daily Throughput of Dish Soap
Dishwashing Release Duration
H -J
1
Saturation Factor
Diameter of Sink (Dishwasher) Opening
Container Residual Loss Paction
Container Size
i
A
Figure Apx F-12. Sensitivity Chart for 8-Hour TWA Occupational Inhalation Exposure to 1,4-
Dioxane at Sites Using Dishwasher Detergents
Page 514 of 570
-------
Appendix G SURFACE WATER CONCENTRATIONS
G.l Surface Water Monitoring Data
G.l.l Monitoring Data Retrieval and Processing
The complete set of 1,4-dioxane monitoring results stored in the WQP was retrieved in July 2022, with
no filters applied other than the chemical name (NWQMC. 2022). This raw dataset included 12,471
samples. To filter down to only the desired surface water samples to include in this analysis, only
samples with the "ActivityMediaSubdivisionName" attribute of "Surface Water" were kept, and among
those, only samples with a "MonitoringLocationTypeName" that was one of the following:
• Spring
• Stream
• Wetland
• Lake
• Reservoir
• Impoundment
• Stream: Canal
• Stream: Ditch
• Facility Other
• Floodwater Urban
• River/Stream
• Great Lake
• Reservoir
• Lake
• River/Stream Intermittent
• River/Stream Perennial
After these steps, 1,449 surface water samples remained in the dataset. Samples flagged as QC blanks in
the "ActivityTypeCode" column were then removed, leaving 1,359 surface water samples for analysis.
Of these remaining samples, only 12 percent were results above the respective reported detection limit.
This monitoring dataset is attached as 1,4-Dioxane Supplemental Information File: WQP Processed
Surface Water Data ( 024w).
Monitoring data from drinking water systems were obtained from state drinking water databases (CA,
MA, NY) and the Third Unregulated Contaminant Monitoring Rule (UCMR3) results database (CA.
Water Board. 2022: NY DOH. 2022: Commonwealth of Massachusetts. 201S; 1 c< « ^ \
-------
water samples remained in the dataset, only 22 percent of which were results above the respective
reported detection limit.
The sampling design of the UCMR3 dataset includes all PWSs serving more than 10,000 people and 800
representative PWSs serving 10,000 or fewer people around the country. In California, monitoring and
reporting for 1,4-dioxane is currently voluntary. In Massachusetts, all community water systems (PWSs
that serve at least 25 people at their primary residences or with at least 15 connections to primary
residences) are required to monitor, while in New York all PWSs are required to monitor.
G.1.2 Raw and Finished Drinking Water
In analyzing drinking water monitoring data in Sections 2, 3, and 5, the conservative approach of
treating both raw water and finished drinking water samples as representing 1,4-dioxane concentrations
that could be served to PWS customers. The reason behind this is that the most common treatment
processes utilized by PWS do not effectively remove 1,4-dioxane. EPA acknowledges that even without
treatment to remove 1,4-dioxane, a PWS may utilize multiple sources of raw water, which could be
combined to dilute concentrations of 1,4-dioxane. An example case is apparent in monitoring data
retrieved from the state of Massachusetts.
Concurrent monitoring of raw and finished water at this PWS show that even with higher 1,4-dioxane
concentrations at a particular intake or source water body, concentrations can be reduced by mixing
(Figure Apx G-l). Despite this treatment facility not utilizing advanced treatment that could remove
1,4-dioxane from the treated water, the finished water contains lower concentrations than what would be
expected from the average concentration of raw water samples. This is due to multiple sources of water,
and a greater portion of the water with a lower concentration being used.
' 0.03 -h
Source
Finished Water
Source 1
Source 2
2013 2014 2015 2016
2017 2018 2019 2020
Sample Date
2021 2022 2023 2024
Figure Apx G-l. Example Raw and Finished Water Concentrations from a PWS Without
Processes to Remove 1,4-Dioxane
Some treatment processes can remove 1,4-dioxane from contaminated water sources (BroughtonetaL
2019; Godri Pollitt et al.. 2019; Otto and Nagaraia. 2007; U.S. EPA. 2006b). Advanced oxidation treatments (e.g.,
hydrogen peroxide with ferrous iron, ozone treatment with ultraviolet [UV] light, etc.) have substantially
lowered concentrations in treated water but may result in the formation of additional byproducts
Page 516 of 570
-------
(bromates) and may increase the cost of water treatment. Granular activated carbon has also lowered
1,4-dioxane concentrations when contaminated water is in the 10 |ig/L to 100 |ig/L range. Due to the
physical-chemical properties of the chemical substance (e.g., water solubility, octanol-water partitioning
coefficient) and the variable characteristics of granulated active carbon (e.g., pore-size distribution,
activation sites, and nonuniformity of lots), this treatment process does not consistently reduce 1,4-
dioxane concentrations in contaminated water (TableApx G-l).
EPA assessed the prevalence of treatment processes that may more consistently remove 1,4-dioxane
using treatment process information contained in the federal SDWIS database (see Table Apx G-2).
Less than one percent of community water systems (CWS) list oxidation processes which could more
reliably reduce 1,4-dioxane concentrations, representing about 14 percent of the population served
drinking water by CWSs.
Table Apx G-l. Summary of Community Water Systems with Treatment Processes Capable of
Removing 1,4-Dioxane
Process
Number of
CWS
Percent of All
CWS
Population Served
Count
Percent of Population
Served by CWS
Ozonation, Post
120
0.22
11,994,890
3.68
Ozonation, Pre
260
0.49
29,357,673
9.00
Peroxide
100
0.19
5,345,429
1.64
Activated Carbon, Granular
1,029
1.93
38,815,800
11.90
G.2 Surface Water Modeling
G.2.1 Hydrologic Flow Data
The NHDPlus V2.1 national seamless flowline network database was used as the source of stream or
river (hereby referred to as stream) flow data for both the facility-specific and aggregate probabilistic
modeling approaches. The NHD dataset is one of the largest national hydrologic datasets, containing
delineated flowline networks, flow sequence data, and associated modeled flow values for >2.7 million
stream segments ( 016c). The joint USGS-EPA data product represents one of the most
comprehensive and functional datasets that can be applied for national-scale hydrologic modeling
studies to date. The Enhanced Runoff Method (EROM) flow database, which includes modeled mean
annual flows, as well as mean monthly flows, for each stream segment in the national flow network, is
developed from multi-step process to estimate and calibrate hydrologic flows. This standard dataset is
incorporated into recordkeeping and modeling across EPA programs, providing consistency and
compatibility with projects such as EPA's ECHO database.
Lists of facilities discharging 1,4-dioxane directly and indirectly via transfers to disposal facilities were
collected from EPA's TRI and DMR databases, as described in Appendix E. For each direct release
facility, NPDES permit information associated with the facility's FRS Identification (FRS ID) was
pulled from the ECHO database API, including the 14-digit NHDPlus reach code. When a facility-
assigned reach code is missing in the ECHO database, the nearest neighboring NHD flowline and
associated reach code within a 2 km radius was identified using GIS software. This process was repeated
for the facilities reported as receiving indirect releases. The QE flow metrics from the EROM database
were used, which represent modeled flows adjusted according to observed flows at USGS flow
monitoring gages. QE values are reported by the user manual to be the "best EROM estimate of actual
mean flow." These modeled flows are based on observed flows from the years 1971 to 2000. The mean
annual and mean monthly modeled QE flows (QE) were extracted from the NHDPlus V2.1 database for
Page 517 of 570
-------
the identified reaches. An individual reach code may be associated with multiple individual flowline
segments in the NHDPlus V2.1 database, each with its own unique COMID identifier. Typically, there
is very little variation in flow metrics between segments that share a reach code. When multiple
segments were with the same reach code were returned during this process, the mean of each flow
metric was calculated and applied to the associated facility. In two instances, the water body associated
with a releasing facility was a lake or coastal water body, without a flow metric. In these cases, the
facility flow (described below) was used, if available. For facilities with no available hydrologic or
facility flow rates (or a modeled flow rate of zero), the lowest non-zero flow within the COU was used.
In addition to the receiving water body information, the Pollutant Loading Tool API was also queried
for available records of water discharge rates from each facility, for each year of release records. The
following facility flow data fields were acquired from the database: Facility Design Flow, Actual
Average Facility Flow, Average Facility Flow. The Average Facility Flow record is most commonly
available, and is preferentially selected to represent the facility flow, followed by the Actual Average
Facility Flow, and finally the Facility Design Flow.
For both the facility-specific and probabilistic modeling approaches, the flow of the receiving water
body is combined with a daily pollutant loading value to estimate a surface water concentration. For
each modeled scenario, before calculating this concentration, the hydrologic flow value is checked
against the best available facility flow. The modeled concentration is sensitive to the flow used in the
calculation, particularly when that flow is very small. In reality, a small stream receiving a large volume
of discharge would have its flow increased substantially by the facility flow rate and modeling the
concentration using only the small stream's flow rate would result in erroneously high concentrations.
When the facility flow is greater than the stream flow, the facility flow is used to calculate the resulting
concentration instead of the stream flow. If a facility flow is not available, the modeled stream flow is
used.
G.2.2 Facility-Specific Release Modeling
In previous TSCA risk evaluations, EPA applied the E-FAST 2014 tool to conduct facility-specific
modeling. In an effort to make the calculations more flexible and rapidly repeatable, rather than using
the E-FAST model directly, the formulas employed in E-FAST were written into an Excel workbook.
This allowed for the incorporation of the NHDPlus V2.1 flow data as a refinement of the methodology,
and for manual adjustments to parameters as needed. Therefore, facility-specific modeling was
conducted using the methodology and logic of the E-FAST 2014 tool, but in a deconstructed form that
provided an opportunity to update flow metrics to improve overall confidence in the resulting
concentrations.
In the past, E-FAST modeling for risk evaluations have used several flow metrics: the arithmetic mean
flow, the harmonic mean flow, the 30Q5 (lowest 30-day average flow that occurs in a 5-year period),
and 7Q10 (lowest 7-day average flow that occurs in a 10-year period). Of these flow metrics, only a
modeled arithmetic mean flow can be obtained from the EROM flow database. Without a national
dataset of these additional flow statistics with the resolution and reliability of the EROM dataset, due to
the challenges of modeling these values across the national dataset, an alternative method to estimate
these metrics consistent with our application of the E-FAST methodology was adapted for this modeling
effort. Regression equations from the E-FAST technical manual relating the arithmetic mean, harmonic
mean, 30Q5, and 7Q10 flows were used to solve for the desired metrics. In addition to an annual
arithmetic mean flow, the EROM database provides modeled monthly average flows for each month of
the year. While the EROM flow database represents averages across a 30-year time period, the lowest of
the monthly average flows was selected as a substitute for the 30Q5 flow used in modeling, as both
Page 518 of 570
-------
approximate the lowest observed monthly flow at a given location. The arithmetic mean and substitute
30Q5 flow were then plugged into the regression equations used by E-FAST to convert between flow
metrics and solved for the remaining terms:
7Q10 =
cfs 30Q5
MLD * 1.782
^1.0352
Where:
7Q10 = the modeled 7Q10 flow, in MLD
30Q5 = the lowest monthly average flow from NHD, in MLD
HM = 1.194 *
/ f \ 0.473 / f \ 0.552
[0A09WLD*AM) * (°-409 mid * 7
-------
Inputs:
Figure Apx G-2. Schematic of the General Fit-for-Purpose EWISRD-XL Model
The model produces an estimation of surface water concentrations at the downstream end of a stream
segment, by combining the total upstream mass flux and dividing by the downstream flow rate:
Cone,
down
(FlllXup + Fluxtrib + FluXfiTf) + FlV-Xreiease)
Flow,
down
Where:
Concdown = the 1,4-dioxane concentration at the downstream end (|ig/L)
Fluxup = the mass flux into the stream at the upstream end (|ig/day)
Fluxtrib = the mass flux into the stream from a tributary (|ig/day)
FluxDTD = the mass flux into the stream from DTD loading (|ig/day)
FluxReiease = the mass flux into the stream from a direct release (|ig/day)
Flowdown = the stream flow at the downstream end (L/day)
The EWISRD-XL model assumes that 1,4-dioxane stays within the water column as it travels
downstream, with no partitioning to sediment or air, and no biological uptake. The total mass flux into
the modeled reach is conserved and assumed to be equal to the mass flux out at the downstream end.
These assumptions are based on the physical chemistry properties (e.g., water solubility, Henry's Law
constant) and fate characteristics (e.g., biodegradability) and appear to represent the behavior of the
chemical fairly well over the relatively small distances covered by most of the case studies.
The mass flux from the upstream end of the segment, or a tributary, is calculated from a known flow rate
and concentration at that location:
Page 520 of 570
-------
FlllXupjrib CoYlCup^trib ( ^ * ^^^up,tribi.d^y^
The mass flux from DTD loading is calculated using a per capita estimation of loading and the estimated
population contributing to DTD loading of the stream segment. Although the upstream flux incorporates
all expected DTD loading from upstream of the segment being modeled, the DTD loading estimated by
the model captures the expected loading between the upstream and downstream ends of the modeled
segment. Per capita estimates of DTD loading are derived from SHEDS-HT model output, for nine
product types (TableApx G-2). The product mass ratios described in Appendix E were used as inputs to
the SHEDS-HT modeling, along with the default model parameters. The DTD component of the
SHEDS-HT output was isolated and evaluated for use in the EWISRD model. SHEDS-HT models non-
commercial consumer product use and reports a distribution of per capita DTD loading values. The
mean DTD loading value was applied in the EWISRD model to represent general non-commercial uses,
while the 90th percentile DTD loading value was applied to represent commercial uses of the same
products.
Table Apx G-2. Summary of per Capita DTD Loading Estimates from SHEDS-HT Modeling
Consumer Products
Non-commercial DTD Loading
(g/day per Capita)
Commercial DTD Loading
(g/day per Capita)
Antifreeze
0.0000
0.0000
Dish Soap
0.0235
0.2076
Dishwasher Detergent
0.0003
0.0046
Spray Polyurethane
0.0000
0.0000
Laundry Detergent
0.0004
0.0035
Surface Cleaner
0.0014
0.0209
Textile Dye
0.0000
0.0000
Floor Lacquer
0.0000
0.0000
Latex Wall Paint
0.0008
0.0000
For case study applications of the EWISRD-XL model, populations contributing to DTD loading within
the case study area were estimated using the 2020 Census Designated Places polygons and
accompanying population records ( sus Bureau. 2015). By visual inspection. Census places
alongside water bodies contributing flow to the stream segment of interest were identified, and the total
population was summed and entered into the EWISRD-XL model. The entirety of the estimated
population was assumed to be contributing to non-commercial DTD loading. The commercial DTD
loading was calculated using average proportions of the population expected to have occupations
resulting in commercial use of the consumer products, derived from the 2020 U.S. Bureau of Labor
Statistics Current Population Survey ( 12).
G.2.3.2 Case Studies to Validate Aggregate Model
Case studies of locations with adequate 1,4-dioxane surface water monitoring data were conducted with
the EWISRD-XL model, to validate the performance of the fit-for-purpose model (Table Apx G-3).
Rather than targeting a conservative estimate of release concentrations, the intention was to best
reproduce the observed monitored concentrations. Therefore, the modeled concentrations within the case
studies represent more average conditions for the time periods modeled. Overall, the application of the
EWISRD-XL model, which incorporated facility releases combined with DTD loading estimations
Page 521 of 570
-------
derived from SHEDS-HT, resulted in reasonable, if not conservative, estimates of average aggregate
concentrations of 1,4-dioxane downstream of multiple sources.
TableApx G-3. Summary of Case Study Locations Including Modeled and Observed Surface
Water Concentrations
Location
Modeled
Water Body
Modeled
Concentration (jig/L)
Observed
Concentration (jig/L)
Inputs Included
Brunswick
County, NC
Cape Fear
River
(range, see below)
(range, see below)
Direct industrial release,
DTD, and upstream
concentration
Columbia, TN
Duck River
0.35
<0.07-0.22
Only DTD
East Liverpool,
OH
Ohio River
0.61
<0.07
Direct industrial release,
DTD, and upstream
concentration
Brunswick County, NC - Cape Fear River
The Cape Fear River upstream of the Brunswick County, NC drinking water intake was selected as a
case study to test the model due to abundant monitoring data in the region (Figure Apx G-3). At the
upstream boundary of the modeled reach, approximately monthly monitoring data from 2017 to 2021 at
the Cape Fear River intake of the PWS in Fayetteville, NC was used to provide the concentration of 1,4-
dioxane at the upstream end of the model. The direct release from the DAK Americas LLC plant in
Fayetteville was included in the modeling (green dot in Figure Apx G-3). The daily loading from this
direct release was calculated as the average daily release from 250 days of operation, using the TRI
annual release records from 2017 to 2021, which ranged from 173 to 7,965 kg/year. In this case study,
the availability of concentrations of 1,4-dioxane in the Cape Fear River at Fayetteville, NC meant that
any DTD contributions from further upstream were already accounted for in the modeling, and therefore
only DTD loading between Fayetteville and Brunswick County needed to be quantified. The population
contributing to DTD loading was calculated by visually approximating the drainage area contributing to
the modeled segment using the NHD flowline network, from the upstream point near Fayetteville, NC to
the downstream endpoint near the Brunswick County intake and summing the 2020 Census populations
for the Census Designated Places within the boundary. At the downstream end, monitoring data,
reported as a minimum, average, and maximum concentration, from the Brunswick County drinking
water plant on the Cape Fear River were collected from Consumer Confidence Reports released by the
county for 2017 to 2021 (Brunswick Con 22).
Page 522 of 570
-------
Fayetteville
Vander
Raeford
Rockfish
'OAK
AMERICAS
LLC\
White Oak
Dublin
Elizabethtown
Brunswick
County
Kilometers
Census Designated Places
with Population Data
O Water Releasing Facilities
Modeled Stream Segment
Additional Tributaries
I I County Boundary
r 1 Approximate Contributing
Drainage Area
Figure Apx G-3. Map of Brunswick County, NC Model Case Study
Note: The downstream end of the modeled reach coincides with the location of the Brunswick County drinking
water intake on the Cape Fear River, which is located near where the Cape Fear River enters Brunswick County.
A separate model run was conducted for each measurement of 1,4-dioxane concentration in the Cape
Fear River near Fayetteville, NC (66 total), to incorporate more temporally-specific flow data and
produce a corresponding downstream modeled concentration at the Brunswick County intake. For each
year, the corresponding calculated average daily release from the DAK Americas LLC plant was
included as an input. For each month, the average corresponding monthly flow from NHDPlus V2.1 was
used for the upstream and downstream hydrologic flow inputs to the model. A static total contributing
population of 191,201 for the DTD component was used. Results from the 66 model runs were
compared with the values reported by Brunswick County (Figure_Apx G-4). The EWISRD-XL model
file used for this case study included as 1,4-Dioxane Supplemental Information File: EWISRDXL
BrunswickConntyNC Case Study (U.S. EPA. 2024p).
Page 523 of 570
-------
10.0 -
O)
3
•S 7.5
03
c
CD
O
c
o
O
CD
C
03
X
o
5.0-
2.5-
0.0-
Modeled Concentrations near
Brunswick County Intake
Range of Monitored Concentrations
at Brunsick County Treatment Plant
Max
o o
o
o
*> °o ^
O oQ
cP-
CP(P
p
'Mean
Min
2017
2018
2019
2020
2021
2022
Date
FigureApx G-4. Plot Comparing Results from Brunswick County Case Study Modeling
with Observed Concentrations
Modeled surface water concentrations generally fell within the ranges reported from monitored
concentrations. Wide ranges of both monitored and modeled values were noted, indicating variability
among inputs to the system. Although the direct discharge, DTD and flow components of the model
represent average daily or monthly values, finer-scale variations in these values could account for the
variability in monitored observations. In this case study, the upstream input concentration ranged from
less than 0.07 to 5.9 |ig/L, and the output was sensitive to this upstream concentration. Modeled
downstream concentrations could only be produced for days with available upstream concentrations, so
the full range of variability could not be captured in this approach. The overall modeled average
concentration from 2017 to 2021 was 1.35 |ig/L, and the annual averages for 2017 to 2021 reported by
Brunswick County ranged from 0.8 to 1.85 |ig/L. The general tendency of the model results to follow
the mean observed values reported from Brunswick County indicate that the assumptions of the model
and inputs effectively approximate resulting downstream concentrations of 1,4-dioxane resulting from
aggregate down the drain and facility releases.
Columbia, TN - Duck River
The Columbia, TN case study was selected because of available monitoring data from the Columbia
PWS located on the Duck River (Figure Apx G-5) with monitored detections of 1,4-dioxane reported
under UCMR3 (U.S. EPA 2017d). Its location near the headwaters of the Duck River meant that there
were no known upstream direct facility releases of 1,4-dioxane into this water body. Therefore, it was
assumed that any 1,4-dioxane in surface water detected in the Duck River at Columbia, TN, would be
due to the DTD contribution from the upstream population.
Page 524 of 570
-------
The upstream drainage area contributing flow to the Duck River at Columbia was visually approximated
from the NHD flowline network, and the 2020 populations of the identified Census Designated Places
were summed as inputs to the DTD component of the model (totaling 70,974 people). The mean annual
flow at the downstream end from NHDPlus V2.1 was used. The four reported measurements of 1,4-
dioxane at the Columbia PWS ranged from less than 0.07 (not detected) to 0.22 |ig/L. Because of the
static DTD inputs, a single model run was conducted using a mean annual flow rate, resulting in a
modeled concentration at the downstream end of 0.35 |ig/L. The intent of this case study was to target
the effectiveness of the model to estimate the DTD contribution to instream concentrations, and the
results suggest that the model assumptions for DTD loading are a reasonable but conservative estimate
of downstream concentrations. The EWISRD-XL model file used for this case study included as 1,4-
Dioxane Supplemental Information File: EWISRDXL ColumbiaTN Case Study (U.S. EPA. 2024q).
East Liverpool, OH
The case study for the Ohio River at East Liverpool, OH, was selected due to the availability of
coincident UCMR3 monitoring data (U.S. EPA. 2017d) and a known direct release from a facility
(Figure Apx G-6). For the sake of averaging reported monitoring measurements, half of the reported
detection limit of 0.07 |ig/L was applied for non-detects. At the upstream end of the model, the average
concentration measured at the Pittsburgh, PA, PWS of eight samples from 2014 to 2015 (via UCMR3)
was used (0.23 |ig/L). An additional tributary, the Beaver River, was included in the model using
UCMR3 monitoring data from Beaver Falls, PA. The average concentration of four samples from 2013
to 2014 reported from the Beaver Falls PWS was 2.66 |ig/L. In this case study, the availability of
concentrations of 1,4-dioxane in the Ohio River at Pittsburgh, PA, and the Beaver River at Beaver Falls
meant that any DTD contributions from further upstream were already accounted for in the modeling,
Page 525 of 570
-------
and therefore only DTD loading between these locations and East Liverpool, OH, needed to be
quantified. These PWS locations can be approximated by the points representing the respective towns
and cities in FigureApx G-5. The 2020 population from the Census Designated Places within the
approximate drainage area were summed for the DTD component, totaling 559,505 people. Annual
releases were only available for 2018 and 2019 from the BASF Corp facility, ranging from 2.98 to 3.66
kg. The average daily loading from this facility was calculated from the greater of these two numbers
divided by 250 days of operation.
Figure Apx G-6. Map of the East Liverpool, OH, Case Study
All four of the reported sample results at East Liverpool, OH, from 2013, were below the detection limit
of 0.07 |ig/L. The modeled concentration from all of the inputs resulted in 0.61 |ig/L at the downstream
end, which appears to be an overestimation for this system, based on the monitoring data. Due to the
timing of samples at the upstream and downstream ends not aligning, average values were used in this
case study, but some temporal variation may still be missed by these values. Additionally, results of this
case study appeared to be sensitive to the high concentrations reported for the Beaver River tributary as
well as the high population estimated to be contributing to the DTD component. The DTD component
was found to result in a small overestimation in the second case study, where the contributing population
was nearly an order of magnitude lower. The EWISRD-XL model file used for this case study is
included as 1,4-Dioxcme Supplemental Information File: EWISRDXL LiverpoolOH Case Study (U.S.
EPA. 2024r).
G.2.3.3 The Probabilistic Model
The probabilistic EWISRD-XL-R model was developed by creating an R script that interfaces with the
EWISRD-XL document (via the XLConnect R library (Mirai Solutions GmbH 2021)). In this
arrangement, the underlying modeling and calculation process is handled within an EWISRD-XL
document. The accompanying R script handles the loading and arrangement of input data, then
Page 526 of 570
-------
iteratively feeds inputs to the model and retrieves the outputs, which are then summarized and
visualized. This allows the EWISRD-XL-R model to run thousands of iterations rapidly.
Although the individual facility modeling focused on only the highest releases per facility, using the
EWISRD-XL-R model allowed the inclusion multiple years of release data available for each facility,
and inclusion of multiple flow estimations, to produce a range of estimated concentrations resulting
from releases. The ability to aggregate multiple inputs within the model also allowed the inclusion of
background 1,4-dioxane concentrations expected to be present in waterways from DTD or other
unregulated sources.
The EWISRD-XL-R model, as applied for the COU-specific probabilistic model, has four major
components:
1. Load and prepare the background concentration data.
Although the model is capable of estimating DTD loading directly from contributing populations,
there is some uncertainty about the distances over which the assumptions inherent in this calculation
remain accurate (including assumptions of persistence in the water column, the rates of DTD
loading, and that the entire upstream population contributes to the DTD loading). Furthermore,
although estimating the population contributing to specific reaches is viable for a case study, that
information is not readily available for each facility release. For these reasons, the background
component of the probabilistic modeling is estimated using the concentrations detected at PWSs.
The background data used to inform this estimation (Figure 2-9) only includes monitoring data for
PWSs that were not found to be located downstream from known 1,4-dioxane releasing facilities, in
order to represent only concentrations from DTD loading and other unregulated releases.
To appropriately pair background data with releasing facilities, the background concentrations and
facilities were stratified by the Strahler stream order of the associated NHDPlus stream reach. For
each stream reach, an empirical cumulative distribution function (ECDF) was created using the
Kaplan-Meier method, which has been recommended for estimating the distributions of datasets,
particularly with a high percent censored data (Gillespie et at.. ). The ecdfPlotCensored function
within the EnvStats R library is called to develop each ECDF (Millard. 2013). which is then wrapped
in a solver function for the inverse of the ECDF. The inverse ECDF solver function can then receive
an input of a percentile and return the corresponding background concentrations from the
distribution. A random value from the stream-order-specific background distribution can be
generated by calling the inverse ECDF solver function with a single input value from a random
uniform distribution between 0 and 1.
2. Load and combine the facility release and flow data.
As described in Section B.2.1, stream flow data (mean annual, and mean monthly for each month of
the year) are retrieved for each releasing facility, as well as facility flow data. For the probabilistic
modeling, all available years of release data, from both TRI and DMR, are loaded into the model,
and the monthly flow averages from NHDPlus are joined to them.
3. Perform a loop of model runs per COU.
The Monte Carlo simulations are then conducted with 10,000 model iterations per COU. In each
model iteration, a random facility within the COU group and a random year of release is selected. Of
the 12 available monthly average flows associated with that facility, one is randomly selected. If the
selected flow rate is less than the facility flow rate, the facility flow rate is used instead. For the
stream order of the reach associated with the releasing facility, a random background concentration
Page 527 of 570
-------
is generated. The resulting combination of daily loading from a direct release, flow, and upstream
concentration are then passed to the EWISRD-XL document as inputs. For the probabilistic
modeling, the resulting concentration is calculated at the point of release, so the EWISRD-XL model
is configured in an arrangement where the downstream flow is equal to the upstream flow. The
resulting stream concentration, as well as the percent contributions of the direct release and the
background loading to that concentration, are retrieved from the model outputs and logged. The
resulting output table records 10,000 combinations of modeled concentrations from different flow,
release, and background concentration combinations. For each iteration, the total stream
concentration (facility release + background) and the stream concentration due to only the facility
release are recorded. A schematic of the flow of data within the probabilistic model is presented in
Figure_Apx G-7.
1. Within COU, Select Facility
1
Year
Release
(kg)
2014
100
2015
S-L
1 —
2016
1
75
2017
90
Facilities
Stream
Order
A
2
B
3
C
7
D
5
Month
Flow
(cfs)
1 1200
2
1400
3
1700
3. Generate Background Concentration 4. Send Inputs to EWISRD-XL
Based on Stream Order
Concentration
sr
OES
Facility
Release (kg)
Flow
(cfs)
Background
Concentration
(ms/l)
Total
Concentration
(HS/L)
1
B
125
1700
0.15
0.27
1
C
20
10000
0.01
0.01
1
A
500
8000
0.2
0.30
1
D
100
2500
0.5
0.57
1
A
650
5000
0.4
0.61
2. From Facility Data, Select Release Amount and Flow Rate
5. Compile Results from Each Iteration
FigureApx G-7. Schematic of the Flow of Data within the EWISRD-XL-R Probabilistic Model
4. Summarize and visualize the model output.
The model outputs are then summarized as percentiles and visualized as histograms. A comparison of
the modeled facility release and the randomly generated background concentration is conducted for each
iteration and summarized. This additional check can indicate whether, within a given COU, the expected
concentrations in surface water due to permitted releases from facilities are typically greater than the
expected background concentration from DTD and other non-regulated releases. The EWISRD-XL-R
script is included as 1,4-Dioxcme Supplemental Information File: EWISRD-XL-R Probabilistic Model
Code (U.S. EPA 2024a).
G.2.3.4 Modeling Ranges of DTD Contributions
The SHEDS-HT model was applied to generate distributions of DTD loading per capita resulting from
products listed in Table Apx G-4. The default scenarios and variables included with version 0.1.9 of
SHEDS-HT were used. Product weight fractions generated during the engineering phase of this risk
evaluation were used as inputs to the modeling. For each product, 10,000 iterations of the model were
Page 528 of 570
-------
run to generate a distribution of results. Only the DTD component of each set of exposure results was
pulled from the generated results, and products of the same type were summed together to summarize
the per capita DTD loading by product type (TableApx G-4).
TableApx G-4. Distribution of per Capita DTD Loading, in G/Day, by Product, for Non-
commercia
Uses Modeled by SHE]
DS-HT
Product
Q10%
Q25%
Q50%
Q75%
Q90%
Q99%
Mean
SD
Overall Relative
Contribution to
DTD Loading
Antifreeze
0
0
0
0
0
0
0
0
0%
Dish Soap
0
0
9.97E-03
2.70E-02
5.81E-02
2.08E-01
2.35E-02
5.04E-02
88%
Dishwashing
Detergent
0
0
0
5.33E-05
8.65E-04
4.63E-03
3.06E-04
1.06E-03
1%
SPF
0
0
0
0
0
0
0
0
0%
Surface
Cleaner
0
0
0
4.57E-04
4.09E-03
2.09E-02
1.43E-03
4.84E-03
6%
Laundry
Detergent
0
0
1.50E-04
4.57E-04
1.03E-03
3.53E-03
4.01E-04
8.86E-04
2%
Dye
0
0
0
0
0
0
0
0
0%
Floor
Lacquer
0
0
0
0
0
0
0
0
0%
Paint
0
0
0
0
0
0
7.87E-04
2.88E-02
3%
SHEDS-HT models consumer (non-commercial) uses of products, so the mean per capita DTD loading
output from the model was applied to represent the average non-commercial per capita DTD loading. To
represent increased usage by commercial applications, the 99th percentile per capita DTD loading was
applied for commercial uses. The number of commercial users of products was determined using the
national average proportion of the population expected to be employed in the following occupations,
based on the 2020 U.S. Bureau of Labor Statistics Current Population Survey (TableApx G-5) (
22).
Table Apx G-5. Proportions of Population Expected to Contribute to DTD Loading through
Commercial Activities and Product Uses
Product
Occupation
Proportion of Population
Antifreeze
Automotive service technicians and mechanics
0.00225
Dish Soap
Dishwashers
0.00055
Dishwasher Detergent
Dishwashers
0.00055
Spray Polyurethane
Insulation workers
0.00015
Surface Cleaner
Janitors and building cleaners
0.00615
Laundry Detergent
Laundry and dry-cleaning workers
0.00036
Surface Cleaner
Maids and housekeeping cleaners
0.00350
Textile Dye
Textile machine setters, operators, and tenders
4.82E-05
Floor Lacquer
Carpet, floor, and tile installers and finishers
0.00051
Latex Wall Paint
Painters and paperhangers
0.00157
Page 529 of 570
-------
To assess the potential range of concentrations resulting from DTD loading from various populations,
the above DTD loading values were applied to a range of population sizes, from 100 to 1,000,000
people. These loadings were applied to a range of mean annual flows, from 300 to 30,000 cfs, as
individual runs of the EWISRD-XL-R model. No other input sources were including in the modeling, so
that the resulting surface water concentrations were entirely due to the DTD loading. Although the
largest populations would be expected to discharge wastewater (i.e., from a POTW), to a larger
receiving water body, the full range of combinations of flow and contributing populations was analyzed.
G.2.3.5 Modeling Concentrations in Surface Water from Hydraulic Fracturing
The potential concentrations in surface water adjacent to hydraulic fracturing operations were modeled
from the distribution of loadings to surface water and stream flow data for reaches located near
hydraulic fracturing operations.
A set of 10,000 random values from the Monte Carlo distribution described in Appendix F.7, was
generated to represent the range of loading values to surface water. These values were generated by
employing a method similar to the generation of random values from background distributions described
in Appendix G.2.3.4. The paired percentile and loading values from the Monte Carlo results were used
to establish an empirical cumulative distribution function, for which the inverse could then be solved. A
uniform distribution of percentile values between 0 and 1 were input into the resulting function to
generate the 10,000 loading values used for this analysis.
Mapped well locations of hydraulic fracturing operations reporting 1,4-dioxane in the wastewater re
retrieved from the Fracfocus database (GWPC and IOGCC. 2022). To identify stream segments near the
hydraulic fracturing operations, which can take place across large areas, a 5 km buffer was drawn
around each well. Flow data from the 2,053 NHDPlus v2.1 stream segments intersecting these buffers
were collected and reviewed. Of the reaches identified, 76 percent were found to have modeled mean
annual flows less than 10 cfs (Figure Apx G-8).
300-
200 -
O
Ł=
CD
=3
O*
CD
^ 1 oo H
o-
10
-3
L±l
lL
_Li
1—1
I I II I III I I L
I....|
10
10
-1
10u 10 10
Mean Annual Flow (cfs)
10J
10"
10"
Figure Apx G-8. Distribution of Mean Annual Modeled Flow Rates for NHDPlus V2.1
Reaches Identified Within 5 km of Hydraulic Fracturing Wells Reporting 1,4-Dioxane
Although the volumetric rate of discharge from hydraulic fracturing operations to surface water were not
readily available, it was assumed that the concentrations in receiving streams with flows less than 10 cfs
Page 530 of 570
-------
would be substantially impacted by the volume of wastewater from the operation. For this analysis,
flows below 10 cfs were excluded from the pool of flows selected for modeling. Based on the
assumption that the rates of discharge from hydraulic fracturing operations are greater, these lower flows
would result in unrealistically high estimates of stream concentrations resulting from these releases.
From the remaining 486 flow rates, 10,000 values were randomly sampled with replacement.
The EWISRD-XL-R model was used to model the concentrations resulting from the 10,000 generated
loading values paired with the 10,000 stream flow rates (Figure Apx G-9). Due to the nature of using a
Monte Carlo distribution to generate the release loadings, and the sensitivity to the results of handling
the nearby stream flow data, the tails of this distribution (i.e., the highest and lowest percentiles) have a
high degree of uncertainty.
1500-
c 1000-
(D
=3
o-
-------
FigureApx G-10. Generic Schematic of Hypothetical Release Point
with Surface Water Intakes for Drinking Water Systems Located
Downstream
An R script was developed to search downstream from the reach codes with facilities, using the node
and reach code sequence information within NHDPlus. The script functions by incrementally stepping
downstream to the next reach and evaluating whether a surface water intake is associated with the reach
code. When a reach with an intake is identified, the details of the PWS and the distance traveled
downstream are recorded, and the script continues until a dead end, or a maximum search distance in
achieved for each release. For this assessment, a maximum search length of 500 reaches (approximately
1,000 km) was used.
Overall, about 31 percent of individual facilities found to have an adult lifetime cancer risk for drinking
water above 1 x 10~6 were located within 250 km upstream from a known DW). It should be noted that
risk estimates are calculated for concentrations in the receiving water at the point of release, and some
decrease in concentration due to dilution would be expected at the location of a DWI further
downstream. For all OESs other than Functional fluids and Printing inks, at least one facility was located
within 250 km upstream of a known DWI. Among Industrial Uses, Manufacture, and Remediation, five
facilities were located within 10 km upstream of a known DWI. For most facilities identified as being
located upstream from any DWI, multiple downstream DWIs were identified.
Page 532 of 570
-------
TableApx G-6. Summary of Proximity of Downstream Drinking Water Intakes to Releasing
Facilities Resulting in Modeled Risk above 1E-06
Facilities with Risk above 1 E-06 and DWI
Downstream
OES
Total
Facilities
Evaluated
Facilities with Lifetime
Adult Cancer Risk
above 1E-06
Within
250 km
Within
100 km
Within
50 km
Within
25 km
Within
10 km
Disposal
25
9
4
4
2
1
0
Ethoxylation
byproduct
8
4
2
1
0
0
0
Functional fluids
(open-system)
6
2
0
0
0
0
0
Import and
repackaging
12
11
3
3
3
2
0
Industrial uses
32
21
3
3
2
1
1
Manufacture
2
2
2
2
2
2
2
PET
manufacturing
23
18
5
5
4
1
0
Printing inks
1
1
0
0
0
0
0
Remediation
16
3
3
3
3
2
2
Total
125
71
22
21
16
9
5
To consider the types of waterways potentially used as source water and susceptible to contamination,
an additional assessment of reaches associated with intakes was conducted. This simple assessment
examined the mean annual flow in NHDPlus V2.1 for each of the reaches matched as being the closest
to a drinking water intake. The resulting distribution (Figure Apx G-l 1)
Page 533 of 570
-------
TO
O
O
<1)
C
S m
400-
o _c
O "g 200
o
O"
<1)
=!=~
10 10 10 10 10 10 10 10"
Annual Average NHDPIus Flow (cfs)
10
10
FigureApx G-ll. Summary Distribution of Mean Annual Flow at Stream Reaches Matched with
Drinking Water Intakes
As described in Section 5.2.2.1.2, the degree of dilution between the initial receiving water body at the
point of release and a downstream drinking water intake was estimated by calculating the ratio of mean
annual NHDPIus flows at both locations. The ranges of dilution (as a percent of the concentration at the
point of release) ranged from much less than 1 to 100 percent and are presented in TableApx G-7
alongside the ranges of diluted downstream harmonic mean concentrations, which ranged from
1.63xl0~4 to 1.27xl04. These diluted concentrations were used to develop exposure and risk estimates,
presented in Table Apx G-8.
Table Apx G-7. Ranges of Dilution and Diluted 1,4-Dioxane Concentrations Modeled at Drinking
Water Intakes Downstream of Industrial Releases
Diluted Concentration as a
Percent of Concentration at
Point of Release (%)
Modeled Harmonic Mean
Concentrations at Downstream
Intakes (jug/L)
Distance
Range
(km)
Number of
Facilities" with
DWI
Downstream
Number of
PWS with
Downstream
Intakes
Min.
Median
Max
Min.
Median
Max
0-10
4
4
<1
1
100
1.63E-02
3.92E-01
1.27E04
10-25
4
7
<1
<1
68
4.42E-02
1.51E-01
8.28E00
25-50
7
8
<1
<1
92
1.81E-03
2.74E-02
3.03E00
50-100
10
15
<1
<1
31
4.42E-03
1.50E-01
2.07E02
100-250
15
57
<1
<1
100
1.63E-04
7.47E-02
1.52E02
11 Only facilities with an adult lifetime cancer risk for drinking water above 1 in a million were included in this analysis.
DWI = drinking water intake; PWS = public water system
Page 534 of 570
-------
TableApx G-8. Ranges of LADD and Adult Lifetime Cancer Risk Estimates for Diluted 1,4-
Dioxane Concentrations Modeled at Drinking Water Intakes Downstream of Industrial Releases
Diluted LADD (mg/kg-day)
Diluted Adult Lifetime Cancer Risk
Distance
Range (km)
Minimum
Median
Maximum
Minimum
Median
Maximum
0-10
7.6E-08
1.8E-06
4.1E-02
9.1E-09
2.1E-07
4.9E-03
10-25
1.4E-07
4.8E-07
2.6E-05
1.7E-08
5.8E-08
3.2E-06
25-50
5.8E-09
9.7E-08
9.7E-06
6.9E-10
1.2E-08
1.2E-06
50-100
1.4E-08
4.8E-07
6.6E-04
1.7E-09
5.7E-08
7.9E-05
100-250
5.2E-10
2.4E-07
4.8E-04
6.2E-11
2.9E-08
5.8E-05
There are important limitations and uncertainties in this analysis. The extent of dilution is highly
variable and is driven by site-specific factors that cannot be fully captured in this national-scale analysis.
This analysis is based on the conservative assumption that the only decrease in concentration is due to
dilution, and the effects of diffusion, advection, or dispersion are not modeled. Additionally, while flows
within a river or stream generally increase in the downstream direction, infrastructure like dams and
water withdrawal activities can lead to decreases in downstream flows. In lieu of a more robust model to
assess each release on a case-by-case basis, this approach allows a rapid assessment of estimated ranges
of dilution. Overall confidence in risk estimates is high for drinking water intakes located at or near the
point of release, but confidence decreases substantially with increasing distance downstream. This
analysis does not provide a comprehensive survey of modeled 1,4-dioxane concentrations at all drinking
water intakes. There may be additional drinking water intakes downstream of facilities releasing 1,4-
dioxane that are not accounted for in the intake database used in this analysis.
Page 535 of 570
-------
Appendix H GROUNDWATER CONCENTRATIONS AND
DISPOSAL PATHWAYS FROM LAND RELEASES
H.l Groundwater Monitoring Data Retrieval and Processing
The complete set of 1,4-dioxane monitoring results stored in the WQP was retrieved in July 2022, with
no filters applied other than the chemical name (NWQMC. 2022). This raw dataset included 12,471
samples. To filter down to only the desired groundwater samples to include in this analysis, only
samples with the "ActivityMediaSubdivisionName" attribute of "Groundwater" were kept, and among
those, only samples with a "MonitoringLocationTypeName" that was one of the following:
• well;
• subsurface;
• subsurface: groundwater drain; and
• well: multiple wells.
After these steps, 8,046 groundwater samples remained in the dataset. Samples flagged as QC blanks in
the "ActivityTypeCode" column were then removed, leaving 7,583 groundwater samples for analysis.
Of these remaining samples, only 30 percent (n = 2,284) were results above the respective reported
detection limit.
H.2 Review of Land Release Permits
EPA reviewed all Underground Injection Class I Permits to understand if sites were in accordance with
regulations. The sites and the corresponding release year, registry number, and disposal weight is
available in TableApx H-l for on-site disposal and TableApx H-2 for off-site.
TableApx H-l. Release Year, TRI Facility ID, Facility Name, State, Registry Number, Disposal
Type, and Disposal Weight for
Dn-Site Class I
Jnderground Injection Wells According to TRI
Release
Year
TRI Facility ID
Facility Name
State
Registry
Number
Disposal
Type
Disposal
Weight (lb)
2019
77536DSPSL2525B
TM DEER
PARK
SERVICES LP
Texas
RN100209568
On-Site
23,098
2018
77536DSPSL2525B
TM DEER
PARK
SERVICES LP
Texas
RN100209568
On-site
23,604
2017
77536DSPSL2525B
TM DEER
PARK
SERVICES LP
Texas
RN 100209568
On-site
23,024
2016
77536DSPSL2525B
TM DEER
PARK
SERVICES LP
Texas
RN 100209568
On-site
12,867
2015
77536DSPSL2525B
TM DEER
PARK
SERVICES LP
Texas
RN 100209568
On-site
94,304
2014
77536DSPSL2525B
TM DEER
PARK
SERVICES LP
Texas
RN 100209568
On-site
731,892
Page 536 of 570
-------
Release
Year
TRI Facility ID
Facility Name
State
Registry
Number
Disposal
Type
Disposal
Weight (lb)
2013
77536DSPSL2525B
TM DEER
PARK
SERVICES LP
Texas
RN100209568
On-site
371,877.95
TableApx H-2. Release Year, Source TRI Facility ID, Source State, Receiving Facility RCRA ID,
State, Disposal Type, and Disposal Weight for Off-Site Class I Underground Injection Wells
According to TRI and RCRAInfo Databases
Release
Year
Source TRI Facility
ID
Source
State
Receiving Facility
RCRA ID
Receiving
State
Disposal
Type
Disposal
Weight (lb)
2019
44044RSSNC36790
Ohio
OHD020273819
Ohio17
Off-site
0.009
2019
29448GNTCMPOBOX
South
Carolina
OHD020273819
Ohio17
Off-site
2
2018
29448GNTCMPOBOX
South
Carolina
OHD020273819
Ohio17
Off-site
23
" The state of Ohio provides an overview of its underground iniection wells via the Ohio Environmental Protection Asencv.
EPA reviewed all RCRA Subtitle C Permits to understand if sites were in accordance with regulations.
The sites and the corresponding release year, registry number, and disposal weight is available in
Table Apx H-3 for on-site disposal and Table Apx H-4 for off-site.
Table Apx H-3. Release Year, TRI Facility ID, Facility Name, State, CERCLIS ID, Disposal Type,
and Disposal Weight for RCRA Subtitle C Landfills According to TRI
Release
Year
TRI Facility ID
Facility Name
State
FRS ID
Disposal
Type
Disposal
Weight (lb)
2015
97812CHMCL17629
Chemical Waste
Management of the
Northwest INC
Oregon
110002059904'7
On-site
13,368.40
2014
97812CHMCL17629
Chemical Waste
Management of the
Northwest INC
Oregon
110002059904'7
On-site
16,108.10
2013
97812CHMCL17629
Chemical Waste
Management of the
Northwest INC
Oregon
110002059904'7
On-site
15,400.30
" This facility has several violation and compliance issues. The facility was fined $25,000 in 2020 for non-compliance
activities. The fine is attributed to inadequate coverage for third party bodily injury and property damage claims. The
facility self-reported in 2021 that another compliance issue had been detected.
Page 537 of 570
-------
TableApx H-4. Release Year, Source TRI Facility ID, Source State, Receiving Facility RCRA ID,
State, Disposal Type, and Disposal Weight for Off-Site Class I Underground Injection Wells
Release
Year
Source TRI Facility ID
Source
State
Receiving Facility
RCRA ID
Receiving
State
Disposal
Type
Disposal
Weight (lb)
2019
84029SFTYK11600
Utah
UTD991301748fl
Utah
Off-site
0.08
2018
84029SFTYK11600
Utah
UTD991301748fl
Utah
Off-site
0.01
2015
84029SFTYK11600
Utah
UTD991301748fl
Utah
Off-site
0.1488
2016
77536SFTYK2027B
Texas
OKD065438376
Oklahoma
Off-site
0.03
2015
77536SFTYK2027B
Texas
OKD065438376
Oklahoma
Off-site
0.16
2019
69145CLNHR5MISO
Nebraska
COD9913004846
Colorado
Off-site
0.29
2018
69145CLNHR5MISO
Nebraska
COD9913004846
Colorado
Off-site
13.29
2017
69145CLNHR5MISO
Nebraska
COD9913004846
Colorado
Off-site
55.49
2019
66736SYSTCCEMEN
Kansas
OKD065438376
Oklahoma
Off-site
750
2019
66736SYSTCCEMEN
Kansas
ALD0006224640
Alabama
Off-site
750
2019
44044RSSNC36790
Ohio
MID048090633d
Michigan
Off-site
0.011
2015
44044RSSNC36790
Ohio
MID000724831d
Michigan
Off-site
0.005
2014
44044RSSNC36790
Ohio
MID000724831d
Michigan
Off-site
0.008
2014
43920VNRLL1250S
Ohio
MID000724831d
Michigan
Off-site
30.2
2013
43920VNRLL1250S
Ohio
MID048090633d
Michigan
Off-site
17
2015
44044RSSNC36790
Ohio
OHD0452437068
Ohio
Off-site
0.001
2014
44044RSSNC36790
Ohio
OHD0452437068
Ohio
Off-site
0.002
2014
43920VNRLL1250S
Ohio
IND093219012f
Indiana
Off-site
72.6
2013
43920VNRLL1250S
Ohio
IND093219012/
Indiana
Off-site
44
"This facility was found to be non-compliant by the state in 2021 and was fined $20,575. The fine was associated with a
formal administrative enforcement action asserting that a remedial action is required.
b This facility was found to have significant non-compliance from 2020 to 2021. The facility was fined $12,000 in 2021.
The fine was associated with a formal administrative enforcement action asserting that a remedial action is required.
c This facility was found to be a significant non-complier by the state in 2020, 2021, and 2022. The facility has been fined a
total of $22,650. The fine was associated with a formal administrative enforcement action asserting that a remedial action is
required.
''These two facilities are likely the same as they have the same address.
e This facility has received written informal notices in 2017, 2018, 2019, and 2021. No enforcement actions have occurred.
' This facility was found to be a significant non-complier by the state from 2015 to 2021; the facility has been fined a total
of $77,385. The fine is associated with a Consent Agreement and Final Order between Region 5 and Heritage
Environmental Services, LLC. Heritage violated its permit, the Indiana Administrative Code, and RCRA and its
implementing regulations by (1) disposing of hazardous waste in the Roachdale landfill without meeting certain land
disposal restriction (LDR) treatment standards; (2) failing to conduct post-treatment verification sampling and analysis of
certain waste streams from two stabilization/LDR treatment processes; (3) failing to obtain a detailed chemical and physical
analysis of representative samples from such waste streams; (4) failing to follow the acceptable analytical methods in its
waste analysis plan (WAP); and (5) failing to determine the proper extraction fluid for TCLP analysis.
H.3 Landfill Analysis Using DRAS
DRAS is an efficient tool developed by EPA Region 6 to provide a multipath risk assessment for the
evaluation of Resource Conservation and Recovery Act (RCRA) hazardous waste delisting. For the
Supplemental Evaluation to the 1,4-dioxane Risk Evaluation, DRAS was specifically applied to model
Page 538 of 570
-------
groundwater concentration estimates from disposing 1,4-dioxane to a hypothetical RCRA Subtitle D
landfill at a range of loading rates and leachate concentrations. A comprehensive description of the
assumptions and calculations applied in DR.AS can be found in the Technical Support Document for the
Hazardous Waste Delisting Risk Assessment Software. It is worth noting that the underlying
assumptions for DRAS are the same as those for EPA's Composite Model for Leachate Migration with
Transformation Products (EPACMTP) described in Appendix Section H.4.
Because DRAS derives calculations based on a survey of drinking water wells located downgradient
from waste management units (U.S. EPA. 1988). the model may provide the closest estimate to real
world scenarios available. Though there is some uncertainty inherent to applying the model as an
assessment tool under the Toxic Substance Control Act (TSCA) for risk evaluations, few other tools are
available to effectively address this pathway. This appendix will provide the input variables and
calculations used to apply the model determine potential groundwater concentrations. TableApx H-5
and Table Apx H-6 provide the input values used for each parameter in the model. Note that loading
volumes were based on the range of TRI release weights and were calculated based on the density of
1,4-dioxane at 20 °C (1.0329 g/cm3). For each loading volume, the range of leachate concentrations was
applied.
Table Apx H-5. Input Variables for Chemical of Concern
Input Variable for Chemical of Concern
Value
Chem Name
1,4-Dioxane
CASRN
123-91-1
Maximum Contaminant Level
0
Oral Slope Cancer Factor
0.1a
Inhalation Slope Cancer Factor (1/mg kg day)
0.018a
Oral Reference Dose (mg/kg day)
0.03a
Inhalation Reference Dose (mg/kg day)
0.03a
Bioconcentration Factor (1/kg)
0.3698
Soil Saturation Level
0
Toxicity Regulatory Rule regulatory level (mg/L)
0 a
Henry's Law Constant (atm -m3/mol)
4.25E-06
Diffusion coefficient in Water (cm2/s)
1.05E-05
Diffusion coefficient in Air (cm2/s)
0.092a
Water Solubility (mg/L)
1,000,000
Landfill Dilution Attenuation Factor
15.4
Surface Impoundment Dilution Attenuation Factor
3.18
Time to Skin Attenuation (hr/event)
0.72a
Skin permeability constant (cm/hour )
0.00029a
Lag time (hr)
0.3a
Bunge constant
4.1E-05a
Organic
Yes
Bioaccumulation Factor (L/kg)
0 a
Chronic Ecological Value (mg/L)
0 a
Page 539 of 570
-------
Input Variable for Chemical of Concern
Value
Carcinogen
Yes
Molecular Weight (g/mol)
88.1
Vapor Pressure (atm)
0.05
Suspended sediment-surface water partitioning
coefficient (mg/L)
0.0549
log Kow (log[mg/l])
-0.27
Chemical Class
voca
Analytical Method
8260Da
Version Description
Nonea
Create Date
Nonea
Creator
Nonea
Cancer Risk Level
1.00E-06a
Hazard Quotient
^a
a Input variables do not directly or indirectly affect groundwater concentrations
Table Apx H-6. Waste Management Unit (WMU) Properties
Input Variable for WMU Properties
Value(s)
Waste Management Unit Type
Landfill
4.39E-07
4.39E-06
4.39E-05
4.39E-04
Loading Volume (m3)
4.39E-03
4.39E-02
4.39E-01
4.39E00
4.39E01
4.39E02
Cancer Risk Level
1.00E-06
Hazard Quotient
1.0
Detection Limit
0.5
Waste Management Active Life (Years)
20
0.0001
0.001
TCLP Concentration (mg/L)/ Total
0.01
Concentration (mg/kg)
0.1
1
10
Page 540 of 570
-------
Input Variable for WMU Properties
Value(s)
100
1,000
10,000
Once the model was executed for each loading rate and leachate concentration scenario, the groundwater
concentration was calculated using the leachate concentration and the 90th percentile weight-adjusted
dilatation attenuation factor using the equation:
Qyy _ Leachate Concentration
c Weight Adjusted DAF '
Where:
GWC = Groundwater concentration
Leachate concentration = Input variable for the waste management unit
Weight Adjusted DAF = Weight adjusted dilution attenuation factor.
The results of these analyses are provided in Table 2-14.
H.4 Landfill Analysis Using EPACMTP
EPACMTP is a fate and transport model developed by EPA to simulate the release of constituents from
waste managed in land disposal units, and the subsequent impacts of these constituents to the subsurface
environment. The model combines two modules to simulate one-dimensional downward flow and
transport of constituents in the unsaturated zone beneath a waste disposal unit, as well as ground water
flow and three-dimensional constituent transport in the underlying saturated zone. The model is
designed to run in a probabilistic or deterministic mode and comes with built-in distributions of national
and regional modeling parameters. The output of the model includes estimated concentrations of
constituents arriving at a downgradient well under steady-state conditions or as a function of time.
Because EPACMTP derives calculations from based on a survey of drinking water wells located
downgradient from a waste management unit ( 988). the model may provide the closest
estimate to real world scenarios available. Though there is some uncertainty inherent to applying the
model as an assessment tool under TSCA for risk evaluations, few other tools are available to effectively
address this pathway. This appendix will provide the input variables and calculations used to apply the
model determine potential groundwater concentrations. More comprehensive information about the
assumptions and calculation embedded in the EPACMTP model can be found online.
EPA ran the model under two scenarios. In one scenario, it is assumed that the waste management unit is
an unlined landfill. In the other, it is assumed the waste management unit is a clay-lined landfill. In
addition to these details, chemical specific input variables are required. For 1,4-dioxane, these included
molecular weight (88.1 g/mole), water solubility (10,000 mg/L), Koc (17.0 g/L), rate of abiotic
hydrolysis (0.0 mortar-1), rate of biodegradation (0.0 mol 'year 'X and temperature (25 °C).
Similarly, initial concentration of the chemical substance was an input and ranged from 1 x 10~4 to 1 x 104
(Table Apx H-7). All other variables in the input files were left in their defaults. Each scenario requires
a separate input file provided with the executable file package. All files for running the executable
model were stored in same folder.
Page 541 of 570
-------
TableApx H-7. Potential Groundwater Concentrations (mg/L) Based on Disposal of 1,4-Dioxane
to Unlined and Clay-Lined Landfills as Assessed by Applying the EPACMTP Model
Leachate
Concentration
(mg/L)
Type of Liner
No Liner
With Clay Liner
Percentile
(n = 10,000)
Average Groundwater
Concentration (mg/L)
Percentile
(n = 10,000)
Average Groundwater
Concentration (mg/L)
0.0001
0
0
0
0
10
0
10
0
25
0
25
0
50
0
50
0
75
3.79E-10
75
7.37E-11
80
6.83E-09
80
4.04E-09
85
3.3E-08
85
2.55E-08
90
1.29E-07
90
8.92E-08
95
7.93E-07
95
7.41E-07
100
3.42E-05
100
3.34E-05
0.001
0
0
0
0
10
0
10
0
25
0
25
0
50
0
50
0
75
3.35E-09
75
8.57E-10
80
2.75E-08
80
1.71E-08
85
1.63E-07
85
8.29E-08
90
1.4E-06
90
7.64E-07
95
8.01E-06
95
7.43E-06
100
0.000342
100
0.000334
0.01
0
0
0
0
10
0
10
0
25
0
25
0
50
0
50
0
75
1.02E-08
75
4.14E-09
80
1.07E-07
80
5.23E-08
85
1.62E-06
85
6.82E-07
90
1.4E-05
90
7.64E-06
95
8.01E-05
95
7.43E-05
100
0.003415
100
0.00334
0.1
0
0
0
0
10
0
10
0
25
0
25
0
50
0
50
0
75
3.57E-08
75
1.38E-08
80
1.05E-06
80
2.77E-07
85
1.62E-05
85
6.82E-06
90
0.00014
90
7.64E-05
95
0.0008
95
0.000743
100
0.03415
100
0.0334
Page 542 of 570
-------
Leachate
Concentration
(mg/L)
Type of Liner
No Liner
With Clay Liner
Percentile
(n = 10,000)
Average Groundwater
Concentration (mg/L)
Percentile
(n = 10,000)
Average Groundwater
Concentration (mg/L)
1
0
0
0
0
10
0
10
0
25
0
25
0
50
0
50
0
75
1.18E-07
75
4E-08
80
1.02E-05
80
2.76E-06
85
0.000161
85
6.81E-05
90
0.001395
90
0.000764
95
0.00793
95
0.007429
100
0.3415
100
0.334
10
0
0
0
0
10
0
10
0
25
0
25
0
50
0
50
0
75
1.22E-06
75
1.89E-07
80
0.000105
80
2.77E-05
85
0.001622
85
0.000682
90
0.01395
90
0.00764
95
0.08004
95
0.07429
100
3.415
100
3.394
100
0
0
0
0
10
0
10
0
25
0
25
0
50
0
50
0
75
1.21E-05
75
1.89E-06
80
0.001046
80
0.000277
85
0.01622
85
0.006816
90
0.1442
90
0.0764
95
0.8499
95
0.7429
100
34.15
100
34.77
1,000
0
0
0
0
10
0
10
0
25
0
25
0
50
0
50
0
75
0.000121
75
1.89E-05
80
0.01046
80
0.002773
85
0.1692
85
0.06816
90
1.537
90
0.764
95
9.076
95
7.589
100
341.5
100
347.7
10,000
0
0
0
0
Page 543 of 570
-------
Leachate
Concentration
(mg/L)
Type of Liner
No Liner
With Clay Liner
Percentile
(n = 10,000)
Average Groundwater
Concentration (mg/L)
Percentile
(n = 10,000)
Average Groundwater
Concentration (mg/L)
10
0
10
0
25
0
25
0
50
0
50
0
75
0.00123
75
0.000189
80
0.1066
80
0.02773
85
1.819
85
0.6816
90
15.78
90
7.851
95
90.89
95
76.19
100
3,415
100
3,477
Note: The results are a product of Monte Carlo analysis and are organized by leachate concentration (mg/L),
percentile, and average concentration of 1,4-dioxane at a well within 1 mile of the disposal facility.
H.5 Surface Impoundment Analysis for the Disposal of Hydraulic
Fracturing Produced Water Using DRAS
The Delisting Risk Assessment Software (DRAS) is an efficient tool developed by U.S. Environmental
Protection Agency (EPA) region 6 to provide a multipath risk assessment for the evaluation of Resource
Conservation and Recovery Act (RCRA) hazardous waste delisting. For the Supplemental Evaluation to
the 1,4-dioxane Risk Evaluation, DRAS was specifically applied to model groundwater concentration
estimates from disposing 1,4-dioxane in produced waters from a hydraulic fracturing operation to a
hypothetical RCRA Surface Impoundment at a range of loading rates and leachate concentrations. A
comprehensive description of the assumptions and calculations applied in DRAS can be found in the
Technical Support Document for the Hazardous Waste Delisting Risk Assessment Software.
Because the model derives calculations from based on a survey of drinking water wells located
downgradient from a waste management unit ( 988). the model may provide the closest
estimate to real world scenarios available. Although there is some uncertainty inherent to applying the
model as an assessment tool under TSCA for risk evaluations, few other tools are available to effectively
address this pathway. This appendix will provide the input variables and calculations used to apply the
model determine potential groundwater concentrations. TableApx H-8 and TableApx H-9 provide the
input values used for each parameter in the model. Note that loading volume were based on the range of
TRI release weights and was calculated based on the density of 1,4-dioxane at 20 °C (1.0329 g/cm3). For
each loading volume, only one potential concentration was applied.
Table Apx H-8. Input Variables for Chemical of Concern
Input Variable for Chemical of Concern
Value
Chem Name
1,4-Dioxane
Chem CASRN
123-91-1
Maximum Contaminant Level
0
Oral Slope Cancer Factor
0.1
Inhalation Slope Cancer Factor (1/mg kg day)
0.018
Oral Reference Dose (mg/kg day)
0.03
Page 544 of 570
-------
Input Variable for Chemical of Concern
Value
Inhalation Reference Dose (mg/kg day)
0.03
Bioconcentration Factor (L/kg)
0.369
Soil Saturation Level
0
Toxicity Regulatory Rule regulatory level (mg/L)
0
Henry's Law Constant (atm -m3/mol)
4.25E-06
Diffusion coefficient in Water (cm2/s)
1.05E-05
Diffusion coefficient in Air (cm2/s)
0.092
Water Solubility (mg/L)
1,000,000
Landfill Dilution Attenuation Factor
15.4
Surface Impoundment Dilution Attenuation Factor
3.18
Time to Skin Attenuation (hour/event)
0.72
Skin permeability constant (cm/hour )
0.00029
Lag time (hours)
0.3
Bunge constant
4.1E-05
Organic
Yes
Bioaccumulation Factor (L/kg)
0
Chronic Ecological Value (mg/L)
0
Carcinogen
Yes
Molecular Weight (g/mol)
88.1
Vapor Pressure (atm)
0.05
Suspended sediment-surface water partitioning
coefficient (mg/L)
0.0549
log Kow (log[mg/L])
-0.27
Chemical Class
voc
Analytical Method
8260D
Version Description
None
Create Date
None
Creator
None
Cancer Risk Level
1.00E-06
Hazard Quotient
1
Page 545 of 570
-------
Table Apx H-9. Waste Management Unit
Input Variable for WMU Properties
Value(s)
Waste Management Unit Type
Surface Impoundment
Loading Volume (m3)
1734
193
67.1
15.1
3.48
0.0334
1.09E-08
Cancer Risk Level
1.00E-06
Hazard Quotient
1.0
Detection Limit
0.5
Waste Management Active Life (Years)
50
TCLP Concentration (mg/L)/Total
Concentration (mg/kg)
0.06
Once the model was executed for each loading rate and leachate concentration scenario, the groundwater
concentration was calculated using the leachate concentration and the 90th percentile weight-adjusted
dilatation attenuation factor using the equation:
Qyy _ Leachate Concentration
c Weight Adjusted DAF '
Where:
GWC = Groundwater concentration
Leachate concentration = Input variable for the waste management unit
Weight Adjusted DAF = Weight-adjusted dilution attenuation factor.
The results of these analyses are provided in Table 2-15.
Page 546 of 570
-------
Appendix I DRINKING WATER EXPOSURE ESTIMATES
Potential acute and chronic drinking water exposures were estimated based on surface water
concentrations estimated in Section 2.3.1 and groundwater concentrations estimated in Section 2.3.2.
Acute and chronic drinking water exposures used to evaluate non-cancer risks are estimated as an Acute
Dose Rate (ADR) or Average Daily Dose (ADD), respectively. Lifetime exposures used to evaluate
cancer risks are estimated as a Lifetime Average Daily Dose (LADD). The equations used to calculate
each of these exposure values are:
ADR =
( DWT\
SWC x (l - x IRdw x RD x CF1
BW x AT
ADD =
( DWT\
SWC x (l -x IRdw x ED x RD x CF1
BW x AT x CF2
LADD =
f DWT\
SWC x (l - x IRdw x ED x RD x CF1
BW x AT x CF2
Where:
SWC
DWT
IRdw
RD
ED
BW
AT
CF1
CF2
Surface water concentration (ppb or |ig/L)
Removal during drinking water treatment (°A
Drinking water intake rate (L/day)
Release days (days/year for ADD, LADD and LADC; 1 day for ADR)
Exposure duration (years for ADD, LADD and LADC; 1 day for ADR)
Body weight (kg)
Exposure duration (years for ADD, LADD and LADC; 1 day for ADR)
Conversion factor (l.OxlCT3 mg/|ig)
Conversion factor (365 days/year)
Inputs for body weight, averaging time (AT), and exposure duration were applied the same across the
evaluation of drinking water, incidental oral exposure, and incidental dermal exposure, but are described
here. For all calculations, mean body weight data were used from Chapter 8, Table 8-1 in the Exposure
Factors Handbook (EFH) ( ). To align with the age groups of interest, weight averages
were calculated for the infant age group (birth to <1 year) and toddlers (1-5 years). The ranges given in
the EFH were weighted by their fraction of the age group of interest. For example, the EFH provides
body weight for 0 to 1 month, 1 to 3 months, 3 to 6 months, and 6 to 12 months. Each of those body
weights were weighted by their number of months out of 12 to determine the weighted average for an
infant 0 to 1 year old. For all ADR calculations, the AT is 1 day, and the days of release are assumed to
be 1 according to the methodology used in E-FAST 2014 (I. c. « ^ \ JO I I). For all ADD calculations,
the AT and the ED are both equal to the number of years in the relevant age group up to the 95th
percentile of the expected duration at a single residence, 33 years ( ). For example,
estimates for a child between 6 and 10 years old would be based on an AT and ED of 5 years. For all
LADD and LADC calculations, the AT is based on a lifetime of 78 years, and the ED is the number of
years of exposure in the relevant age group, up to 33 years. EPA considered the impact of assuming a
longer exposure duration and determined that LADDs for a full 78 years of exposure would be 2.26
Page 547 of 570
-------
times greater than those calculated for 33 years (after accounting for age-specific differences in drinking
water intakes).
Drinking water exposure was estimated for the following age groups: Adult (21+ years), Youth (16-20
years), Youth (10-15 years), Child (6-10 years), Toddler (1-5 years), and infant (birth to <1 year).
Drinking water intake rates are provided in the 2019 update of Chapter 3 of the Exposure Factors
Handbook(U ,S. EPA. 2019a). Weighted averages were calculated for acute and chronic drinking water
intakes for adults 21+ years and toddlers 1 to 5 years. From Table 3-17 in the Handbook, 95th percentile
consumer data were used for acute drinking water intake rates. From Table 3-9 in the Handbook, mean
per capita data were used for chronic drinking water intake rates. The 95th percentile water intake values
from Table 3-9 of the Handbook vary by age group and range up to approximately 3 to 4 times higher
ingestion than the mean values used. Averaged across all age groups, the 95th percentile ingestion rates
averaged across all ages are 3.7 times greater than mean ingestion rates.
1.1 Surface Water Sources of Drinking Water
To estimate drinking water exposures that may result from surface water contamination, EPA used water
concentrations estimated in Section 2.3.1. Concentrations in estuaries or bays are not considered as they
are unlikely to be potable waters. Drinking water exposures are also not considered for large lakes due to
high uncertainty in the applicable dilution factors. This is in alignment with the methodology used in E-
FAST 2014 (U.S. EPA. 2014)
ADR or acute exposure concentrations used the modeled stream concentrations with the lowest monthly
flow rate while the ADD, LADD, and LADC or chronic calculations used the modeled harmonic mean
stream concentrations. Drinking water treatment removal (DWT) was set to 0 percent to represent a
conservative estimate of possible drinking water exposures.
1.2 Groundwater Sources of Drinking Water
To estimate drinking water exposures that may result from groundwater contamination, EPA used
groundwater concentrations estimated in Section 2.3.2.
Chronic and lifetime exposures (ADD and LADD) were calculated based on groundwater concentrations
estimated using the DRAS model. Acute exposures to groundwater were not calculated because the
available models EPA used for estimating groundwater concentrations are designed to predict long-term
trends rather than short peaks in exposure. DWT was set to 0 percent for groundwater under the
assumption that home wells are unlikely to remove 1,4-dioxane.
Page 548 of 570
-------
Appendix J AIR EXPOSURE PATHWAY
J.l Ambient Air Concentrations and Exposures
EPA applied a tiered approach to estimate ambient air concentrations and exposures for members of the
general population that are in proximity (between 5-10,000 m) to emissions sources emitting the
chemicals being evaluated to the ambient air (Figure Apx J-l). All exposures were assessed for the
inhalation route only.
Figure Apx J-l. Summary of Methodologies Used to Estimate Ambient Air
Concentrations and Exposures
J.l.l Ambient Air: Screening Methodologies and Results Summary - Fenceline
The Ambient Air: Screening Methodology identifies, at a high level, if there are inhalation exposures to
select populations from a chemical undergoing risk evaluation which indicates a potential risk. This
methodology inherently includes both estimates of exposures as well as estimates of risks to inform the
need, or potential need, for further analysis. If findings from the Ambient Air: Screening Methodology
indicate any potential risk (acute non-cancer, chronic non-cancer, or cancer) for a given chemical above
(or below as applicable) typical Agency benchmarks, EPA generally will conduct a higher-tier analysis
of exposures and associated risks for that chemical. If findings from the Ambient Air: Screening
Methodology do not indicate any potential risks for a given chemical above (or below as applicable)
typical agency benchmarks, EPA would not expect a risk would be identified with higher-tier analyses,
but may still conduct a limited higher-tier analysis at select distances to ensure potential risks are not
missed (e.g., at distances <100 m to ensure risks do not appear very near a facility where people may be
exposed).
Page 549 of 570
-------
Model
The Ambient Air: Screening Methodology utilizes EPA's IIOAC model to estimate high-end and central
tendency (mean) exposures to select receptors at three pre-defined distances from a facility releasing a
chemical to the ambient air (100, 100 to 1,000, and 1,000 m). IIOAC is an Excel-based tool that
estimates indoor and outdoor air concentrations using pre-run results from a suite of dispersion scenarios
run in a variety of meteorological and land-use settings within EPA's AERMOD. As such, IIOAC is
limited by the parameterizations utilized for the pre-run scenarios within AERMOD (meteorologic data,
stack heights, distances, receptors, etc.) and any additional or new parameterization would require
revisions to the model itself. Readers can learn more about the IIOAC model, equations within the
model, detailed input and output parameters, pre-defined scenarios, default values used, and supporting
documentation by reviewing the IIOAC users guide ( 2).
Releases
EPA modeled exposures from two release values for 1,4-dioxane. These values were extracted from
2019 TRI data as follows:
1. The maximum individual facility 1,4-dioxane release value among all facilities reporting releases
of 1,4-dioxane to TRI.
2. The average (mean) 1,4-dioxane release value across all facilities reporting 1,4-dioxane to TRI.
A summary of the releases evaluated for TRI reporting facilities is provided in TableApx J-l.
TableApx J-l. Release Estimates from 2019 TRI Used for Ambient Air: Screening Methodology
for 1,4-Dioxane
Number of
Operating Days
Maximum Facility Release
Average Facility Release
Pounds
(lb)
Kilograms
(kg)
kg/site-
day
Pounds
(lb)
Kilograms
(kg)
kg/site-day
365
10,442
4,735.601
12.97
792
359.184
0.98
260
18.21
1.38
Exposure Scenarios
EPA developed and evaluated a series of exposure scenarios for the max and mean 1,4-dioxane release
values identified above. The scenarios were designed to capture a variety of release types, topography,
meteorological conditions, and release scenarios as presented in Figure Apx J-2. It includes a total of 16
different exposure scenarios, each of which is applied to both the maximum and mean 1,4-dioxane
release value resulting in a total of 32 exposure scenarios modeled.
Page 550 of 570
-------
Release Type
CD
>
61)
3
U.
Topography
Urban
Rural
Urban
Stack
Rural
L West North Central
South Coastal
West North Central
South Coastal
West North Central
Meteorological Data
Release Scenario
365 days, 24/7
South Coastal
260 days, 8/5
365 days, 24/7
260 days, 8/5
365 days, 24/7
South Coastal
260 days, 8/5
365 days, 24/7
West North Central
260 days, 8/5
365 days, 24/7
260 days, 8/5
365 days, 24/7
260 days, 8/5
365 days, 24/7
260 days, 8/5
365 days, 24/7
260 days, 8/5
FigureApx J-2. Exposure Scenarios Modeled for Max and Mean Release Using IIOAC Model for
Ambient Air: Screening Methodology
EPA modeled exposure scenarios for two source types: stack (point source) and fugitive (area source)
releases. These source types have different plume and dispersion characteristics accounted for
differently within the IIOAC model. The topography represents an urban or rural population density and
certain boundary layer effects (like heat islands in an urban setting) that can affect turbulence and
resulting concentration estimates at certain times of the day.
IIOAC includes 14 pre-defined climate regions (each with a surface station and upper-air station). Since
release data used for the Ambient Air: Screening Methodology was not facility or location specific, EPA
selected 2 of the 14 climate regions to represent a central tendency (West North Central) and high-end
(South [Coastal]) climate region. This selection was based on a sensitivity analysis of the average
concentration and deposition predictions. The two climate regions selected represent meteorological data
sets that tended to provide high-end and central tendency concentration estimates relative to the other
stations within IIOAC. The meteorological data within the IIOAC model are from years 2011 to 2015 as
that is the meteorological data utilized in the suite of pre-run AERMOD exposure scenarios during
development of the IIOAC model (see IIOAC users guide ( )). While this is older
meteorological data, sensitivity analyses related to different years of meteorological data found that
although the data does vary, the variation is minimal across years so the impacts to the model outcomes
remain relatively unaffected.
The release scenarios consider two potential facility operating conditions. The first represents a facility
that operates year-round (365 days/year, 24/7). The second represents a facility that operates generally
on a Monday through Friday schedule (260 days/year) for 8 hours per day, 5 days per week. The
Page 551 of 570
-------
difference between the two release scenarios is the resulting total daily release, frequency of release, and
duration of release. These conditions result in a different exposure pattern that is captured by modeling
both release scenarios. As an example, if a facility has a total annual release of 10,000 lb/year, then the
daily release from a facility operating 365 days/year, 7 days per week, and 24 hours per day would be
27.4 lb per day for every day of the year over a 24-hour period. If the facility operates 260 days per year,
5 days per week, for 8 hours per day, the daily release would be 38.5 lb per day, but only Monday
through Friday and only over an 8-hour period.
Exposure Results and Risks
Modeled exposure concentration results from the Ambient Air: Screening Methodology modeling effort
were reviewed and summarized for each scenario modeled. To ensure potential risks were not missed,
EPA selected the highest estimated exposure concentrations from the 32 scenarios modeled for 1,4-
dioxane for use in risk calculations. These values were used to estimate the MOE and excess cancer risk.
The calculated risks were then compared to screening level benchmarks (POD-specific benchmark
MOEs for non-cancer risks and 1x 10~6 for general population cancer risk). Overall, the Ambient Air:
Screening Methodology did not identify risk relative to benchmark values for non-cancer risks but did
identify risk estimates above the benchmark value for cancer for three of the four release scenarios
summarized. Because the results from this methodology indicate potential risks to people near a
releasing facility, EPA conducted additional, higher-tier analyses to apply more COU and site-specific
data and results to further analyze exposures and associated potential risks resulting from such
exposures.
TableApx 3-2. Exposure and Risk Estimates from the Ambient Air: Screening Methodology for
Receptor
(Distance in m)
Release
Scenario
Maximum, High-End Exposure
Concentration (ppm)
Risk Estimates" - Inhalation Exposure
AC
ADC
LADC
Non-cancer
Cancer6
Acute MOE
Chronic MOE
Chronic IUR
Liver Effect
Respiratory
Respiratory
Fenceline
(100 m)
Max
6.2E-03
6.2E-03
2.6E-03
4,239
137
4.19E-05
Mean
4.7E-04
4.7E-04
2.0E-04
56,238
1,815
3.16E-06
Community Avg.
(100-1,000 m)
Max
7.2E-04
7.2E-04
3.0E-04
36,432
1,175
4.87E-06
Mean
5.4E-05
5.4E-05
2.3E-05
483,282
15,593
3.67E-07
"Details on the methods used to calculate risks are described in Section 5. Shading indicates risk relative to screening level
benchmarks.
h Lifetime cancer risks based on 33 years of continuous inhalation exposure averaged over a 78-year lifetime. Lifetime
cancer risks for a full lifetime (78 years) of continuous inhalation exposure would be 2.36 times greater than the risk
estimates presented here.
J.1.2 Ambient Air: IIOAC Methodology and Results for COUs Without Site-Specific
Data (Hydraulic Fracturing, Industrial, and Institutional Laundry Facilities)
The Ambient Air: IIOAC Methodology for COUs without Site-Specific Data was utilized to evaluate
exposures from three new COUs for the ambient air pathway (hydraulic fracturing, industrial laundry,
and institutional laundry) previously not included in the published risk evaluation or draft fenceline
report. The methodology utilizes IIOAC to estimate high-end and central tendency exposure
concentrations at three pre-defined distances from a releasing facility. This methodology is a higher-tier
methodology which integrates additional data provided as part of the release assessment. In particular,
this additional data included
Page 552 of 570
-------
1. Source attribution (fugitive and stack release types),
2. Days of release,
3. Multiple release percentiles, and
4. Chemical phase/form of release (vapor and particulate phase releases).
Other input parameters like release duration, meteorology, and topography were varied across the
scenarios outlined in Figure Apx J-3. A summary of the various input parameters is provided in
TableApx J-3. Modeling consisted of evaluating all possible iterations/combinations of the input
parameters listed resulting in the following total exposure and release scenarios:
1. Hydraulic Fracturing (fugitive releases only): 8 Exposure Scenarios, each with 28 release
scenarios;
2. Industrial Laundry (liquid): 8 exposure scenarios, each with 56 release scenarios for each of two
release types (fugitive and stack) and 1 chemical release form (vapor only);
3. Institutional Laundry (liquid): 8 exposure scenarios, each with 56 release scenarios for each of
two release types (fugitive and stack) and 1 chemical release form (vapor only);
4. Industrial Laundry (powder): 8 exposure scenarios, each with 56 release scenarios for each of
two release types (fugitive and stack) and 3 chemical release form (vapor only, PM10, PM2.5);
and
5. Institutional Laundry (powder): 8 exposure scenarios, each with 56 release scenarios for each of
two release types (fugitive and stack) and 3 chemical release form (vapor only, PM10, PM2.5).
Table Apx J-3. Exposure Scenarios and Inputs Utilized for Pre-screening Analysis of Hydraulic
Fracturing, Industrial Laundry, and Institutional Laundry CPU
Release
Perecntilc
Release
Type
Release
Release
Chemieal
cou
Duration
(h/day)
Frequeney
(Days)
Phase/Form of
Release
Meteorology
Topography
Hydraulic
fracturing
Maximum
99th
95th
50th
5th
Minimum
Mean
Fugitive
24
8
72
16
1
15
Vapor Only
South (Coastal)-
HE
West North
Central-CT
Rural
Urban
Industrial
Maximum
Fugitive
24
365
Vapor Only
South (Coastal)-
Rural
laundry -
liquid
99th
95th
50th
5th
Minimum
Mean
Stack
Unknown
(Fugitive,
Stack,
Other)
8
223
20
260
HE
West North
Central-CT
Urban
Industrial
Maximum
Fugitive
24
365
Vapor
South (Coastal)-
Rural
laundry -
powder
99th
95th
50th
5th
Minimum
Mean
Stack
Unknown
(Fugitive,
Stack,
Other)
8
223
20
260
Particulate
(Coarse)
Particulate
(Fine)
HE
West North
Central-CT
Urban
Institutional
laundry -
liquid
Maximum
99th
95th
50th
Fugitive
Stack
24
8
365
287
250
260
Vapor Only
South (Coastal)-
HE
Rural
Urban
Page 553 of 570
-------
cou
Release
Pcreentilc
Release
Type
Release
Duration
(h/day)
Release
Frequeney
(Days)
Chemieal
Phase/Form of
Release
Meteorology
Topography
5th
Unknown
West North
Minimum
(Fugitive,
Central-CT
Mean
Stack,
Other)
Institutional
laundry -
powder
Maximum
99th
95th
50th
5th
Minimum
Mean
Fugitive
Stack
Unknown
(Fugitive,
Stack,
Other)
24
8
365
287
250
260
Vapor
Particulate
(Coarse)
Particulate
(Fine)
South (Coastal)-
HE
West North
Central-CT
Rural
Urban
Results
Results for the Ambient Air: IIOAC Methodology for COUs without Site-Specific Data for these three
new COUs are summarized in Section 3.2.3.2 for exposure and Section 5.2.2.3.2 for estimated risks.
Complete results are presented in 1,4-Dioxane Supplemental Information File: Air Exposure and Risk
Estimates for 1,4-Dioxane Emissions from Hydraulic Fracturing Operations ( >24b) and
1,4-Dioxane Supplemental Information File: Air Exposures and Risk Estimates for Industrial Laundry
( 2024c). Generally, results from application of this methodology found the following:
1. Hydraulic Fracturing: Lifetime cancer risk estimates for distance within 1000 m of hydraulic
fracturing operations range from 1.7/10 3 to 7,7/10 6 across a range of high-end and central
tendency release and exposure scenarios; and
2. Industrial and Institutional Laundry: Lifetime cancer risk estimates for distances within 1,000 m
of laundry facilities range from 1.5 x 10_11 to 3.8x10~8 across a range of high-end and central
tendency release and exposure scenarios.
J.1.3 Ambient Air: Single Year Methodology (AERMOD)
AERMOD was developed to allow EPA to conduct a higher-tier analysis of releases, exposures, and
associated risks to people around releasing facilities at multiple distances when EPA has site-specific
data like reported releases, facility locations (for local meteorological data), source attribution, and other
data, when reasonably available. This methodology can also incorporate additional site-specific
information like stack parameters (stack height, stack temperature, plume velocity, etc.), building
characteristics, release patterns, different terrains, and other parameters when reasonably available.
AERMOD can be performed independent of the Ambient Air: Screening Methodology described above,
provides a more thorough analysis, can include wet and dry deposition estimates, and allows EPA to
fully characterize identified risks for chemicals undergoing risk evaluation. While the application of this
methodology in this supplemental risk evaluation focuses on a single year of data, the methodology can
be expanded to include multiple years of data.
Model
AERMOD for this supplemental risk evaluation estimated 1,4-dioxane exposures to fenceline
communities at user-defined distances from a facility releasing 1,4-dioxane. AERMOD is a steady-state
Gaussian plume dispersion model that incorporates air dispersion based on planetary boundary layer
turbulence structure and scaling concepts, including treatment of both surface and elevated sources and
both simple and complex terrain. AERMOD can incorporate a variety of emission source characteristics,
chemical deposition properties, complex terrain, and site-specific hourly meteorology to estimate air
Page 554 of 570
-------
concentrations and deposition amounts at user-specified receptor distances and at a variety of averaging
times. Readers can learn more about AERMOD, equations within the model, detailed input and output
parameters, and supporting documentation by reviewing the AERMOD users guide (U.S. EPA. 2018d).
Releases
EPA modeled exposures using the release data developed as described in Section 2.1.1.2 and
summarized below. Release data was provided (and modeled) on a facility-by-facility basis:
1. Facility-specific chemical releases (fugitive and stack releases) as reported to the 2019 TRI,
where available.
2. Alternative release estimates as described in the decision tree for estimating air releases, where
facility specific 2019 TRI data were not available. Alternative release estimates may include
facility specific releases reported in previous TRI reporting years (2016 to 2018) or modeled
release estimates using existing EPA models or other surrogate data.
Exposure Scenarios
AERMOD evaluated exposures at eight finite distances (5, 10, 30, 60, 100, 2,500, 5,000, and 10,000 m)
and one area distance (100 to 1,000 m) from each releasing facility (or generic facility for alternative
release estimates). Receptors for each of the eight finite distances were placed in a polar grid every 22.5
degrees around the respective distance ring. This results in a total of 16 receptors around each finite
distance ring for which exposures are modeled. FigureApx J-3 provides a visual depiction of the
placement of receptors around a finite distance ring. Although the visual depiction only shows receptor
locations around a single finite distance ring, the same placement of receptors occurred for all eight
finite distance rings
Receptor Locations around each Finite Distance Ring
100 -1,000 m
X
U
m
m
xl
2,500 m
10,000 m
22.5 '
Releasing Facility
Figure Apx J-3. Modeled Receptor Locations for Finite Distance Rings
Receptors for the area distance evaluated were placed in a cartesian grid at equal distances between 200
and 900 m around each releasing facility (or generic facility for alternative release estimates). Receptors
were placed at 100-meter increments. This results in a total of 456 receptors for which exposures are
Page 555 of 570
-------
modeled. Figure Apx J-4 provides a visual depiction of the placement of receptors (each dot) around the
area distance ring.
Exposure Concentration Outputs
Hourly-average concentration outputs were provided from AERMOD for each receptor around each
distance ring (i.e., each of 16 receptors around a finite distance ring or each receptor within the area
distance ring). Daily and Period averages were then calculated from the modeled hourly data. Daily
averages for the finite distance rings were calculated as arithmetic averages of all hourly data for each
day modeled for each receptor around each ring. Daily averages for the area distance ring were
calculated as the arithmetic average of the hourly data for each day modeled across all receptors within
the area distance ring. This results in the following number of daily average concentrations at each
distance modeled.
1. Daily averages for TRI reporting facilities (using 2016 calendar year meteorological data): One
daily average concentration for each of 366 days for each of 16 receptors around each finite
distance ring. This results in a total of 5,856 daily average concentration values for each finite
distance modeled (366 x 16 = 5,856).
2. Daily averages for EPA estimated releases (using 2011 to 2015 meteorological data): Five daily
average concentrations (for each year of meteorological data) for each of 365 (or 366) days for
each of 16 receptors around each finite distance ring. This results in a total of 29,216 daily
average concentration values for each finite distance modeled.
3. Daily averages for both TRI reporting facilities and EPA estimated releases: One daily average
concentration for each of 365 or 366 days across all receptors within the area distance ring. This
results in a total of 365 or 366 daily average concentration values for the area distance.
Period averages were calculated from all the daily averages for each receptor for each distance ring over
1 year for TRI reporting facilities and 5 years for facilities where releases were estimated. This results in
a total of 16 period average concentration values for each finite distance ring. This is derived from either
averaging the daily averages across the single year of meteorological data used (2016) for TRI reporting
facilities or across the multi-year meteorological data used (2011 to 2015) for EPA estimated releases.
Page 556 of 570
-------
Daily and period average Outputs were stratified by different source scenarios, such as urban/not urban
setting or emission-strengths, where needed. Outputs from AERMOD are provided in units of
micrograms per cubic meter (|ig/m3) requiring conversion to parts per million (ppm) for purposes of
calculating risk estimates for 1,4-dioxane. The following formula was used for this conversion:
Cppm = (24.45* (Caermod)/1,000)/MW
Where:
Cppm = Concentration (ppm)
24.45 = Molar volume of a gas at 25 °C and 1 atmosphere pressure
Caermod = Concentration from AERMOD (|ig/m3)
MW = Molecular weight of the chemical of interest (g/mole).
Post-processing scripts were used to extract and summarize the output concentrations for each facility,
release, and exposure scenario. The following statistics for daily- and period-average concentrations
were extracted or calculated from the results for each of the modeled distances (i.e., each ring or grid of
receptors) and scenarios (also see Table Apx J-4):
• Minimum;
• Maximum;
• Average;
• Standard deviation; and
• 10th, 25th, 50th, 75th, and 95th percentiles.
Table Apx J-4. Description of Daily or Period Average and Air Concentration Statistics
Statistic
Description
Minimum
The minimum daily or period average concentration estimated at any receptor location on any day
at the modeled distance.
Maximum
The maximum daily or period average concentration estimated at any receptor location on any day
at the modeled distance.
Average
Arithmetic mean of all daily or period average concentrations estimated at all receptor locations on
all days at the modeled distance. This incorporates lower values (from days when the receptor
location largely was upwind from the facility) and higher values (from days when the receptor
location largely was downwind from the facility).
Percentiles
The daily or period average concentration estimate representing the numerical percentile value
across the entire distribution of all concentrations at all receptor locations on any day at the
modeled distance. The 50th percentile represents the median of the daily or period average
concentration across all concentration values for all receptor locations on any day at the modeled
distance.
J.1.4 Ambient Air: Multi-Year Analysis Methodology (IIOAC)
The multi-year analysis incorporates SACC recommendations by evaluating multiple years of chemical
release data to estimate exposures and associated risks to fenceline communities. This is achieved by
conducting a facility-by-facility evaluation of all 1,4-dioxane releases reported to TRI over six reporting
years (2015 through 2020). Data for these 6 years were obtained from the TRI database (TRI basic plus
files downloaded on August 5, 2022). Annual release data for 1,4-dioxane were extracted from the entire
TRI data set for all facilities reporting air releases of 1,4-dioxane for one or more years between 2015
Page 557 of 570
-------
and 2020. Facilities were categorized into occupational exposure scenarios for modeling purposes and
later cross-walked to COUs for risk management purposes.
The TRI data extracted for the multi-year analysis were used as direct inputs to the IIOAC model. An
additional arithmetic average of the TRI data for each facility was also calculated when the facility
reported releases to TRI for two or more of the years evaluated and used as a direct input to the IIOAC
model. EPA then evaluated the more "conservative exposure scenario" of the 16 scenarios evaluated for
the Ambient Air: Screening Methodology described above to estimate exposure concentrations. This
more conservative exposure scenario consists of a facility that operates year-round (365 days per year,
24 hours per day, 7 days per week), a South Coastal meteorologic region, and a rural topography setting.
The Ambient Air: Multi-Year Analysis Methodology includes a land-use analysis utilizing the same
visual methodology described for the 2022 fenceline analysis and the Ambient Air: Single Year
Methodology (AERMOD). However, the land use analysis was limited those facilities where the multi-
year analysis (1) found risk estimates above the benchmark value extending farther out when compared
to the 2022 fenceline analysis, or (2) identified a new facility with risk estimates above the benchmark
that was not captured by the 2022 fenceline analysis. Using this methodology, EPA identified if there is
an expected exposure for people in fenceline communities to releases from the facility of interest within
the distances where the benchmark was exceeded.
J.2 Inhalation Exposure Estimates for Fenceline Communities
Acute and chronic inhalation exposures were estimated based on air concentrations estimated in Section
2.3.3 using the methodologies described above.
Acute and chronic inhalation exposures used to evaluate non-cancer risks are estimated as an Acute
Concentration (AC) or Average Daily Concentration (ADC), respectively. Lifetime exposures used to
evaluate cancer risks are estimated as a Lifetime Average Daily Concentration (LADC). Methods
adequate to quantify the impact of lifestage differences on 1,4-dioxane exposure are not available (see
Section 4.3) and air concentration is used as the exposure metric for all lifestages per EPA guidance
( 2012. 1994b).
The equations used to calculate each of the exposure values are:
DAC x ET
AC
AT
AAC x ET x EF x ED
ADC
AT
AAC x ET x EF x ED
LADC =
AT
Where:
AC
DAC
Acute Concentration (ppm)
Daily Average Air Concentration, model output reflecting average concentrations
over a 24-hour period (ppm)
Exposure Time (24 hours/day)
Annual Average Air Concentration, model output reflecting average
concentrations over a year (ppm)
ET
AAC
Page 558 of 570
-------
EF
ED
AT
Exposure Frequency (365 days/year)
Exposure Duration (1 year for non-cancer ADC; 33 years for cancer LADC)
Averaging Time
Averaging time for AC = 24 hours
Averaging time for ADC = 24 hours/day x 365 days/year x 1 year
Averaging time for LADC = 24 hours/day x 365 days/year x 78 years
For fenceline communities, all exposure estimates assume continuous exposure (24 hours/day)
throughout the duration of exposure. The exposure duration used to calculate the LADC is based on the
95th percentile of the expected duration at a single residence, 33 years ( ) and the
averaging time is based on a 78-year lifetime. To determine the exposures for 78 years exposure
duration, presented results should be multiplied by 2.36.
Detailed reporting of modeled air concentrations and corresponding AC, ADC, and LADC estimates for
33 years exposure duration are provided in 1,4-Dioxane Supplemental Information File: Air Exposures
and Risk Estimates for Single Year Analysis ( 24e).
J.3 Land Use Analysis
As described in Section 5.2.2.3, EPA conducted a review of land use patterns around facilities where
cancer risk exceeded 1 x 1CT6. The methodology for this analysis is consistent with what was previously
described in the Draft TSCA Screening Level Approach for Assessment Ambient Air and Water
Exposures to Fenceline Communities Version 1.0. This review was limited to those facilities with real
Global Information System (GIS) locations that showed risk. The land use analysis does not include
generic facilities (since there is no real location around which to conduct the land use analysis) where
alternative release estimates were modeled to estimate exposures. The purpose of this review was to
determine if EPA can reasonably expect an exposure to fenceline communities to occur within the
modeled distances where there was an indication of risk. This detailed review consisted of visual
analysis using aerial imagery and interpreting land use/zoning practices around the facility. More
specifically, EPA used ESRI ArcGIS (Version 10.8) and Google maps to characterize land use patterns
within the radial distances evaluated where there was an indication of risk. For locations where
residential or industrial/commercial businesses or other public spaces are present within those radial
distances indicating risk, EPA includes those locations within the fenceline communities category and
reasonably expects an exposure and therefore an associated potential risk. Where the radial distances
showing an indication of risk occur within the boundaries of the facility or is limited to uninhabited
areas, EPA does not reasonably expect an exposure to fenceline communities to occur and therefore
does not expect an associated risk.
Page 559 of 570
-------
TableApx J-5. Summary of Fenceline Community Exposures Expected near Facilities Where
Modeled Air Concentrations Indicated Risk for 1,4-1
>ioxane
OES
cou
Total
Number of
Facilities
Evaluated
Number of
Facilities
with Risk
Indicated"
Number of Facilities with
Risk Indicated and
Fenceline Community
Exposures Expected"
Percent of Total Facilities
with Risk Indicated and
Fenceline Community
Exposures Expected"
Disposal
Disposal
15
4
1
7%
Ethoxylation
byproduct
Ethoxylation
byproduct
6
3
2
33%
Industrial uses
Industrial uses
12
4
1
8%
Manufacturing
Manufacture
1
1
1
100%
PET
manufacturing
PET
manufacturing
13
10
6
46%
" Only includes facilities with TRI ID
Individual facility summaries are available in 1,4-Dioxane Supplemental Information File: Air
Exposures and Risk Estimates for Single Year Analysis ( s24e).
J.4 Aggregate Analysis across Facilities
A conservative screening method for aggregated risk within the air pathway is included to address
whether the combined general population exposures to emissions from nearby facilities present any
additional risk not represented by the individual facility analysis. By taking a conservative approach, this
methodology can effectively screen out aggregate concerns where no additional air risk is identified, and
flag groups of facilities that demonstrate the potential for additional aggregate air risk.
The aggregate air approach utilized the existing modeling results from the single year analysis (2019
TRI data) for individual facilities to estimate aggregate exposures from facilities within proximity to
other facilities releasing 1,4-dioxane within a 10 km buffer. Facilities with releases to air were mapped
using location coordinates from the TRI database. A 10 km buffer was drawn around each facility, and
groups of facilities were identified by any overlap between these buffers {i.e., any facilities within 20 km
of another facility, even if not all of the facilities have overlapping buffers) (Figure Apx J-5).
Page 560 of 570
-------
feroupd]
RS.QvEGRRNfli:
m5X6 |j)'555»Z5-2-5 Bl
F^53ffS &^k*2W7,B]
Kilometers
f 1,4-Dioxane Facilities
1\ Facility Groups
FigureApx J-5. Example of Group of Air Releasing Facilities with Overlapping 10 km
Buffers for Aggregate Air Risk Screening
Next, the modeled air concentrations from each facility in the group were combined to generate
hypothetical "worst-case scenario" aggregate air concentrations for the facility group. Due to the
modeling methodology for individual facilities producing resulting air concentrations at discrete
distances from each facility, the aggregate screening analysis also assesses concentrations and risk at
discrete distances. For the sake of the analysis, the facilities are treated as if they are all releasing from
the same point. This is a conservative approach, since the facilities with each group all have some
distance between them, and the air concentrations tend to decrease with greater distance from the source
facility. Within each facility group, the 95th percentile total (stack and fugitive) air concentrations for
each facility were summed for each modeled distance interval. Cancer risk levels were similarly added
together for each modeled distance interval, due to their proportional relationship to concentration, and
non-cancer MOE values were combined using the equation below for each distance interval.
1
M0 Efotal ~ J J J
WOEl+ WOE~z + MOE3 +
Where:
MOEtotai = The aggregated MOE value for the group
MOE(ij2,3„.) = The individual MOE values for each facility in the group
Aggregated risk values were then compared against cancer and non-cancer benchmarks to identify
values indicating risk relative to benchmarks. For each facility included in an aggregated group, it was
noted whether the individual risk calculation results indicated risk relative to cancer or non-cancer
benchmarks before aggregating. Additionally, for each facility group the relative contribution of each
facility to the 95th percentile cancer risk was calculated, by dividing the individual facility risk by the
aggregated group risk, to determine whether the resulting numbers may be disproportionately due to
only one or more facilities. The resulting aggregate risk calculations were reviewed to determine where
the numerical results suggested a concern for aggregate air risk that had not been represented by the
Page 561 of 570
-------
individual facility risk analysis. Where this additional risk was flagged, the mapped locations of the
facilities were then inspected to confirm that the distances between the facilities supported aggregating
releases from the facilities at the flagged distance interval. The review of the aggregated results and
facility locations was applied to characterize whether aggregate air risk relative to benchmarks is
expected for each group.
For example, if the aggregate risk calculations for a group of two facilities indicated cancer risk greater
than 1 in 1 million (1 x 1CT6) at the 100 m distance, and the individual facilities only showed that level of
risk up to 60 m, the map would be inspected. If the facilities were found to be located 1,000 m apart, the
group would be characterized as not showing risk relative to a 1 in 1 million benchmark beyond what
was captured by the individual analysis. However, if the facilities were located within 200 m of one
another, such that their 100 m distance intervals would intersect, the group would be characterized as
showing potential for aggregated air risk beyond what was captured by the individual analysis.
If aggregate air risk relative to benchmarks is identified, then an additional land use check is performed
to confirm the potential for a general population exposure at the new distance. In some cases, no
additional aggregate air risk is identified, because no distance intervals present risk relative to
benchmarks.
Page 562 of 570
-------
Start Here
Did any of the
facilities in the
) >
group show air
nsk before
aggregating?
Does aggregate
air risk extend to
distances beyond
air risk for any of
the individual
facilities?
additional aggregate air nsk
ified (facillUes further apart than
distance at which additional nsk
would be suggested)"
7 zzfzz
facilities individually not showing j
( cm/
Is the aggregate
air risk entirely
>
due to facilities
already showing
risk at those
distances?
/ ^Potential aggregate air nsk out to f
/ f distances beyond individual facility f
f air nsk"
¦ ^
FigureApx J-6. Decision Tree for Characterizing Aggregate Air Risk for Multiple Facilities
Page 563 of 570
-------
Table Apx J-6. Summary of Groups of Facilities Considered in Aggregate Analysis
Total Air Facilities
with Release Data
Number of Facilities in
Groups
Number of Groups
Number of Groups
with Additional
Aggregate Risk
50
12
5
0
The grouping analysis for 1,4-dioxane resulted in five groups of nearby facilities, ranging from two to
four facilities per group. No additional aggregate air risk relative to benchmarks was identified for each
of the five groups. Where three groups each contained a single facility showing risk out to some
distance, there was no additional distance interval showing risk from the aggregate calculation.
Although the proximity of the facilities may indicate a reality of greater localized air concentrations than
are represented in the individual facility analysis, the aggregated concentrations did not cross any
additional risk benchmarks, so any determinations of risk are already accounted for by the individual
facility analysis. For the remaining two groups, no aggregated or individual risks were present.
Therefore, further inspection and additional land use analysis were not warranted for these facility
groups.
Maps of the five facility groups, with the 10 km buffers used to define them are provided below in
Figure_Apx 1-7 through Figure_Apx J-l I.
Figure Apx J-7. Map of Aggregated Air Facilities, Group 1
Page 564 of 570
-------
^ 1,4-Dioxane Facilities
Facility Groups
Kilometers
FigureApx J-8. Map of Aggregated Air Facilities, Group 2
f 1,4-Dioxane Facilities
I Facility Groups
K292Ti2'C'RIINSlDS^IGj
Kilometers
Figure Apx J-9. Map of Aggregated Air Facilities, Group 3
Page 565 of 570
-------
^ 1,4-Dioxane Facilities
Facility Groups
LUnoupM]
[s;i(3;;ri^>iNiTK»gi;)^5^i
Kilometers
FigureApx J-10. Map of Aggregated Air Facilities, Group 4
~ 1,4-Dioxane Facilities
Facility Groups
LCj no.u pis]
Kilometers
Figure Apx J-ll. Map of Aggregated Air Facilities, Group 5
Page 566 of 570
-------
Appendix K SUMMARY OF REVISED ANALYSES COMPLETED
IN RESPONSE TO SACC AND PUBLIC COMMENT
As described in Section 1.1, EPA revised elements of the analysis presented in in this revised
supplement based on SACC and public comments on the draft.
Specifically, EPA revised occupational exposure and risk estimates for some COUs based on additional
information, alternate input assumptions, or modifications to Monte Carlo models. In some cases, these
revisions increased or decreased risk estimates by up to an order of magnitude. For other COUs, these
revisions had no quantitative impact on risk estimates.
EPA also revised release assessments for some COUs based on revised Monte Carlo models and
alternate input assumptions. For hydraulic fracturing releases to surface water, the revised release
estimates were used to generate revised exposure and risk estimates. For other revised release estimates,
The Agency did not revise the corresponding exposure and risk estimates because the magnitude of the
change was not expected to be sufficient to alter overall risk conclusions.
In addition, EPA considered the quantitative impact of certain assumptions. For example, while EPA
retained risk estimates based on original exposure assumptions, the revised supplement discusses the
extent to which alternate assumptions about exposure amount and duration would increase risk estimates
(Section 5.2.2, Appendix I, Appendix J.2). EPA also considered the magnitude of impact of aggregating
risk across routes. Although the Agency retained risk estimates based individual routes, this revised
supplement discusses the extent to which aggregation across routes would alter overall risk (Section
5.2.4).
Table Apx K-l. Summary of Changes to Occupational Exposure and Risk Estimates
OES
Changes to Occupational Inhalation Exposures
Changes to Occupational Dermal
Exposures
Antifreeze
Incorporates updated Monte Carlo modeling resulting
in increased inhalation exposure estimates. Exposures
and risk estimates went up by almost an order of
magnitude due to increased use rate. However, the
exposure levels remain small (E-07).
No change.
Surface
cleaner
Updated exposure calculations for a higher exposure
duration per SACC comment (from 4- to 8-hours,
thereby doubling the exposure estimate). However,
updated risk estimates stayed within an order of
magnitude.
Incorporates NY waiver data (product
concentration), resulting in an order
of magnitude increase in dermal
exposure estimates and risk estimates.
Textile dyes
Updated exposure calculations per public comments,
which increased CT exposure and risk estimates by an
order of magnitude and reduced HE exposure
estimates within the same order of magnitude. HE risk
estimates were reduced by an order of magnitude.
No change.
Dish soap
Incorporates Monte Carlo modeling resulting in
decreased inhalation exposure estimates by two (CT)
to three (HE) orders of magnitude. Risk estimates
decreased by two to three orders of magnitude
depending on the risk category. The modeled
exposures are lower than the original exposure
Incorporates NY waiver data, but this
resulted in no change to exposure and
risk estimates.
Page 567 of 570
-------
OES
Changes to Occupational Inhalation Exposures
Changes to Occupational Dermal
Exposures
estimates in the Supplemental RE (which were based
on old monitoring data).
Dish
detergent
Incorporates Monte Carlo modeling resulting in
decreased inhalation exposure estimates by two (CT)
and three (HE) orders of magnitude. Risk estimates
decreased by three orders of magnitude. The modeled
exposures are lower than the original exposure
estimates in the Supplemental RE (which were based
on old monitoring data).
Incorporates NY waiver data,
resulting in slight increase in dermal
exposure and risk estimates within the
same order of magnitude.
Laundry
detergent
Incorporates updated Monte Carlo modeling resulting
in increased inhalation exposure estimates. Exposures
and risk estimates went up by up to an order of
magnitude due to the higher temperature and product
concentrations. Exposures are still small (E-02 to
E-04 level).
Incorporates NY waiver data,
resulting in an order of magnitude
increase in dermal exposure and risk
estimates.
PET
byproduct
Incorporates monitoring data from public comments.
This reduced worker CT by an order of magnitude but
had little impact on the HE exposure estimate. Risk
estimates for both CT and HE decreased by and order
of magnitude. With this data, we were also able to
estimate ONU exposures, which we were not able to
do in the published draft Supplemental RE.
Incorporates data from public
comments, which resulted in an order
of magnitude increase in dermal
exposure and risk estimates.
Ethoxylation
byproduct
Incorporates monitoring data from public comments.
Previously, we only had a single exposure estimate.
Now we have CT and HE. The new HE exposure and
risk estimates are very similar to the single estimate
we originally had. The CT exposure and risk estimates
are lower by an order of magnitude.
No change.
Hydraulic
fracturing
Incorporates updated Monte Carlo modeling per
SACC comments, including fixes to fugitive release/
exposure equations, resulting in decreased inhalation
exposure estimates. Exposures and risk estimates went
down by one to two orders of magnitude once the
model updates were made and equations were fixed.
No change.
Page 568 of 570
-------
Table Apx K-2. Summary of Revisions to Release Assessments
Model
Changes Made
Impaet on Release
Assessment
Impaet on General
Population
Exposure and Risk
Estimates
Basis for Final Risk Estimates
Presented in the Revised Supplement
Textile Dyes
Updated operating
days from
triangular to
discrete
distribution.
The daily and annual
releases are
unchanged. The
number of release
days has changed.
Change not carried
through risk estimates
as overall conclusion
was not expected to
change.
N/A (this analysis does not directly
inform general population exposures
because we rely on SHEDS-HT for
down-the-drain).
Laundry
Detergent -
Institutional
and
Industrial
Ran the model with
an updated wash
water temperature.
Incorporated NY
waiver data
(product
concentration) for
product
concentration.
Most release points
increased by up to
an order of
magnitude due to the
higher temperature
and product
concentrations.
Change not carried
through risk estimates
because conclusion of
no risk from air
emissions from
laundries is not
expected to change;
magnitude of impact
of temperature
assumption discussed
qualitatively.
For the ambient air pathway, risk
estimates are based on the original
release assessment presented in the
draft supplement and retained in the
supplemental Excel file; see tabs
labeled:
"Release_Results_(Liquid/Powder)_Ori
ginal")
For the surface water pathway, down-
the-drain releases are estimated using
SHED-HT.
Hydraulic
Fracturing
Updated to reflect
Revised hydraulic
fracturing emission
scenario document
(ESD). Added spill
release and
adjusted some of
the release media
partitioning. Fixed
error in one
fugitive release
calculation and
exposure
calculations.
Overall releases
decreased by one
order of magnitude
once the fugitive
release equations
were revised.
Release media
partitioning changed
due to the Revised
ESD changes to
incorporate spills.
Change carried
through surface water
risk estimates
because high-end of
the distribution is
now lower.
For releases to surface water, risk
estimates are based on the revised
release estimates available in the
supplemental Excel file (see tab labeled
"ReleaseResultsUpdated")
For releases to groundwater and air,
risk estimates are based on the original
releases available in the supplemental
Excel file (see tab labeled
"ReleaseResultsOriginal")
Dish Soap
and
Detergent
Developed a Monte
Carlo model to
assess releases and
exposures with data
from standard
sources and public
comments.
Modeled release
results are difficult
to compare to those
in the Supplemental
RE, since the results
in the Supplemental
RE are for the
Liverpool OH case
study and the
modeled results are
per site.
Change not carried
through risk
estimates.
N/A (this analysis does not directly
inform general population exposures
because we rely on SHEDS-HT for
down-the-drain).
Page 569 of 570
-------
Appendix L OCCUPATIONAL EXPOSURE VALUE
EPA has calculated a draft 8-hour existing chemical occupational exposure value. That value was
previously published in a memo posted to the docket (EPA-HQ-OPPT-2022-0905-0039) August 2023,
titled "Draft Existing Chemical Exposure Limit (ECEL) for Occupational Use of 1,4-Dioxane
Although the 2023 memo refers to the calculated value as an "ECEL", EPA has updated the terminology
and now considers this value to reflect an "Occupational Exposure Value" calculated without
consideration of costs or other non-risk factors. The calculated Occupational Exposure Value
(previously referred to as an ECEL in the 2023 memo) is 0.055 ppm (0.20 mg/m3) based on chronic
cancer risk.
The calculated draft occupational exposure value for 1,4-dioxane represents the exposure concentration
below which workers and occupational non-users are not expected to exhibit any appreciable risk of
adverse toxicological outcomes, accounting for potentially exposed and susceptible populations (PESS).
It is derived based on the most sensitive human health effect relative to benchmarks and standard
occupational scenario assumptions of 8 hours/day, 5 days/week exposures for a total of 250 days
exposure per year, and a 40-year working life.
TSCA requires risk evaluations to be conducted without consideration of costs and other non-risk
factors, and thus this draft occupational exposure value represents a risk-only number. In risk
management, EPA may consider costs and other non-risk factors, such as technological feasibility, the
availability of alternatives, and the potential for critical or essential uses. Any existing chemical
exposure limit (ECEL) used for occupational safety risk management purposes could differ from the
draft occupational exposure value based on additional consideration of exposures and non-risk factors
consistent with TSCA section 6(c).
Page 570 of 570
------- |