PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

vvEPA

April 2020

United States	Office of Chemical Safety and

Environmental Protection Agency	Pollution Prevention

Assessment of Occupational Exposure and Environmental

Releases for
Perchloroethylene
(Ethene, l,l»2,2-Tetrachloro)

CASRN: 127-18-4

ci ci

M

CI CI

April 2020


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

TABLE OF CONTENTS

ABBREVIATIONS	17

EXECUTIVE SUMMARY	20

1	INTRODUCTION	22

1.1	Overview	22

1.2	Scope	22

1.3	Components of the Occupational Exposure and Environmental Release Assessment	28

1.4	General Approach and Methodology for Occupational Exposures and Environmental Releases
28

1.4.1	Estimates of Number of Facilities	28

1.4.2	Process Description	29

1.4.3	Worker Activities	29

1.4.4	Number of Workers and Occupational Non-Users	29

1.4.5	Inhalation Exposure Assessment Approach and Methodology	30

1.4.5.1	General Approach	30

1.4.5.2	Approach for thi s Ri sk Evaluati on	31

1.4.6	Dermal Exposure Assessment Approach	32

1.4.7	Consideration of Engineering Controls and Personal Protective Equipment	32

1.4.8	Water Release Sources	35

1.4.9	Water Release Assessment Approach and Methodology	35

2	ENGINEERING ASSESSMENT	37

2.1	Manufacturing	37

2.1.1	Estimates of Number of Facilities	37

2.1.2	Process Description	38

2.1.3	Exposure Assessment	39

2.1.3.1	Worker Activities	39

2.1.3.2	Number of Potentially Exposed Workers	39

2.1.3.3	Occupational Exposure Results	40

2.1.4	Water Release Assessment	42

2.1.4.1	Water Release Sources	42

2.1.4.2	Water Release Assessment Results	42

2.2	Repackaging	46

2.2.1	Estimates of Number of Facilities	46

2.2.2	Process Description	47

2.2.3	Exposure Assessment	47

2.2.3.1	Worker Activities	47

2.2.3.2	Number of Potentially Exposed Workers	47

2.2.3.3	Occupational Exposure Results	49

2.2.4	Water Release Assessment	50

2.2.4.1	Water Release Sources	50

2.2.4.2	Water Release Assessment Results	50

2.3	Processing as a Reactant	51

Page 2 of316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

2.3.1	Estimates of Number of Facilities	51

2.3.2	Process Description	52

2.3.3	Exposure Assessment	52

2.3.3.1	Worker Activities	52

2.3.3.2	Number of Potentially Exposed Workers	53

2.3.3.3	Occupational Exposure Results	54

2.3.4	Water Release Assessment	55

2.3.4.1	Water Release Sources	55

2.3.4.2	Water Release Assessment Results	55

2.4	Incorporation into Formulation, Mixture, or Reaction Product	58

2.4.1	Estimates of Number of Facilities	58

2.4.2	Process Description	58

2.4.3	Exposure Assessment	59

2.4.3.1	Worker Activities	59

2.4.3.2	Number of Potentially Exposed Workers	59

2.4.3.3	Occupational Exposure Results	61

2.4.3.3.1	Inhalation Exposure Results for Aerosol Packing Formulation Sites Using
Monitoring Data	61

2.4.3.3.2	Inhalation Exposure Results for Non-Aerosol Packing Formulation Sites Using
Modeling 62

2.4.4	Water Release Assessment	65

2.4.4.1	Water Release Sources	65

2.4.4.2	Water Release Assessment Results	65

2.5	Batch Open-Top Vapor Degreasing	66

2.5.1	Estimates of Number of Facilities	66

2.5.2	Process Description	67

2.5.3	Exposure Assessment	70

2.5.3.1	Worker Activities	70

2.5.3.2	Number of Potentially Exposed Workers	70

2.5.3.3	Occupational Exposure Results	71

2.5.4	Water Release Assessment	73

2.5.4.1	Water Release Sources	73

2.5.4.2	Water Release Assessment Results	73

2.6	Batch Closed-Loop Vapor Degreasing	76

2.6.1	Estimates of Number of Facilities	76

2.6.2	Process Description	76

2.6.3	Exposure Assessment	77

2.6.3.1	Worker Activities	77

2.6.3.2	Number of Potentially Exposed Workers	78

2.6.3.3	Occupational Exposure Results	78

2.6.4	Water Release Assessment	80

2.6.4.1	Water Release Sources	80

2.6.4.2	Water Release Assessment Results	80

2.7	Conveyorized Vapor Degreasing	80

2.7.1 Estimates of Number of Facilities	80

Page 3 of316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

2.7.2	Process Description	80

2.7.3	Exposure Assessment	85

2.7.3.1	Worker Activities	85

2.7.3.2	Number of Potentially Exposed Workers	85

2.7.3.3	Occupational Exposure Results	85

2.7.4	Water Release Assessment	87

2.7.4.1	Water Release Sources	87

2.7.4.2	Water Release Assessment Results	87

2.8	Web Degreasing	87

2.8.1	Estimates of Number of Facilities	87

2.8.2	Process Description	88

2.8.3	Exposure Assessment	88

2.8.3.1	Worker Activities	88

2.8.3.2	Number of Potentially Exposed Workers	89

2.8.3.3	Occupational Exposure Results	89

2.8.4	Water Release Assessment	91

2.8.4.1	Water Release Sources	91

2.8.4.2	Water Release Assessment Results	92

2.9	Cold Cleaning	92

2.9.1	Estimates of Number of Facilities	92

2.9.2	Process Description	92

2.9.3	Exposure Assessment	93

2.9.3.1	Worker Activities	93

2.9.3.2	Number of Potentially Exposed Workers	93

2.9.3.3	Occupational Exposure Results	94

2.9.3.3.1	Inhalation Exposure Assessment Results Using Monitoring Data	95

2.9.3.3.2	Inhalation Exposure Assessment Results Using Modeling	96

2.9.4	Water Release Assessment	99

2.9.4.1	Water Release Sources	99

2.9.4.2	Water Release Assessment Results	99

2.10	Aerosol Degreasing and Aerosol Lubricants	99

2.10.1	Estimates of Number of Facilities	99

2.10.2	Process Description	101

2.10.3	Exposure Assessment	101

2.10.3.1	Worker Activities	101

2.10.3.2	Number of Potentially Exposed Workers	102

2.10.3.3	Occupational Exposure Results	103

2.10.3.3.1	Inhalation Exposure Assessment Results Using Monitoring Data	103

2.10.3.3.2	Inhalation Exposure Assessment Results Using Modeling	104

2.10.4	Water Release Assessment	106

2.11	Dry Cleaning and Spot Cleaning	106

2.11.1	Estimates of Number of Facilities	106

2.11.2	Process Description	107

2.11.3	Exposure Assessment	109

Page 4 of316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

2.11.3.1	Worker Activities	109

2.11.3.2	Number of Potentially Exposed Workers	109

2.11.3.3	Occupational Exposure Results	110

2.11.3.3.1	Inhalation Exposure Assessment Results Using Monitoring Data	110

2.11.3.3.2	Inhalation Exposure Assessment Results Using Modeling	114

2.11,4 Water Release Assessment	119

2.11.4.1	Water Release Sources	119

2.11.4.2	Water Release Assessment Results	120

2.12	Adhesives, Sealants, Paints, and Coatings	123

2.12.1	Estimates of Number of Facilities	123

2.12.2	Process Description	127

2.12.3	Exposure Assessment	128

2.12.3.1	Worker Activities	128

2.12.3.2	Number of Potentially Exposed Workers	128

2.12.3.3	Occupational Exposure Results	131

2.12.4	Water Release Assessment	133

2.12.4.1	Water Release Sources	133

2.12.4.2	Water Release Assessment Results	133

2.13	Maskant for Chemical Milling	141

2.13.1	Estimates of Number of Facilities	141

2.13.2	Process Description	141

2.13.3	Exposure Assessment	141

2.13.3.1	Worker Activities	141

2.13.3.2	Number of Potentially Exposed Workers	142

2.13.3.3	Occupational Exposure Results	144

2.13.4	Water Release Assessment	145

2.13.4.1	Water Release Sources	145

2.13.4.2	Water Release Assessment Results	146

2.14	Industrial Processing Aid	148

2.14.1	Estimates of Number of Facilities	148

2.14.2	Process Description	148

2.14.3	Exposure Assessment	149

2.14.3.1	Worker Activities	149

2.14.3.2	Number of Potentially Exposed Workers	149

2.14.3.3	Occupational Exposure Results	150

2.14.4	Water Release Assessment	151

2.14.4.1	Water Release Sources	151

2.14.4.2	Water Release Assessment Results	152

2.15	Metalworking Fluids	153

2.15.1	Estimates of Number of Facilities	153

2.15.2	Process Description	154

2.15.3	Exposure Assessment	154

2.15.3.1	Worker Activities	154

2.15.3.2	Number of Potentially Exposed Workers	155

2.15.3.3	Occupational Exposure Results	155

Page 5 of316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

2.15,4 Water Release Assessment	156

2.15.4.1	Water Release Sources	156

2.15.4.2	Water Release Assessment Results	157

2.16	Wipe Cleaning and Metal/Stone Polishes	157

2.16.1	Estimates of Number of Facilities	157

2.16.2	Process Description	157

2.16.3	Exposure Assessment	157

2.16.3.1	Worker Activities	157

2.16.3.2	Number of Potentially Exposed Workers	157

2.16.3.3	Occupational Exposure Results	157

2.16.4	Water Release Assessment	159

2.17	Other Spot Cleaning/Spot Removers (Including Carpet Cleaning)	159

2.17.1	Estimates of Number of Facilities	159

2.17.2	Process Description	160

2.17.3	Exposure Assessment	160

2.17.3.1	Worker Activities	160

2.17.3.2	Number of Potentially Exposed Workers	160

2.17.3.3	Occupational Exposure Results	160

2.17.4	Water Release Assessment	161

2.18	Other Industrial Uses	161

2.18.1	Estimates of Number of Facilities	161

2.18.2	Process Description	161

2.18.3	Exposure Assessment	162

2.18.3.1	Worker Activities	162

2.18.3.2	Number of Potentially Exposed Workers	162

2.18.3.3	Occupational Exposure Results	165

2.18.4	Water Release Assessment	167

2.18.4.1	Water Release Sources	167

2.18.4.2	Water Release Assessment Results	167

2.19	Other Commercial Uses	168

2.19.1	Estimates of Number of Facilities	168

2.19.2	Process Description	169

2.19.3	Exposure Assessment	169

2.19.3.1	Worker Activities	169

2.19.3.2	Number of Potentially Exposed Workers	169

2.19.3.3	Occupational Exposure Results	169

2.19.4	Water Release Assessment	171

2.19.4.1	Water Release Sources	171

2.19.4.2	Water Release Assessment Results	171

2.20	Laboratory Chemicals	172

2.20.1	Estimates of Number of Facilities	172

2.20.2	Process Description	172

2.20.3	Exposure Assessment	173

2.20.3.1 Worker Activities	173

2.20.3.1	Number of Potentially Exposed Workers	173

2.20.3.2	Occupational Exposure Results	173

Page 6 of316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

2.20,4 Water Release Assessment	174

2.20.4.1	Water Release Sources	174

2.20.4.2	Water Release Assessment Results	174

2.21	Waste Handling, Disposal, Treatment, and Recycling	174

2.21.1	Estimates of Number of Facilities	174

2.21.2	Process Description	174

2.21.3	Exposure Assessment	179

2.21.3.1	Worker Activities	179

2.21.3.2	Number of Potentially Exposed Workers	180

2.21.3.3	Occupational Exposure Results	182

2.21.4	Water Release Assessment	183

2.21.4.1	Water Release Sources	183

2.21.4.2	Water Release Assessment Results	183

2.22	Other Department of Defense Uses	184

2.22.1	Data Overview	185

2.22.2	Process Description and Worker Activities	186

2.22.3	Occupational Exposure Assessment	186

2.22.4	Water Release Assessment	187

2.23	Dermal Exposure Assessment	187

3 DISCUSSION OF UNCERTAINTIES AND LIMITATIONS	193

3.1	Variability	193

3.2	Uncertainties and Limitations	193

3.2.1	Number of Workers	193

3.2.2	Analysis of Exposure Monitoring Data	194

3.2.3	Near-Field/Far-Field Model Framework	194

3.2.3.1	Tank Truck and Railcar Loading and Unloading Release and Inhalation Exposure
Model	195

3.2.3.2	EPA AP-42 Loading Model and EPA Mass Balance Inhalation Model	196

3.2.3.3	Vapor Degreasing and Cold Cleaning Models	196

3.2.3.4	Brake Servicing Model	197

3.2.3.5	Dry Cleaning Model	197

3.2.4	Modeled Dermal Exposures	198

REFERENCES	199

Appendix A Approach for Estimating Number of Workers and Occupational Non-Users	200

Appendix B Equations for Calculating Acute and Chronic (Non-Cancer and Cancer) Inhalation
Exposures 206

Appendix C Sample Calculations for Calculating Acute and Chronic (Non-Cancer and Cancer)
Inhalation Exposures	212

C.l Example High-End AC, ADC, and LADC Calculations						.212

C,2 Example Central. Tendency AC, ADC, and LADC Calculations 						212

Appendix D Approach for Estimating Water Releases from Manufacturing Sites Using Effluent
Guidelines 214

Page 7 of316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Appendix E Tank Truck and Railcar Loading and Unloading Release and Inhalation Exposure
Model Approach and Parameters	218

E.l Displacement of Saturated Air Inside Tank Trucks and Railcars				.,218

E.2 Emissions of Saturated Air that Remain in Transfer Hoses/Loading Arm				......219

E,3 Emissions from Leaks													..221

E.4	Exposure Estimates																					....224

Appendix F Drum Loading and Unloading Release and Inhalation Exposure Model Approach
and Parameters	227

F.l	Model Design Equations..........................									..227

F.2	Model Parameters															.........229

F.2.1 Saturation Factor	232

F.2.2 Volume of Container	232

1 .2.3 Fill Rate	232

F.2.4 Vapor Pressure Correction Factor and Mole Fraction	232

F.2.5 Ventilation Rate	232

F.2.6 Mixing Factor	233

F.2.7 Weight Fraction of Chemical	233

F.2.8 Production Volume	233

F.2.9 Number of Sites	234

F.2.10 Operating Days	234

F.2.11 Exposure Frequency	234

F.2.12	Exposure Duration	234

Appendix G Vapor Degreasing and Cold Cleaning Near-Field/Far-Field Inhalation Exposure
Models Approach and Parameters	236

G.l	Model Design Equations															.....236

G.2	Model Parameters											.......................241

G.2.1	Far-Field Volume	245

G.2.2 Air Exchange Rate	245

G. 2.3 N ear-F i el d Indoor AirSpeed	245

G.2.4 Near-Field Volume	246

G.2.5 Exposure Duration	246

G.2.6 Averaging Time	246

G.2.7 Vapor Generation Rate	246

G.2.8	Operating Hours	248

Appendix H Brake Servicing Near-Field/Far-Field Inhalation Exposure Model Approach and
Parameters 250

H.l	Model Design Equations...........................										250

H.2 Model Parameters..............							256

H.2.1	Far-Field Volume	260

H.2.2 Air Exchange Rate	260

H.2.3 Near-Field Indoor Air Speed	260

H.2.4 Near-Field Volume	261

H.2.5 Application Time	261

Page 8 of316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

H.2.6 Averaging Time	261

H.2.7 Perchloroethylene Weight Fraction	261

H.2.8 Volume of Degreaser Used per Brake Job	262

H.2.9 Number of Applications per Brake Job	262

H.2.10 Amount of Perchloroethylene Used per Application	262

H.2.11 Operating Hours per Week	263

H.2.12	Number of Brake Jobs per Work Shift	263

Appendix I Dry Cleaning Multi-Zone Inhalation Exposure Model Approach and Parameters 264

1.1	Model Design Equations											...265

1.2	Model Parameters...																	.........................273

I.2.1	Facility Parameters	276

1.2.1.1	Far-Field Volume	276

1.2.1.2	Air Exchange Rate	276

1.2.1.3	Near-Field Indoor Air Speed	277

1.2.2	Dry Cleaning Machine Parameters	277

1.2.2.1	Machine Door Diameter	277

1.2.2.2	Number of Loads per Day	277

1.2.2.3	Load Time	277

1.2.2.4	Machine Cylinder Concentration	278

1.2.2.5	Cylinder Volume	278

1.2.2.6	Exposure Duration	278

1.2.3	Finishing and Pressing Parameters	278

1.2.3.1	Near-Field Volume	278

1.2.3.2	Residual Solvent	278

1.2.3.3	Load Size	279

1.2.3.4	Exposure Duration	281

1.2.4	Spot Cleaning Parameters	281

1.2.4.1	Near-Field Volume	281

1.2.4.2	Spot Cleaning Use Rate	281

1.2.4.3	Exposure Duration	282

1.2.5	Other Parameters	282

1.2.5.1	Operating Hours	282

1.2.5.2	Operating Days per Year	282

1.2.5.3	Fractional Number of Operating Days that a Worker Works	282

Appendix J Solvent Releases in Water Discharge from Dry Cleaning Machines Model Approach
and Parameters	283

J.l Model Design Equations								....284

J.2 Model Parameters..........															...........285

J.2.1 Solubility in Water	285

J.2.2 Produced Water	285

J.2.3 Load Size	287

J.2.4 Number of Loads per Day	289

J.2.5 Number of Machines per Site	289

J.2.6 Operating Days per Year	290

Page 9 of316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Appendix K Dermal Exposure to Volatile Liquids Model Approach and Parameters	292

K.1 Incorporating the Effects of Evaporation....					..........292

K. 1.1 Modification of EPA/OPPT Models	292

K.2 Calculation of/abs																	292

K.2.1 Small Doses (Case 1: Yl„ Msat)	293

K.2.2 Large Doses (Case 2: Mo > Msat)	294

K.3 Comparison of/abs to Experimental Values for 1-BP												.295

K.4 Potential for Occlusion............................									296

K.5 Incorporating Glove Protection										297

K.6 Proposed Dermal Dose Equation 																....298

K.7 Equations for Calculating Acute and Chronic (Non-Cancer and Cancer) Dermal Dose		299

Page 10 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

LIST OF TABLES

Table 1-1. Crosswalk of Subcategories of Use Listed in the Problem Formulation Document to

Conditions of Use Assessed in the Risk Evaluation	23

Table 1-2. Assigned Protection Factors for Respirators in OSHA Standard 29 CFR 1910.134	33

Table 1-3. Number and Percent of Establishments and Employees Using Respirators Within 12 Months

Prior to Survey	35

Table 2-1. List of Assessed Perchloroethylene Manufacturing Sites	38

Table 2-2. Estimated Number of Workers Potentially Exposed to Perchloroethylene During

Manufacturing	40

Table 2-3. Summary of Worker Inhalation Monitoring Data for the Manufacture of Perchloroethylene 41

Table 2-4. Summary of OCPSF Effluent Guidelines for Perchloroethylene	43

Table 2-5. Reported Wastewater Discharges of Perchloroethylene from Manufacturing Sites Reporting

to 2016 TRI	44

Table 2-6. Estimated Wastewater Discharges of Perchloroethylene from Manufacturing Sites Not

Reporting to 2016 TRI	45

Table 2-7. Crosswalk of Repackaging SIC Codes in DMR to NAICS Codes	48

Table 2-8. Estimated Number of Workers Potentially Exposed to Perchloroethylene During

Repackaging	48

Table 2-9. Summary of Worker Inhalation Exposure Monitoring Data for Repackaging of

Perchloroethylene	49

Table 2-10. Reported Wastewater Discharges of Perchloroethylene from Repackaging Sites	50

Table 2-11. Crosswalk of Reactant SIC Codes in DMR to NAICS Codes	53

Table 2-12. Estimated Number of Workers Potentially Exposed to Perchloroethylene During Processing

as a Reactant	54

Table 2-13. Summary of Worker Inhalation Monitoring Results for Processing Perchloroethylene as a

Reactant	55

Table 2-14. Reported Wastewater Discharges of Perchloroethylene from Sites Processing

Perchloroethylene as a Reactant	56

Table 2-15. Crosswalk of Formulation SIC Codes in DMR to NAICS Codes	60

Table 2-16. Estimated Number of Workers Potentially Exposed to Perchloroethylene During

Formulation	61

Table 2-17. Summary of Worker Inhalation Exposure Monitoring Data for Aerosol Packing Formulation

Sites	62

Table 2-18. Estimated Throughputs of Perchloroethylene by Formulated Product Type	63

Table 2-19. Summary of Exposure Modeling Results for Formulation of Perchloroethylene-Based

Products	65

Table 2-20. Reported Wastewater Discharges of Perchloroethylene from Formulation of

Perchloroethylene-Containing Products	66

Table 2-21. Estimated Number of Workers Potentially Exposed to Perchloroethylene During Use in

Open-Top Vapor Degreasing	71

Table 2-22. Summary of Worker Inhalation Exposure Monitoring Data for Open-Top Vapor Degreasing

	72

Table 2-23. Reported Wastewater Discharges of Perchloroethylene from Sites Using Perchloroethylene
in Open-Top Vapor Degreasing	74

Page 11 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Table 2-24. Estimated Number of Workers Potentially Exposed to Perchloroethylene During Use in

Closed-Loop Vapor Degreasing	78

Table 2-25. Summary of Worker Inhalation Exposure Monitoring Data for Closed-Loop Vapor

Degreasing	79

Table 2-26. Estimated Number of Workers Potentially Exposed to Perchloroethylene During Use in

Conveyorized Vapor Degreasing	85

Table 2-27. Summary of Exposure Modeling Results for Use of Perchloroethylene in Conveyorized

Vapor Degreasing	87

Table 2-28. Estimated Number of Workers Potentially Exposed to Perchloroethylene During Use in

Web Degreasing	89

Table 2-29. Unit Emission Rates Used to Model Perchloroethylene Web Degreasing Systems	90

Table 2-30. Summary of Exposure Modeling Results for Use of Perchloroethylene in Web Degreasing91
Table 2-31. Estimated Number of Workers Potentially Exposed to Perchloroethylene During Use in

Cold Cleaning	94

Table 2-32. Summary of Worker Inhalation Exposure Monitoring Data for Use of Perchloroethylene in

Cold Cleaning	96

Table 2-33. Unit Emission Rates Used to Model Perchloroethylene Cold Cleaning	97

Table 2-34. Unit Operating Hours Used to Model Perchloroethylene Cold Cleaning	98

Table 2-35. Summary of Exposure Modeling Results for Use of Perchloroethylene in Cold Cleaning.. 99

Table 2-36. NAICS Codes for Aerosol Degreasing and Lubricants	100

Table 2-37. Estimated Number of Workers Potentially Exposed to Perchloroethylene During Use of

Aerosol Degreasers and Aerosol Lubricants	103

Table 2-38. Summary of Worker Inhalation Exposure Monitoring Data for Aerosol Degreasing	104

Table 2-39. Summary of Worker and Occupational Non-User Inhalation Exposure Modeling Results for

Aerosol Degreasing	106

Table 2-40. Estimated Number of Workers Potentially Exposed to Perchloroethylene During Dry

Cleaning	110

Table 2-41. Summary of Survey Responses for Dry Cleaning Machine Generations	Ill

Table 2-42. Summary of Worker Inhalation Exposure Monitoring Data for Dry Cleaning	114

Table 2-43. Summary of Worker and Occupational Non-User Inhalation Exposure Modeling Results for

Dry Cleaning	119

Table 2-44. Reported Wastewater Discharges of Perchloroethylene for Industrial Launderers in 2016

DMR	121

Table 2-45. Model Results for Perchloroethylene Discharges to POTW from Dry Cleaning Sites	122

Table 2-46. Summary of Direct Discharge Data for Commercial Dry Cleaning Reporters in the 2016

DMR	122

Table 2-47. Perchloroethylene Use Rate at Coating and Adhesive Application Sites	124

Table 2-48. Estimated Number of Workers Potentially Exposed to Perchloroethylene During Use of

Adhesives and Sealants	129

Table 2-49. Estimated Number of Workers Potentially Exposed to Perchloroethylene During Use of

Paints and Coatings	129

Table 2-50. Summary of Inhalation Exposure Monitoring Data for Use of Perchloroethylene-Based

Adhesives	131

Table 2-51. Summary of Inhalation Exposure Monitoring Data for Use of Perchloroethylene-Based

Paints/Coatings	133

Table 2-52. Parameters for Estimating Water Discharges from Spray Coating Applications	136

Page 12 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Table 2-53. Estimated Wastewater Discharges of Perchloroethylene from Coating and Adhesive

Application Sites	137

Table 2-54. Crosswalk of Maskant SIC Codes in DMR to NAICS Codes	142

Table 2-55. Estimated Number of Workers Potentially Exposed to Perchloroethylene During Use of

Chemical Maskants	144

Table 2-56. Summary of Worker Inhalation Exposure Monitoring Data for Chemical Maskants	145

Table 2-57. Reported Wastewater Discharges of Perchloroethylene from Chemical Maskant Sites	147

Table 2-58. Crosswalk of Processing Aid SIC Codes in DMR to NAICS Codes	149

Table 2-59. Estimated Number of Workers Potentially Exposed to Perchloroethylene During Use as a

Processing Aid	150

Table 2-60. Summary of Worker inhalation Monitoring Data for Use of Perchloroethylene as a

Processing Aid	151

Table 2-61. Reported Wastewater Discharges of Perchloroethylene from Processing Aid Sites	152

Table 2-62. ESD Exposure Estimates for Metalworking Fluids Based on Monitoring Data	156

Table 2-63. Summary of Exposure Results for Use of PCE in Metalworking Fluids Based on ESD

Estimates	156

Table 2-64. Summary of Worker Inhalation Monitoring Data for Use of Perchloroethylene as a Wipe

Cleaning Solvent and Metal/Stone Polishes	159

Table 2-65. Summary of Worker Inhalation Exposure Monitoring Data for Other Spot Cleaning/Spot

Removers (Including Carpet Cleaning)	161

Table 2-66. Crosswalk of Other Industrial Use SIC Codes in DMR to NAICS Codes	162

Table 2-67. Estimated Number of Workers Potentially Exposed to Perchloroethylene During Other

Industrial Uses	164

Table 2-68. Summary of Exposure Modeling Results for Other Industrial Uses of Perchloroethylene 167

Table 2-69. Reported Wastewater Discharges of Perchloroethylene from Other Industrial Uses	168

Table 2-70. Summary of Exposure Monitoring Data for Other Commercial Uses of Perchloroethylene

	171

Table 2-71. Reported Wastewater Discharges of Perchloroethylene from Other Commercial Uses in the

2016 DMR	172

Table 2-72. Crosswalk of Disposal SIC Codes in DMR to NAICS Codes	181

Table 2-73. Estimated Number of Workers Potentially Exposed to Perchloroethylene During

Disposal/Treatment	182

Table 2-74. Reported Wastewater Discharges of Perchloroethylene from Disposal/Treatment of

Perchloroethylene-Containing Wastes	183

Table 2-75. Summary of DoD Inhalation Monitoring Data Not Included in Assessments for Other

Conditions of Use	185

Table 2-76. Summary of Worker Inhalation Monitoring Data for Other DoD Uses of Perchloroethylene

	187

Table 2-77. Glove Protection Factors for Different Dermal Protection Strategies	188

Table 2-78. Estimated Dermal Acute Retained Dose for Workers in All Conditions of Use	191

LIST OF FIGURES

Figure 2-1. Use of Vapor Degreasing in a Variety of Industries	68

Figure 2-2. Open-Top Vapor Degreaser	69

Figure 2-3. Open-Top Vapor Degreaser with Enclosure	70

Page 13 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Figure 2-4. Closed-Loop/Vacuum Vapor Degreaser	77

Figure 2-5. Monorail Conveyorized Vapor Degreasing System (U.S. EPA, 1977)	 81

Figure 2-6. Cross-Rod Conveyorized Vapor Degreasing System (U.S. EPA, 1977)	 82

Figure 2-7. Vibra Conveyorized Vapor Degreasing System (U.S. EPA, 1977)	 83

Figure 2-8. Ferris Wheel Conveyorized Vapor Degreasing System (U.S. EPA, 1977)	 84

Figure 2-9. Belt/Strip Conveyorized Vapor Degreasing System (U.S. EPA, 1977)	 84

Figure 2-10. Schematic of the Conveyorized Degreasing Near-Field/Far-Field Inhalation Exposure

Model	86

Figure 2-11. Web Degreasing System	88

Figure 2-12. Schematic of the Web Degreasing Near-Field/Far-Field Inhalation Exposure Model	90

Figure 2-13. Typical Batch-Loaded, Maintenance Cold Cleaner (U.S. EPA, 1981)	93

Figure 2-14. Schematic of the Cold Cleaning Near-Field/Far-Field Inhalation Exposure Model	96

Figure 2-15. Overview of Aerosol Degreasing	102

Figure 2-16. Schematic of the Brake Servicing Near-Field/Far-Field Inhalation Exposure Model	105

Figure 2-17. Overview of Dry Cleaning Process	108

Figure 2-18. Overview of Use of Spot Cleaning at Dry Cleaners	109

Figure 2-19. Illustration of the Dry Cleaning Multi-Zone Inhalation Exposure Model	116

Figure 2-20. General Laboratory Use Process Flow Diagram	173

Figure 2-21. Typical Waste Disposal Process	176

Figure 2-22.Typical Industrial Incineration Process	177

Figure 2-23. General Process Flow Diagram for Solvent Recovery Processes (U.S. EPA, 1980)	 179

LIST OF APPENDIX TABLES

TableApx A-l. SOCs with Worker and ONU Designations for All Conditions of Use Except Dry

Cleaning	201

Table_Apx A-2. SOCs with Worker and ONU Designations for Dry Cleaning Facilities	202

Table_Apx A-3. Estimated Number of Potentially Exposed Workers and ONUs under NAICS 812320

	204

Table Apx B-l. Parameter Values for Calculating Inhalation Exposure Estimates	207

TableApx B-2. Overview of Average Worker Tenure from U.S. Census SIPP (Age Group 50+)	210

Table Apx B-3. Median Years of Tenure with Current Employer by Age Group	211

Table Apx D-l. Summary of OCPSF Effluent Guidelines for Perchloroethylene	214

Table Apx D-2. Default Parameters for Estimating Water Releases of Perchloroethylene from

Manufacturing Sites	215

TableApx D-3. Summary of Facility Perchloroethylene Production Volumes and Wastewater Flow

Rates	216

Table Apx E-l. Example Dimensions and Volume of Loading Arm/Transfer System	220

Table Apx E-2. Default Values for Calculating Emissions Rate of Perchloroethylene from

Transfer/Loading Arm	221

TableApx E-3. Parameters for Calculating Emission Rate of Perchloroethylene from Equipment Leaks

	222

Table_Apx E-4. Default Values for Fa and N	223

Table Apx E-5. Parameters for Calculating Exposure Concentration Using the EPA/OPPT Mass

Balance Model	225

Page 14 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

TableApx E-6. Calculated Emission Rates and Resulting Exposures from the Tank Truck and Railcar
Loading and Unloading Release and Inhalation Exposure Model for Perchloroethylene

	226

TableApx F-l. Summary of Parameter Values and Distributions Used in the Inhalation Exposure

Model	230

TableApx G-l. Summary of Parameter Values and Distributions Used in the Conveyorized Degreasing

Near-Field/Far-Field Inhalation Exposure Model	242

Table Apx G-2. Summary of Parameter Values and Distributions Used in the Web Degreasing Near-

Field/Far-Field Inhalation Exposure Model	243

TableApx G-3. Summary of Parameter Values and Distributions Used in the Cold Cleaning Near-

Field/Far-Field Inhalation Exposure Model	244

Table Apx G-4. Summary of Perchloroethylene Vapor Degreasing and Cold Cleaning Data from the

2014 Mil	246

TableApx G-5. Distribution of Perchloroethylene Conveyorized Degreasing Unit Emissions	247

Table Apx G-6. Distribution of Perchloroethylene Web Degreasing Unit Emissions	247

Table Apx G-l. Distribution of Perchloroethylene Cold Cleaning Unit Emissions	248

Table Apx G-8. Distribution of Perchloroethylene Conveyorized Degreasing Operating Hours	249

Table Apx G-9. Distribution of Perchloroethylene Web Degreasing Operating Hours	249

Table Apx G-10. Distribution of Perchloroethylene Cold Cleaning Operating Hours	249

TableApx H-l. Summary of Parameter Values and Distributions Used in the Brake Servicing Near-

Field/Far-Field Inhalation Exposure Model	257

Table Apx H-2. Summary of Perchloroethylene-Based Aerosol Degreaser Formulations	262

Table Apx 1-1. Summary of Survey Responses for Dry Cleaning Machine Generations	272

Table Apx 1-2. Summary of Parameter Values and Distributions Used in the Dry Cleaning Multi-Zone

Inhalation Exposure Model	274

TableApx 1-3. Composite Distribution of Dry Cleaning Facility Floor Areas	276

Table Apx 1-4. Survey Responses of Actual Pounds Washed per Load for Primary Machine (if more

than one machine) from 2010 King County Survey	279

Table Apx 1-5. Distribution of Actual Load Sizes from 2010 King County Survey	280

Table Apx J-l. Summary of Parameter Values and Distributions for the Solvent Release in Water

Discharge from Dry Cleaning Machines Model	285

Table Apx J-2. Distribution of Machine Types Based on 2010 King County Survey Results	286

TableApx J-3. Distribution of Produced Separator Water Emission Factors by Machine Type Used in

Model	287

Table Apx J-4. Survey Responses of Actual Pounds Washed per Load for Primary Machine (if more

than one machine) from 2010 King County Survey	287

Table Apx J-5. Distribution of Actual Load Sizes from 2010 King County Survey	288

Table_Apx J-6. Survey Results of Number of Machines per Facility	290

Table_Apx J-7. Distribution of Number of Machines per Facility Used in the Model	290

Table Apx K-l. Estimated Fraction Evaporated and Absorbed (/abs) using Equation Apx K-8	294

Table Apx K-2. Exposure Control Efficiencies and Protection Factors for Different Dermal Protection

Strategies from ECETOC TRA v3	298

Table Apx K-3. Parameter Values for Calculating Dermal Dose Estimates	300

TableApx K-4. Overview of Average Worker Tenure from U.S. Census SIPP (Age Group 50+)	 303

Table_Apx K-5. Median Years of Tenure with Current Employer by Age Group	304

Page 15 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

LIST OF APPENDIX FIGURES

FigureApx E-l. Illustration of Transfer Lines Used During Tank Truck Unloading and Associated

Equipment Assumed by EPA	224

Figure Apx G-l. The Near-Field/Far-Field Model as Applied to the Cold Cleaning Near-Field/Far-Field

Inhalation Exposure Model	237

Figure Apx G-2. The Near-Field/Far-Field Model as Applied to the Conveyorized Degreasing Near-

Field/Far-Field Inhalation Exposure Model	238

Figure Apx G-3. The Near-Field/Far-Field Model as Applied to the Web Degreasing Near-Field/Far-

Field Inhalation Exposure Model	238

Figure Apx H-l. The Near-Field/Far-Field Model as Applied to the Brake Servicing Near-Field/Far-

Field Inhalation Exposure Model	251

Figure Apx 1-1. Illustration of the Dry Cleaning Multi-Zone Inhalation Exposure Model	266

Figure Apx 1-2. Fit Comparison of Beta Cumulative Distribution Function to Load Size Survey Results

	281

Figure Apx J-l. Process Flow Diagram of a 5th Generation Dry Cleaning Machine (NIOSH, 1997a) 283
Figure Apx J-2. Fit Comparison of Beta Cumulative Distribution Function to Load Size Survey Results
	289

Page 16 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

ABBREVIATIONS

AC

Acute Concentration

ACGM

Association Advancing Occupational and Environmental Health

ACP

AC Products

ADC

Average Daily Concentration

AFPM

American Fuel and Petrochemical Manufacturers

AIHA

American Industrial Hygiene Association

AISE

Association for Soaps, Detergents and Maintenance Products

AL

Alabama

AR

Arizona

ATSDR

Agency for Toxic Substances and Disease Registry

BAT

Best Available Technology Economically Achievable

BLS

Bureau of Labor Statistics

BPT

Best Practicable Control Technology Currently Available

CA

California

CARB

California Air Resources Board

CBI

Confidential Business Information

CDC

Center for Disease Control

CDR

Chemical Data Reporting

CFC

Chi orofluorocarb on

Cff

Concentration (Far-Field)

Cnf

Concentration (Near-Field)

CO

Colorado

CT

Connecticut

CTFE

Chi orotrifluoroethy 1 ene

DEP

Department of Environmental Protection

DEQ

Department of Environmental Quality

DIY

Do It Yourself

DLI

Dry Cleaning and Laundry Institute

DMR

Discharge Monitoring Report

ECHO

Enforcement and Compliance History Online

EDC

Ethylene Dichloride

EG

Effluent Guidelines

EPA

Environmental Protection Agency

ERG

Eastern Research Group

ESD

Emission Scenario Document

EU

European Union

FL

Florida

G

Evaporation rate of PCE

GS

Generic Scenario

HCFC

Hy drochl orofluorocarb on

HERO

Health and Environmental Research Online

HF

Hydrofluoric Acid

HHE

Health Hazard Evaluation

HSIA

Halogenated Solvents Industry Alliance

Page 17 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

IL

Illinois

IN

Indiana

ISOR

Initial Statement of Reasons

KS

Kansas

KY

Kentucky

LA

Louisiana

LADC

Lifetime Average Daily Concentration

lb

Pound

LEV

Local Exhaust Ventilation

LOD

Limit of Detection

LPG

Liquified Petroleum Gas

MI

Missouri

MN

Minnesota

MP&M

Metal Products and Machinery

NAICS

North American Industry Classification System

NCA

National Cleaner's Association

NEI

National Emissions Inventory

NESHAP

National Emission Standards for Hazardous Air Pollutants

NEWMOA

Northeast Waste Management Officials' Association

NIOSH

National Institute for Occupational Safety and Health

NJ

New Jersey

NKRA

Not Known or Reasonably Ascertainable

NPDES

National Pollutant Discharge Elimination System

NSPS

New Source Performance Standards

NV

Nevada

NY

New York

OAQPS

Office of Air Quality Planning and Standards

OARS

Occupational Alliance for Risk Science

OCPSF

Organic Chemicals, Plastics and Synthetic Fibers

OECD

Organization for Economic Co-operation and Development

OEL

Occupational Exposure Limit

OES

Occupational Employment Statistics

OH

Ohio

OK

Oklahoma

ONU

Occupational Non-User

OPPT

Office of Pollution Prevention and Toxics

OSHA

Occupational Safety and Health Administration

OTVD

Open-Top Vapor Degreaser

PA

Pennsylvania

PBZ

Personal Breathing Zone

PCE

Perchloroethylene

PEL

Permissible Exposure Limit

PERC

Perchloroethylene

PF

Protection Factor

POTW

Publicly-Owned Treatment Works

PPE

Personal Protective Equipment

Page 18 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

PSES

Pretreatment Standards for Existing Sources

PSNS

Pretreatment Standards for New Sources

QC

Quality Control

Qff

Ventilation rate (Far-Field)

Qnf

Ventilation rate (Near-Field)

RAR

Risk Assessment Report

RCRA

Resource Conservation and Recovery Act

RDF

Refuse-Derived Fuel

REL

Recommended Exposure Limit

RFI

Reporting Forms and Instructions

SC

South Carolina

SDS

Safety Data Sheet

SIC

Standard Industrial Classification

SOC

Standard Occupational Classification

SpERC

Specific Environmental Release Category

SUSB

Statistics of U.S. Businesses

TCE

T ri chloroethy 1 ene

TLV

Threshold Limit Value

TRI

Toxic Release Inventory

TSCA

Toxic Substances Control Act

TSDF

Treatment, Storage, and Disposal Facility

TTO

Total Toxic Organics

TWA

Time-Weighted Average

TX

Texas

US

United States

USA

United States of America

UT

Utah

Vff

Volume (Far-Field)

Vnf

Volume (Near-Field)

voc

Volatile Organic Compound

VT

Vermont

WA

Washington

WEEL

Workplace Environmental Exposure Level

WI

Wisconsin

WV

West Virginia

WWT

Wastewater Treatment

yr

Year

Page 19 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

EXECUTIVE SUMMARY

TSCA § 6(b)(4) requires the United States Environmental Protection Agency (EPA) to establish a risk
evaluation process. In performing risk evaluations for existing chemicals, EPA is directed to "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." In December of 2016, EPA published a list of 10 chemical substances that
are the subject of the Agency's initial chemical risk evaluations (81 FR 91927), as required by TSCA §
6(b)(2)(A). Perchloroethylene (PCE) was one of these chemicals.

PCE, also known as ethene, 1,1,2,2-tetrachloro, tetrachloroethylene, and PERC, is a colorless volatile
liquid with a mildly sweet odor that is used primarily as a reactant, a dry cleaning solvent, a vapor
degreasing solvent, and aerosol degreasing solvent and is subject to federal and state regulations and
reporting requirements. PCE is a Toxics Release Inventory (TRI)-reportable substance effective January
1, 1987.

Focus of this Risk Evaluation

During scoping and problem formulation, EPA considered all known TSCA uses for PCE. PCE has been
manufactured and imported in the U.S. in large volumes with the most recently available data from the
2016 Chemical Data Reporting (CDR) indicating approximately 324 million pounds were either
manufactured or imported in the U.S. in 2015 (	). The largest use of PCE are as a

reactant/intermediate in the production of fluorinated compounds, such as hydrofluorocarbons (HFCs)
and hydrochlorofluorocarbons (HCFCs). The second largest use of PCE is as a dry cleaning solvent;
however, in recent years, there appears to be a trend towards alternatives to PCE in the dry cleaning
industry. The third most prevalent use of PCE is as a degreasing solvent for vapor degreasing machines,
cold cleaning machines, and aerosol degreasing products (e.g., brake cleaners) that are used to clean
contaminated metal parts or other fabricated materials.

Exposures to workers, consumers, general populations, and ecological species may occur from
industrial, commercial, and consumer uses of PCE and releases to air, water or land. Workers and
occupational non-users may be exposed to PCE during conditions of use such as manufacturing, import,
processing, distribution, repackaging, dry cleaning, degreasing, recycling, disposal, and other
miscellaneous uses of PCE. Consumers and bystanders may also be exposed to PCE via inhalation of
PCE that volatizes during use of consumer products or dermal contact with products containing PCE.
Exposures to the general population and ecological species may occur from industrial releases related to
the manufacture, import, processing, distribution, and use of PCE.

Risk Evaluation Approach

EPA evaluated acute and chronic exposures to workers and occupational non-users in association with
PCE conditions of use. EPA used inhalation monitoring data from literature sources where available and
exposure models where monitoring data were not available or were deemed insufficient for capturing
actual exposure within the condition of use. EPA also used modeling approaches to estimate dermal
exposures. EPA evaluated releases to water from the conditions of use assessed in this risk evaluation.
EPA used release data from literature sources where available and used modeling approaches where
release data were not available.

Page 20 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Uncertainties of this Risk Evaluation

There are a number of uncertainties associated with the monitoring and modeling approaches used to
assess PCE exposures and releases. For example, the sites used to collect exposure monitoring and
release data were not selected randomly, and the data reported therein may not be representative of all
sites pertaining to the exposure and release scenarios. Further, of necessity, modeling approaches
employed knowledge-based assumptions that may not apply to all use scenarios. Because site-specific
differences in use practices and engineering controls exist, but are largely unknown, this represents
another source of variability that EPA could not quantify in the assessment.

Human and Ecological Populations Considered in this Risk Evaluation

EPA assessed risks for acute and chronic exposure scenarios in workers (those directly handling PCE)
and occupational non-users (workers not directly involved with the use of PCE) for PCE in the uses
outlined under Focus of this Risk Evaluation. EPA assumed that workers and occupational non-users
would be individuals of both sexes (age 16 years and older, including pregnant workers) based upon
occupational work permits, although exposures to younger workers in occupational settings cannot be
ruled out. An objective of the monitored and modeled inhalation data was to provide separate exposure
level estimates for workers and occupational non-users.

EPA assessed releases to water to estimate exposures to aquatic species. The water release estimates
developed by EPA are used to estimate the presence of PCE in the environment and biota and evaluate
the environmental hazards. The release estimates were used to model exposure to aquatic species where
environmental monitoring data were not available.

Page 21 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

1 INTRODUCTION

1.1	Overview

TSCA § 6(b)(4) requires the United States Environmental Protection Agency (EPA) to establish a risk
evaluation process. In performing risk evaluations for existing chemicals, EPA is directed to "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." In December of 2016, EPA published a list of 10 chemical substances that
are the subject of the Agency's initial chemical risk evaluations (81 FR 91927), as required by TSCA §
6(b)(2)(A). Perchloroethylene (PCE) was one of these chemicals.

PCE, also known as ethene, 1,1,2,2-tetrachloro, tetrachloroethylene, and PERC, is a colorless volatile
liquid with a mildly sweet odor that is used primarily as a reactant, a dry cleaning solvent, a vapor
degreasing solvent, and aerosol degreasing solvent and is subject to federal and state regulations and
reporting requirements. PCE is a TRI-reportable substance effective January 1, 1987.

1.2	Scope

Workplace exposures and releases to water have been assessed for the following industrial and
commercial conditions of use of PCE:

1.	Manufacturing;

2.	Repackaging;

3.	Processing as a Reactant;

4.	Incorporation into Formulation, Mixture, or Reactant Product;

5.	Batch Open-Top Vapor Degreasing;

6.	Batch Closed-Loop Vapor Degreasing;

7.	Conveyorized Vapor Degreasing;

8.	Web Degreasing;

9.	Cold Cleaning;

10.	Aerosol Degreasing and Aerosol Lubricants;

11.	Dry Cleaning and Spot Cleaning;

12.	Adhesives, Sealants, Paints, and Coatings;

13.	Maskant for Chemical Milling;

14.	Industrial Processing Aid;

15.	Metalworking Fluids;

16.	Wipe Cleaning and Metal/Stone Polishes;

17.	Other Spot Cleaning/Spot Removers (Including Carpet Cleaning);

18.	Other Industrial Uses;

19.	Other Commercial Uses; and

20.	Waste Handling, Disposal, Treatment, and Recycling.

Page 22 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

For work place exposures, EPA considered exposures to both workers who directly handle PCE and
occupational non-users (ONUs) who do not directly handle PCE but may be exposed to vapors or mists
that enter their breathing zone while working in locations in close proximity to where PCE is being used.

For purposes of this report, "releases to water" include both direct discharges to surface water and
indirect discharges to publicly-owned treatment works (POTW) or non-POTW wastewater treatment
(WWT). It should be noted that for purposes of risk evaluation, discharges to POTW and non-POTW
WWT are not evaluated the same as discharges to surface water. EPA considers removal efficiencies of
POTWs and WWT plants and environmental fate and transport properties when evaluating risks from
indirect discharges. The purpose of this report is only to quantify direct and indirect discharges;
therefore, these factors are not discussed. The details on how these factors were considered when
determining risk are described in the Risk Evaluation for Perchloroethylene (Ethene, 1,1,2,2-
Tetrachloro).

The assessed conditions of use were described in Table 2-3 of the Problem Formulation of the Risk
Evaluation for Perchloroethylene (Ethene, 1,1,2,2-Tetrachloro) (Problem Formulation Document) (U.S.

); however, due to expected similarities in both processes and exposures/releases several of
the subcategories of use in Table 2-3 were grouped and assessed together during the risk evaluation
process. A crosswalk of the conditions of use in Table 2-3 to the conditions of use assessed in this report
is provided in Table 1-1.

Table 1-1. Crosswalk of Subcategories of Use Listed in the Problem Formulation Document to
Conditions of Use Assessed in the Risk Evaluation

Life Cycle Stage

Category1

Subcategory1'

Assessed Condition of I se

Manufacture

Domestic Manufacture

Domestic
Manufacture

Section 2.1- Manufacturing

Import

Import

Section 2.2 - Repackaging51

Processing

Processing as a
Reactant/Interm edi ate

Intermediate in
industrial gas
manufacturing

Section 2.3 - Processing as a
Reactant

Intermediate in basic
organic chemical
manufacturing

Intermediate in
petroleum refineries

Residual or
byproduct

Incorporated into
formulation mixture or
reaction product

Cleaning and
degreasing products

Section 2.4 - Incorporation
into Formulation, Mixture, or
Reactant Product

Adhesive and sealant
products

Page 23 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Life ( vole Stage

Category1

Subcategory1'

Assessed Condition of I se





Paint and coating
products







Other chemical
products and
preparations





Incorporated into
articles

Plastic and rubber
products

After further review, EPA
determined that PCE is not
incorporated into plastic
articles but rather is used as a
degreasing solvent at plastic
manufacture sites; therefore,
no exposure scenario was
developed for incorporation
into articles. Use of PCE as a
degreasing solvent at plastic
manufacture sites is assessed
with other degreasing
scenarios in Sections 2.5 to 2.9



Repackaging

Solvent for cleaning
or degreasing

Section 2.2 - Repackaging





Intermediate





Recycling

Recycling

Section 2.21 - Waste
Handling, Disposal, Treatment,
and Recycling

Distribution in
commerce

Distribution

Distribution

Activities related to
distribution (e.g., loading,
unloading) are considered
throughout the life cycle,
rather than using a single
distribution scenario."

Industrial use

Solvents (for cleaning
or degreasing)

Solvents and/or
Degreasers (cold,
aerosol spray or
vapor degreaser; not
specified in
comment)

See sections for specified
degreasing and cleaning
operations.





Batch vapor
degreaser (e.g.,
open-top, closed-
loop)

Section 2.5 - Batch Open-Top
Vapor Degreasing;

Section 2.6 - Batch Closed-
Loop Vapor Degreasing

Page 24 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Life ( vole Stage

Category1

Subcategory1'

Assessed Condition of I se





In-line vapor
degreaser (e.g.,
conveyorized, web
cleaner)

Section 2.7 - Conveyorized
Vapor Degreasing;

Section 2.8 - Web Degreasing

Cold cleaner

Section 2.9 - Cold Cleaning

Aerosol spray
degreaser/cleaner

Section 2.10- Aerosol
Degreasing and Aerosol
Lubricants

Dry cleaning solvent

Section 2.11- Dry Cleaning
and Spot Cleaning

Spot cleaner

Lubricants and greases

Lubricants and
greases (e.g.,
penetrating
lubricants, cutting
tool coolants, aerosol
lubricants)

Section 2.10- Aerosol
Degreasing and Aerosol
Lubricants;

Section 2.15 - Metalworking
Fluids

Adhesive and sealant
chemicals

Solvent-based
adhesives and
sealants

Section 2.12 - Adhesive,
Sealants, Paints, and Coatings

Paints and coatings
including paint and
coating removers

Solvent-based paints
and coatings,
including for
chemical milling

Section 2.12 - Adhesive,
Sealants, Paints, and Coatings;
Section 2.13- Maskant for
Chemical Milling

Processing aids, not
otherwise listed

Pesticide, fertilizer
and other agricultural
chemical
manufacturing

Section 2.14 - Industrial
Processing Aid

Processing aids,
specific to petroleum
production

Catalyst regeneration
in petrochemical
manufacturing

Section 2.14 - Industrial
Processing Aid

Other uses

Textile processing

Section 2.17- Other Spot
Cleaning/Spot Removers
(Including Carpet Cleaning);
Section 2.18 - Other Industrial
Uses

Wood furniture
manufacturing

Section 2.18 - Other Industrial
Uses

Page 25 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Life ( vole Stage

Category1

Subcategory1'

Assessed Condition of I se





Laboratory
chemicals

Section 2.20 - Laboratory
Chemicals

Foundry applications

Section 2.18 - Other Industrial
Uses

Commercial/consumer
use

Cleaning and furniture
care products

Cleaners and
degreasers (other)

Section 2.16- Wipe Cleaning
and Metal/Stone Polishes;
Section 2.17- Other Spot
Cleaning/Spot Removers
(Including Carpet Cleaning);
Section 2.19- Other
Commercial Uses

Dry cleaning solvent
Spot cleaner

Section 2.11- Dry Cleaning
and Spot Cleaning

Automotive care
products (e.g.,
engine degreaser and
brake cleaner)

Aerosol cleaner

Section 2.10- Aerosol
Degreasing and Aerosol
Lubricants

Non-aerosol cleaner

Section 2.16- Wipe Cleaning
and Metal/Stone Polishes

Lubricants and greases

Lubricants and
greases (e.g.,
penetrating
lubricants, cutting
tool coolants, aerosol
lubricants)

Section 2.10- Aerosol
Degreasing and Aerosol
Lubricants;

Section 2.15 - Metalworking
Fluids

Adhesives and sealant
chemicals

Adhesives for arts
and crafts

Not assessed in occupational
settings - consumer use only

Light repair
adhesives

Section 2.12 - Adhesive,
Sealants, Paints, and Coatings

Paints and coatings

Solvent-based paints
and coatings

Section 2.12 - Adhesive,
Sealants, Paints, and Coatings

Other Uses

Carpet cleaning

Section 2.17- Other Spot
Cleaning/Spot Removers
(Including Carpet Cleaning)

Laboratory
chemicals

Section 2.20 - Laboratory
Chemicals

Page 26 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Life ( vole Stage

Category1

Subcategory1'

Assessed Condition of I se





Metal (e.g., stainless
steel) and stone
polishes

Section 2.16- Wipe Cleaning
and Metal/Stone Polishes

Inks and ink removal
products

Section 2.19- Other
Commercial Uses

Welding

Section 2.10- Aerosol
Degreasing and Aerosol
Lubricants'3

Photographic film

Section 2.19- Other
Commercial Uses

Mold cleaning,
release and
protectant products

Section 2.19- Other
Commercial Uses

Disposal

Disposal

Industrial pre-
treatment

Section 2.21 - Waste
Handling, Disposal, Treatment,
and Recycling0

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

a The repackaging scenario covers only those sites that purchase PCE or PCE containing products from domestic and/or
foreign suppliers and repackage the PCE from bulk containers into smaller containers for resale. Sites that import and directly
process/use PCE are assessed in the relevant condition of use. Sites that that import and either directly ship to a customer site
for processing or use or warehouse the imported PCE and then ship to customers without repackaging are assumed to have no
exposures or releases and only the processing/use of PCE at the customer sites are assessed in the relevant conditions of use.
b Identified welding products were anti-spatter aerosol products; therefore, the assessment is included with the assessment of
other aerosol products.

Page 27 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

0 Each of the conditions of use of PCE may generate waste streams of the chemical that are collected and transported to third-
party sites for disposal, treatment, or recycling. Industrial sites that treat, dispose, or directly discharge onsite wastes that they
themselves generate are assessed in each condition of use assessment. This section only assesses wastes of PCE that are
generated during a condition of use and sent to a third-party site for treatment, disposal, or recycling.

1.3	Components of the Occupational Exposure and Environmental

Release Assessment

The occupational exposure and environmental release assessment of each condition of use comprises the
following components:

•	Estimates of Number of Facilities: An estimate of the number of sites that use PCE for the
given condition of use.

•	Process Description: A description of the condition of use, including the role of the chemical in
the use; process vessels, equipment, and tools used during the condition of use.

•	Worker Activities: A descriptions of the worker activities, including an assessment for potential
points of worker and occupational non-user (ONU) exposure.

•	Number of Workers and Occupational Non-Users: An estimate of the number of workers and
occupational non-users potentially exposed to the chemical for the given condition of use.

•	Occupational Inhalation Exposure Results: Central tendency and high-end estimates of
inhalation exposure to workers and occupational non-users. See Section 1.4.5 for a discussion of
EPA's statistical analysis approach for assessing inhalation exposure.

•	Water Release Sources: A description of each of the potential sources of water releases in the
process for the given condition of use.

•	Water Release Assessment Results: Estimates of chemical released into water (surface water,
POTW, or non-POTW WWT).

In addition to the above components for each condition of use, a separate dermal exposure section is
included that provides estimates of the dermal exposures for all the assessed conditions of use.

1.4	General Approach and Methodology for Occupational Exposures and

Environmental Releases

1.4.1 Estimates of Number of Facilities

Where available, EPA used 2016 CDR (	), 2016 TRI(	), 2016

Discharge Monitoring Report (DMR) (I v >1 P \ i b) and 2014 National Emissions Inventory (NEI)
(	2016a) data to provide a basis to estimate the number of sites using PCE within a condition

of use. Generally, information for reporting sites in CDR and NEI was sufficient to accurately
characterize each reporting sites condition of use. However, information for determining the condition
of use for reporting sites in TRI and DMR is typically more limited.

In TRI, sites submitting a Form R indicate whether they perform a variety of activities related to the
chemical including, but not limited to: produce the chemical; import the chemical; use the chemical as a
reactant; use the chemical as a chemical processing aid; and ancillary or other use. In TRI, sites
submitting Form A are not required to designate an activity. For both Form R and Form A, TRI sites are
also required to report the primary North American Industry Classification System (NAICS) code for

Page 28 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

their site. For each TRI site, EPA used the reported primary NAICS code and activity indicators to
determine the condition of use at the site. For instances where EPA could not definitively determine the
condition of use because: 1) the report NAICS codes could include multiple conditions of use; 2) the site
report multiple activities; and/or 3) the site did not report activities due to submitting a Form A, EPA
had to make an assumption on the condition of use to avoid double counting the site. For these sites,
EPA supplemented the NAICS code and activity information with the following information to
determine a "most likely" or "primary" condition of use:

1.	Information on known uses of the chemical and market data identifying the most prevalent
conditions of use of the chemical.

2.	Information obtained from public comments and/or industry meetings with EPA that provided
specific information on the site.

In DMR, the only information reported on condition of use is each site's Standard Industrial
Classification (SIC) code. EPA could not determine each reporting site's condition of use based on SIC
code alone; therefore, EPA supplemented the SIC code information with the same supplementary
information used for the TRI sites (market data, public comments, and industry meetings).

Where the number of sites could not be determined using CDR/TRI/DMR/NEI or where
CDR/TRI/DMR/NEI data were determined to not capture the entirety of sites within a condition of use,
EPA supplemented the available data with U.S. economic data using the following method:

1.	Identify the NAICS codes for the industry sectors associated with these uses.

2.	Estimate total number of sites using the U.S. Census' Statistics of US Businesses (SUSB) (SUSB
Data) data on total establishments by 6-digit NAICS.

3.	Use market penetration data to estimate the percentage of establishments likely to be using PCE
instead of other chemicals.

4.	Combine the data generated in Steps 1 through 3 to produce an estimate of the number of sites
using PCE in each 6-digit NAICS code, and sum across all applicable NAICS codes for the
condition of use to arrive at a total estimate of the number of sites within the condition of use.

1.4.2	Process Description

EPA performed a literature search to find descriptions of processes involved in each condition of use.
Where process descriptions were unclear or not available, EPA referenced relevant Emission Scenario
Documents (ESD) or Generic Scenarios (GS). Process descriptions for each condition of use can be
found in Section 2.

1.4.3	Worker Activities

EPA performed a literature search to identify worker activities that could potentially result in
occupational exposures. Where worker activities were unclear or not available, EPA referenced relevant
ESD's or GS's. Worker activities for each condition of use can be found in Section 2.

1.4.4	Number of Workers and Occupational Non-Users

Where available, EPA used CDR data to provide a basis to estimate the number of workers and ONUs.
EPA supplemented the available CDR data with U.S. economic data using the following method:

Page 29 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

1.	Identify the North American Industry Classification System (NAICS) codes for the industry
sectors associated with these uses.

2.	Estimate total employment by industry/occupation combination using the Bureau of Labor
Statistics' Occupational Employment Statistics (OES) data (BLS Data).

3.	Refine the OES estimates where they are not sufficiently granular by using the U.S. Census'
Statistics of US Businesses (SUSB) (SUSB Data) data on total employment by 6-digit NAICS.

4.	Use market penetration data to estimate the percentage of employees likely to be using PCE
instead of other chemicals.

5.	Where market penetration data are not available, use the estimated workers/ONUs per site in the
6-digit NAICS code and multiply by the number of sites estimated from CDR, TRI, DMR and/or
NEI. In DMR data, sites report Standard Industrial Classification (SIC) codes rather than NAICS
codes; therefore, EPA mapped each reported SIC code to a NAICS code for use in this analysis.

6.	Combine the data generated in Steps 1 through 5 to produce an estimate of the number of
employees using PCE in each industry/occupation combination, and sum these to arrive at a total
estimate of the number of employees with exposure within the condition of use.

1.4,5 Inhalation Exposure Assessment Approach and Methodology

1.4.5.1 General Approach

EPA provided occupational exposure results representative of central tendency conditions and high-end
conditions. A central tendency is assumed to be representative of occupational exposures in the center of
the distribution for a given condition of use. For risk evaluation, EPA used the 50th percentile (median),
mean (arithmetic or geometric), mode, or midpoint values of a distribution as representative of the
central tendency scenario. EPA's preference is to provide the 50th percentile of the distribution.

However, if the full distribution is not known, EPA may assume that the mean, mode, or midpoint of the
distribution represents the central tendency depending on the statistics available for the distribution.

A high-end is assumed to be representative of occupational exposures that occur at probabilities above
the 90th percentile but below the exposure of the individual with the highest exposure (	).

For risk evaluation, EPA provided high-end results at the 95th percentile. If the 95th percentile is not
available, EPA used a different percentile greater than or equal to the 90th percentile but less than or
equal to the 99.9th percentile, depending on the statistics available for the distribution. If the full
distribution is not known and the preferred statistics are not available, EPA estimated a maximum or
bounding estimate in lieu of the high-end.

For occupational exposures, EPA used measured or estimated air concentrations to calculate exposure
concentration metrics required for risk assessment, such as average daily concentration (ADC) and
lifetime average daily concentration (LADC). These calculations require additional parameter inputs,
such as years of exposure, exposure duration and frequency, and lifetime years. EPA estimated exposure
concentrations from monitoring data, modeling, or occupational exposure limits.

For the final exposure result metrics, each of the input parameters (e.g., air concentrations, working
years, exposure frequency, lifetime years) may be a point estimate (i.e., a single descriptor or statistic,
such as central tendency or high-end) or a full distribution. EPA considered three general approaches for
estimating the final exposure result metrics:

Page 30 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

•	Deterministic calculations: EPA used combinations of point estimates of each parameter to
estimate a central tendency and high-end for each final exposure metric result. EPA documented
the method and rationale for selecting parametric combinations to be representative of central
tendency and high-end in Appendix B.

•	Probabilistic (stochastic) calculations: EPA used Monte Carlo simulations using the full
distribution of each parameter to calculate a full distribution of the final exposure metric results
and selecting the 50th and 95th percentiles of this resulting distribution as the central tendency and
high-end, respectively.

•	Combination of deterministic and probabilistic calculations: EPA had full distributions for some
parameters but point estimates of the remaining parameters. For example, EPA used Monte
Carlo modeling to estimate exposure concentrations, but only had point estimates of exposure
duration and frequency, and lifetime years. In this case, EPA documented the approach and
rationale for combining point estimates with distribution results for estimating central tendency
and high-end results in Appendix B.

EPA follows the following hierarchy in selecting data and approaches for assessing inhalation
exposures:

1.	Monitoring data:

a.	Personal and directly applicable

b.	Area and directly applicable

c.	Personal and potentially applicable or similar

d.	Area and potentially applicable or similar

2.	Modeling approaches:

a.	Surrogate monitoring data

b.	Fundamental modeling approaches

c.	Statistical regression modeling approaches

3.	Occupational exposure limits:

a.	Company-specific OELs (for site-specific exposure assessments, e.g., there is only one
manufacturer who provides to EPA their internal OEL but does not provide monitoring data)

b.	OSHA PEL

c.	Voluntary limits (ACGIH TLV, NIOSH REL, Occupational Alliance for Risk Science
(OARS) workplace environmental exposure level (WEEL) [formerly by AIHA])

1.4.5.2 Approach for this Risk Evaluation

EPA reviewed workplace inhalation monitoring data collected by government agencies such as OSHA
and NIOSH, monitoring data found in published literature (i.e., personal exposure monitoring data and
area monitoring data), and monitoring data submitted via public comments. Studies were evaluated
using the evaluation strategies laid out in the Application of Systematic Review in TSCA Risk
Evaluations (U.S. EPA. 2018a).

Exposures are calculated from the datasets provided in the sources depending on the size of the dataset.
For datasets with six or more data points, central tendency and high-end exposures were estimated using
the 50th percentile and 95th percentile. For datasets with three to five data points, central tendency

Page 31 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

exposure was calculated using the 50th percentile and the maximum was presented as the high-end
exposure estimate. For datasets with two data points, the midpoint was presented as a midpoint value
and the higher of the two values was presented as a higher value. Finally, data sets with only one data
point presented the single exposure value. For datasets including exposure data that were reported as
below the limit of detection (LOD), EPA estimated the exposure concentrations for these data, following
EPA's Guidelines for Statistical Analysis of Occupational Exposure Data (	b) which

recommends using the ^j=- if the geometric standard deviation of the data is less than 3.0 and if the

geometric standard deviation is 3.0 or greater. Specific details related to each condition of use can be
found in Section 2. For each condition of use, these values were used to calculate acute and chronic
(non-cancer and cancer) exposures. Equations and sample calculations for chronic exposures can be
found in Appendix B and Appendix C, respectively.

EPA used exposure monitoring data or exposure models to estimate inhalation exposures for all
conditions of use. Specific details related to the use of monitoring data for each condition of use can be
found in Section 2. Descriptions of the development and parameters used in the exposure models used
for this assessment can be found in Appendix E through Appendix I.

1.4.6	Dermal Exposure Assessment Approach

Dermal exposure data was not readily available for the conditions of use in the assessment. Because
PCE is a volatile liquid that readily evaporates from the skin, EPA estimated dermal exposures using the
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 PCE based on a theoretical framework provided by Kasting (2006). The amount of liquid
on the skin is adjusted by the weight fraction of PCE in the liquid to which the worker is exposed.
Specific details of the dermal exposure assessment can be found in Section 2.23 and equations and
sample calculations for estimate dermal exposures can be found in Appendix K.

1.4.7	Consideration of Engineering Controls and Personal Protective Equipment

OSHA and NIOSH recommend employers utilize the hierarchy of controls to address hazardous
exposures in the workplace. The hierarchy of controls strategy outlines, in descending order of priority,
the use of elimination, substitution, engineering controls, administrative controls, and lastly personal
protective equipment (PPE). The hierarchy of controls prioritizes the most effective measures first which
is to eliminate or substitute the harmful chemical (e.g., use a different process, substitute with a less
hazardous material), thereby preventing or reducing exposure potential. Following elimination and
substitution, the hierarchy recommends engineering controls to isolate employees from the hazard,
followed by administrative controls, or changes in work practices to reduce exposure potential (e.g.,
source enclosure, local exhaust ventilation systems). Administrative controls are policies and procedures
instituted and overseen by the employer to protect worker exposures. As the last means of control, the
use of personal protective equipment (e.g., respirators, gloves) is recommended, when the other control
measures cannot reduce workplace exposure to an acceptable level.

OSHA's Respiratory Protection Standard (29 CFR § 1910.134) requires employers in certain industries
to address workplace hazards by implementing engineering control measures and, if these are not
feasible, provide respirators that are applicable and suitable for the purpose intended. Respirator
selection provisions are provided in § 1910.134(d) and require that appropriate respirators are selected

Page 32 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

based on the respiratory hazard(s) to which the worker will be exposed and workplace and user factors
that affect respirator performance and reliability. Assigned protection factors (APFs) are provided in
Table 1 under § 1910.134(d)(3)(i)(A) (see below in Table 1-2) and refer to the 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
OSHA's Respiratory Protection Standard.

If respirators are necessary in atmospheres that are not immediately dangerous to life or health, workers
must use NIOSH-certified air-purifying respirators or NIOSH-approved supplied-air respirators with the
appropriate APF. Respirators that meet these criteria include air-purifying respirators with organic vapor
cartridges. Respirators must meet or exceed the required level of protection listed in Table 1-2. Based on
the APF, inhalation exposures may be reduced by a factor of 5 to 10,000, if respirators are properly
worn and fitted.

Table 1-2. Assigned Protection Factors for Respirators in OSHA Standard 29 CFR 1910.134

Type of Uespiralor

Quarter
Mask

Half
Mask

1 Mil

Kacepiece

Helmet/
Mood

Loose-fit ling
I'acepiece

1. Air-Purifying Respirator

5

10

50





2. Power Air-Purifying Respirator (PAPR)



50

1,000

25/1,000

25

3. Supplied-Air Respirator (SAR) or Airline Respirator

• Demand mode



10

50





• Continuous flow mode



50

1,000

25/1,000

25

• Pressure-demand or other positive-pressure
mode



50

1,000





4. Self-Contained Breathing Apparatus (SCBA)

• Demand mode



10

50

50



• Pressure-demand or other positive-pressure
mode (e.g., open/closed circuit)





10,000

10,000



Source: 29 CFR § 1910.134(d)(3)(i)(A)

The National Institute for Occupational Safety and Health (NIOSH) and the U.S. Department of Labor's
Bureau of Labor Statistics (BLS) conducted a voluntary survey of U.S. employers regarding the use of
respiratory protective devices between August 2001 and January 2002. The survey was sent to a sample
of 40,002 establishments designed to represent all private sector establishments. The survey had a 75.5%
response rate (NIOSH. 2003). A voluntary survey may not be representative of all private industry
respirator use patterns as some establishments with low or no respirator use may choose to not respond
to the survey. Therefore, results of the survey may potentially be biased towards higher respirator use.

NIOSH and BLS estimated about 619,400 establishments used respirators for voluntary or required
purposes (including emergency and non-emergency uses). About 281,800 establishments (45%) were
estimated to have had respirator use for required purposes in the 12 months prior to the survey. The

Page 33 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

281,800 establishments estimated to have had respirator use for required purposes were estimated to be
approximately 4.5% of all private industry establishments in the U.S. at the time (NIOSH. 2003).

The survey found that the establishments that required respirator use had the following respirator
program characteristics (NIOSH. 2003):

•	59% provided training to workers on respirator use.

•	34% had a written respiratory protection program.

•	47% performed an assessment of the employees' medical fitness to wear respirators.

•	24%) included air sampling to determine respirator selection.

The survey report does not provide a result for respirator fit testing or identify if fit testing was included
in one of the other program characteristics.

Of the establishments that had respirator use for a required purpose within the 12 months prior to the
survey, NIOSH and BLS found (NIOSH. 2003):

•	Non-powered air purifying respirators are most common, 94%> overall and varying from 89%> to
100%) across industry sectors.

•	Powered air-purifying respirators represent a minority of respirator use, 15% overall and varying
from 7% to 22% across industry sectors.

•	Supplied air respirators represent a minority of respirator use, 17% overall and varying from 4%
to 37%) across industry sectors.

Of the establishments that used non-powered air-purifying respirators for a required purpose within the
12 months prior to the survey, NIOSH and BLS found (NIOSH. 2003):

•	A high majority use dust masks, 76% overall and varying from 56% to 88% across industry
sectors.

•	A varying fraction use half-mask respirators, 52% overall and varying from 26% to 66% across
industry sectors.

•	A varying fraction use full-facepiece respirators, 23% overall and varying from 4% to 33%
across industry sectors.

Table 1-3 summarizes the number and percent of all private industry establishments and employees that
used respirators for a required purpose within the 12 months prior to the survey and includes a
breakdown by industry sector (NIOSH. 2003).

Page 34 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Table 1-3. Number and Percent of Establishments and Employees Using Respirators Within 12

Months Prior to Survey

Industry

r.slnbli>
Number

It m en Is

Percent of All
Kslahlishmcnls

Km
Number

plovers

Percenl of All
r.mployccs

Total Private Industry

281,776

4.5

3,303,414

3.1

Agriculture, forestry, and fishing

13,186

9.4

101,778

5.8

Mining

3,493

11.7

53,984

9.9

Construction

64,172

9.6

590,987

8.9

Manufacturing

48,556

12.8

882,475

4.8

Transportation and public utilities

10,351

3.7

189,867

2.8

Wholesale Trade

31,238

5.2

182,922

2.6

Retail Trade

16,948

1.3

118,200

0.5

Finance, Insurance, and Real Estate

4,202

0.7

22,911

0.3

Services

89,629

4.0

1,160,289

3.2

1.4.8	Water Release Sources

EPA performed a literature search to identify process operations that could potentially result in direct or
indirect discharges to water for each condition of use. Where release sources were unclear or not
available, EPA referenced relevant ESD's or GS's. Water release sources for each condition of use can
be found in Section 2.

1.4.9	Water Release Assessment Approach and Methodology

Where available, EPA used 2016 TR1 (	) and 2016 DMR(	) data to

provide a basis for estimating releases. Facilities are only required to report to TRI if the facility has 10
or more full-time employees, is included in an applicable NAICS code, and manufactures, processes, or
uses the chemical in quantities greater than a certain threshold (25,000 pounds for manufacturers and
processors of PCE and 10,000 pounds for users of PCE). Due to these limitations, some sites that
manufacture, process, or use PCE may not report to TRI and are therefore not included in these datasets.

For the 2016 DMR (	>), EPA used the Water Pollutant Loading Tool within EPA's

Enforcement and Compliance History Online (ECHO) to query all PCE point source water discharges in
2016. DMR data are submitted by National Pollutant Discharge Elimination System (NPDES) permit
holders to states or directly to the EPA according to the monitoring requirements of the facility's permit.
States are only required to load major discharger data into DMR and may or may not load minor
discharger data. The definition of major vs. minor discharger is set by each state and could be based on
discharge volume or facility size. Due to these limitations, some sites that discharge PCE may not be
included in the DMR dataset.

Page 35 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Where releases are expected but TRI and DMR data were not available or where EPA determined TRI
and DMR data did not capture the entirety of water releases for a condition of use, releases were
estimated using release data from literature, relevant ESD's or GS's, existing EPA models, and/or
relevant Effluent Guidelines (EG). EG are national regulatory standards set forth by EPA for wastewater
discharges to surface water and municipal sewage treatment plants. Specific details related to the use of
release data or models for each condition of use can be found in Section 2.

For each condition of use EPA estimated annual releases, average daily releases, and number of release
days per year. Where TRI and/or DMR were available, EPA used the reported annual releases for each
site and estimated the daily release by averaging the annual release over the expected release days per
year. Where ESDs, GSs, existing models, or EGs were used EPA estimated a daily release and
calculated the annual release by multiplying the daily release by the number of release days per year.

Page 36 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

2 Engineering Assessment

The following sections contain process descriptions and the specific details (worker activities, analysis
for determining number of workers, exposure assessment approach and results, release sources, media of
release, and release assessment approach and results) for the assessment for each condition of use.

EPA assessed the conditions of use as stated in the Problem Formulation of the Risk Evaluation for
Perchloroethylene (Ethene, 1,1,2,2-Tetrachloro) published by EPA in May 2018 (	b).

2.1 Manufacturing

2.1.1 Estimates of Number of Facilities

The 20l6CDR(tetashowirto^

2016d). In the 2016 CDR, there are four sites that domestically manufacture PCE and eight sites where
the domestic manufacture/import activity field is either claimed as confidential business information
(CBI) or withheld (U.S. EPA. ). Of the eight sites, four reported 0 lb of PCE imported or
manufactured for reporting year 2015 (	d). EPA assumed manufacture/import of PCE at

these sites has ceased.

To determine whether the remaining four CDR sites were manufacturers or importers, EPA mapped the
sites to 2016 TRI data using the facility names and addresses and found that two of the sites reported
manufacturing PCE in TRI (	). EPA assumed the other two sites for which the activity

could not be determined through CDR or TRI may import or manufacture PCE. Therefore, there may be
up to eight sites that domestically manufacture PCE. It should be noted that EPA only considered sites
reporting to the 2016 CDR for the universe of manufacturing sites and supplemented the CDR data with
TRI data to overcome CBI claims or withheld data in the 2016 CDR. Other sites in TRI may have
reported "producing the chemical" for PCE; however, based on the process described in Section 1.4.1,
EPA assessed a different condition of use at these sites and did not consider them for manufacturing to
avoid double counting.

In the 2016 CDR, one site reported 131,453 lb of PCE manufactured in 2015 (I v \ _
-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Table 2-1. List of Assessed Perchloroethylene Manufacturing Sites

Site

Basis lor
Manufacturing
Determination

Assessed
Production

Volume
(Ih/site-vr)

Production Volume Basis

Axiall Corporation,
Westlake, LA

2016 CDR

45,732,418

Average of unallocated 2015 national
domestically manufactured volume in
CDR

Blue Cube Operations
LLC - Plaquemine Site,
Plaquemine, LA

2016 CDR

45,732,418

Average of unallocated 2015 National
Production Volume

Geon Oxy Vinyl
Laporte Plant,
Laporte, TXa

2016 TRI

45,732,418

Average of unallocated 2015 National
Production Volume

Greenchem,

West Palm Beach, FL

Activity unknown;
assumed manufacturer

45,732,418

Average of unallocated 2015 National
Production Volume

Occidental Chemical
Corp Geismar Plant,
Geismar, LA

2016 TRI

45,732,418

Average of unallocated 2015 National
Production Volume

Olin Blue Cube,
Freeport, TX

2016 CDR

45,732,418

Average of unallocated 2015 National
Production Volume

Solvents & Chemicals,
Pearland, TX

2016 CDR

131,453

2015 reported production volume in
CDR

Univar USA Inc,
Redmond, WA

Activity unknown;
assumed manufacturer

45,732,418

Average of unallocated 2015 National
Production Volume

a The site name listed here is based on its 2016 CDR reported site name. In the 2016 TRI, the site is listed as "Oxy Vinyls LP
La Porte VCM Plant". EPA determined they are the same site as the address in each database is the same and in 1999 the site
became a part of the newly formed Oxy Vinyls, LP which is a joint venture of the Occidental Petroleum Corporation and The
Geon Company (Hydrocarbon Online. 1999).

2.1.2 Process Description

PCE was previously produced through chlorination of acetylene to tetrachloroethane, then
dehydrochlorination to trichloroethylene (TCE), followed by chlorination of TCE to pentachloroethane
and finally dehydrochlorination to PCE (Snedecor et at.. 2004). The last U.S. plant using the acetylene
process was shut down in 1978 (Snedecor et at.. 2004). Currently, most PCE is manufactured using one
of three methods: chlorination of ethylene dichloride (EDC); chlorination of hydrocarbons containing
one to three carbons (CI to C3) or their partially chlorinated derivatives; or oxy chlorination of two-
carbon (C2) chlorinated hydrocarbons ( PR. 2014; Snedecor et at.. 2004; U.S. EPA. 1985).

The chlorination of EDC involves a non-catalytic reaction of chlorine and EDC or other C2 chlorinated
hydrocarbons to form PCE and TCE as co-products and hydrochloric acid (HC1) as a byproduct
(ATSDR. 2014; Snedecor et at.. 2^' 1,1 v « « \ l )85). The chlorination of C1-C3 hydrocarbons

Page 38 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

involves the reaction of chlorine with a hydrocarbon such as methane, ethane, propane, propylene or
their chlorinated derivatives, at high temperatures (550-700°C), with or without a catalyst, to form PCE
and carbon tetrachloride (CCU) as co-products and HC1 as a byproduct ( DR. 2014; Snedecor et at..
2004; U.S. EPA. 1985). The oxychlorination of C2 chlorinated hydrocarbons involves the reaction of
either chlorine or HC1 and oxygen with EDC in the presence of a catalyst to produce PCE and TCE as
co-products (ATSDR. 201 i; Snedecor et at.. 2004). In all three processes the product ratio of PCE to
TCE/CCU products are controlled by adjusting the reactant ratios (Snedecor et at.. 2004).

2.1.3 Exposure Assessment

2.1.3.1	Worker Activities

During manufacturing, workers are potentially exposed while connecting and disconnecting hoses and
transfer lines to containers and packaging to be loaded with PCE product (e.g., railcars, tank trucks,
totes, drums, bottles) and intermediate storage vessels (e.g., storage tanks, pressure vessels). Workers
near loading racks and container filling stations are potentially exposed to fugitive emissions from
equipment leaks and displaced vapor as containers are filled. These activities are potential sources of
worker exposure through dermal contact with liquid and inhalation of vapors.

ONUs include employees that work at the site where PCE 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. ONUs for manufacturing include supervisors, managers, and tradesmen that
may be in the manufacturing area but do not perform tasks that result in the same level of exposures as
manufacturing workers.

2.1.3.2	Number of Potentially Exposed Workers

EPA estimated the number of workers and occupational non-users potentially exposed to PCE at
manufacturing sites using 2016 CDR (	I) data (where available), Bureau of Labor

Statistics' OES data (r S HI S. 2016) and the U.S. Census' SUSB (\ S t ^iisus Bureau. 2015). The
method for estimating number of workers from the Bureau of Labor Statistics' OES data and U.S.
Census' SUSB data is detailed in Section 1.4.4 and Appendix A. These estimates were derived using
industry- and occupation-specific employment data from the BLS and U.S. Census.

2016 CDR data for number of workers are available for four manufacturing sites. Of the four sites, two
sites reported 100 to 500 workers, one site reported 50 to 100 workers, and one site reported 25 to 50
workers (	). For the other four manufacturing sites, the number of workers in CDR is

either claimed as CBI or withheld (	1).

EPA identified the NAICS code 325199, All Other Basic Organic Chemical Manufacturing, as the code
expected to include sites manufacturing PCE. Based on data from the BLS for this NAICS code and
related SOC codes, there are an average of 39 workers and 18 ONUs per site, or a total of 57 potentially
exposed workers and ONUs, for sites under this NAICS code (\ c. c. , ,	nsus Bureau.

2015). This is consistent with the one site reporting 50 to 100 workers and only slightly higher than the
one site reporting 25 to 50 workers.

To determine the average number of workers, EPA used the average of the ranges reported in CDR for
the four sites where data were available, and the average worker and ONUs estimates from the BLS

Page 39 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

analysis for the other four sites. CDR data do not differentiate between workers and ONUs; therefore,
EPA assumed the ratio of workers to ONUs would be similar as determined in the BLS data where
approximately 68% of the exposed personnel are workers and 32% are ONUs (\] S HI S. 2016; U.S.
Census Bureau. ). This resulted in approximately 640 workers and 300 ONUs (see Table 2-2).

Table 2-2. Estimated Number of Workers Potentially Exposed to Perchloroethylene During

Manufacturing

Number of
Sites

Kxposed
Workers per
Site

Kxposed
Occupational
Non-l sers per
Site

Total Kxposcd
Workers

Total Kxposcd
Occupational
Non-l sers

Total Kxposcd

4a

39

18

154

73

227

2b

204

96

408

192

600

lc

51

24

51

24

75

ld

25

12

25

12

38

Total®

80

38

640

300

940

a For the sites using values from the BLS analysis, the total number of workers and occupational non-users are calculated
using the number of workers and occupational non-users per site and estimated from BLS and multiplying by the four sites.
The number of workers and occupational non-users per site presented in the table round the values estimated from the BLS
analysis to the nearest integer.

b Number of workers and occupational non-users per site estimated by taking the average of 100 and 499 (per 2016 CDR) and
multiplying by 68% and 32%, respectively. Values are rounded to the nearest integer.

0 Number of workers and occupational non-users per site estimated by taking the average of 50 and 99 (per 2016 CDR) and
multiplying by 68% and 32%, respectively. Values are rounded to the nearest integer.

d Number of workers and occupational non-users per site estimated by taking the average of 25 and 49 (per 2016 CDR) and
multiplying by 68% and 32%, respectively. Values are rounded to the nearest integer.
e Totals have been rounded to two significant figures. Totals may not add exactly due to rounding.

2.1.3.3 Occupational Exposure Results

EPA assessed inhalation exposures during manufacturing using identified monitoring data. Table 2-3
summarizes 8-hr and 12-hr TWA samples obtained from data submitted by the Halogenated Solvents
Industry Alliance (HS1A) via public comment for three companies (HSIA. 2018). Data were not
available to estimate ONU exposures; EPA estimates that ONU exposures are lower than worker
exposures, since ONUs do not typically directly handle the chemical.

Three additional studies with monitoring data for manufacturing were identified; however, the data from
these studies were not used in the assessment. Two of these studies were from China and almost 30
years old and are unlikely to be representative of current conditions at U.S. manufacturing sites ("Seiji et
at.. 1990; Seiii et at.. 1989). The third study provides data collected in 1982 from a Dow Chemical site
manufacturing PCE and carbon tetrachloride; however, this site was not identified as a current
manufacturer of PCE (see Table 2-1) (Dow Chei	5c). Due to the age of the collected data (over

30 years old) and the fact the site is no longer identified as manufacturing PCE coupled with the
availability of more recent monitoring data from current manufacturing sites, EPA did not include the
data from the Dow Chemical site in this analysis.

Page 40 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

HSIA (2018) provided monitoring data for PCE collected by three companies listed as "Company A",
"Company B", and "Company C". The data were collected between 2006 and 2018 with full-shift data
collected over 8 to 12 hours during which workers engaged in a variety of activities including collecting
catch samples; performing filter changes; line and equipment opening; loading and unloading; process
sampling; and transferring of hazardous wastes (HSIA. 2018).

EPA assessed exposures for both 8-hr and 12-hr exposures separately. The high-ends for the 15-min, 30-
min, 8-hr, and 12-hr TWAs are the 95th percentile of the respective data sets and the central tendencies
are the 50th percentile. It should be noted that approximately 65% of the 8-hr TWA exposure data and
73% of the 12-hr TWA exposure data were below the limit of detection (LOD). To estimate exposure
concentrations for these data, EPA followed the Guidelines for Statistical Analysis of Occupational
Exposure Data (	94b) as discussed in Section 1.4.5.2. The geometric standard deviation for

both 8-hr TWA data and 12-hr TWA data were both above 3.0; therefore, EPA used the to estimate

'	'	2

the exposure value as specified in the guidelines (	3). Because over 50% of the data are

below the LOD, calculating statistics from this data does present the potential to introduce biases into
the results. Estimation of exposure values for results below the LOD may over- or under-estimate actual
exposure thus skewing the calculated statistics higher or lower, respectively. The overall directional bias
of the exposure assessment, accounting for both the overestimate and underestimate, is not known.

It should also be noted that 18 8-hr TWA exposure data points from Company C were not included in
the results as they were reported as being below the detection limit, but the company did not provide the
value of the LOD. Therefore, EPA could not estimate a value for these data using the guidelines
described above.

Table 2-3. Summary of Worker Inhalation Monitoring Data for the Manufacture of
Perchloroethylene	

Scenario

S- or 12-
lir TW A
(ppm)

AC
(ppm)

A IK

(ppm)

I.A IK'
(ppm)

Nil m her
of Data
Points

15- or 30-
in imile
TWA
(ppm)

N il in her of
Shorl-lerm
Data Points

8-hr TWA Results

15-minute TWA Results

High-End

2.6

0.9

0.6

0.3

75a

15

161

Central Tendency

3.25E-02

1.08E-02

7.42E-03

2.95E-03

2.0

12-hr TWA Results

30-minute TWA Results

High-End

0.2

0.1

7.26E-02

3.72E-02

77

12

38b

Central Tendency

2.05E-02

1.03E-02

7.02E-03

2.79E-03

0.7

AC = Acute Concentration; ADC = Average Daily Concentration; and LADC = Lifetime Average Daily Concentration.
Equations and parameters for calculation of the AC, ADC, and LADC are described in Appendix B.
a Data does not include 18 data points that were reported as being below the detection limit, but for which the company did
not provide the LOD for use in estimating an exposure value.

b Data does not include five data points that were reported as being below the detection limit, but for which the company did
not provide the LOD for use in estimating an exposure value.

Sources: (HSIA. 2018)

Page 41 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

2.1.4 Water Release Assessment

2.1.4.1	Water Release Sources

In general, potential sources of water releases in the chemical industry may include the following:
equipment cleaning operations, aqueous wastes from scrubbers/decanters, reaction water, process water
from washing intermediate products, and trace water settled in storage tanks (OE	). Based on

the process for manufacturing PCE, EPA expects the sources of water releases to be from aqueous
wastes from decanters used to separate catalyst fines, caustic neutralizer columns and caustic scrubbers;
and water removed from the PCE product in drying columns (	9). Additional water

releases may occur if a site uses of water to clean process equipment; however, EPA does not expect this
to be a primary source of water releases from manufacturing sites as equipment cleaning is not expected
to occur daily and manufacturers are expected to use an organic solvent to clean process equipment.

2.1.4.2	Water Release Assessment Results

Of the eight manufacturing sites assessed, four reported in the 2016 TRI (	). For these

sites, EPA assessed water releases as reported in the 2016 TRI (	) For the remaining

four sites, EPA assessed water releases at the maximum daily and maximum average monthly
concentrations allowed under the Organic Chemicals, Plastics and Synthetic Fibers (OCPSF) Effluent
Guidelines and Standards (	). Effluent Guidelines (EGs) are national regulatory

standards set forth by EPA for wastewater discharges to surface water and municipal sewage treatment
plants. The OCPSF EG applies to facilities classified under the following SIC codes:

•	2821—Plastic Materials, Synthetic Resins, and Nonvulcanizable Elastomers;

•	2823—Cellulosic Man-Made Fibers;

•	2865—Cyclic Crudes and Intermediates, Dyes, and Organic Pigments; and

•	2869—Industrial Organic Chemicals, Not Elsewhere Classified (	1019b).

Manufacturers of PCE would typically be classified under SIC code 2869; therefore, the requirements of
the OCPSF EG apply to these sites. Subparts I, J, and K of the OCPSF EG set limits for the
concentration of PCE in wastewater effluents for industrial facilities that are direct discharge point
sources using end-of-pipe biological treatment, direct discharge point sources that do not use end-of-
pipe biological treatment, and indirect discharge point sources, respectively (	2019b). Direct

dischargers are facilities that discharge effluents directly to surface waters and indirect dischargers are
facilities that discharge effluents to publicly-owned treatment works (POTW). The OCPSF limits for
PCE are provided in Table 2-4.

Page 42 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Table 2-4. Summary of OCPSF Effluent Guidelines for Perchloroethylene

OCPSF Subpart

.Maximum
for Any One
l)a\

W-)

.Maximum for
Any Monthly
Average
(ws/i.)

Basis

Subpart I - Direct Discharge
Point Sources That Use End-of-
Pipe Biological Treatment

56

22

BAT effluent limitations and
NSPS

Subpart J - Direct Discharge
Point Sources That Do Not Use
End-of-Pipe Biological Treatment

164

52

BAT effluent limitations and
NSPS

Subpart K - Indirect Discharge
Point Sources

164

52

Pretreatment Standards for
Existing Sources (PSES) and
Pretreatment Standards for New
Sources (PSNS)

BAT = Best Available Technology Economically Achievable; NSPS = New Source Performance Standards; PSES =
Pretreatment Standards for Existing Sources; PSNS = Pretreatment Standards for New Sources.

Source: 40 C.F.R. 414

EPA did not identify PCE-specific information on the amount of wastewater produced per day. The
Specific Environmental Release Category (SpERC) developed by the European Solvent Industry Group
for the manufacture of a substance estimates 10 m3 of wastewater generated per metric ton of substance
produced (European Solvents Industry Group. 2012). In lieu of PCE-specific information, EPA
estimated water releases using the SpERC specified wastewater production volume and the annual PCE
production rates from each facility as shown in Table 2-1 in Section 2.1.1.

EPA estimated both a maximum daily release and an average daily release using the OCPSF EG
limitations for PCE for maximum on any one day, and maximum for any monthly average, respectively.
Prevalence of end-of-pipe biological treatment at PCE manufacturing sites is unknown; therefore, EPA
used limitations for direct discharges with no end-of-pipe biological treatment and indirect dischargers
to give most protective estimate. EPA estimated annual releases from the average daily release and
assuming 350 days/yr of operation1. Details of the approach and sample calculations for estimating
water release using the OCPSF EG limitations are provided in Appendix D.

Table 2-5 summarizes water releases from the manufacturing process for sites reporting to the 2016 TRI
and Table 2-6 summarizes water releases from sites not reporting to the 2016 TRI.

1 Due to large throughput, manufacturing sites are assumed to operate seven days per week and 50 weeks per year with two
weeks per year for shutdown activities.

Page 43 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Table 2-5. Reported Wastewater Discharges of Perchloroethy

ene from Manufacturing Sites Reporting to 2016 TRI

Site

Annual Release3
(kg/site-yr)

Annual Release
Days (days/yr)

Average Daily

Release3
(kg/site-day)

NPDES Code

Release Media/

Treatment
Facility Type

Blue Cube Operations LLC - Plaquemine
Site,

Plaquemine, LA

0

N/A

0

Not available

N/A

Geon Oxy Vinyl Laporte Plant,
Laporte, TX

0

N/A

0

TX0070416

N/A

Occidental Chemical Corp Geismar Plant,
Geismar, LA

0.6

350

1.68E-03

LA0002933

Surface Water

Olin Blue Cube, Freeport, TX

15

350

4.15E-02

Not available

Non-POTW
WWT

POTW = Publicly-Owned Treatment Works; WWT = Wastewater Treatment; N/A = Not applicable

a Annual release amounts are based on the site reported values. Therefore, daily releases are calculated from the annual release rate and assuming 350 days of operation
per year.

Source: (U.S. EPA 2017d)


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Table 2-6. Estimated Wastewater Discharges of Perchloroethylene from Manufacturing Sites Not Reporting to 2016 TRI

Site

Annual
Operating

Days
(days/yr)

Daily
Production

Volume11
(kg/site-day)

Daily
Wastewater

Mow1'
(1./site-day)

Maximum
Daily
Release'
(kg/site-day)

Average

Daily
Release'1
(kg/site-day)

Average
Annual
Release''
(kg/site-yr)

NPDKS
Code

Release
Media/
Treat incnl
l-'acility
Type

Axiall Corporation,
Westlake, LA

350

59,268

592,682

0.1

3.08E-02

11

Not available

Surface
Water or
POTW

Greenchem,

West Palm Beach, FL

350

59,268

592,682

0.1

3.08E-02

11

Not available

Surface
Water or
POTW

Solvents & Chemicals,
Pearland, TX

350

170

1,704

2.79E-04

8.86E-05

3.10E-02

Not available

Surface
Water or
POTW

Univar USA Inc,
Redmond, WA

350

59,268

592,682

0.1

3.08E-02

11

Not available

Surface
Water or
POTW

POTW = Publicly-Owned Treatment Works

a Daily production volume calculated using the annual production volume provided in Table 2-1 and dividing by the annual operating days peryear (350 days/yr).
b The estimated wastewater flow rate is calculated assuming 10 m3 of wastewater is produced per metric ton of PCE produced (equivalent to 10 L wastewater/kg of PCE)
based on the SpERC for the manufacture of a substance.

0 The maximum daily release is calculated using the maximum daily concentration from the OCPSF EG, 164 |.ig/L. and multiplying by the daily wastewater flow.
d The average daily release is calculated using the maximum monthly average concentration from the OCPSF EG, 52 |.ig/L. and multiplying by the daily wastewater flow.
e The average annual release is calculated as the maximum monthly average concentration multiplied by the daily wastewater production, and 350 operating days/year.

Page 45 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

2.2 Repackaging

2.2.1_ Estimates of Number of Facilities

The repackaging scenario covers only those sites that purchase PCE or PCE containing products from
domestic and/or foreign suppliers and repackage the PCE from bulk containers into smaller containers
for resale. It does not include sites that import PCE and either: (1) store the chemical in a warehouse and
resell directly without repackaging; (2) act as the importer of record for PCE but PCE is never present at
the site2; or (3) import the chemical and process or use the chemical directly at the site. Case #1 presents
only a de minimus exposure or release potential as the containers are never opened. In case #2, the
potential for exposure and release is at the site receiving PCE, not the "import" site and
exposures/releases at the site receiving PCE are assessed in the relevant scenario based on the condition
of use for PCE at the site. Similarly, for case #3, the potential for exposure and release at these sites are
evaluated in the relevant scenario depending on the condition of use for PCE at the site.

To determine the number of sites that may repackage PCE, EPA considered 2016 CDR (U.S. EPA.
2016d), 2016 TRI data (	|), and 2016 DMR (	j) data. In the 2016 CDR,

two manufacturing facilities reported downstream repackaging processes in the industrial processing and
use section with one reporting the number of sites as CBI and one reporting 25 to 100 sites (
2016d). There are also two import sites and one manufacturing site in the 2016 CDR that report uses that
are "not known or reasonably ascertainable" (NKRA) which may include repackaging activities (

).

In the 2016 TRI, 27 facilities report a repackaging activity; however, 16 of these sites either report other
activities to TRI or report under a NAICS related to disposal/recycling of PCE (1 c. « ^ \ 101 <{). As
described in Section 1.4.1, EPA determined that the other reported activities or activities related to
disposal/recycling are the "primary" condition of use for PCE. Therefore, the evaluation of these 16 sites
are included in the evaluation of the scenario related to the primary condition of use and are not included
in the repackaging scenario. In addition to the sites discussed above, there are 19 sites in the 2016 TRI
that report under the NAICS code 424690, Other Chemical and Allied Products Merchant Wholesalers,
that reported on a Form A and, therefore, were not required to designate an activity (	).

EPA assumes that these sites may perform repackaging activities as well resulting in a total of 30 sites in
the 2016 TRI where the repackage of PCE is the primary condition of use.

In the 2016 DMR data, there are two sites that reports under the SIC code 4225, General Warehousing
and Storage; 10 sites that report under the SIC code 4226, Special Warehousing and Storage; two sites
that report under SIC code 4491, Marine and Cargo Handling; seven sites that report under the SIC code
5169, Chemical and Allied Products, Not Elsewhere Classified; and 1 site reporting under SIC code
5172, Petroleum and Petroleum Products Wholesalers, Except Bulk Stations and Terminals, with 1 site
reporting under SIC code 5169 being the same as one of the identified TRI sites (	)3.

EPA assumes the primary condition of use at these sites is repackaging. Therefore, EPA assesses a total
of 51 sites (30+2+10+2+7+1 = 52 sites - 1 duplicate site = 51 sites) for the repackaging of PCE.

2	In CDR, the reporting site is the importer of record which may be a corporate site or other entity that facilitates the import
of the chemical but never actually receives the chemical. Rather, the chemical is shipped directly to the site processing or
using the chemical.

3	Although the name of the SIC code 5169 (Chemical and Allied Products, Not Elsewhere Classified) does not indicate it, the
"51" group of SIC codes refers to the wholesale trade of non-durable goods. EPA assumed the primary activity at a
wholesaler is repackaging.


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

2.2.2	Process Description

In general, commodity chemicals are imported into the United States in bulk via water, air, land, and
intermodal shipments (Tomer and Kane. 2015). These shipments take the form of oceangoing chemical
tankers, railcars, tank trucks, and intermodal tank containers. Chemicals shipped in bulk containers may
be repackaged into smaller containers for resale, such as drums or bottles. Domestically manufactured
commodity chemicals may be shipped within the United States in liquid cargo barges, railcars, tank
trucks, tank containers, intermediate bulk containers (IBCs)/totes, and drums. Both imported and
domestically manufactured commodity chemicals may be repackaged by wholesalers for resale; for
example, repackaging bulk packaging into drums or bottles.

The exact shipping and packaging methods specific to PCE are not known. For this risk evaluation, EPA
assesses the repackaging of PCE from bulk packaging to drums and bottles at wholesale repackaging
sites.

2.2.3	Exposure Assessment

2.2.3.1	Worker Activities

During repackaging, workers are potentially exposed while connecting and disconnecting hoses and
transfer lines to containers and packaging to be unloaded (e.g., railcars, tank trucks, totes), intermediate
storage vessels (e.g., storage tanks, pressure vessels), and final packaging containers (e.g., drums,
bottles). Workers near loading racks and container filling stations are potentially exposed to fugitive
emissions from equipment leaks and displaced vapor as containers are filled. These activities are
potential sources of worker exposure through dermal contact with liquid and inhalation of vapors.

ONUs include employees that work at the site where PCE is repackaged, 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. ONUs for repackaging include supervisors, managers, and tradesmen that may be in
the repackaging area but do not perform tasks that result in the same level of exposures as repackaging
workers.

2.2.3.2	Number of Potentially Exposed Workers

EPA estimated the number of workers and occupational non-users potentially exposed during
repackaging of PCE using Bureau of Labor Statistics" OES data (	,S. 2016) and the U.S. Census"

SUSB (U.S. Census Bureau. 2015) as well as the primary NAICS and SIC code reported by each site in
the 2016 TR1 (	) or 2016 DMR (	), respectively. The method for

estimating number of workers is detailed in Section 1.4.4 and Appendix A. These estimates were
derived using industry- and occupation-specific employment data from the BLS and U.S. Census. The
employment data from the U.S. Census SUSB and the Bureau of Labor Statistics' OES data are based
on NAICS code; therefore, SIC codes reported in DMR had to be mapped to a NAICS code to estimate
the number of workers. A crosswalk of the SIC codes to the NAICS codes used in the analysis are
provided in Table 2-7. Sites from TRI report NAICS codes; therefore, these codes were used directly in
the analysis.

Page 47 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Table 2-7. Crosswalk of Repackaging SIC Codes in DMR to NAICS Codes

Sl( ( ode

Corresponding NAICS Code

4225 - General Warehousing and Storage51

493100 - Warehousing and Storage

4226 - Special Warehousing and Storage, Not
Elsewhere Classified51

493100 - Warehousing and Storage

4491 - Marine Cargo Holdingb

488300 - Support Activities for Water
Transportation

5169 - Chemicals and Allied Products, Not
Elsewhere Classified

424690 - Other Chemical and Allied Products
Merchant Wholesalers

5172 - Petroleum and Petroleum Products
Wholesalers, Except Bulk Stations and Terminals

424720 - Petroleum and Petroleum Products
Merchant Wholesalers (except Bulk Stations and
Terminals)

a The SIC codes 4225 and 4226 may map to any of the following NAICS codes: 493110, 493120, or 493190. There is not
enough information in the DMR data to determine the appropriate NAICS for each site; therefore, EPA uses data for the 4-
digit NAICS, 493100, rather than a specific 6-digit NAICS.

b The SIC codes 4491 may map to any of the NAICS codes 488310 or 488320. There is not enough information in the DMR
data to determine the appropriate NAICS for each site; therefore, EPA uses data for the 4-digit NAICS, 488300, rather than a
specific 6-digit NAICS.

Table 2-8 provides a summary of the reported NAICS codes (or NAICS identified in the crosswalk), the
number of sites reporting each NAICS code, and the estimated number of workers and ONUs for each
NAICS code as well as an overall total for repackaging of PCE. There are approximately 210 workers
and 75 ONUs potentially exposed during repackaging of PCE.

Table 2-8. Estimated Number of Workers Potentially Exposed to Perchloroethylene During
Repackaging	

NAICS
Code

Number of
Shes

Exposed
Workers
per Site11

Kxposed
Occupational
Non-l sers
per Site"

Total
Kxposed
Workers

Total
Kxposed
Occupational
Non-l sers

Total
Kxposed

325199

2

39

18

77

36

114

325211

1

27

12

27

12

40

325611

1

19

4

19

4

23

424690

32

1

0.4

40

14

55

424720

1

1

0.1

1

0.1

1

488300

2

3

0.5

7

1

8

493100

12

3

1

37

7

44

Totalb

51

4

1

210

75

280

a Number of workers and occupational non-users per site are calculated by dividing the exposed number of workers or
occupational non-users by the number of establishments in the relevant NAICS codes. The workers/ONUs per site are then

Page 48 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

multiplied by the number of sites within that NAICS to get the total exposed. The number of workers/ONUs per site is
rounded to the nearest integer. The number of occupational non-users per site for NAICS 424690, 424720, and 488300 are
shown as 0.4, 0.1, and 0.5, respectively, as they round down to zero.

b Totals have been rounded to two significant figures. Totals may not add exactly due to rounding.

2.2.3.3 Occupational Exposure Results

EPA assessed inhalation exposures during repackaging using identified monitoring data. Table 2-9
summarizes 8-hr, 30-min and 15-min TWA samples obtained from data submitted to EPA by Dow
Chemical under TSCA (Dow Chem Co. 1984). The data were collected by Dow Chemical at the Joliet,
IL marine terminal during the loading of PCE into trucks and sampling activities as part of an industrial
hygiene (1H) study (Dow Chem Co. 1984). Ten full-shift samples were collected with sample times
ranging from approximately 4.5 to 8.5 hour (Dow Chem Co. 1984). EPA converted to 8-hr TWAs
assuming exposures outside the sample time were zero. The 95th percentile and 50th percentile are
presented as the high-end and central tendency exposure values, respectively, in Table 2-9. Data were
not available to estimate ONU exposures; EPA estimates that ONU exposures are lower than worker
exposures, since ONUs do not typically directly handle the chemical.

The study also collected two approximately 15-min TWA samples and five approximately 30-min TWA
samples (Dow Chem Co. 1984). For the 15-min TWA, only two data points were available; therefore,
EPA presents two scenarios: 1) using the maximum as a "higher value"; and 2) using the midpoint as a
"midpoint value". These scenarios are plausible, but EPA cannot determine the statistical
representativeness of the value. For the 30-min TWA, only five data points were available; therefore, the
maximum is presented as the high-end and the median is presented as the central tendency. It should be
noted that two of the 30-min TWA samples measured below the LOD (Dow Chem Co. 1984). To
estimate exposure concentrations for these data, EPA followed the Guidelines for Statistical Analysis of
Occupational Exposure Data (	Mb) as discussed in Section 1.4.5.2. The geometric standard

deviation for was above 3.0; therefore, EPA used the to estimate the exposure value as specified in

the guidelines (U.S. EPA. 1994b).

Table 2-9. Summary of Worker Inhalation Exposure Monitoring Data for Repackaging of
Perchloroethylene	

Scenario

8-hr
TWA

(ppm)

AC
(ppm)

A IK

(ppm)

I.AIK
(ppm)

Nil m her
of Data
Points

30-min
TWA
(ppm)

N il in her
of Data
Points

15-min
TWA
(ppm)11

N il in her
of Data
Points

High-End

0.8

0.3

0.2

9.59E-02

10

5.7

5

1.6

2

Central
Tendency

0.4

0.1

9.94E-02

3.95E-02

8.00E-02

0.9

AC = Acute Concentration; ADC = Average Daily Concentration; and LADC = Lifetime Average Daily Concentration.
Equations and parameters for calculation of the AC, ADC, and LADC are described in Appendix B.
a Due to only two data points identified, EPA presents two scenarios: 1) using the higher of the two values; and 2) using the
midpoint of the two values.

Sources: (Dow Chem Co. 1984)

Page 49 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

2.2.4 Water Release Assessment

2.2.4.1	Water Release Sources

EPA expects the primary source of water releases from repackaging activities to be from the use of
water or steam to clean bulk containers used to transport PCE or products containing PCE. EPA expects
the use of water/steam for cleaning containers to be limited at repackaging sites as PCE is an organic
substance and classified as a hazardous waste under RCRA. EPA expects the majority of sites to use
organic cleaning solvents which would be disposed of as hazardous waste (incineration or landfill) over
water or steam.

2.2.4.2	Water Release Assessment Results

EPA assessed water releases using the values reported to the 2016 TR1 (	) and the 2016

DMR (	b) by the 51 repackaging sites. In the 2016 TRI, all 30 sites reported zero direct

discharges to surface water and zero indirect discharges to POTW (	1). One site reported

an indirect discharge of 615 lb/yr (-279 kg/yr) to non-POTW WWT and the other 29 sites reported zero
indirect discharges to non-POTW WWT (	[). In the 2016 DMR, one site reported a

direct discharge of 2.64 lb/yr (1.20 kg/yr), one site reported 0.66 lb/yr (0.30 kg/yr), one site reported
0.05 lb/yr (0.02 kg/yr), and the remaining sites all report zero direct discharges (indirect discharges not
reported in DMR) (	) To estimate the daily release, EPA used a default assumption of

250 days/yr of operation (assumes operation 5 days/week and 50 weeks/year) and averaged the annual
release over the operating days. Table 2-10 summarizes the releases from sites with non-zero discharges

Table 2-10. Reported Wastewater Discharges of Perchloroethylene from Repackaging Sites

She

Annual
Release"
(kg/silc-year)

Annual
Release Days
(days/yr)

Daily
Release
(kg/si to-
day)11

\pi)i:s

Code

Release
Media/
Treatment
l-'acility
Type

Source

Chemtool,
Rockton, IL

0.3

250

1.20E-03

IL0064564

Surface
Water

(U.S.
EPA,
2016b")

Harvey
Terminal,
Harvey, LA

2.28E-02

250

9.14E-05

LA0056600

Surface
Water

(U.S.
EPA,
2016b")

Hubbard-Hall
Inc, Waterbury,
CT

279

250

1.1

Not
available

Non-POTW
WWT

(U.S.
EPA,
2017d")

Vopak
Terminal
Westwego Inc,
Westwego, LA

1.2

250

4.79E-03

LA0124583

Surface
Water

(U.S.
EPA,
2016b)

POTW = Publicly-Owned Treatment Works; WWT = Wastewater Treatment

a Annual release amounts are based on the site reported values. Therefore, daily releases are calculated from the annual
release rate and assuming 250 days of operation per year.

Page 50 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Sources: (U.S. EPA. 2017d. 2016b)

2.3 Processing as a Reactant

2.3.1 Estimates of Number of Facilities

To determine the number of sites that process PCE as a reactant, EPA considered 2016 CDR (
2016d), 2016 TRI (U.S. EPA. ), and 2016 DMR (	) data. In the 2016 CDR, five

sites reported at least one downstream processing as a reactant activity in the industrial processing and
use section (U.S. EPA. 2016d)4. There are three reports of processing as a reactant in the "all other
organic chemical manufacturing" industry sector, one in the "industrial gas manufacturing" industry
sector, one in the "petroleum refineries" industry sector, and two CBI industry sectors (

2016d)5. There are also two reports where the submitter reports processing as a reactant but reports the
function as either "solvents (for cleaning and degreasing)" or "Solvents (which become part of the
product formulation or mixture)"; EPA assumes the reported processing as a reactant is an error based
on the functional codes reported (	2016d). Of the seven reported instances of industrial

processing as a reactant, four reported fewer than 10 sites, one reported 10 to 25 sites, and two reported
the number of sites as CBI (	1).

In the 2016 TRI, 16 facilities reported use of PCE as a reactant; however, three of these sites also
reported as manufacturers of PCE in the 2016 CDR (	). The manufacturing sites are not

included in the assessment for reactant uses as exposures and releases from these sites have already been
assessed in Section 2.1. Some of the sites in TRI also reported other activities such as processing aids,
manufacturing aids, and/or ancillary use; however, based on the reported NAICS codes and the fact that
65 to 70% of the total annual U.S. production volume is expected to be used for reactant uses, EPA
expects the primary condition of use at these sites to be for reactant uses (NTP. 2014; H.SIA. 2008).
Therefore, there are a total of 13 sites in the 2016 TRI where the processing of PCE as a reactant is the
primary condition of use.

In the 2016 DMR data, there are five sites that report under the SIC code 2812, Alkalies and Chlorine;
one site that reports under the SIC code 2816, Inorganic Pigments; 12 sites that report under the SIC
code 2819, Industrial Inorganic Chemicals, Not Elsewhere Classified; 86 sites that report under the SIC
code 2869, Industrial Organic Chemicals, Not Elsewhere Classified; and 3 sites that did not report a SIC
code6, with three sites: 1) Eagle US 2, LLC; 2) Honeywell International Baton Rouge Plant, and 3)
Westlake Vinyls, Inc. being the same as three of the identified TRI sites (	6b). These SIC

codes include sites that are engaged in the manufacture of organic and inorganic chemicals for which
PCE may be used as a reactant to create, including various organic and inorganic chlorinated
compounds. Additional information for conditions of use is not provided in the DMR data; therefore,
EPA assumes the primary condition of use at these sites is processing as a reactant based solely on the
SIC code. Based on the DMR and TRI data, EPA assesses a total of 117 sites (13+5+1+12+86+3 = 120
sites - 3 duplicate site =117 sites) for the processing of PCE as a reactant.

4	In CDR, only manufacturers and importers report; therefore, "downstream" processing and use activities may refer to
additional processing/use at the reporting site or the processing/use activities of the reporting sites' customers.

5	The number of industry sectors reported is greater than the number of sites reporting processing as a reactant as each site
may report multiple industry sectors.

6	These sites were assumed to be processing PCE as a reactant based on the company name and the fact that 70% of the
national PCE production volume is used as a reactant.

Page 51 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

2.3.2	Process Description

Processing as a reactant or intermediate is the use of PCE as a feedstock in the production of another
chemical product via a chemical reaction in which PCE is consumed to form the product. In the past,
PCE was used as feedstock (with chlorine) for the manufacture of one- and two-carbon (CI and C2)
CFCs (Smart and Fernandez. 2000). However, due to discovery that CFCs contribute to stratospheric
ozone depletion, the use of CFCs was phased-out by the year 2000 to comply with the Montreal Protocol
(Smart a landez. 2000). Since the phase-out of CFCs, PCE has been used to manufacture the CFC
alternatives, HCFCs, specifically the HCFC-123 alternative to CFC-11 (Smart, and Fernandez. 2000).
PCE is also used as a feedstock in the production of trichloroacetyl chloride (Smart and Fernandez.
2000).

HCFC-123 is produced by fluorination of PCE with liquid or gaseous hydrofluoric acid (HF). The
manufacture of HCFC is more complex than the manufacture of CFCs due to potential byproduct
formation or catalyst inactivation caused by the extra hydrogen atom in the HCFCs (Smart, and
Fernandez. 2000). Therefore, the process involved in the manufacture of HCFCs requires additional
reaction and distillation steps as compared to the CFC manufacturing process (Smart and Fernandez.
2000).

PCE is also used by Honeywell International Inc. in the manufacture of HFC-125 (R-125), HCFC-124
(R-124), and CFC-1 13 (R-l 13) (Honeywell. ). In 2016, Honeywell used approximately 65 million
pounds of PCE to manufacture R-125 and R-l24 and approximately 20 million pounds to manufacture
R-1 13 (Honeywell. 2017). The majority of the R-1 13 is used as an intermediate for manufacture of
chlorotrifluoroethylene (CTFE) monomer; however, a small portion is used in exempted applications
vital to U.S. security (Honeywell. ). PCE is received at the Honeywell facilities in rail cars and
trucks and is transferred into storage vessels with a pump and vapor balance (Honeywell. 2017). Some
PCE is lost when disconnecting the hose; however, the storage tank is pressurized so there are no point
emissions or breathing losses (Honeywell. 2017). The primary emission of PCE at Honeywell facilities
are from fugitive emissions (Honeywell. 2017). The facilities utilize a fugitive emissions monitoring
program and leak detection program to reduce fugitive emissions (Honeywell 2017).

Honeywell representatives indicated that the R-125/R-124 processes achieve a once through PCE
conversion of 95% and the remaining 5% is recovered and recycled back into the process (Honeywell
2017). For the R-1 13 process, the once through conversion rate is 99% and the remaining 1% is
recovered and recycled back into the process (Honeywell 2017). The ultimate conversion from both
processes is 100%. Honeywell indicated they do not detect any PCE in their products (Honeywell
2017).

2.3.3	Exposure Assessment

2.3.3.1 Worker Activities

At industrial facilities, workers are potentially exposed when unloading PCE from transport containers
into intermediate storage tanks and process vessels. Workers may be exposed via inhalation of vapor or
via dermal contact with liquids while connecting and disconnecting hoses and transfer lines. Once PCE
is unloaded into process vessels, it is consumed as a chemical intermediate.

Page 52 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

ONUs are employees who work at the facilities that process and use PCE, but who do not directly
handle the material. ONUs may also be exposed to PCE but are expected to have lower inhalation
exposures and are not expected to have dermal exposures. ONUs for this condition of use may include
supervisors, managers, engineers, and other personnel in nearby production areas.

2.3.3.2 Number of Potentially Exposed Workers

EPA estimated the number of workers and occupational non-users potentially exposed during processing
of PCE as a reactant using Bureau of Labor Statistics' OES data (il$	i ) and the U.S. Census'

SUSB (U.S. Census Bureau. 2015) as well as the primary NAICS and SIC code reported by each site
identified in Section 2.3.1 in the 2016 TRI (	)or2016DMR(	>),

respectively. The method for estimating number of workers is detailed above in Section 1.4.4 and
Appendix A. These estimates were derived using industry- and occupation-specific employment data
from the BLS and U.S. Census. The employment data from the U.S. Census SUSB and the Bureau of
Labor Statistics' OES data are based on NAICS codes; therefore, SIC codes reported in the DMR had to
be mapped to a NAICS code to estimate the number of workers. A crosswalk of the SIC codes to the
NAICS codes used in the analysis are provided in Table 2-11. In the 2016 DMR there were three sites
that did not report a SIC code; for these sites, EPA used the average workers and ONUs per site
calculated from the other sites with reported NAICS or SIC codes (	>)• Sites from TRI

report NAICS codes; therefore, these codes were used directly in the analysis.

Table 2-11. Crosswalk of Reactant SIC Codes in DMR to NAICS Codes

Sl( ( ode

Corresponding NAICS Code

2812 - Alkalies and Chlorine

325180 - Other Basic Inorganic Chemical
Manufacturing

2816 - Inorganic Pigments51

325100 - Basic Chemical Manufacturing

2819 - Industrial Inorganic Chemicals, Not
Elsewhere Classified

325180 - Other Basic Inorganic Chemical
Manufacturing

2869 - Industrial Organic Chemicals, Not
Elsewhere Classified13

325100 - Basic Chemical Manufacturing

a The SIC code 2812 may map to any of the following NAICS codes: 325130 or 325180. There is not enough information in
the DMR data to determine the appropriate NAICS for each site; therefore, EPA uses data for the 4-digit NAICS, 325100,
rather than a specific 6-digit NAICS.

b The SIC code 2869 may map to any of the following NAICS codes: 325110, 325120, 325193, 325194, or 325199. There is
not enough information in the DMR data to determine the appropriate NAICS for each site; therefore, EPA uses data for the
4-digit NAICS, 325100, rather than a specific 6-digit NAICS.

Table 2-12 provides a summary of the NAICS codes reported in the 2016 TRI and the NAICS identified
in the crosswalk from the SIC codes reported in the 2016 DMR, the number of sites reporting each
NAICS code or corresponding SIC code, and the estimated number of workers and ONUs for each
NAICS code as well as an overall total for processing of PCE as a reactant. There are approximately
4,200 workers and 1,900 ONUs potentially exposed during processing of PCE as a reactant.

Page 53 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Table 2-12. Estimated Number of Workers Potentially Exposed to Perchloroethylene During
Processing as a Reactant	

NAICS
Code

Number of
Sites

Kxposed
Workers
per Site11

Kxposed
Occupational
Non-l sers
per Site"

Total
Kxposed
Workers

Total
Kxposed
Occupational
Non-l sers

Total
Kxposed

324110

6

170

75

1,021

453

1,474

325100

86

29

13

2,454

1,156

3,610

325120

2

14

7

28

13

41

325180

16

25

12

403

190

592

325199

3

39

18

116

55

170

325211

1

27

12

27

12

40

Unknown
NAICS

3

51

23

152

69

221

Totalb

117

36

17

4,200

1,900

6,100

a Number of workers and occupational non-users per site are calculated by dividing the exposed number of workers or
occupational non-users by the number of establishments in the relevant NAICS codes. The workers/ONUs per site are then
multiplied by the number of sites within that NAICS to get the total exposed. The number of workers/ONUs per site is
rounded to the nearest integer. Number of workers and occupational non-users per site for sites with unknown NAICS codes
are calculated by averaging the values of the known sites.

b Totals have been rounded to two significant figures. Totals may not add exactly due to rounding.

2.3.3.3 Occupational Exposure Results

EPA identified inhalation monitoring data at a Dow Chemical site for a "Phase Separation Facility" that
may be related to processing PCE as a reactant (Dow Chemical. 1983). However, the data were not used
in the assessment as details of the facility were not provided in the report to confirm the specific
condition of use of PCE. It is also unclear if PCE is meant to be in the phase separation area or if it is
only present as an impurity in a refrigerant product after the reaction has complete. In such a case, the
low concentration of PCE as an impurity in the refrigerant product would limit potential exposures and
thus not be representative of exposures of handling bulk liquid PCE at the same facility (e.g., during
unloading of tank trucks or rail cars of raw PCE). Additionally, the sample times for these data are all
less than three hours and, therefore, may not be representative of full-shift exposures.

EPA assumes that potential sources of exposure at sites using PCE as a reactant are similar to sites
manufacturing raw PCE. Therefore, EPA assessed inhalation exposures during processing PCE as a
reactant using monitoring data from manufacturing sites as a surrogate for sites processing PCE as a
reactant. For a discussion of these data see Section 2.1.3.3. The data are summarized in Table 2-13,
where the 50th percentile is presented as the central tendency and the 95th percentile is presented as the
high-end.

Page 54 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Table 2-13. Summary of Worker Inhalation Monitoring Results for Processing Perchloroethylene
as a Reactant

Scenario

S- or 12-
lir TW A
(ppm)

AC
(ppm)

A IK

(ppm)

I.AIK
(ppm)

Nil m her
of Data
Points

15- or 30-
miniilc
TWA
(ppm)

N il in her of
Shorl-lcrm
Data Points

8-hr TWA Results

15-minute TWA Results

High-End

2.6

0.9

0.6

0.3

75a

15

161

Central Tendency

3.25E-02

1.08E-02

7.42E-03

2.95E-03

2.0

12-hr TWA Results

30-minute TWA Results

High-End

0.2

0.1

7.26E-02

3.72E-02

77

12

38b

Central Tendency

2.05E-02

1.03E-02

7.02E-03

2.79E-03

0.7

AC = Acute Concentration; ADC = Average Daily Concentration; and LADC = Lifetime Average Daily Concentration.
Equations and parameters for calculation of the AC, ADC, and LADC are described in Appendix B.
a Data does not include 18 data points that were reported as being below the detection limit, but for which the company did
not provide the LOD for use in estimating an exposure value.

b Data does not include five data points that were reported as being below the detection limit, but for which the company did
not provide the LOD for use in estimating an exposure value.

Sources: (HSIA. 20.1.8').

2.3.4 Water Release Assessment

2.3.4.1	Water Release Sources

Potential sources of water releases are expected to be similar to those described in Section 2.1.4.1 for
manufacturing and may include the following: equipment cleaning operations, aqueous wastes from
scrubbers/decanters, reaction water, process water from washing intermediate products, and trace water
settled in storage tanks (01	).

2.3.4.2	Water Release Assessment Results

EPA assessed water releases using the values reported to the 2016 TR1 (	) and the 2016

DMR (	b) by the 117 sites using PCE as a reactant. Note: Eagle US 2, LLC reported to

both the 2016 TRI and 2016 DMR; EPA assessed using the reported discharge value from DMR as it is
more protective than the value reported in TRI (I v U \ ). In the 2016 TRI, seven sites
reported non-zero direct discharges to surface water, one site reported indirect discharges to POTW, and
all the sites reported zero indirect discharges to non-POTW WWT (	1). In the 2016

DMR, 12 sites reported non-zero direct discharges to surface water and the remainder report zero
discharges to surface water (indirect discharges not reported in DMR data) (	).

To estimate the daily release, EPA assumed 350 days/yr of operation7 and averaged the annual release
over the operating days. Table 2-14 summarizes the water releases from the 2016 TRI and DMR for
sites with non-zero discharges.

7 Similar to manufacturing, sites using PCE as a reactant are expected to have high throughputs and as such are assumed to
operate seven days per week and 50 weeks per year with two weeks per year for shutdown activities.

Page 55 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Table 2-14. Reported Wastewater Discharges of Perchloroethylene from Sites Processing
Perchloroethylene as a Reactant	

Site

Annual

Release11
(kg/site-
year)

Annual
Release

Days
(days/yr)

Daily
Release
(kg/siie-
(iay)11

NPDKS
Code

Release
Media/
Treatment
Kacility
Type

Source

Akzo Nobel Surface
Chemistry LLC,
Morris, IL

4.82E-02

350

1.38E-04

IL0026069

Surface
Water

(

EPA,
2016b")

Atkemix Ten Inc,
Louisville, KY

26

350

7.39E-02

KY0002780

Surface
Water

(11 S
EPA,
2016b")

Bayer Corporation,
Haledon, NJ

1.37E-02

350

3.92E-05

NJG104451

Surface
Water

(11 S
EPA,
2016b")

Bayer

Material Science, New
Martinsville, WV

0.2

350

7.11E-04

WV0005169

Surface
Water

(

EPA,
2016b")

Chemtura North and
South Plants,
Morgantown, WV

8.28E-03

350

2.37E-05

WV0004740

Surface
Water

(

EPA,
2016b")

Dupont-Chemours
Montague Site,
Montague, MI

5.9

350

2.37E-05

MI0000884

Surface
Water

(11 S

EPA,
2016b")

Eagle US 2 LLC -
Lake Charles
Complex, Lake
Charles, LA

465

350

1.3

LA0000761

Surface
Water

(11 S

EPA,
2016b)

Flint Hills Resources
Corpus Christi LLC -
West Plant, Corpus
Christi, TX

24

350

6.87E-02

TXU001146

Surface
Water

(11 S

EPA,

)

Flint Hills Resources
Pine Bend LLC,
Rosemount, MN

4.1

350

1.17E-02

MN0070246

Surface
Water

(11 S

EPA,

)

Honeywell
International Inc -
Geismar Complex,
Geismar, LA

7.1

350

2.03E-02

LA0006181

Surface
Water

(

EPA,
2016b)

Page 56 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Site

Annusil

Release"
(kg/sile-
vesir)

Annual
Release

Days
(davs/vr)

Daily
Release
(kg/sile-
day)11

NPDKS
Code

Release
.Media/
Treatment
l-'acilily
Type

Source

Honeywell
International Inc
Geismar Plant,
Carville, LA

7.3

350

2.07E-02

LA0006181

Surface
Water

(U.S.
EPA,

)

Honeywell
International Inc-
Baton Rouge Plant,
Baton Rouge, LA

17

350

4.92E-02

LAR10E873

Surface
Water

(11 S

EPA,

)

Indorama Ventures
Olefins, LLC,
Sulphur, LA

4.07E-03

350

1.16E-05

LA0069850

Surface
Water

(

EPA,
2016b")

Keeshan And Bost
Chemical Co., Inc.,
Manvel, TX

1.66E-02

350

4.73E-05

TX0072168

Surface
Water

(11 S

EPA,
2016b")

Phillips 66 Lake
Charles Refinery,
Westlake, LA

21

350

5.87E-02

LAR05P540

Surface
Water

(11 S

EPA,

)

Phillips 66 Los
Angeles Refinery
Wilmington Plant,
Wilmington, CA

38

350

0.1

CA0000035

POTW

(

EPA,

)

Premcor Refining
Group Inc Port
Arthur, Port Arthur,
TX

45

350

0.1

TX0005991

Surface
Water

(11 S

EPA,

)

Solutia Nitro Site,
Nitro, WV

5.76E-02

350

1.64E-04

WV0116181

Surface
Water

(11 S

EPA,
2016b")

Solvay - Houston
Plant, Houston, TX

8.3

350

2.36E-02

TX0007072

Surface
Water

(

EPA,
2016b")

POTW = Publicly-Owned Treatment Works; WWT = Wastewater Treatment

a Annual release amounts are based on the site reported values. Therefore, daily releases are calculated from the annual
release rate and assuming 350 days of operation per year.

Sources: (U.S. EPA. 2017d. 2016b)

Page 57 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

2.4 Incorporation into Formulation, Mixture, or Reaction Product

2.4.1	Estimates of Number of Facilities

To determine the number of sites that incorporate PCE into a formulation, mixture or reaction product,
EPA considered 2016 CDR (	), 2016 TRI data (	), and 2016 DMR

(	2016b) data. In the 2016 CDR, 10 sites reported at least one downstream incorporation of

PCE into formulation activity in the industrial processing and use section (as described in Section 2.3.1,
"downstream" may refer to activities at the reporting site or activities at the reporting site's customers)
(	2016d). The industry sectors reported include: soap, cleaning compound, and toilet

preparation manufacturing; paint and coating manufacturing; petroleum refineries; fabricated metal
product manufacturing, all other chemical product and preparation manufacturing; wholesale and retail
trade; adhesive manufacturing; and one sector claimed as CBI (U.S. EPA. 2016d). Of the 10 reported
instances of incorporation, seven reported fewer than 10 sites, one claimed the number of sites as CBI,
and two reported the number of sites at not known or reasonably ascertainable (	I016d)-

EPA identified 23 facilities in the 2016 TRI where the primary condition of use is expected to be
incorporation into formulation based on the site reporting "processing as a formulation component" and
the reported NAICS codes (	). Note: Additional sites may have reported processing as a

formulation component that are not included in the 23 sites used for this scenario because they were
determined to fit best in another condition of use based on other processing activities and/or NAICS
codes reported in the 2016 TRI (see Section 1.4.1 for details of this process).

In the 2016 DMR data, there is one site that reported SIC code 2841, Soap and Other Detergents, Except
Specialty Cleaners; one site that reported SIC code 2843, Surface Active Agents, Finishing Agents,
Sulfonated Oils, and Assistants; two sites that reported SIC code 2851, Paints, Varnishes, Lacquers,
Enamels, and Allied Products; one site that reported 2891, Adhesives and Sealants; eight sites that
reported 2899, Chemicals and Chemical Preparations, Not Elsewhere Classified; and three sites that
reported SIC code 2992, Lubricating Oils and Greases (U.S. EPA. 2016b). There are an additional two
sites in DMR that were the same as formulation sites identified in TRI; therefore, they were not included
in these counts. These SIC codes cover the manufacture of various products in which PCE is a
formulation component, including degreasing and cleaning solvents, aerosol degreasers and lubricants,
paints, coatings, adhesives, and sealants. Therefore, EPA assumes sites reporting these SIC codes are
primarily engaged in formulation activities. Additional information for conditions of use is not provided
in the DMR data; therefore, EPA assessed the primary condition of use at this site based solely on the
SIC code. Based on the DMR and TRI data, EPA assesses a total of 39 sites (23+1+1+2+1+8+3 = 39
sites) for the incorporation of PCE into formulations.

2.4.2	Process Description

After manufacture, PCE may be supplied directly to end-users, or may be incorporated into various
products and formulations at varying concentrations for further distribution. Incorporation into a
formulation, mixture, or reaction product refers to the process of mixing or blending several raw
materials to obtain a single product or preparation. For example, formulators may mix PCE with other
additives to formulate adhesives, coatings, inks, aerosols, and other products.

The formulation of coatings and inks typically involves dispersion, milling, finishing and filling into
final packages (OE< 10, 2009b). Adhesive formulation involves mixing together volatile and non-

Page 58 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

volatile chemical components in sealed, unsealed or heated processes (OECD. 2009a). Sealed processes
are most common for adhesive formulation because many adhesives are designed to set or react when
exposed to ambient conditions (OECD. 2009a). Lubricant formulation typically involves the blending of
two or more components, including liquid and solid additives, together in a blending vessel (OECD.

2004).

Aerosol packing involves first adding PCE and other components into a mixing vessel and blending to
create the final formulation (NIO| lib). The formulation is then gravity filled into the cans and the
dispensing valves are placed and crimped on the can (NIOS Sib). Then the propellent is injected
into the cans and buttons are placed on top of the valves (NIOS Ob). Finally, the cans are passed
through a tank of heated water to check for leaks and weighed to insure the proper level of contents
CNIQSH. 1981b).

2,4.3 Exposure Assessment

2.4.3.1	Worker Activities

At formulation facilities, workers are potentially exposed when unloading PCE into mixing vessels,
taking QC samples, and packaging formulated products into containers and tank trucks. The exact
activities and associated level of exposure will differ depending on the degree of automation, presence
of engineering controls, and use of PPE at each facility.

2.4.3.2	Number of Potentially Exposed Workers

EPA estimated the number of workers and occupational non-users potentially exposed during
formulation of PCE-containing products using Bureau of Labor Statistics" OES data (	1. 2016)

and the U.S. Census" SUSB (U.S. Census Bureau. 2015) as well as the primary NAICS and SIC code
reported by each site in the 2016 TRI (U.S. EPA. 2017d) or 2016 DMR (U.S. EPA. 2016b).
respectively. The method for estimating number of workers is detailed above in Section 1.4.4 and
Appendix A. These estimates were derived using industry- and occupation-specific employment data
from the BLS and U.S. Census. The employment data from the U.S. Census SUSB and the Bureau of
Labor Statistics' OES data are based on NAICS codes; therefore, SIC codes reported in the DMR had to
be mapped to a NAICS code to estimate the number of workers. A crosswalk of the SIC codes to the
NAICS codes used in the analysis are provided in Table 2-15. Sites from TRI report NAICS codes;
therefore, these codes were used directly in the analysis.

Page 59 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Table 2-15. Crosswalk of Formulation SIC Codes in DMRto NAICS Codes

Sl( ( ode

Corresponding NAICS Code

2841 - Soap and Other Detergents, Except
Specialty Cleaners

325611 - Soap and Other Detergent
Manufacturing

2843 - Surface Active Agents, Finishing Agents,
Sulfonated Oils, and Assistants

325613 - Surface Active Agent Manufacturing

2851 - Paints, Varnishes, Lacquers, Enamels, and
Allied Products

325510 - Paint and Coating Manufacturing

2891 - Adhesives and Sealants

325520 - Adhesive Manufacturing

2899 - Chemicals and Chemical Preparations, Not
Elsewhere Classified

325998 - All Other Miscellaneous Chemical
Product and Preparation Manufacturing

2992 - Lubricating Oils and Greases

324191 - Petroleum Lubricating Oil and Grease
Manufacturing

Table 2-16 provides a summary of the reported NAICS codes (or NAICS mapped to the reported SIC
code), the number of sites reporting each NAICS code, and the estimated number of workers and ONUs
for each NAICS code as well as an overall total for formulation of PCE-containing products. There are
approximately 800 workers and 310 ONUs potentially exposed during formulation of PCE-containing
products.

Page 60 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Table 2-16. Estimated Number of Workers Potentially Exposed to Perchloroethylene During
Formulation

NAICS
Code

Number of
Sites

Exposed
Workers
per Site11

Kxposed
Occupational
Non-l sers
per Site"

Total
Kxposed
Workers

Total
Kxposed
Occupational
Non-l sers

Total
Kxposed

324110

1

170

75

170

75

246

324191

3

20

9

61

27

87

325212

1

25

11

25

11

36

325510

4

14

5

57

21

79

325520

5

18

7

90

34

124

325611

2

19

4

37

9

46

325612

2

17

4

33

8

41

325613

1

22

5

22

5

27

325998

18

14

5

253

84

337

326150

1

15

4

15

4

19

336413

1

41

35

41

35

76

Totalb

39

21

8

800

310

1,100

a Number of workers and occupational non-users per site are calculated by dividing the exposed number of workers or
occupational non-users by the number of establishments in the relevant NAICS codes. The workers/ONUs per site are then
multiplied by the number of sites within that NAICS to get the total exposed. The number of workers/ONUs per site is
rounded to the nearest integer.

b Totals have been rounded to two significant figures. Totals may not add exactly due to rounding.

2.4.3.3 Occupational Exposure Results

EPA identified inhalation exposure monitoring data related to the aerosol packing of PCE-containing
products. However, no monitoring data was identified for other formulation sites and it is unlikely
aerosol packing is representative of other formulation sites where workers are exposed during unloading
of bulk containers (i.e., tank trucks and rail cars) and loading of formulated products into smaller
containers (e.g., drums). Therefore, EPA used the monitoring data to assess exposures at aerosol packing
facilities and the EPA/OAQPS AP-42 Loading Model, EPA/OPPTMass Balance Model and Monte
Carlo simulation to assess exposures at other non-aerosol packing facilities. The modeling approach is
presented in Appendix F.

2.4.3.3.1 Inhalation Exposure Results for Aerosol Packing Formulation Sites Using Monitoring
Data

Table 2-17 summarizes 8-hr TWA PBZ monitoring data for aerosol packing formulation sites. The data
were obtained by NIOSH during an inspection at a facility that packages commercial aerosol spot
removers containing PCE and methyl chloroform (NIOSH. 1981b). The report indicates that local
exhaust ventilation was present at the filling, button tipper, and hot tank locations (NIOSH. 1981b). The
report also indicated that administrative controls requiring employees to rotate through various positions

Page 61 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

throughout each workday with each employee working at four different locations during an eight-hour
day (NIOSH. IS ). TWA exposures were calculated by combining short-term samples collected from
each employee at each position throughout the day (NIOSH. 1' ). Total sample times ranged from
6.5 to 8 hours; for sample times less than eight hours, the 8-hr TWAs were calculated assuming
exposure to be zero outside the sampling time (NIOSH. 1981b). Due to the limited number of data
points (five), EPA used the maximum value as the high-end and the 50th percentile as the central
tendency. Data were not available to estimate ONU exposures; EPA estimates that ONU exposures are
lower than worker exposures, since ONUs do not typically directly handle the chemical.

Table 2-17. Summary of Worker Inhalation Exposure Monitoring Data for Aerosol Packing
Formulation Sites

Scenario

8-hr TWA

AC

A IK

I.AIK

Number of

(ppm)

(ppm)

(ppm)

(ppm)

Data Points

High-End

13

4.4

3.0

1.5

r

Central Tendency

8.3

2.8

1.9

0.8

3

AC = Acute Concentration; ADC = Average Daily Concentration; and LADC = Lifetime Average Daily Concentration.
Equations and parameters for calculation of the AC, ADC, and LADC are described in Appendix B.

Sources: fNIOSH. .1.98.1.51

2.4.3.3.2 Inhalation Exposure Results for Non-Aerosol Packing Formulation Sites Using
Modeling

The modeling approach used to assess exposures at non-aerosol packing formulation sites estimates
exposures to workers loading formulated PCE-based products into drums. Inhalation exposure to
chemical vapor during loading is a function of physical properties of PCE, various EPA default
constants, and other model parameters. While physical properties are fixed for a substance, some model
parameters, such as weight fraction of PCE in the product, ventilation rate, mixing factor, and vapor
saturation factor, are expected to vary from one facility to another. This approach addresses variability
for these parameters using a Monte Carlo simulation.

The modeling approach requires an input on the number of containers loaded per day which is
determined based on the throughput of PCE at each site and the weight fraction of PCE in the product.
To determine these values EPA divide each site identified in Section 2.4.1 into one of the following
categories: 1) sites formulating degreasing solvents; 2) sites formulating dry cleaning solvents, and 3)
sites formulating "miscellaneous" PCE-containing products, including coatings, adhesives,
metalworking fluids, and other niche use PCE-based products. Note: Market data for the third group
were not available at a detailed level; therefore, EPA could not divide the PCE production volume
amongst the product types to calculate per site throughputs. Each site was categorized based on its
NAICS code. EPA categorized the NAICS codes as follows:

• Degreasing solvent formulation NAICS codes:
o 324110 - Petroleum Refineries8; and

8 EPA does not typically expect petroleum refineries to formulate degreasing solvents; however, the one site reporting this
NAICS code to the 2016 TRI also reported as an importer to the 2016 CDR and reported its entire import volume as used on-
site and reported formulation of solvents for cleaning and degreasing.

Page 62 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

o 325998 - All Other Miscellaneous Chemical Product and Preparation Manufacturing9.

•	Dry cleaning solvent formulation NAICS codes:

o 325611 - Soap and Other Detergent Manufacturing;
o 325612 - Polish and Other Sanitation Good Manufacturing; and
o 325613 - Surface Active Agent Manufacturing.

•	Miscellaneous formulation NAICS codes:

o All NAICS codes reported not listed above.

The categorization resulted in 19 formulation sites for degreasing solvents, five for dry cleaning
solvents, and 15 for miscellaneous products. EPA then used market data to estimate the throughput at
each site by dividing the estimated percentage of PCE used in each formulation type by the number of
formulation sites for that product. To estimate daily throughputs, EPA assumed 300 days/yr10 of
operation as given in the SpERC developed by the European Solvent Industry Group for the formulation
and (re)packing of substances and mixtures and averaged the annual throughput over the operating days
(European Solvents Industry Group. 2019a). The market data estimated 7-10% of the national PCE
production volume is used for degreasing, 10-15% is used for dry cleaning, and 3-10% is used for
miscellaneous uses ( x JO I I; H\ J008). EPA used 7% for degreasing, 10% for dry cleaning, and
3% for miscellaneous because these values represent more recent data. Table 2-18 summarizes the
estimated per site PCE-throughputs for each category.

Table 2-18. Estimated Throughputs of Perchloroethylene by Formulated Product Type

l-'ormulalion
Type

Percent of
National
Prod ucl ion
Volume

Annual PCK
I se Kate
(Ib/vr)

Total
I'orm illation
Sites

Annual Per
Site PCK-
Throiighput
(Ih/site-vr)

Operating

Days
(davs/yr)

Daily PC'K-
Throiighpul
(Ih/site-day)

Degreasing
Solvent

7%

22,696,852

19

1,194,571

300

3,982

Dry Cleaning
Solvent

10%

32,424,074

5

6,484,815

300

21,616

Miscellaneous

3%

9,727,222

15

648,481

300

2,162

EPA assumed formulated products were loaded into 55-gallon drums. It is possible that some formulated
products, such as coatings and adhesives, may be loaded into smaller containers (e.g., pails) for smaller
commercial and consumer applications; however, EPA does not have information to estimate the
volume packaged into drums versus smaller containers. Therefore, EPA assessed the entire throughput
as packaged into drums to give the most protective worker exposure estimates.

9	This NAICS codes may also include sites manufacturing aerosol products; therefore, the total number of sites for
formulating degreasing solvents may be overestimated.

10	EPA uses 300 days per year rather than 350 as used in the manufacturing and reactant scenarios because it is likely that
formulation sites make multiple products not all of which will contain PCE. Drum loading of PCE-based products is only
expected to occur on days were PCE-containing products are produced.

Page 63 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

To estimate the number of drums loaded per day EPA used the per site daily throughput of PCE and the
expected weight fraction of PCE in the formulated product to estimate the total volume of PCE loaded
into each drum. For degreasing and dry cleaning solvents EPA assumed the PCE weight fraction to be
100%. Typically, the only materials expected to be added to degreasing and dry cleaning solvents are
stabilizers used to prevent decomposition during storage and use (European Chlorinated Solvents
Association. 2011). PCE generally requires less stabilizers than other chlorinated solvents with weight
fractions of stabilizers expected to be less than 0.5% in degreasing solvents, and less than 0.05% in dry
cleaning solvents. (European Chlorinated Solvents Association. 2011). Therefore, the assumption of
100%) PCE in the model is not expected to significantly impact exposure results.

For miscellaneous products, the concentration of PCE can vary greatly depending on the product being
formulated. For modeling purposes, EPA assessed used a uniform distribution of 30 to 80% PCE in the
formulated product based on the expected concentrations of solvents in organic solvent-borne coatings
estimated by the OECD ESD (OE 09b). This range was used as it is expected to encompass the
range of compositions for the majority of PCE-based products in this category (e.g., per the OECD ESD
(OECD. 2009a) typical organic solvent concentrations in adhesives is estimated to be between 60 to
75%) which falls within the range used in the model). While it is possible that some of the products
contain PCE concentrations outside this range, the error from this is expected to be small as, based on
the reported NAICS codes, 10 of the 15 formulation sites assessed in this category are either coatings
(including maskants) or adhesive formulation sites.

Model results for each category of formulation site are presented in Table 2-19 with the 50th percentile
presented as the central tendency and the 95th percentile presented as the high-end. It should be noted
that an additional exposure for workers may occur during unloading of raw PCE from bulk containers
(tank trucks and rail cars) into formulation equipment and is not accounted for in the results presented in
Table 2-19. Although EPA can estimate exposures during this unloading activity using the Tank Truck
and Railcar Loading and Unloading Release and Inhalation Exposure Model (see Appendix E for model
description), it is unclear if the same workers will perform both unloading and loading activities in the
same day. Therefore, it may not be accurate to combine estimates from each model to estimate a total
exposure.

In the case where a worker is both unloading bulk containers and loading products into drums on the
same day, the overall error from not including exposures during unloading in the results is expected to
be small as the Tank Truck and Railcar Loading and Unloading Release and Inhalation Exposure Model
estimates an 8-hr TWA exposure of 0.01 ppm for tank truck unloading and an 8-hr TWA of 0.04 ppm
for railcar unloading whereas the model for drum loading estimates 8-hr TWAs ranging from 0.60 to
14.1 ppm.

The results show that exposures at sites formulating dry cleaning solvents are an order of magnitude
higher than other formulation sites. This due to the fact that dry cleaning solvents are a larger use than
the other assessed categories and have the fewest number of formulation sites resulting in larger
numbers of drums loaded per day at each site.

Page 64 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Table 2-19. Summary of Exposure Modeling Results for Formulation of Perchloroethylene-Based
Products

l-'ormiilalion Type

Scenario

8-hr TWA

(ppm)

AC
(ppm)

A IK

(ppm)

LADC
(ppm)

Degreasing Solvent

High-End

2.6

0.4

5.66E-02

8.35E-03

Central Tendency

0.7

0.1

1.59E-02

2.43E-03

Dry Cleaning
Solvent

High-End

14

2.1

0.3

4.53E-02

Central Tendency

4.0

0.6

8.61E-02

1.27E-02

Miscellaneous

High-End

1.4

0.2

3.07E-02

4.53E-03

Central Tendency

0.4

5.86E-02

8.64E-03

1.27E-03

AC = Acute Concentration; ADC = Average Daily Concentration; and LADC = Lifetime Average Daily Concentration.
Equations and parameters for calculation of the AC, ADC, and LADC are described in Appendix B.

2.4.4 Water Release Assessment

2.4.4.1	Water Release Sources

The primary source of water releases from the formulation of PCE-containing products is from water
used to clean the formulation equipment (OECD. 2010. 2009a. b, 2004). There is also potential for water
releases from cleaning of containers used to transport raw PCE (OECD. 2009b). For organic solvent-
based products such as PCE-based products, EPA expects the majority of container and equipment
cleaning to be performed using organic solvents that are not discharged to water. However, there is the
potential for sites to use water as a cleaning solvent that is subsequently discharged directly to surface
water or indirectly to POTWs or non-POTW WWT.

2.4.4.2	Water Release Assessment Results

EPA assessed water releases using the values reported to the 2016 TRI (	) and the 2016

DMR (U.S. EPA. 2016b) by the 39 formulation sites. In the 2016 TRI, one site reported indirect
discharges to POTW, one site reported indirect discharges to non-POTW WWT, and the remaining sites
reported zero discharges (\ c. < ^ \ T0i ,;). In the 2016 DMR, one site reported non-zero direct
discharges to surface water and the remaining sites reported zero discharges to surface water (indirect
discharges not reported in DMR data) (	b).

To estimate the daily release, EPA assumed 300 days/yr of operation as given in the SpERC developed
by the European Solvent Industry Group for the formulation and (re)packing of substances and mixtures
and averaged the annual release over the operating days (European Solvents Industry Group. 2019a).
Table 2-20 summarizes the water releases from the 2016 TRI and DMR for sites with non-zero
discharges.

Page 65 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Table 2-20. Reported Wastewater Discharges of Perchloroethylene from Formulation of
Perchloroethylene-Containing Products	

Site

Annual
Release"
(kg/sile-
year)

Annual
Release

Days
(davs/vr)

Daily
Release
(kg/sile-
day)11

NPDKS
Code

Release
Media/
Treatment
Kacilily Type

Source

Lord Corp,
Saegertown, PA

1,579

300

5.3

PA0101800

Non-POTW
WWT

(U.S. EPA.
2017d)

Stepan Co
Millsdale Road,
Elwood, IL

0.5

300

1.66E-03

IL0002453

Surface Water

(U.S. EPA.
2016b)

Weatherford
Aerospace LLC,
Weatherford, TX

0.5

300

1.51E-03

Not available

POTW

(U.S. EPA.
2017d)

POTW = Publicly-Owned Treatment Works; WWT = Wastewater Treatment

a Annual release amounts are based on the site reported values. Therefore, daily releases are calculated from the annual
release rate and assuming 300 days of operation per year.

Sources: (U.S. EPA. 2017d. 2016b)

2.5 Batch Open-Top Vapor Degreasing

2.5.1 Estimates of Number of Facilities

To determine the number of sites that use PCE in batch open-top vapor degreasers (OTVD), EPA
considered 2014 NEI (U.S. EPA. 2016a). 2016 TRI (	I), and 2016 DMR (U.S. EPA.

2016b) data. However, due to the various reporting thresholds and requirements for each of the above
sources, EPA does not expect the sites from these sources to represent the entirety of sites operating
OTVDs. Therefore, EPA used methods presented in the 2017 Draft ESD on Vapor Degreasing to
estimate the number of sites (	). Based on market data from HSIA (2008) and NTP (2014).

EPA expects 7 to 10% of the production volume of PCE to be used in vapor degreasing. Due to data
limitations, this portion of the production volume cannot be further divided into different degreasing
types (OTVDs, closed-loop degreasing, conveyorized degreasing, web degreasing, and cold cleaning).
Therefore, EPA had to perform bounding estimates on the number of sites, using the full portion of the
production volume used in metal degreasing for each degreaser type. Bounding estimates may
overestimate actual number of sites. To estimate the number of sites for OTVDs, EPA assessed 7% of
the national production volume (10,295,119 kg/yr) as used in OTVDs. EPA used 7% rather than 10%
because the 7% value is more recent and to reduce the degree of overestimation from the bounding
calculation.

The ESD estimates a 50th percentile use-rate for OTVDs of 2,083 kg/site-yr and a 95th percentile use rate
of 25,852 kg/site-yr (O	). EPA calculated the number of sites corresponding to both the 50th

and 95th percentile use-rates using the following equation:

Equation 2-1

PV

Ns~ur

Page 66 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Where:

Ns

PV

UR

Number of Sites

Annual PCE Production Volume Used in Degreasing (kg/yr)
Annual use-rate of PCE (kg/site-yr)

This resulted in 398 sites using the 95th percentile use-rate and 4,942 sites using the 50th percentile use-
rate.

2.5.2 Process Description

Vapor degreasing is a process used to remove dirt, grease, and surface contaminants in a variety of
industries, including but not limited to:

•	Electronic and electrical product and equipment manufacturing;

•	Metal, plastic, and other product manufacturing, including plating;

•	Aerospace manufacturing and maintenance cleaning;

•	Cleaning skeletal remains; and

•	Medical device manufacturing (Morford. 2017).

PCE is typically chosen as a degreasing solvent for applications where flammability is a concern as PCE
has no flash point and no upper and lower explosive limits (Rudnick.: ). Figure 2-1 is an illustration
of vapor degreasing operations, which can occur in a variety of industries.

Page 67 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Fabrication
Shops

Metal
Plating
Shops

Electronics
Assembly
Shops

Repair
Shops



<&£>

2.

Vafoh Of gas ammo

0	lk

mm

X

Vuoa Dkhuw

mm

X

V«K« 0(tf(UM6

M

% t
¦>

I i

i i

X *

I I

o

Vwo* Dickum

;5\

O



# flj-. — ...

I

# ad

Electroplating

%,5,

4 jjJ

tT #

V

Figure 2-1. Use of Vapor Degreasing in a Variety of Industries

Vapor degreasing may take place in batches or as part of an in-line (i.e., continuous) system. In batch
machines, each load (parts or baskets of parts) is loaded into the machine after the previous load is
completed. With in-line systems, parts are continuously loaded into and through the vapor degreasing
equipment as well as the subsequent drying steps. Vapor degreasing equipment can generally be
categorized into one of the three categories: (1) batch vapor degreasers, (2) conveyorized vapor
degreasers and (3) web vapor degreasers.

In batch open-top vapor degreasers (OTVDs), a vapor cleaning zone is created by heating and
volatilizing the liquid solvent in the OTVD. Workers manually load or unload fabricated parts directly
into or out of the vapor cleaning zone. The tank usually has chillers along the side of the tank to prevent
losses of the solvent to the air. However, these chillers are not able to eliminate emissions, and
throughout the degreasing process significant air emissions of the solvent can occur. These air emissions
can cause issues with both worker health and safety as well as environmental issues. Additionally, the
cost of replacing solvent lost to emissions can be expensive ("Jewmoa. 2001). Figure 2-2 illustrates a
standard OTVD.

Page 68 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

/

Condensing Coils

Water Jacket

Vapor Zone

O ^

'Water Separator

Boiling sump

Heat Source

Figure 2-2. Open-Top Vapor Degreaser

OTVDs with enclosures operate the same as standard OTVDs except that the OTVD is enclosed on all
sides during degreasing. The enclosure is opened and closed to add or remove parts to/from the machine,
and solvent is exposed to the air when the cover is open. Enclosed OTVDs may be vented directly to the
atmosphere or first vented to an external carbon filter and then to the atmosphere (EPA and Consulting.
2004). Figure 2-3 illustrates an OTVD with an enclosure. The dotted lines in Figure 2-3 represent the
optional carbon filter that may or may not be used with an enclosed OTVD.

Page 69 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Carbon Filter

vent

-II

Loading/
unloading
lock

M

Condensing Coils
^Wate Jacket

Vapor Zone

er Separator

O

Boiling sump'

Heat Sou ce

I

Figure 2-3. Open-Top Vapor Degreaser with Enclosure

2.5.3 Exposure Assessment

2.5.3.1 Worker Activities

The EPA defined a vapor degreasing "worker" as an employee who operates or performs maintenance
tasks on the degreaser, such as draining, cleaning, and charging the degreaser bath tank. When operating
OTVD, workers manually load or unload fabricated parts directly into or out of the vapor cleaning zone.
Worker exposure can occur from solvent dragout or vapor displacement when the substrates enter or exit
the equipment, respectively (Kanegsberg and Kanegsberg. 2011). The amount of time a worker spends
at the degreaser can vary depending on the number of workloads needed to be cleaned. Reports from
NIOSH at three sites using OTVDs found degreaser operators may spend 0.5 to 2 hours per day at the
degreaser (NIOSH. 2002a. b, d).

Worker exposure is also possible while charging new solvent or disposing spent solvent. The frequency
of solvent charging can vary greatly from site-to-site and is dependent on the type, size, and amount of
parts cleaned in the degreaser. NIOSH investigations found that one site added a 55-gallon drum of new
solvent to the degreaser unit every one to two weeks; another site added one 55-gallon drum per month;
and another site added two 55-gallon drums per month to its large degreaser and three 55 gallon drums
per year to its small degreaser (NIOSH. 2002a. b, d).

EPA defined "occupational non-user" as an employee who does not regularly handle PCE or operate the
degreaser but performs work in the area around the degreaser.

EPA estimated the number of workers and occupational non-users potentially exposed during use of
PCE in OTVDs using the Draft ESD on the Use of Vapor Degreasers (OECD. 2017a). The ESD
estimates seven workers and four ONUs per site (OECD. 2017a). EPA multiplied these values by the

2.5.3.2 Number of Potentially Exposed Workers

Page 70 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

number of sites estimated in Section 2.5.1. This resulted in approximately 2,800 workers and 1,600
ONUs using the number of sites estimated from the 95th percentile use-rate and 35,000 workers and
20,000 ONUs using the number of sites estimated from the 50th percentile use-rate. Table 2-21
summarizes these results. Note: As described in Section 2.5.1, these are bounding estimates and may
overestimate actual number of workers.

Table 2-21. Estimated Number of Workers Potentially Exposed to Perchloroethylene During Use
in Open-Top Vapor Degreasing	

I se-Uale
Scenario

Number of
Sites

Exposed
Workers
per Site

Kxposed
Occupational
Non-l sers
per Site

Total
Kxposed
Workers"

Total
Kxposed
Occupational
Non-l sers"

Total
Kxposed"

95th
Percentile

398

7

4

2,800

1,600

4,400

50th
Percentile

4,942

7

4

35,000

20,000

54,000

a Totals have been rounded to two significant figures. Totals may not add exactly due to rounding.

2.5.3.3 Occupational Exposure Results

EPA assessed exposures using identified inhalation exposure monitoring data from NIOSH
investigations at five sites using PCE as a degreasing solvent in OTVDs. Table 2-22 summarizes the 8-
hr TWA monitoring data, 4-hr TWA monitoring data, and 15-minute TWA monitoring data for the use
of PCE in OTVDs. The high-end and central tendency values for the 8-hr TWA data represent the 95th
and 50th percentile, respectively. Due to the limited number of data points (three samples), the 4-hr
TWA high-end is the maximum value and the central tendency is the median. There is only a single 15-
min TWA sample. Results based on a single value are considered plausible, but EPA cannot determine
the statistical representativeness of the value.

EPA recognizes that worker job titles and activities may vary significantly from site to site; therefore,
EPA typically identified samples as worker samples unless it was explicitly clear from the job title (e.g.,
inspectors) and the description of activities in the report that the employee was not operating the
degreaser during the sampling period. Samples from employees determined not to be operating the
degreasing equipment were designated as ONU samples.

The data were obtained from NIOSH Health Hazard Evaluation reports (HHEs) and NIOSH In-Depth
Survey Reports. NIOSH HHEs are conducted at the request of employees, employers, or union officials,
and provide information on existing and potential hazards present in the workplaces evaluated. The
NIOSH In-Depth Surveys were conducted as part of an interagency agreement with OSHA to evaluate
the extent of employee exposure to PCE at sites using it as a solvent in vapor degreaser and to document
engineering controls and work practices at the workplace affecting exposures (NIOSH. 2002a. b, d).

Data from these sources cover exposures at several industries including aerospace parts manufacturing
and repair/refurbishment, parts manufacturing for surgical implants, and brazed aluminum heat
exchanger and cooling system manufacturing (NIOSH. 2002a. b, d, 1984b. 1982b). Except for one site,
sample times ranged from approximately two to eight hours (NIOSH. 2002a. d, 1984b. 1982b). The

Page 71 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

other site worked on two 10-hr shifts; therefore, the majority of samples were taken for over 8.5 hours,
with only five samples 8 hours or less (NIOSH. 2002b). Where sample times were less than eight hours,
EPA converted to an 8-hr TWA assuming exposure outside the sample time was zero. For sample times
greater than eight hours, EPA left the measured concentration as is. It should be noted that additional
sources for degreasing were identified but were not used in EPA's analysis as they either: 1) did not
specify the machine type in use; or 2) only provided a statistical summary of worker exposure
monitoring.

Gold (2008) completed a comprehensive literature review of studies evaluating PCE exposures from a
variety of uses in the U.S. The study complied data for degreasing from studies completed from 1944 to
2001 and provided the general sample times (either as <1, 1-6 or >6 hours), overall range and mean for
the data as well as ranges and means for each decade and each job title (overall for the job title and by
decade) identified in the studies (Gold et at.. 2008). The most recent data for vapor degreasing
referenced in the article were from studies completed in the 2000s (Gold et al. 2008). The overall
arithmetic mean and maximum from these studies for samples where the sampling time was greater than
six hours was 0.4 ppm and 0.9 ppm, respectively, for degreaser operators (Gold et al.. 2008). The mean
is an order of magnitude lower than the central tendency in EPA's analysis and the maximum is two
orders of magnitude lower than the high-end in EPA's analysis. The difference in results is likely due to
the increased number samples, the data from Gold (2008) only included nine samples, whereas, the
worker data used in this analysis includes 63 samples from multiple sites (number of sites from the Gold
(2008) study is unknown). It should be noted that Gold (2008) does not separate by machine type;
therefore, it may include closed-loop or conveyorized systems, thus further impacting the results.

Table 2-22. Summary of Worker Inhalation Exposure Monitoring Data for Open-Top Vapor
Degreasing	

Scenario

8-hr
TWA
(ppm)

AC
(ppm)

A IK

(ppm)

I.ADC
(ppm)

Nil m her
of Dala
Points

4-hr
TWA
(ppm)

N il in her
of Dala
Points

15-
Minule
TWA
(ppm)11

N il in her
of Dala
Points

Worker Monitoring Data

High-End

32

11

7.3

3.8

63

1.6

3

17

1

Central
Tendency

2.1

0.7

0.5

0.2

1.3

Occupational Non-User Monitoring Data

High-End

5.2

1.7

1.19

0.6

12

No 4-hr or 15-minute data identified
for ONUs

Central
Tendency

0.6

0.2

0.14

5.50E-02

AC = Acute Concentration; ADC = Average Daily Concentration; and LADC = Lifetime Average Daily Concentration.
Equations and parameters for calculation of the AC, ADC, and LADC are described in Appendix B.
a Only one data point identified for 15-min TWAs.

Source: (NIOSH. 2002a. b, d, 1984b. 1982b')

Page 72 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

2.5.4 Water Release Assessment

2.5.4.1	Water Release Sources

The primary source of water releases from OTVDs is wastewater from the water separator. Water in the
OTVD may come from two sources: 1) Moisture in the atmosphere that condenses into the solvent when
exposed to the condensation coils on the OTVD; and/or 2) steam used to regenerate carbon adsorbers
used to control solvent emissions on OTVDs with enclosures (Durkee. 2014; Kameesbere and
Kameesbere. 2011; NIOSH. 2002a. b, c, d). The water is removed in a gravity separator and sent for
disposal (NIOSH. 2002a. b, c, d). The current disposal practices of the wastewater are unknown;
however, a U.S. EPA (1982) report estimated 20% of water releases from metal cleaning (including
batch systems, conveyorized systems, and vapor and cold systems) were direct discharges to surface
water and 80% of water releases were discharged indirectly to a POTW.

2.5.4.2	Water Release Assessment Results

Water releases for OTVDs were assessed using data reported by sites in the 2016 TRI (

2017d) and 2016 DMR (	b). EPA identified 123 sites between the 2016 TRI and 2016

DMR data that, based on activities reported in TRI and/or the facilities' reported NAICS/SIC codes are
likely performing degreasing operations. It should be noted that sites in TRI and DMR do not report
information to differentiate between sites with different degreasing machine types and/or sites using
PCE to perform metalworking activities instead of degreasing activities. Therefore, it is possible the
actual condition of use at these sites is not OTVD but rather a different type of solvent cleaning (e.g.,
closed-loop degreasing, conveyorized degreasing, web cleaning, or cold cleaning) or use of PCE as a
metalworking fluid. These sites are assessed as OTVD based on the fact that 7-10% of the production
volume of PCE is used in metal cleaning/degreasing (compared to <3-10% for all other miscellaneous
uses including metalworking) and, based on NEI reporting, OTVDs are expected to be the primary
cleaning machines used in industry (23 OTVDs reported compared to 1 closed-loop system, 1
conveyorized system, and 10 web cleaning systems11) (I v << \ ¦>., x h I, ti 308).

Only a subset of the 123 sites reported discharges to water. This is likely due to different waste handling
procedures at each site. For instance, some sites may collect wastewater and send to an off-site waste
handling facility that does not discharge the wastewater to WWT or surface waters. EPA assessed
annual releases as reported in the 2016 TRI (	) or 2016 DMR (U.S. EPA. 2016b) and

assessed daily releases by assuming 260 days of operation per year, as recommended in the Draft ESD
on Use of Vapor Degreasers (OEC	) and averaging the annual releases over the operating days.

A summary of the water releases reported to TRI and DMR can be found in Table 2-23.

11 Based on the NEI reporting requirements, the counts of machine types may not be representative of the overall machine
type distribution. However, EPA expects the OTVDs to be the most prevalent type of system as closed-loop systems have
longer cleaning cycles that limit part throughputs and increased cost compared to OTVDs; conveyorized systems are
generally limited to sites with high part throughputs; and web cleaning systems are limited to parts that are coiled or on
spools such as films, wires and metal strips.

Page 73 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Table 2-23. Reported Wastewater Discharges of Perchloroethylene from Sites Using
Perchloroethylene in Open-Top Vapor Degreasing	

Site

Annual
Release"
(kg/sile-
year)

Annual
Release

Days
(days/yr)

Daily
Release
(kg/sile-
(lay)11

m»di:s

Code

Release
Media/
Treatment
l-'acilily Type

Source

601 Nassau St Assoc
LLC, North
Brunswick Twp, NJ

2.44E-03

260

9.39E-06

NJG129127

Surface
Water

(

2016b")

ASCO Valve
Manufacturing,
Aiken, SC

3.70E-02

260

1.42E-04

SC0049026

Surface
Water

(

2016b")

Chemours -
Beaumont Works,
Beaumont, TX

1.7

260

6.49E-03

TX0004669

Surface
Water

(

2016b")

Delphi Harrison
Thermal Systems,
Dayton, OH

1.7

260

6.46E-03

OH0009431

Surface
Water

(

2016b")

Equistar Chemicals
LP, La Porte, TX

3.2

260

1.25E-02

TX0119792

Surface
Water

(

2016b")

Fairfield Works,
Fairfield, AL

1.1

260

4.09E-03

AL0003646

Surface
Water

(

2016b")

Gayston Corp,
Dayton, OH

0.8

260

3.12E-03

OHO 127043

Surface
Water

(

2016b")

Getzen Co Inc,
Elkhorn, WI

9.07E-02

260

3.49E-04

Not available

POTW

(

2017d")

GM Components
Holdings LLC,
Lockport, NY

18

260

7.08E-02

NY0000558

Surface
Water

(

2016b")

HB Fuller Co,
Morris, IL

0.2

260

7.90E-04

IL0079758

Surface
Water

(

2016b")

Hyster-Yale Group,
Inc, Sulligent, AL

2.35E-04

260

9.03E-07

AL0069787

Surface
Water

(

2016b")

MEMC Electronic
Materials
Incorporated,
Moore, SC

6.79E-02

260

2.61E-04

SC0036145

Surface
Water

(

2016b")

Page 74 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Site

Anniiiil

Release"
(k«/site-
vcsir)

A ii ii n:il

Release

Days
(dsivs/vr)

Daily
Release
(kg/sile-
(isiv)11

m»di:s

Code

Release
Media/
Treat nienl
I'lK'ililv Type

Source

Piano Factory-Grand
Haven, Grand
Haven, MI

0.2

260

7.17E-04

MI0054399

Surface
Water

(

2016b)

Rex Heat Treat
Lansdale Inc,
Lansdale, PA

0.5

260

1.94E-03

PA0052965

Surface
Water

(

2016b)

Styrolution America
LLC, Channahon, IL

0.2

260

6.37E-04

IL0001619

Surface
Water

(

2016b)

Trane Residential
Solutions - Fort
Smith, Fort Smith,
AR

3.41E-03

260

1.31E-05

AR0052477

Surface
Water

(

2016b)

US Steel Fairless
Hills Facility,
Fairless Hills, PA

0.3

260

1.01E-03

PA0013463

Surface
Water

(

2016b)

POTW = Publicly-Owned Treatment Works; WWT = Wastewater Treatment

a Annual release amounts are based on the site reported values. Therefore, daily releases are back-calculated from the annual
release rate and assuming 260 days of operation per year.

Sources: (U.S. EPA. 2017d. 2016b)

As discussed in Section 2.5.1, data from TRI and DMR may not represent the entirety of sites using PCE
in OTVDs. EPA did not identify other data sources to estimate water releases from sites not reporting to
TRI or DMR. However, sites operating degreasers are regulated by the following national EGs:

•	Electroplating Point Source Category Subparts A, B, D, E, F, G, and H (1; S 1 P \ 2019a)12;

•	Iron and Steel Manufacturing Point Source Category Subpart J (U.S. EPA. 2019a);

•	Metal Finishing Point Source Category Subpart A (	2019c)13;

•	Coil Coating Point Source Category Subpart D (\ v H* \ >i);

•	Aluminum Forming Point Source Category Subparts A, B, C, D, E, and F (U.S. EPA. 2019D;
and

•	Electrical and Electronic Components Point Source Category Subparts A and B (U.S. EPA.
2019e).

12	The Electroplating EG applies only to sites that discharge to POTW (indirect discharge) that were in operation before July
15, 1983. Processes that began operating after July 15, 1983 and direct dischargers are subject to the Metal Finishing EG

(U.S. EPA. 2019c).

13	The Metal Finishing EG do not apply when wastewater discharges from metal finishing operations are already regulated by
the Iron and Steel, Coil Coating, Aluminum Forming, or Electrical and Electronic Components EGs.

Page 75 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Except for the Iron and Steel EG, the above EGs set discharges limits based on the total toxic organics
(TTO) concentration in the wastewater stream and not a specific PCE limit. TTO is the summation of
the concentrations for a specified list of pollutants which may be different for each promulgated EG and
includes PCE for the above referenced EGs. Therefore, the concentration of PCE in the effluent is
expected to be less than the TTO limit. The Iron and Steel EG sets discharge limits specifically for PCE
based on the operation PCE is being discharged from.

The operation of the water separator via gravity separation is such that the maximum concentration of
PCE leaving the OTVD is equal to the solubility of PCE in water, 206 mg/L (Durkee. 2014). In cases
where this concentration exceeds the limit set by the applicable EGs, EPA expects sites will perform
some form of wastewater treatment for the effluent stream leaving the OTVD to ensure compliance with
the EG prior to discharge. EPA did not identify information on the amount of wastewater generated
from OTVDs to estimate releases from sites not reporting to TRI or DMR.

2.6 Batch Closed-Loo|) Vapor Decreasing

2.6.1	Estimates of Number of Facilities

EPA estimated the number of sites using closed-loop vapor degreasers using the draft ESD on Vapor
Degreasing (OE	) using the same methodology as described for OTVDs in Section 2.5.1. The

ESD estimates a 50th percentile use-rate of 403 kg/site-yr and a 95th percentile use-rate of 740 kg/site-yr
(	). EPA calculated bounding estimates for number sites using the ESD use-rates and the

total 7% of the national production volume reported as used in metal degreasing by HS1A (2008). This
resulted in 13,912 sites using the 95th percentile use-rate and 25,546 sites using the 50th percentile use-
rate. Note: Bounding estimates may overestimate actual number of sites.

2.6.2	Process Description

In closed-loop degreasers, parts are placed into a basket, which is then placed into an airtight work
chamber. The door is closed, and solvent vapors are sprayed onto the parts. Solvent can also be
introduced to the parts as a liquid spray or liquid immersion. When cleaning is complete, vapors are
exhausted from the chamber and circulated over a cooling coil where the vapors are condensed and
recovered. The parts are dried by forced hot air. Air is circulated through the chamber and residual
solvent vapors are captured by carbon adsorption. The door is opened when the residual solvent vapor
concentration has reached a specified level (Kameesbere and Kameesbere. 2011). Figure 2-4 illustrates a
standard closed-loop vapor degreasing system.

Page 76 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

j Vent

Solvent Abatement Loop

Refrigeration

z

Working Chamber
Workload

Electric Heat

Figure 2-4. Closed-Loop/Vacuum Vapor Degreaser

Airless degreasing systems are also sealed, closed-loop systems, but remove air at some point of the
degreasing process. Removing air typically takes the form of drawing vacuum but could also include
purging air with nitrogen at some point of the process (in contrast to drawing vacuum, a nitrogen purge
operates at a slightly positive pressure). In airless degreasing systems with vacuum drying only, the
cleaning stage works similarly as with the airtight closed-loop degreaser. However, a vacuum is
generated during the drying stage, typically below 5 torr (5 mmHg). The vacuum dries the parts and a
vapor recovery system captures the vapors (Kanegsberg and Kanegsberg. 2011; Newmoa. 2001; U.S.
EPA. 200 laV

Airless vacuum-to-vacuum degreasers are true "airless" systems because the entire cycle is operated
under vacuum. Typically, parts are placed into the chamber, the chamber sealed, and then vacuum
drawn within the chamber. The typical solvent cleaning process is a hot solvent vapor spray. The
introduction of vapors in the vacuum chamber raises the pressure in the chamber. The parts are dried by
again drawing vacuum in the chamber. Solvent vapors are recovered through compression and cooling.
An air purge then purges residual vapors over an optional carbon adsorber and through a vent. Air is
then introduced in the chamber to return the chamber to atmospheric pressure before the chamber is
opened (Durkee. 2014; Newmoa. 2001). The general design of vacuum vapor degreasers and airless
vacuum degreasers is similar as illustrated in Figure 2-4 for closed-loop systems except that the work
chamber is under vacuum during various stages of the cleaning process.

2.6.3 Exposure Assessment

2.6.3.1 Worker Activities

For closed-loop vapor degreasing, worker activities can include placing or removing parts from the
basket, as well as general equipment maintenance. Workers can be exposed to residual vapor as the door
to the degreaser chamber opens after the cleaning cycle is completed. The amount of time workers spend
in the degreaser area can vary greatly by site. One NIOSH report (NIOSH. 2002c) reported workers
spent 1.5 to 2 hours per shift at the degreaser and another NIOSH report (NIOSH. 2002a) indicating that

Page 77 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

workers spent over 90% of their day in the degreaser area. Similarly, addition of fresh solvent to the
degreasing machine can vary significantly with one site indicating 50 gallons of PCE per month were
added and another site indicating 10 to 20 gallons of PCE per year were added to the machine (NIQSH.

2002a. c).

2.6.3.2 Number of Potentially Exposed Workers

EPA estimated the number of workers and occupational non-users potentially exposed during use of
PCE in closed-loop vapor degreasing using the Draft ESD on the Use of Vapor Degreasers (OECD.
2017a). The ESD estimates seven workers and four ON Us per site (OECD. 2017a). EPA multiplied
these values by the number of sites estimated in Section 2.6.1. This resulted in approximately 97,000
workers and 56,000 ONUs using the number of sites estimated from the 95th percentile use-rate and
180,000 workers and 100,000 ONUs using the number of sites estimated from the 50th percentile use-
rate. Table 2-24 summarizes these results. Note: As described in Section 2.6.1, these are bounding
estimates and may overestimate actual number of workers.

Table 2-24. Estimated Number of Workers Potentially Exposed to Perchloroethylene During Use
in Closed-Loop Vapor Degreasing	

I se-Uale
Scenario

Number of
Sites

Exposed
Workers
per Site

Kxposed
Occupational
Non-l sers
per Site

Total
Exposed
Workers"

Total
Kxposed
Occupational
Non-l sers"

Total
Kxposed"

95th
Percentile

13,912

7

4

97,000

56,000

150,000

50th
Percentile

25,546

7

4

180,000

100,000

280,000

a Totals have been rounded to two significant figures. Totals may not add exactly due to rounding.

2.6.3.3 Occupational Exposure Results

EPA identified inhalation exposure monitoring data from NIOSH investigations at two sites using PCE
as a degreasing solvent in batch closed-loop vapor degreasers (NIQSH. 2002a. c). Due to the large
variety in shop types that may use PCE as a vapor degreasing solvent, it is unclear how representative
these data are of a "typical" shop. EPA does not have a model for estimating exposures from closed-loop
degreasers; therefore, the assessment is based on the identified monitoring data.

Table 2-25 summarizes the 8-hr TWA and 4-hr TWA monitoring data for the use of PCE in closed-loop
vapor degreasers. For workers, the 8-hr TWA high-end and central tendency are based on the 95th and
50th percentiles, respectively. Due to the limited data points for worker 4-hr TWAs, EPA used the
maximum and median as the high-end and central tendency, respectively. For ONUs, only two data
points were available; therefore, EPA presents two scenarios: 1) using the maximum as a "higher value";
and 2) using the midpoint as a "midpoint value". These scenarios are plausible, but EPA cannot
determine the statistical representativeness of the value.

The data were obtained from NIOSH In-Depth Survey Reports conducted as part of an interagency
agreement with OSHA to evaluate the extent of employee exposure to PCE at sites using it as a solvent
in vapor degreaser and to document engineering controls and work practices at the workplace affecting

Page 78 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

exposures (NIQSH. 2002a. c). Workers and ON Us were differentiated by the job titles provided in the
data, degreaser operators and assistant operators (or other similar job title assumed to be operating the
degreasing machine based on worker activities described in the studies) were assigned the worker
designation and non-operators were assigned the ONU designation.

Data from these sources cover exposures at a parts cleaning job site that had both a vacuum degreaser
and a cold cleaner and an aircraft manufacturer that had one vacuum degreaser and two OTVDs
(NIQSH. 2002a. c). Sample times at the two sites ranged from approximately 1.5 to 8 hours (NIQSH.
2002a. c). Where sample times were less than eight hours, EPA converted to an 8-hr TWA assuming
exposure outside the sample time was zero. Similarly, where sample times were less than four hours,
EPA converted to 4-hr TWAs assuming exposure outside the sample time was zero. It should be noted
that additional sources for degreasing were identified but were not used in EPA's analysis as they either:
1) did not specify the machine type in use; or 2) only provided a statistical summary of worker exposure
monitoring.

When comparing to monitoring data from OTVDs, the data show a decrease in worker exposure of
99.2% at the 95th percentile and 96.6% at the 50th percentile and a decrease in ONU exposure of 98.2%
at the 95th percentile and 89.2% at the 50th percentile. This is generally consistent with data in literature
which found that solvent purchases for closed-loop systems were reduced by 83% to over 98% as
compared to OTVDs and air emissions were reduced from 95% to over 99% as compared to OTVDs
(Dufkee. 2014; Newmoa. 2001).

Table 2-25. Summary of Worker Inhalation Exposure Monitoring Data for Closed-Loop Vapor
Degreasing	

Scenario

8-hr
TWA
(ppm)

AC
(ppm)

A IK

(ppm)

I.AIK
(ppm)

Nil m her
of Data
Points

4-hr
TWA
(ppm)

Nil in her
of Data
Points

Worker Monitoring Data

High-End

0.3

8.43E-02

5.78E-02

2.96E-02

13

8.56E-02

3

Central Tendency

7.22E-02

2.41E-02

1.65E-02

6.55E-03

1.97E-02

Occupational Non-User Monitoring Data

Higher Valuea

0.1

3.19E-02

2.19E-02

1.12E-02

2

No 4-hr data identified
for ONUs

Midpoint Valuea

6.54E-02

2.18E-02

1.49E-02

5.94E-03

AC = Acute Concentration; ADC = Average Daily Concentration; and LADC = Lifetime Average Daily Concentration.
Equations and parameters for calculation of the AC, ADC, and LADC are described in Appendix B.
a Due to only two data points identified, EPA presents two scenarios: 1) using the higher of the two values; and 2) using the
midpoint of the two values.

Source: fNIOSH. 2002a. c)

Page 79 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.
2.6.4 Water Release Assessment

2.6.4.1	Water Release Sources

Similar to OTVDs, the primary source of water releases from closed-loop systems is wastewater from
the water separator. However, unlike OTVDs, no water is expected to enter the system through
condensation (Durkee. 2014). The reason for this is that enclosed systems flush the work chamber with
water-free vapor (typically nitrogen gas) after the parts to be cleaned are added to the chamber and the
chamber is sealed but before the solvent enters (Durkee. 2014). Multiple flushes can be performed to
reduce the concentration of water to acceptable levels prior to solvent cleaning (Durkee. 2014).
Therefore, the primary source of water in closed-loop systems is from steam used to regenerate carbon
adsorbers (Durkee. JO I I; Kameesbere and Kameesbere. 2011; NIOSH. 2002a. b, c, d). Similar to
OTVDs, the water is removed in a gravity separator and sent for disposal (NIOSH. 2002a. b, c, d). As
indicated in the OTVD assessment, current disposal practices of the wastewater are unknown with the
latest available data from a U.S. EPA( 1982) report estimating 20% of water releases were direct
discharges to surface water and 80% of water releases were discharged indirectly to a POTW.

2.6.4.2	Water Release Assessment Results

EPA assesses water releases using TRI and DMR data. However, EPA cannot distinguish between
degreaser types in TRI and DMR data; therefore, a single set of water release for all degreasing
operations is presented in Section 2.5.4.2 for OTVDs.

2.7 Conveyorized Vapor Degreasing

2.7.1	Estimates of Number of Facilities

EPA estimated the number of sites using conveyorized degreasers using the draft ESD on Vapor
Degreasing (OEt O 1 j) using the same methodology as described for OTVDs in Section 2.5.1. The
ESD estimates a 50th percentile use-rate of 18,112 kg/site-yr and a 95th percentile use-rate of 26,060
kg/site-yr (OECD. 2017a). EPA calculated bounding estimates for number sites using the ESD use-rates
and the total 7% of the national production volume reported as used in metal degreasing by HSIA
(2008). This resulted in 395 sites using the 95th percentile use-rate and 568 sites using the 50th percentile
use-rate. Note: Bounding estimates may overestimate actual number of sites.

2.7.2	Process Description

In conveyorized degreasers, parts are cleaned in a continuous stream using an automated parts handling
system, typically a conveyor, to continuously loads parts into and through the vapor degreasing
equipment and the subsequent drying steps. Conveyorized degreasing systems are usually fully enclosed
except for the conveyor inlet and outlet portals. Conveyorized degreasers are likely used in shops where
there are a large number of parts being cleaned. There are seven major types of conveyorized
degreasers: monorail degreasers; cross-rod degreasers; vibra degreasers; ferris wheel degreasers; belt
degreasers; strip degreasers; and circuit board degreasers (\ v < < \ I 7).

• Monorail Degreasers - Monorail degreasing systems are typically used when parts are already
being transported throughout the manufacturing areas by a conveyor (	). They use

a straight-line conveyor to transport parts into and out of the cleaning zone. The parts may enter
one side and exit and the other or may make a 180° turn and exit through a tunnel parallel to the
entrance (	). Figure 2-5 illustrates a typical monorail degreaser (	J).

Page 80 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Con veyop^
Path

Monora11

Water
Jacket

Figure 2-5. Monorail Conveyorized Vapor Degreasing System (XLS. EPA, 1977)

• Cross-rod Degreasers - Cross-rod degreasing systems utilize two parallel chains connected by a
rod that support the parts throughout the cleaning process. The parts are usually loaded into
perforated baskets or cylinders and then transported through the machine by the chain support
system. The baskets and cylinders are typically manually loaded and unloaded ( '.S. EPA. 1977).
Cylinders are used for small parts or parts that need enhanced solvent drainage because of
crevices and cavities. The cylinders allow the parts to be tumbled during cleaning and drying and
thus increase cleaning and drying efficiency. Figure 2-6 illustrates a typical cross-rod degreaser
(U.S. EPA. 1977).

Page 81 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Figure 2-6. Cross-Rod Convevorized Vapor Degreasing System (J.S. EPA, 1977)

Vibra Degreasers - In vibra degreasing systems, parts are fed by conveyor through a chute that leads to
a pan flooded with solvent in the cleaning zone. The pan and the connected spiral elevator are
continuously vibrated throughout the process causing the parts to move from the pan and up a spiral
elevator to the exit chute. As the parts travel up the elevator, the solvent condenses and the parts are
dried before exiting the machine ( J.S. EPA. 1977). Figure 2-7 illustrates a typical vibra degreaser (U.S.
EPA 1977).

Page 82 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Workload Discharger Chute

Vibrating
Trough

Distillate Return
For Counter
flow Wash

Workload
Entry Chute

Steam Coils

Figure 2-7. Vibra Conveyorized Vapor Degreasing System (J.S. EPA, 1977)

Ferris wheel degreasers - Ferris wheel degreasing systems are generally the smallest of all the
conveyorized degreasers ( J.S. EPA. 1977). In these systems, parts are manually loaded into perforated
baskets or cylinders and then rotated vertically through the cleaning zone and back out. Figure 2-8
illustrates a typical ferris wheel degreaser ( J.S. EPA, 1977).

Page 83 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Work
Basket

Sear to tumble
baskets

Boiling
Chamber

Figure 2-8. Ferris Wheel Conveyorized Vapor Degreasing System ( J.S. EPA, 1977)

• Belt degreasing systems (similar to strip degreasers; see next bullet) are used when simple and
rapid loading and unloading of parts is desired ( IS. EPA, 1977). Parts are loaded onto a mesh
conveyor belt that transports them through the cleaning zone and out the other side. Figure 2-9
illustrates a typical belt or strip degreaser ( J.S. EPA. 1977).

Figure 2-9. Belt/Strip Conveyorized Vapor Degreasing System ( J.S. EPA, 1977)

Page 84 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

•	Strip degreasers - Strip degreasing systems operate similar to belt degreasers except that the belt
itself is being cleaned rather than parts being loaded onto the belt for cleaning. Figure 2-9
illustrates a typical belt or strip degreaser (	).

•	Circuit board cleaners - Circuit board degreasers use any of the conveyorized designs. However,
in circuit board degreasing, parts are cleaned in three different steps due to the manufacturing
processes involved in circuit board production (	7).

2.7.3 Exposure Assessment

2.7.3.1	Worker Activities

For conveyorized vapor degreasing, worker activities can include placing or removing parts from the
basket, as well as general equipment maintenance. Depending on the level of enclosure and specific
conveyor design, workers can be exposed to vapor emitted from the inlet and outlet of the conveyor
portal.

2.7.3.2	Number of Potentially Exposed Workers

EPA estimated the number of workers and occupational non-users potentially exposed during use of
PCE in conveyorized degreasing using the Draft ESD on the Use of Vapor Degreasers (	).

The ESD estimates seven workers and four ONUs per site (OECD. 2017a). EPA multiplied these values
by the number of sites estimated in Section 2.7.1. This resulted in approximately 2,800 workers and
1,600 ONUs using the number of sites estimated from the 95th percentile use-rate and 4,000 workers and
2,300 ONUs using the number of sites estimated from the 50th percentile use-rate. Table 2-26
summarizes these results. Note: As described in Section 2.7.1, these are bounding estimates and may
overestimate actual number of workers.

Table 2-26. Estimated Number of Workers Potentially Exposed to Perchloroethylene During Use
in Conveyorized Vapor Degreasing	

I se-Uale
Scenario

Number of
Sites

Exposed
Workers
per Site

Kxposed
Occupational
Non-l sers
per Site

Total
Exposed
Workers"

Total
Kxposed
Occupational
Non-l sers"

Total
Kxposed"

95th
Percentile

395

7

4

2,800

1,600

4,300

50th
Percentile

568

7

4

4,000

2,300

6,200

a Totals have been rounded to two significant figures. Totals may not add exactly due to rounding.

2.7.3.3 Occupational Exposure Results

EPA did not identify any inhalation exposure monitoring data related to the use of PCE in conveyorized
degreasing. Therefore, EPA assessed inhalation exposures during conveyorized degreasing using the
Conveyorized Degreasing Near-Field/Far-Field Inhalation Exposure Model.

A more detailed description of the modeling approach is provided 0. Figure 2-10 illustrates the near-
field/far-field model that can be applied to conveyorized vapor degreasing. As the figure shows, PCE

Page 85 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

vapors evaporate into the near-field (at evaporation rate G), resulting in near-field exposures to workers
at a concentration Cnf. The concentration is directly proportional to the evaporation rate of PCE, G, into
the near-field, whose volume is denoted by Vnf. The ventilation rate for the near-field zone (Qnf)
determines how quickly PCE dissipates into the far-field (i.e., the facility space surrounding the near-
field), resulting in occupational bystander exposures to PCE at a concentration Cff. Vff denotes the
volume of the far-field space into which the PCE dissipates out of the near-field. The ventilation rate for
the surroundings, denoted by Qff, determines how quickly PCE dissipates out of the surrounding space
and into the outdoor air. 0 outlines the equations uses for this model.

Far-Field	i

Figure 2-10. Schematic of the Conveyorized Degreasing Near-Field/Far-Field Inhalation Exposure

Model

0 presents the model parameters, parameter distributions, and assumptions for the PCE Conveyorized
Degreasing Near-Field/Far-Field Inhalation Exposure Model. To estimate the PCE vapor generation
rate, the model uses the annual emission rate and annual operating time from the single conveyorized
degreasing unit reported in the 2014 NEI. The calculated emission rate used in the model is 4.08 lb/unit-
hr and the operating hours used was 13 hr/day (U.S. EPA. 2016a). Because the vapor generation rate and
operating hours are based on a single data point and not a distribution of data it is unknown how
representative the model is of a "typical" conveyorized degreasing site.

EPA performed a Monte Carlo simulation with 100,000 iterations and the Latin Hypercube sampling
method in @Risk to calculate 8-hour TWA near-field and far-field exposure concentrations. Near-field
exposure represents exposure concentrations for workers who directly operate the vapor degreasing
equipment, whereas far-field exposure represents exposure concentrations for occupational non-users
(i.e., workers in the surrounding area who do not handle the degreasing equipment). The modeled 8-hr
TWA results and the values in Appendix B are used to calculate 24-hr AC, ADC, and LADC.

Page 86 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Table 2-27 presents a statistical summary of the exposure modeling results. Estimates of AC, ADC, and
LADC for use in assessing risk were made using the approach and equations described in Appendix B.
These exposure estimates represent modeled exposures for the workers and occupational non-users. For
workers, the 50th percentile exposure is 78.09 ppm 8-hr TWA, with a 95th percentile of 186 ppm 8-hr
TWA.

Table 2-27. Summary of Exposure Modeling Results for Use of Perchloroethylene in Conveyorized
Vapor Degreasing	

Scenario

8-hr TW A
(ppm)

AC
(ppm)

A IK

(ppm)

I.AIK
(ppm)

Worker Model Results

High-End

186

62

42

17

Central Tendency

78

26

18

6.7

Occupational Non-User Model Results

High-End

126

42

29

12

Central Tendency

41

14

9.3

3.5

AC = Acute Concentration; ADC = Average Daily Concentration; and LADC = Lifetime Average Daily Concentration.
Equations and parameters for calculation of the AC, ADC, and LADC are described in Appendix B.

2,7,4 Water Release Assessment

2.7.4.1	Water Release Sources

Similar to OTVDs, the primary source of water releases from conveyorized systems is expected to be
from wastewater from the water separator with the primary sources of water being: 1) Moisture in the
atmosphere that condenses into the solvent when exposed to the condensation coils on the system;
and/or 2) steam used to regenerate carbon adsorbers used to control solvent emissions (Durkee. 2014;
Kaneesbere and Kanegsberg. I I, \HQSH. 2002a. b, c, d). The current disposal practices of the
wastewater are unknown; however, a U.S. EPA( 1982) report estimated 20% of water releases from
metal cleaning (including batch systems, conveyorized systems, and vapor and cold systems) were direct
discharges to surface water and 80% of water releases were discharged indirectly to a POTW.

2.7.4.2	Water Release Assessment Results

EPA assesses water releases using TRI and DMR data. However, EPA cannot distinguish between
degreaser types in TRI and DMR data; therefore, a single set of water release for all degreasing
operations is presented in Section 2.4.4.2 for OTVDs.

2.8 Web Degreasing

2.8.1 Estimates of Number of Facilities

EPA estimated the number of sites using web degreasers using the draft ESD on Vapor Degreasing
((	) using the same methodology as described for OTVDs in Section 2.5.1. The ESD does

not present separate use-rates for web degreasers; therefore, EPA estimates the number of sites using the

Page 87 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

use-rates for conveyorized degreasers. The ESD estimates a 50th percentile use-rate of 18,112 kg/site-yr
and a 95th percentile use-rate of 26,060 kg/site-yr (OECD. 2017a). EPA calculated bounding estimates
for number sites using the ESD use-rates and the total 7% of the national production volume reported as
used in metal degreasing by HSIA (2008). This resulted in 395 sites using the 95th percentile use-rate
and 568 sites using the 50th percentile use-rate. Note: Bounding estimates may overestimate actual
number of sites.

2,8.2 Process Description

Continuous web cleaning machines (also called reel-to-reel systems) are a subset of conveyorized
degreasers but differ in that they are specifically designed for cleaning parts that are coiled or on spools
such as films, wires and metal strips (Kanegsberg and Kanegsberg. 2011; U.S. EPA 2006b). The part to
be cleaned is a continuous object uncoiled from one spool and fed onto rollers that transport it from end-
to-end through a cleaning solution, a drier, and then recoiled onto another spool (Kanegsberg and
Kanegsberg. 2011; U.S. EPA 2006b). They are generally classified as transporting the coiled part
through the cleaning machine at speeds greater than 11 feet per minute (U.S. EPA 2006b). Parts can
also be cut after exiting the cleaning machine (Kanegsberg and Kanegsberg. 2011; U.S. EPA. 2006b).
Figure 2-11 illustrates a typical continuous web cleaning machine.

Figure 2-11. Web Degreasing System

2.8.3 Exposure Assessment

2.8.3.1 Worker Activities

Worker activities for web degreasing are expected to be similar to other degreasing uses and can include
placing or removing parts from the degreasing machine, as well as general equipment maintenance.
Depending on the level of enclosure and specific design, workers can be exposed to vapor emitted from
the inlet and outlet of the conveyor portal.

Page 88 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

2.8.3.2 Number of Potentially Exposed Workers

EPA estimated the number of workers and occupational non-users potentially exposed during use of
PCE in web degreasing using the Draft ESD on the Use of Vapor Degreasers (OECD. 2017a). The ESD
estimates seven workers and four ON Us per site (OECD. 2017a). EPA multiplied these values by the
number of sites estimated in Section 2.8.1. This resulted in approximately 2,800 workers and 1,600
ONUs using the number of sites estimated from the 95th percentile use-rate and 4,000 workers and 2,300
ONUs using the number of sites estimated from the 50th percentile use-rate. Table 2-28 summarizes
these results. Note: As described in Section 2.8.1, these are bounding estimates and may overestimate
actual number of workers.

Table 2-28. Estimated Number of Workers Potentially Exposed to Perchloroethylene During Use
in Web Degreasing	

I se-Uale
Scenario

Number of
Sites

Exposed
Workers
per Site

Kxposed
Occupational
Non-l sers
per Site

Total
Exposed
Workers"

Total
Kxposed
Occupational
Non-l sers"

Total
Kxposed"

95th
Percentile

395

7

4

2,800

1,600

4,300

50th
Percentile

568

7

4

4,000

2,300

6,200

a Totals have been rounded to two significant figures. Totals may not add exactly due to rounding.

2.8.3.3 Occupational Exposure Results

EPA did not identify any inhalation exposure monitoring data related to the use of PCE in web
degreasing. Therefore, EPA assessed inhalation exposures during web degreasing using the Web
Degreasing Near-Field/Far-Field Inhalation Exposure Model.

A more detailed description of the modeling approach is provided 0. Figure 2-12 illustrates the near-
field/far-field model that can be applied to web degreasing. As the figure shows, PCE vapors evaporate
into the near-field (at evaporation rate G), resulting in near-field exposures to workers at a concentration
Cnf. The concentration is directly proportional to the evaporation rate of PCE, G, into the near-field,
whose volume is denoted by Vnf. The ventilation rate for the near-field zone (Qnf) determines how
quickly PCE dissipates into the far-field (i.e., the facility space surrounding the near-field), resulting in
occupational bystander exposures to PCE at a concentration Cff. Vff denotes the volume of the far-field
space into which the PCE dissipates out of the near-field. The ventilation rate for the surroundings,
denoted by Qff, determines how quickly PCE dissipates out of the surrounding space and into the
outdoor air. 0 outlines the equations uses for this model.

Page 89 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Far-Field

Near-Field

Figure 2-12. Schematic of the Web Degreasing Near-Field/Far-Field Inhalation Exposure Model

0 presents the model parameters, parameter distributions, and assumptions for the PCE Web Degreasing
Near-Field/Far-Field Inhalation Exposure Model. To estimate the PCE vapor generation rate, the model
developed a distribution from the reported annual emission rates and annual operating times reported in
the 2014 NEI (U.S. EPA. 2016a). Although the vapor generation rate is based on a distribution of the
emission rates from 10 web degreasing units, the data is only from web degreasers at two sites;
therefore, it is unknown how representative the model is of a "typical" site (U.S. EPA 2016a). A
summary of the unit emission distribution used in the model for PCE is provided in Table 2-29.

Table 2-29. Unit Emission Rates Used to Model Perchloroethylene Web Degreasing Systems

Unit Emissions
(lb PCE/unit-hr)

Fractional
Probability

0.0495

0.1000

0.0495

0.1000

0.0495

0.1000

0.0495

0.1000

0.0330

0.1000

0.0330

0.1000

0.0200

0.4000

Web degreasers are assumed to operate 24 hours per day, based on NEI data on the reported operating
hours for web degreasers using PCE (U.S. EPA 2016a). EPA performed a Monte Carlo simulation with

Page 90 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

100,000 iterations and the Latin Hypercube sampling method in @Risk to calculate 8-hour TWA near-
field and far-field exposure concentrations. Near-field exposure represents exposure concentrations for
workers who directly operate the vapor degreasing equipment, whereas far-field exposure represents
exposure concentrations for occupational non-users (i.e., workers in the surrounding area who do not
handle the degreasing equipment). The modeled 8-hr TWA results and the values in Appendix B are
used to calculate 24-hr AC, ADC, and LADC.

Table 2-30 presents a statistical summary of the exposure modeling results. Estimates of AC, ADC, and
LADC for use in assessing risk were made using the approach and equations described in Appendix B.
These exposure estimates represent modeled exposures for the workers and occupational non-users. For
workers, the 50th percentile exposure is 0.61 ppm 8-hr TWA, with a 95th percentile of 1.80 ppm 8-hr
TWA.

It showed be noted that results for web degreasing are two orders of magnitude lower than for the related
conveyorized degreasers. This is expected based on the emissions reported in the 2014 NEI as the
conveyorized data resulted in a unit emission of 4.06 lb/unit-hr which is two orders of magnitude greater
than the high-end emission rate for web degreasers (0.0495 lb/unit-hr) (	a). Because the

conveyorized emission rate is based on a single site and the web degreasing emission rate is based on
only two sites it is unclear if this difference in exposure is a function of the available data or an actual
function of the two systems. However, based on the types of parts being cleaned in the two systems,
EPA expects less dragout of solvent vapors (the primary route of exposure) in web degreasing machines
as the parts (e.g., film and metal sheets) are essentially two-dimensional objects compared to the three-
dimensional objects being carried through a conveyorized system. Therefore, these results are in-line
with EPA's expectations of the two systems.

Table 2-30. Summary of Exposure Modeling Results for Use of Perchloroethylene in Web
Degreasing	

Scenario

8-hr TW A
(ppm)

AC
(ppm)

A IK

(ppm)

I.AIK
(ppm)

Worker Model Results

High-End

1.8

0.6

0.4

0.2

Central Tendency

0.6

0.2

0.1

5.25E-02

Occupational Non-User Model Results

High-End

1.2

0.34

0.3

0.1

Central Tendency

0.3

0.1

7.30E-02

2.75E-02

AC = Acute Concentration; ADC = Average Daily Concentration; and LADC = Lifetime Average Daily Concentration.
Equations and parameters for calculation of the AC, ADC, and LADC are described in Appendix B.

2,8.4 Water Release Assessment

2.8.4.1 Water Release Sources

Similar to OTVDs, the primary source of water releases from web systems is expected to be from
wastewater from the water separator with the primary sources of water being: 1) Moisture in the

Page 91 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

atmosphere that condenses into the solvent when exposed to the condensation coils on the system;
and/or 2) steam used to regenerate carbon adsorbers used to control solvent emissions (Durkee. 2014;
Kaneesbere and Kaneesbere. I I, \1IQSH.. 2002a. b, c, d). The current disposal practices of the
wastewater are unknown; however, a U.S. EPA (1982) report estimated 20% of water releases from
metal cleaning (including batch systems, conveyorized systems, and vapor and cold systems) were direct
discharges to surface water and 80% of water releases were discharged indirectly to a POTW.

2.8.4.2 Water Release Assessment Results

EPA assesses water releases using TRI and DMR data. However, EPA cannot distinguish between
degreaser types in TRI and DMR data; therefore, a single set of water release for all degreasing
operations is presented in Section 2.5.4.2 for OTVDs.

2.9 Cold Cleaning

2.9.1	Estimates of Number of Facilities

To determine the number of sites that use PCE in cold cleaning, EPA considered 2014 NEI (
2016a). 2016 TRI (\ ^ \ :01 »!), and 2016 DMR (U.S. EPA. 2016b) data. Sites in TRI and DMR
do not differentiate between vapor degreasers and cold cleaning and are considered to be included in the
bounding estimates for the OTVD assessment and are not considered here. In the 2014 NEI, 17 sites
reported operation of a total of 34 cold cleaning machines (	). Therefore, EPA assesses

17 sites for cold cleaning. It should be noted that this number is expected to underestimate the total
number of sites using PCE in cold cleaners as NEI data does not include cold cleaner operations that are
classified as area sources. Area sources are reported at the county level and do not include site-specific
information. Therefore, any sites operating a cold cleaning machine that is classified as an area source
would not be included in the count of sites in the 2014 NEI. EPA does not have sufficient information to
estimate the number of area sources that may operator cold cleaning machines.

2.9.2	Process Description

Cold cleaners are non-boiling solvent degreasing units. Cold cleaning operations include spraying,
brushing, flushing and immersion. Figure 2-13 shows the design of a typical batch-loaded, maintenance
cold cleaner, where dirty parts are cleaned manually by spraying and then soaking in the tank. After
cleaning, the parts are either suspended over the tank to drain or are placed on an external rack that
routes the drained solvent back into the cleaner. Batch manufacturing cold cleaners could vary widely
but have two basic equipment designs: the simple spray sink and the dip tank. The dip tank design
typically provides better cleaning through immersion, and often involves an immersion tank equipped
with agitation (	1). Emissions from batch cold cleaning machines typically result from (1)

evaporation of the solvent from the solvent-to-air interface, (2) "carry out" of excess solvent on cleaned
parts and (3) evaporative losses of the solvent during filling and draining of the machine (

2006b).

Page 92 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

T

Air

Pump

t

Waste
Solvent

Cold Cleaner

Figure 2-13. Typical Batch-Loaded, Maintenance Cold Cleaner (U.S. EPA, 1981)

Emissions from cold in-line (conveyorized) cleaning machines result from the same mechanisms, but
with emission points only at the parts' entry and exit ports (U.S. EPA. 2006b).

2.9.3 Exposure Assessment

The general worker activities for cold cleaning include placing the parts that require cleaning into a
vessel. The vessel is usually something that will hold the parts but not the liquid solvent (i.e., a wire
basket). The vessel is then lowered into the machine, where the parts could be sprayed, and then
completely immersed in the solvent. After a short time, the vessel is removed from the solvent and
allowed to drip/air dry. Depending on the industry and/or company, these operations may be performed
manually (i.e., by hand) or mechanically. Sometimes parts require more extensive cleaning; in these
cases, additional operations are performed including directly spraying solvent on the part, agitation of
the solvent or parts, wipe cleaning and brushing (NIOSH. 2001; U.S. EPA 1997).

EPA estimated the number of workers and occupational non-users potentially exposed during use of
PCE in cold cleaners using Bureau of Labor Statistics' OES data (U.S. BLS. 2016) and the U.S. Census'
SUSB (U.S. Census Bureau. 2015) as well as the NAICS code reported by the site in the 2014 NEI (U.S.
EPA. 2016a). The method for estimating number of workers is detailed above in Section 1.4.4 and
Appendix A. These estimates were derived using industry- and occupation-specific employment data
from the BLS and U.S. Census. In the 2014 NEI, four sites reported NAICS code for which there was no
Census data available (U.S. EPA. 2016a). To estimate the number of workers/ONUs at these sites, EPA
referenced the Draft Emission Scenario Document (ESD) on the Use of Vapor Degreasers (OECD.

2.9.3.1 Worker Activities

2.9.3.2 Number of Potentially Exposed Workers

Page 93 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

2017a)14. Table 2-31 provides the results of the number of worker analysis. There are approximately 710
workers and 420 ONUs potentially exposed during use of PCE in cold cleaning.

Table 2-31. Estimated Number of Workers Potentially Exposed to Perchloroethylene During Use

in Cold Cleaning

NAICS Code

Nilmher of
Siles

Kxposed
Workers per
Site'1

Kxposed
()cciip;ilion;il
Non-1 sers
per Site11

Tolsil
Kxposed
Workers

Tolsil
Kxposed
Occupiilioiiiil
Non-l sers

Tolal
Kxposed

221112

1

6

8

6

8

13

322130

1

120

18

120

18

139

323111

1

2

1

2

1

3

325180

1

25

12

25

12

37

325211

1

27

12

27

12

40

327331

1

8

1

8

1

10

331110

1

53

18

53

18

71

332117

1

15

5

15

5

20

332812

2

7

2

14

3

18

332912

1

28

11

28

11

38

336414

1

372

314

372

314

686

339920

1

9

2

9

2

11

Subtotal for

Known
NAICS Data

13

52

31

681

405

1,086

No Data

4

7

4

27

17

44

Totalb

17

42

25

710

420

1,100

a Number of workers and occupational non-users per site are calculated by dividing the exposed number of workers or
occupational non-users by the number of establishments in the relevant NAICS codes. The workers/ONUs per site are then
multiplied by the number of sites within that NAICS to get the total exposed. The number of workers/ONUs per site is
rounded to the nearest integer.

b Totals have been rounded to two significant figures. Totals may not add exactly due to rounding.

2.9.3.3 Occupational Exposure Results

EPA identified inhalation exposure monitoring data from aNIOSH investigation at a single site using
PCE as a cold cleaning solvent. Due to the large variety in shop types that may use PCE as a cold
cleaning solvent, it is unclear how representative these data are of a "typical" shop. Therefore, EPA
supplemented the identified monitoring data using the Cold Cleaning Near-Field/Far-Field Inhalation

14 Although the ESD covers vapor degreasers not cold cleaners, the types of industries using cold cleaners are assumed to be
similar to those using vapor degreasers. Therefore, the number of workers/ONUs are assumed to be similar.

Page 94 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Exposure Model. The following subsections detail the results of EPA's occupational exposure
assessment for cold cleaning based on inhalation exposure monitoring data and modeling.

2.9.3.3.1 Inhalation Exposure Assessment Results Using Monitoring Data

Table 2-32 summarizes the 8-hr TWA and 4-hr TWA monitoring data for the use of PCE in cold
cleaners. For the 8-hr TWA, the 95th percentile and 50th percentile of the identified exposure data are
presented as the high-end and central tendency exposure values, respectively. Due to the limited number
of data points for the 4-hr TWA, the maximum and 50th percentile (median) of the data are presented as
the high-end and central tendency, respectively. The data were obtained from two sources: 1) aNIOSH
In-Depth Survey Report (NIOSH. 2002c); and 2) a study submitted to EPA by Vulcan Chemicals
(1994b) under TSCA. . The data only includes values for workers; data for ON Us were not identified.

The NIOSH In-Depth Survey Report was conducted as part of an interagency agreement with OSHA to
evaluate the extent of employee exposure to PCE at sites using it as a solvent in degreasers and to
document engineering controls and work practices at the workplace affecting exposures (NIOSH.
2002c). The cold cleaning data from this study were collected at a parts cleaning job site that had both a
vacuum degreaser and a cold cleaner (NIOSH. 2002c). Sample times for cold cleaning operations were
approximately 3 hours (NIOSH. 2002c). Where sample times were less than eight hours, EPA converted
to an 8-hr TWA assuming exposure outside the sample time was zero. Similarly, where sample times
were less than four hours, EPA converted to 4-hr TWAs assuming exposure outside the sample time was
zero.

The study submitted by Vulcan Chemicals was conducted to evaluate the feasibility of replacing 1,1,1-
trichloroethane (TCA) with two solvent blends of PCE in cold cleaning applications (Vulcan Chemicals.
1994b). The study was conducted at a site that manufactures and repairs small electric motors for the
aircraft industry (Vulcan Chemicals. 1994b). The study evaluated two blends, one containing 28% PCE
and one containing 50% PCE. It should be noted that the PCE can also be used as a pure cold cleaning
solvent (concentration >99%); therefore, results from this study may underestimate exposures from use
of pure PCE cold cleaning solvent. Sample times ranged from two to eight hours; where sample times
were less than eight hours, EPA converted to an 8-hr TWA assuming exposure outside the sample time
was zero (Vulcan Chemicals. 1994b). Similarly, where sample times were less than four hours, EPA
converted to 4-hr TWAs assuming exposure outside the sample time was zero.

In both studies EPA assumed the exposure concentrations outside of the sample times was zero which
may result in underestimates of exposure. However, both studies indicated that cold cleaning operations
are not expected to occur for the duration of the work-shift. Therefore, EPA expects the overall error
from this assumption to be minimal as the exposure potential when not performing cold cleaning
operations is expected to be minimal.

It should be noted that additional sources for solvent cleaning were identified but were not used in
EPA's analysis as they either: 1) did not specify between vapor and cold cleaning machines; or 2) only
provided a statistical summary of worker exposure monitoring.

Page 95 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Table 2-32. Summary of Worker Inhalation Exposure Monitoring Data for Use of

Scenario

8-hr
TWA
(ppm)

AC
(ppm)

ADC
(ppm)

LADC
(ppm)

Number
of Data
Points

4-hr

TWA
(ppm)

Number
of Data
Points

High-End

4.1

1.4

0.9

0.5

29

4.3

5

Central Tendency

1.4

0.5

0.3

0.1

2.9

AC = Acute Concentration; ADC = Average Daily Concentration; and LADC = Lifetime Average Daily Concentration.
Equations and parameters for calculation of the AC, ADC, and LADC are described in Appendix B.

Source: (NIOSH. 2002c: Vulcan Chemicals. 1994b)

2.9.3.3.2 Inhalation Exposure Assessment Results Using Modeling

A more detailed description of the modeling approach is provided 0. Figure 2-14 illustrates the near-
field/far-field model that can be applied to cold cleaning. As the figure shows, PCE vapors evaporate
into the near-field (at evaporation rate G), resulting in near-field exposures to workers at a concentration
Cnf. The concentration is directly proportional to the evaporation rate of PCE, G, into the near-field,
whose volume is denoted by Vnf. The ventilation rate for the near-field zone (Qnf) determines how
quickly PCE dissipates into the far-field (i.e., the facility space surrounding the near-field), resulting in
occupational bystander exposures to PCE at a concentration Cff. Vff denotes the volume of the far-field
space into which the PCE dissipates out of the near-field. The ventilation rate for the surroundings,
denoted by Qff, determines how quickly PCE dissipates out of the surrounding space and into the
outdoor air. 0 outlines the equations uses for this model.

Far-Field

0NF—H

Near-Field

NF

>Q,

NF

Figure 2-14. Schematic of the Cold Cleaning Near-Field/Far-Field Inhalation Exposure Model

0 presents the model parameters, parameter distributions, and assumptions for the PCE Cold Cleaning

Page 96 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Near-Field/Far-Field Inhalation Exposure Model. To estimate the PCE vapor generation rate, the model
developed a distribution from the reported annual emission rates and annual operating times reported in
the 2014 NEI. NEI records where the annual operating time was not reported were excluded from the
distribution. There were also four cold cleaning units at a single site for which the reported emission rate
was zero that were excluded from the distribution (	2016a). The site indicated the use of a

thermal oxidizer with 100% capture efficiency; therefore, the reported emissions are the emissions after
the control device (U.S. EPA. 2016a). Workers/ONlJs would be exposed to PCE that evaporates from
the cold cleaner prior to its capture by the control device. Therefore, only uncontrolled emissions are
used in the model. Uncontrolled emissions from the four cold cleaners from this site cannot be
determined, thus, emissions from these machines are not included in the model. A summary of the unit
emission distribution used in the model for PCE is provided in Table 2-33.

Table 2-33. Unit Emission Rales Used to Alodel Percldoroethylene Cold Cleaning

I nil K missions
(Ih PCT./iinil-hr)

l''raclional
Probability

0.12

0.04

0.08

0.04

0.02

0.04

1.17E-02

0.04

4.02E-03

0.04

8.03E-04

0.04

4.01E-04

0.04

2.67E-04

0.04

2.66E-04

0.04

2.30E-04

0.04

2.01E-04

0.08

1.34E-04

0.04

9.13E-05

0.19

2.77E-05

0.04

2.28E-05

0.04

2.17E-05

0.04

1.83E-05

0.04

1.49E-06

0.04

2.98E-07

0.08

1.13E-07

0.04

Page 97 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Cold cleaners are assumed to operate between 1 to 24 hours per day, based on NEI data on the reported
operating hours for cold cleaners using PCE (U.S. EPA. 2016a). A summary of the unit operating hours
distribution used in the model for PCE is provided in Table 2-34.

Table 2-34. Unit Operating Hours Used to Model Perchloroethylene Cold Cleaning

I nil K missions
(Ih PC E/iinil-hr)

I'raclional
Probability

24

0.70

8

0.26

1

0.04

EPA performed a Monte Carlo simulation with 100,000 iterations and the Latin Hypercube sampling
method in @Risk to calculate 8-hour TWA near-field and far-field exposure concentrations. Near-field
exposure represents exposure concentrations for workers who directly operate the vapor degreasing
equipment, whereas far-field exposure represents exposure concentrations for occupational non-users
(i.e., workers in the surrounding area who do not handle the cold cleaning equipment). The modeled 8-hr
TWA results and the values in Appendix B are used to calculate 24-hr AC, ADC, and LADC.

Table 2-35 presents a statistical summary of the exposure modeling results. Estimates of AC, ADC, and
LADC for use in assessing risk were made using the approach and equations described in Appendix B.
These exposure estimates represent modeled exposures for the workers and occupational non-users. For
workers, the 50th percentile exposure is 0.002 ppm 8-hr TWA, with a 95th percentile of 1.54 ppm 8-hr
TWA. It should be noted that the central tendency exposure estimate is three orders of magnitude less
than the high-end estimate, this is due to the large variation in unit emissions estimated from NEI with
three orders of magnitude separating the median and maximum emission rates from the 2014 NEI.

The high-end results of the model are on the same order of magnitude as the high-end and central
tendency found in the monitoring data. However, the central tendency estimated by the model is three
orders of magnitude lower. This may be due to the limited number of sites from which the monitoring
data were taken whereas the model is meant to capture a broader range of scenarios.

Page 98 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Table 2-35. Summary of Exposure Modeling Results for Use of Perchloroethylene in Cold
Cleaning	

Scenario

8-hr TWA
(ppm)

AC
(ppm)

A IK

(ppm)

I.AIK'
(ppm)

Worker Model Results

High-End

1.5

0.5

0.4

0.1

Central Tendency

2.40E-03

7.99E-04

5.71E-04

2.04E-04

Occupational Non-User Model Results

High-End

0.8

0.3

0.2

6.71E-02

Central Tendency

1.24E-03

4.14E-04

2.83E-04

1.05E-04

AC = Acute Concentration; ADC = Average Daily Concentration; and LADC = Lifetime Average Daily Concentration.
Equations and parameters for calculation of the AC, ADC, and LADC are described in Appendix B.

2.9.4 Water Release Assessment

2.9.4.1	Water Release Sources

Similar to OTVDs, the primary source of water releases from cold cleaners is expected to be from
wastewater from the water separator with the primary source of water expected to be from moisture in
the atmosphere that condenses into the solvent. Water may also enter vapor degreasers via steam used to
regenerate carbon adsorbers; however, it is unclear if carbon adsorbers would be used in conjunction
with cold cleaning equipment. The current disposal practices of the wastewater are unknown; however, a
U.S. EPA (1982) report estimated 20% of water releases from metal cleaning (including batch systems,
conveyorized systems, and vapor and cold systems) were direct discharges to surface water and 80% of
water releases were discharged indirectly to a POTW.

2.9.4.2	Water Release Assessment Results

EPA assesses water release using TRI and DMR data. However, EPA cannot distinguish between
degreasers and cold cleaners in TRI and DMR data; therefore, a single set of water release for all
degreasing and cold cleaning operations is presented in Section 2.5.4.2 for OTVDs.

2.10 Aerosol Degreasing and Aerosol Lubricants

2.10.1 Estimates of Number of Facilities

EPA estimated the number of facilities using aerosol degreasers and aerosol lubricants using data from
the U.S. Census" SUSB (lj_S ('casus Bureau. 2015). The method for estimating number of facilities is
detailed above in Section 1.4.1. These estimates were derived using industry-specific data from the U.S.
Census. Table 2-36 presents the NAICS industry sectors relevant to aerosol degreasing and aerosol
lubricants. For aerosol degreasing, EPA selected all NAICS codes associated with automotive,
electronic equipment, or other machinery/equipment repair. The list of NAICS codes includes the codes
for sporting goods stores and automobile dealers. The sporting goods stores NAICS code includes bike
shops, golf pro shops, and gun shops which may perform aerosol degreasing when performing repairs or
maintenance on the equipment. The automobile dealers NAICS code was included as many automobile
dealers also have repair shops associated with them. For both NAICS codes, EPA does not expect all of

Page 99 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

the establishments within the NAICS to perform aerosol degreasing; however, information was not
available to determine the percentage of sporting goods stores that fall within a category expected to
have repair or maintenance activities or percentage of automobile dealers with associated repair shops.
Therefore, inclusion of these NAICS codes may result in overestimate of sites using PCE-based aerosol
products.

Table 2-36. NAICS Codes for Aerosol Degreasing and Lubricants	

NAICS

Industry

811111

General Automotive Repair

811112

Automotive Exhaust System Repair

811113

Automotive Transmission Repair

811118

Other Automotive Mechanical and Electrical Repair and Maintenance

811121

Automotive Body, Paint, and Interior Repair and Maintenance

811122

Automotive Glass Replacement Shops

811191

Automotive Oil Change and Lubrication Shops

811198

All Other Automotive Repair and Maintenance

811211

Consumer Electronics Repair and Maintenance

811212

Computer and Office Machine Repair and Maintenance

811213

Communication Equipment Repair and Maintenance

811219

Other Electronic and Precision Equipment Repair and Maintenance

811310

Commercial and Industrial Machinery and Equipment (except Automotive and
Electronic) Repair and Maintenance

811411

Home and Garden Equipment Repair and Maintenance

811490

Other Personal and Household Goods Repair and Maintenance

451110

Sporting Goods Stores

441100

Automobile Dealers

There are 256,850 establishments among the industry sectors expected to use aerosol degreasers and/or
aerosol lubricants (U.S. Census Bureau. 2015). A 1997 manufacturer survey from CARB found that
approximately 44% of all aerosol brake cleaning products sold in California contained PCE and
approximately 37% of aerosol brake cleaning products available contained PCE (	MO).

Similarly, a CARB survey of automotive maintenance and repair facilities found, of the 73% of facilities
that use brake cleaning products to perform brake jobs, approximately 38% of these facilities used brake
cleaning products containing chlorinated chemicals (	300).

These data only relate to aerosol brake cleaning products used in the automotive repair industry;
however, aerosol degreasing and lubricant products may also be used in electronics repair, industrial

Page 100 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

equipment repair, home and garden equipment repair, or other similar industries. Market penetration
data for these industries were not identified; therefore, in lieu of other information, EPA assumed a
similar market penetration rate as for brake cleaning products. It is also possible the brake cleaning
product manufacturer and facility surveys completed by CARB underestimate the total number of
establishments that may use a PCE-containing product as some establishments may use an aerosol
lubricant containing PCE but not a brake cleaning product containing PCE. However, EPA expects the
potential error from this to be relatively small as only approximately 0.1% (317,000 lbs) of the total U.S
production volume of PCE is expected to be used in lubricants (	1). For comparison,

based on reported sales in 1996, CARB estimated approximately 2.7 million pounds of PCE were used
in brake cleaning products in California alone (	2000).

EPA assumed the average market penetration for PCE aerosol degreasers and lubricants was the average
of the low- and high-end values found by CARB, or 40.5% multiplied by the 73% of facilities that use
brake cleaning products, or 29.6% (40.5% x 73%=29.6%) (	00). This results in approximately

75,938 establishments using aerosol products containing PCE. It is unclear whether the number of
establishments using PCE-based aerosol degreasers has changed since 2000.

2.10.2	Process Description

EPA's Preliminary Information on Manufacturing, Processing, Distribution, Use, and Disposal for PCE
(Use Document) identified 170 aerosol-based products containing PCE (I v «« \ 201 ). CRC
Industries, Inc., a manufacturer of PCE-based degreasing products, indicates that PCE-based products
are used where flammability is a concern for worker and consumer safety as PCE has no flash point and
no upper and lower explosive limits (Rudnick. 2017). PCE-based aerosol products include degreasers
for applications such as brake cleaning, engine degreasing, electric motor cleaners, cable cleaners, coil
cleaners, and other metal product cleaning (Rudnick JO I ; 1 c. « i1 \ JO I .*). The weight percent of
PCE in these products ranges from 2.5 to 100% (	). Additional aerosol products include

penetrating lubricants and oils, high pressure non-melt red greases, white lithium greases, silicone
lubricants, chain and cable lubricants, vandal mark removers, mold cleaners, and weld anti-spatter
protectants (Rudnick. 201 ;l c. « ^ \ :0l ). The weight percent of PCE in these products ranges
from <1 to 100% (	) EPA expects significant overlap in the industry sectors that use

aerosol-based products; therefore, these uses are combined.

Aerosol degreasing is a process that uses an aerosolized solvent spray, typically applied from a
pressurized can, to remove residual contaminants from fabricated parts. A propellant is used to
aerosolize the formulation, allowing it to be sprayed onto substrates. Similarly, aerosol lubricant
products use an aerosolized spray to help free frozen parts by dissolving rust and leave behind a residue
to protect surfaces against rust and corrosion. Based on the safety data sheets for the identified products,
PCE-based aerosol products generally use carbon dioxide as the propellant, although a vandalism mark
and stain remover was identified that uses liquefied petroleum gas (LPG) as a propellant (i.e., propane
and butane).

2.10.3	Exposure Assessment

2.10.3.1 Worker Activities

Figure 2-15 illustrates the typical process of using aerosol degreasing to clean components in
commercial settings. One example of a commercial setting with aerosol degreasing operations is repair

Page 101 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

shops, where service items are cleaned to remove any contaminants that would otherwise compromise
the service item's operation. Internal components may be cleaned in place or removed from the service

# *
i «

# *

Figure 2-15. Overview of Aerosol Degreasing

Workers at these facilities are expected to be exposed through dermal contact with and inhalation of
mists during application of the aerosol product to the service item. ONUs include employees that work
at the facility but do not directly apply the aerosol product to the service item and are therefore expected
to have lower inhalation exposures and are not expected to have dermal exposures.

2.10.3.2 Number of Potentially Exposed Workers

EPA estimated the number of workers and occupational non-users potentially exposed to aerosol
degreasers and aerosol lubricants containing PCE using Bureau of Labor Statistics' OES data (U.S.
BLS. 2016) and the U.S. Census' SUSB (U.S. Census Bureau. 2015). The method for estimating number
of workers is detailed above in Section 1.4.4 and Appendix A. These estimates were derived using
industry- and occupation-specific employment data from the BLS and U.S. Census.

To estimate the number of workers and ONUs, EPA multiplied the total number of workers and ONUs
for each NAICS code identified in Table 2-36 (derived from the U.S. Census' SUSB and the Bureau of
Labor Statistics' OES data) by the market penetration of 29.6%. EPA then summed the workers and
ONUs for each identified NAICS code to estimate a total number of workers and ONUs exposed. Based
on this analysis, there are approximately 250,000 workers and 29,000 occupational non-users potentially
exposed to PCE as an aerosol degreasing solvent or aerosol lubricant (see Table 2-37) (U.S. BLS. 2016;
U.S. Census Bureau. 2015; CARB. 2000).

item, cleaned, and then re-installed once dry (U.S. EPA. 2014).

Q

Page 102 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Table 2-37. Estimated Number of Workers Potentially Exposed to Perchloroethylene During Use
of Aerosol Degreasers and Aerosol Lubricants	

Number of
Sites

Kxposed
Workers per
Site11

Kxposed
Occupational
\on-l sers per
Site11

Total Kxposcd
Workers1'

Total Kxposcd
Occupational
Non-l sers1'

Total Kxposcd1'

75,938

3

0.4

250,000

29,000

280,000

a Number of workers and occupational non-users per site are calculated by dividing the exposed number of workers or
occupational non-users by the number of establishments. The number of workers per site is rounded to the nearest integer.
The number of occupational non-users per site is shown as 0.4, as it rounds down to zero.
b Totals have been rounded to two significant figures. Totals may not add exactly due to rounding.

2.10.3.3 Occupational Exposure Results

EPA identified inhalation exposure monitoring data related to the use of PCE in aerosol degreasers for
brake servicing and vehicle maintenance. However, as described in Section 2.10.2, PCE is used in a
variety of other aerosol degreasing applications and other aerosol products for which EPA did not
identify any inhalation exposure monitoring data. Therefore, EPA supplemented the identified
monitoring data using the Brake Servicing Near-Field/Far-Field Inhalation Exposure Model. EPA used
the brake servicing model as a representative scenario for this condition of use as there was ample data
describing the brake servicing use and it is a significant use of PCE-based aerosol products. The
following subsections detail the results of EPA's occupational exposure assessment for aerosol
degreasing and aerosol lubricants based on inhalation exposure monitoring data and modeling.

2.10.3.3.1 Inhalation Exposure Assessment Results Using Monitoring Data

Table 2-38 summarizes 8-hr TWA PBZ monitoring data and 15-min TWA PBZ monitoring data for the
use of PCE-based aerosol products. The 95th percentile of the identified monitoring data is presented as
the high-end exposure and the 50th percentile is presented as the central tendency. The data were
obtained from three studies on the use of aerosol brake cleaners during commercial brake servicing and
from data provided to EPA from the Department of Defense (DoD) (Defense Occupational and
Environmental Health Readiness System - Industrial Hygiene. 2018; Cosgrove and Hygiene. 1994;
Vulcan Chemicab rs° ', r^C). One other study with monitoring data was identified; however, the
study states it was performed at two research and development locations with conditions expected to be
more severe than any "worst case scenario" at a normal brake shop (Vulcan Chemicals. 1994a).
Therefore, EPA did not include this data in the analysis. All identified aerosol exposure data are for
workers using the aerosol brake cleaner; data for ONUs were not identified.

One of the studies was performed by Health & Hygiene, Inc. (Cosgrove and Hygiene. 1994) who
collected the samples from five different automotive repair shops during routine cleaning of disc and
drum brakes. Workers at each site were supplied with an extension tube to create a concentrated liquid
stream of product when sprayed on the brake parts (Cosgrove and Hygiene. 1994). Other than the
supplied extensions, workers were instructed to use the aerosol product as they normally would
(Cosgrove and Hygiene. 1994). Health & Hygiene, Inc. (Cosgrove and Hygic H) stated that many
of the shops chose to have the garage doors opened for ventilation purposes. The authors noted that the
natural air current could either direct the mist away from the worker if their back was to the air flow or
towards the worker and potentially increasing exposure if they were facing the air flow (Cosgrove and
Hygiene. 1994).

Page 103 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

The other two studies were submitted to EPA under TSCA by Vulcan Chemicals (Vulcan Chemicals.
1993. 1992). The purpose of both studies was to evaluate exposures to aerosol products proposed as
alternatives to 1,1,1-trichloroethane (methylchloroform) brake cleaners (Vulcan Chemicals. 1993. 1992).
One study evaluated various formulations of aerosol degreasers containing 25% PCE, and the other
study evaluated one formulation containing 30% PCE, and one with 60% PCE. Based on data from
CARB (	X)) and modeling results (See Section 2.10.3.3.2 and Appendix H), PCE

concentration in brake cleaning products ranges from 20% to 99% with a median concentration of
78.4%). The monitoring data collected in these two studies may underestimate "typical" exposures as the
PCE concentration in the evaluated formulations were all below the median concentration.

The data provided by DoD did not explicitly state the use of aerosol degreasers (Defense Occupational
and Environmental Health Readiness System - Industrial Hygiene. 2018). Rather, the data indicated that
samples were collected during vehicle maintenance, which EPA assumed to be related to aerosol
degreasing activities.

The sample times for the identified monitoring data ranged from approximately four to nine hours.
Where sample times were less than eight hours, EPA converted to 8-hr TWAs assuming zero exposures
outside the sample time. It should be noted that approximately 15% of the 8-hr TWA data were
measured below the LOD. To estimate exposure concentrations for data below the LOD, EPA followed
the Guidelines for Statistical Analysis of Occupational Exposure Data (	b) as discussed

in Section 1.4.5.2. The geometric standard deviation for the data was above 3.0; therefore, EPA used the
to estimate the exposure value as specified in the guidelines (	lb).

Table 2-38. Summary of Worker Inhalation Exposure Monitoring Data for Aerosol Degreasing

Scenario

8-hr
TWA

(ppm)

AC
(ppm)

A IK

(ppm)

I.AIK
(ppm)

Nil m her
of Dala
Points

15-
Minule
TWA
(ppm)

Nil m her
of Dala
Points

High-End

7.8

2.6

1.8

0.9

130

123

67

Central Tendency

1.4

0.5

0.3

0.1

29

AC = Acute Concentration; ADC = Average Daily Concentration and LADC = Lifetime Average Daily Concentration.
Equations and parameters for calculation of the AC, ADC, and LADC are described in Appendix B.

Source: (Defense Occupational and Environmental Health Readiness System - Industrial Hygiene. 20.1.8; Cosgrove and
Hygiene. .1.994; Vulcan Chemicals. .1.993. .1.992')

2.10.3.3.2 Inhalation Exposure Assessment Results Using Modeling

A more detailed description of the modeling approach is provided in Appendix H. Figure 2-16 illustrates
the near-field/far-field for the aerosol degreasing scenario. As the figure shows, PCE in aerosolized
droplets immediately volatilizes into the near-field, resulting in worker exposures at a concentration Cnf.
The concentration is directly proportional to the amount of aerosol degreaser applied by the worker, who
is standing in the near-field-zone (i.e., the working zone). The volume of this zone is denoted by Vnf.
The ventilation rate for the near-field zone (Qnf) determines how quickly PCE dissipates into the far-
field (i.e., the facility space surrounding the near-field), resulting in occupational non-user exposures to
PCE at a concentration Cff. Vff denotes the volume of the far-field space into which the PCE dissipates

Page 104 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

out of the near-field. The ventilation rate for the surroundings, denoted by Qff, determines how quickly
PCE dissipates out of the surrounding space and into the outside air.

In this scenario, PCE mists enter the near-field in non-steady "bursts," where each burst results in a
sudden rise in the near-field concentration, followed by a more gradual rise in the far-field
concentration. The near-field and far-field concentrations then decay with time until the next burst
causes a new rise in near-field concentration.

Based on site data from maintenance and auto repair shops obtained by CARB (2000) for brake cleaning
activities, the model assumes a worker will perform 11 applications of the degreaser product per brake
job with five minutes between each application and that a worker may perform one to four brake jobs
per day each taking one hour to complete. EPA modeled two scenarios, one where the brake cleaning
jobs occurred back-to-back and one where braking cleaning jobs occurred one hour apart. Based on data
from CARB (2000). EPA assumes each brake job requires one 14.4-oz can of aerosol brake cleaner. The
model determines the application rate of PCE using the weight fraction of PCE in the aerosol product.
EPA uses uniform distribution of weight fractions for PCE based on facility data for the aerosol products
in use (CARB. 2000). It is uncertain whether the use rate and weight fractions for brake cleaning are
representative of other aerosol degreasing and lubricant applications. Model parameters and assumptions
for aerosol degreasing are presented in Appendix H.

Figure 2-16. Schematic of the Brake Servicing Near-Field/Far-Field Inhalation Exposure Model

EPA performed a Monte Carlo simulation with 100,000 iterations and the Latin hypercube sampling
method to model near-field and far-field exposure concentrations in the aerosol degreasing scenario. The
model calculates both 8-hr TWA exposure concentrations and maximum 1-hr TWA exposure
concentrations. Table 2-39 presents a statistical summary of the exposure modeling results.

For workers, the exposures are 5.48 ppm 8-hr TWA at the 50th percentile and 17.2 ppm 8-hr TWA at the
95th percentile. The model exposure levels at both the central tendency and high-end for workers are

Page 105 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

higher than that found in the monitoring data but are within one order of magnitude of the monitoring
data. This is not unexpected as the model is meant to capture a wider range of shop conditions than is
found in the monitoring data and the monitoring data includes data for sites using brake cleaning
formulations containing concentrations less than the median concentration (78.4%) used in the model.
For occupational non-users, the model exposures are 0.10 ppm 8-hr TWA at the 50th percentile and 0.75
ppm 8-hr TWA at the 95th percentile.

Table 2-39. Summary of Worker and Occupational Non-User Inhalation Exposure Modeling
Results for Aerosol Degreasing	

Scenario

8-hr TW A
(ppm)

AC
(ppm)

A IK

(ppm)

I.AIK
(ppm)

Maximum 1-
lir TWA
Kxposurcs
(ppm)

Workers Model Results

High-End

17

5.7

3.9

1.6

50

Central Tendency

5.5

1.8

1.3

0.5

17

Occupational Non-Users Model Results

High-End

0.7

0.2

0.2

7.00E-02

2.2

Central Tendency

0.1

3.35E-02

2.00E-02

1.00E-02

0.3

AC = Acute Concentration; ADC = Average Daily Concentration and LADC = Lifetime Average Daily Concentration.
Equations and parameters for calculation of the AC, ADC, and LADC are described in Appendix B.

2.10.4 Water Release Assessment

EPA does not expect releases of PCE to water from the use of aerosol products. Due to the volatility of
PCE, the majority of releases from the use of aerosol products will likely be to air as PCE evaporates
from the aerosolized mist and the substrate surface. There is a potential that any PCE that deposits on
shop floors during the application process could possibly end up in a floor drain (if the shop has one) or
could runoff outdoors if garage doors are open. However, EPA expects the potential release to water
from this to be minimal as there would be time for PCE to evaporate before entering one of these
pathways. This is consistent with estimates from the International Association for Soaps, Detergents and
Maintenance Products (AISE) SpERC for Wide Dispersive Use of Cleaning and Maintenance Products,
which estimates 100% of volatiles are released to air (International Association for Soaps Detergents
and Maintenance Products. 2012)5178607. EPA expects residuals in the aerosol containers to be
disposed of with shop trash that is either picked up by local waste management or by a waste handler
that disposes shop wastes as hazardous waste.

2.11 Dry Cleaning and Spot Cleaning

2.11.1 Estimates of Number of Facilities

EPA estimated the number of dry cleaning facilities using PCE as a solvent using data from the U.S.
Census' SUSB (U.S. Census Bureau. ). The method for estimating number of facilities is detailed
above in Section 1.4.1. These estimates were derived using industry-specific data from the U.S. Census.

Page 106 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

PCE may be used as a solvent at small commercial facilities categorized under the NAICS 812320,
Dry cleaning and Laundry Services (except Coin-Operated) and at large industrial dry cleaning facilities
categorized under 812332, Industrial Launderers (U.S. EPA. 2006a). EPA expects the majority of PCE
use to occur at small commercial facilities as large industrial launderers only account for approximately
2% of the total PCE consumption in the dry cleaning industry (U.S. EPA. 2006a).

There are 21,370 establishments in the United States under NAICS 812320, Drycleaning and Laundry
Services (U.S. Census Bureau. 2015). The Dry Cleaning and Laundry Institute (DL1) and the National
Cleaners Association (NCA) estimate approximately 60% of dry cleaning machines now use PCE (DLI
and NCA. 2017). In 1991, EPA estimated that 83% of all dry-cleaning facilities used PCE as solvent
0 v H' \ I In 2008, the Halogenated Solvents Industry Alliance (HSIA) estimated that 70% of
dry cleaners used PCE as a dry-cleaning solvent (HI	). Similarly, a 2010 profile of the dry-

cleaning industry conducted by King County, WA found that 69% of respondents (105 of the 152
respondents) used PCE in their primary machine (Whittaker and Johanson. 2011). Hence, there appears
to be a trend towards alternatives to PCE in dry cleaning. Therefore, EPA uses a market penetration of
60% to be consistent with current conditions reported by the dry-cleaning industry. Using this factor,
EPA estimated that approximately 12,822 small commercial dry cleaning establishments use PCE.

In 2006, EPA/OAQPS estimated 12 large industrial dry cleaners using PCE as a solvent (
2006a). Industrial dry cleaners include facilities that clean heavily stained articles such as work gloves,
uniforms, mechanics' overalls, mops, and shop rags, and facilities that operate as a central plant for a
chain of retail storefronts (U.S. EPA. 2006a). EPA did not identify more recent data for industrial dry
cleaners; therefore, EPA assumes 12 industrial dry cleaners.

2.11.2 Process Description

Dry cleaning machines are typically categorized into five generations of machines. The purchase of new
first generation (transfer machines) and second generation (dry-to-dry, vented machines) dry cleaning
machines were banned in the 1993 Perchloroethylene NESHAP for Dry Cleaning Facilities, and the
2006 Perchloroethylene NESHAP for Dry Cleaning Facilities banned the use of PCE in all first-
generation machines (	3a). The typical useful life of these machines is approximately 15

years; therefore, PCE is only expected to be used in third, fourth, and fifth generation machines
currently (	306a). Figure 2-17 provides an overview of the dry cleaning process.

Page 107 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Receiving Garments	Pre-Spotting	Dry Cleaning

Figure 2-17. Overview of Dry Cleaning Process

Third generation equipment, introduced in the late 1970s and early 1980s, are non-vented, dry-to-dry
machines with refrigerated condensers. These machines are essentially closed systems and are only open
to the atmosphere when the machine door is opened. In third generation machines, heated drying air is
recirculated back to the drying drum through a vapor recovery system (NIOSH. 1997b).

Fourth generation diy cleaning equipment are essentially third-generation machines with added
secondary vapor control. These machines "rely on both a refrigerated condenser and carbon adsorbent to
reduce the PCE concentration at the cylinder outlet below 300 ppm at the end of the dry cycle" and are
more effective at recovering solvent vapors. Fifth generation equipment have the same features as fourth
generation machines, but also have a monitor inside the machine drum and an interlocking system to
ensure that the concentration is below approximately 300 ppm before the loading door can be opened
(MOSI1. 19Q7b).

PCE is also found in products used to spot clean garments. On receiving a garment, dry cleaners inspect
for stains or spots they can remove as much of as possible before cleaning the garment in a dry cleaning
machine. As Figure 2-18 shows, spot cleaning occurs on a spotting board and can involve the use of a
spotting agent containing various solvents, such as PCE. The spotting agent can be applied from squeeze
bottles, hand-held spray bottles, or even from spray guns connected to pressurized tanks. Once applied,
the diy cleaner may come into further contact with the PCE if using a brush, spatula, pressurized air or
steam, or their fingers to scrape or flush away the stain (Young. 2012; NIOSFI. 1997a).

Page 108 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Figure 2-18. Overview of Use of Spot Cleaning at Dry Cleaners

2.11.3 Exposure Assessment

2.11.3.1	Worker Activities

Worker activities at dry cleaning shops can include:

•	Receiving garments and tagging garments for identification;

•	Inspecting and sorting garments by color, weight, finish;

•	Pre-treating any visible stain on the garment with a spotter, typically from a spray or squeeze
bottle;

•	Loading garments into the machine, running the wash cycle, and unloading the cleaned
garments;

•	Post-spotting any stain that was not already removed during the dry cleaning process; and

•	Pressing and finishing, after which the pressed garment is returned to an overhead rack and
wrapped in plastic for customer pickup (NIOSH. 1997a).

EPA expects worker exposure at dry cleaning facilities to primarily occur when workers are: 1)
unloading and loading garments from the machines; 2) performing manual stain removal (i.e., spot
cleaning); and 3) transferring solvent from a storage container to the machine. Workers can also be
exposed during maintenance activities, such as cleaning the machine lint trap, button trap and still,
changing solvent filters, and disposing hazardous wastes. However, these maintenance activities occur
on a much less frequent basis (NIOSH. 1997a).

ONUs at dry cleaning facilities are employees who are not expected to handle PCE, operate dry cleaning
machines, or perform spotting or finishing operations. They include cashiers, counter clerks and other
similar employees.

2.11.3.2	Number of Potentially Exposed Workers

EPA estimated the number of workers and occupational non-users potentially exposed to PCE at dry
cleaners using Bureau of Labor Statistics' OES data (U.S. BLS. 2016) and the U.S. Census' SUSB (U.S.
Census Bureau. 2015). The method for estimating number of workers is detailed above in Section 1.4.4
and Appendix A. These estimates were derived using industry- and occupation-specific employment
data from the BLS and U.S. Census.

Based on a market penetration of 60% for commercial facilities, assuming 12 industrial dry cleaners,
and data from the BLS and U.S. Census, there are approximately 44,000 workers and 14,000
occupational non-users potentially exposed to PCE at dry cleaning facilities (see Table 2-40).

Page 109 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Table 2-40. Estimated Number of Workers Potentially Exposed to Perchloroethylene During Dry
Cleaning	

NAICS
Code

Number of
Sites

Kxposed
Workers per
Site11

Exposed
Occupational
Non-l sers
per Site11

Total
Kxposed
Workers

Total Exposed
Occupational
Non-l sers

Total
Kxposed

812320

12,822

3

1

43,314

13,530

56,844

812332

12

25

3

304

32

336

Totalb

12,834

3

1

44,000

14,000

57,000

a Number of workers and occupational non-users per site are calculated by dividing the exposed number of workers or
occupational non-users by the number of establishments in the relevant NAICS codes. The workers/ONUs per site are then
multiplied by the number of sites within that NAICS to get the total exposed. The number of workers/ONUs per site is
rounded to the nearest integer.

b Totals have been rounded to two significant figures. Totals may not add exactly due to rounding.

2.11.3.3 Occupational Exposure Results

EPA identified inhalation exposure monitoring data related to the use of PCE as a dry cleaning solvent.
However, as estimated in Section 2.11.1, PCE is expected to be used in thousands of dry cleaning shops
throughout the U.S. and the monitoring data only captures a small fraction of those shops. Therefore,
EPA supplemented the identified monitoring data using the Dry Cleaning Multi-Zone Inhalation
Exposure Model to capture variation amongst dry cleaning shops that may not be captured in the
monitoring data. The following subsections detail the results of EPA's occupational exposure
assessment for dry cleaning based on inhalation exposure monitoring data and modeling.

2.11.3.3.1 Inhalation Exposure Assessment Results Using Monitoring Data

Table 2-42 summarizes the 8-hr TWA PBZ monitoring data for workers and ONUs at dry cleaners
obtained from OSHA facility inspections, NIOSH studies and data provided to EPA from DoD (Defense
Occupational and Environmental Health Readiness System - Industrial Hygiene JO IS; \ _ JO I ;

NIOSH 2000h, 1	3a, b, 1995). The data are divided into two categories: 1) statistics for data

collected after the promulgation of the 2006 Perchloroethylene NESHAP for Dry Cleaning Facilities;
and 2) data collected for fourth or fifth generation machines only. For workers, the 95th percentile is
presented as the high-end and the 50th percentile is presented as the central tendency. For the post-2006
NESHAP data, only a single data point was available for ONUs. Results based on a single value are
plausible, but EPA cannot determine the statistical representativeness of the value. For fourth and fifth
generation machines, there was only four ONU data points available; therefore, the maximum is
presented as the high-end and the median as the central tendency.

Approximately 28% of respondents to a 2003 survey of California dry cleaners indicated they used
fourth generation machines and approximately 61% of respondents to a 2010 survey of dry cleaners in
King County, WA reported using fourth or fifth generation machines (Whittaker and Joh an son. 2011;
California Air Resources Board. 2006). Therefore, EPA expects the industry to be trending towards
higher usage of fourth and fifth generation machines as compared to third generation machines. EPA
assumes the post-2006 NESHAP data are representative of the machine type mix provided in the King
County, WA survey (Whittaker and Johanson. 2011) and expects current exposures at dry cleaning

Page 110 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

shops to fall somewhere between the post-2006 exposure concentrations and the concentrations from
fourth and fifth generation machines only. Table 2-41 provides a summary of the trends in dry cleaning
machine types from several surveys.

Table 2-41. Summary of Survey Responses for Dry Cleaning Machine Generations



Percent of Survey Respondents or Projected !•"

icililies

Machine Type

2000 MSI A
Survey (

)

2003 ( A Survey
(California Air

)

2006 Projection
( )

2010 King Cunty
W A Survey

(

)

1st Generation

1.4%

1%

1%

1%

2nd Generation

3%

—

1%

6%

2nd Generation
Retrofitted

—

2%

—

3%

3rd Generation

65%

62%

37%

23%

4th Generation

31%

28%

61%

28%

5th Generation

—

—

—

33%

Other

—

2%

—

6%

Total

100%

95%

100%

100%

The data from OSHA were collected during compliance inspections at nine different facilities occurring
between 2012 and 2016 (OSHA.. 2017). Inspection data are compiled in an agency information system
(OIS) for internal use. Air sampling data records from inspections are entered into the OSHA Chemical
Exposure Health Database (CEHD) that can be accessed on the agency website
(https://www.osha.eov/openeov/healthsamples.html). The OSHA compliance data do not provide the
dry cleaning machine types; however, based on the dates of collection, EPA assumed that these data are
representative of the post-2006 mix of machine types as provided in the 2010 King County, WA Survey
(Whittaker a an son. 2011). Personal air samples for PCE were collected from approximately 2.5 to
8 hours (OS	). Where the air samples were collected for times less than eight hours, EPA

calculated the 8-hr TWAs by assuming exposure to be zero for the unsampled time. Seven samples
calculated 8-hr TWAs based on sample times less than six hours resulting in assumption of zero
exposure for over a quarter of the work shift and thus potentially underestimating actual exposure. The
OSHA air sampling data contain nine short-term PCE air measurements collected over 5 to 15 minutes
(OSHA. ). The short-term exposures are characterized as 15-minute TWAs in Table 2-42. Since
the OSHA data are from compliance inspections often as a result of worker complaints, they may not
necessarily be representative of PCE concentrations encountered in the typical commercial dry cleaning
establishment.



Page 111 of 316

2018)


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

do not provide the dry cleaning machine type; however, based on the dates of collection, EPA assumed
that these data are representative of the post-2006 mix of machine types as provided in the 2010 King
County, WA Survey (Whittaker and Joh an son. 2011). The sample times for the data ranged from 7 to
7.5 hours; where the air samples were collected for times less than eight hours, EPA calculated the 8-hr
TWAs by assuming exposure to be zero for the unsampled time (citation for DoD data). The DoD data
contains one sample that was reported at being less than the LOD (Defense Occupational and
Environmental Health Readiness System - Industrial Hygiene. 2018). To estimate exposure
concentrations for data below the LOD, EPA followed the Guidelines for Statistical Analysis of
Occupational Exposure Data (	Mb) as discussed in Section 1.4.5.2. The geometric standard

deviation for the data was above 3.0; therefore, EPA used the to estimate the exposure value as

specified in the guidelines (	).

The 1995 NIOSH (1995) report summarizes data collected as part of an industry study to evaluate
engineering controls to reduce exposure to PCE at dry cleaners. The 1995 report is part of a series of
studies completed by NIOSH that included data from several sites with first through fifth generation
machines. Only data from this report are included because the other reports either: 1) only included data
for first or second generation machines which are no longer in use; 2) only included area samples rather
that PBZ data; or 3) did not provide full-shift sample results. In this study, the 8-hr TWAs were
constructed from four samples taken for approximately 120 min each over a single day with total sample
times ranging from approximately five to eight hours (NIOSH. 1995). Where samples times were less
than eight hours, EPA converted to 8-hr TWAs assuming zero exposure outside the sample time.

The 1999 and 2000 NIOSH (NIOSH. 2000a. 1999a. b) reports are part of a series of studies conducted
as part of an industry study to evaluate exposures and control technologies for shops with fourth and
fifth generation machines. The studies evaluated exposures to pressers, machine operators, and other dry
cleaning employees at eight different shops (NIOSH. 2000a. 1999a. b). Sample times ranged from
approximately 3 to 10 hours with 18 of the 111 samples exceeding 8.5 hours. Where samples times were
less than eight hours, EPA converted to 8-hr TWAs assuming zero exposure outside the sample time and
where sample times exceeded 8 hours, EPA left the data "as is".

The 2000 NIOSH (2000b) report summarized data collected as part of an industry study to evaluate the
effectiveness of local exhaust ventilation (LEV) to reduce exposures in the shop. The study evaluated
exposures both pre- and post-installation of LEV at a shop utilizing third generation machines (NIOSH.
2000b). Sample times ranged from approximately four to seven hours; where samples times were less
than eight hours, EPA converted to 8-hr TWAs assuming zero exposure outside the sample time.

Additional PCE worker exposure monitoring data from dry cleaners were identified in other studies such
as Brodkin (1995). Gold (2008). Materna (1985). Ludwig (1983). and Solet (1990). However, these
studies are not used in EPA's assessment because they do not provide discrete data points. They are
presented here as a qualitative comparison to the results in Table 2-42.

EPA's systematic review process identified three studies conducted in the U.S. from 1985 to 1995 that
provided arithmetic means for workers ranging from 4.6 to 28.2 ppm and one study conducted in the
U.S. in 1983 that provided a geometric mean of 16 ppm (Brodkin et at.. 1995; Solet et at.. 1990;

Materna. 1985; Ludwig et at.. 1983). The low end of this range of means is generally consistent with

Page 112 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

EPA's 50th percentile for the Post-2006 NESHAP data in Table 2-42; however, the high-end of the
means is significantly greater than any of EPA's 50th percentiles. The difference in these studies from
the results in EPA's assessment may be a result of differences in machine types as the studies only
indicate that the exposures are from "dry-to-dry" machines without further specification of machine
type. Therefore, the results may include second generation machines that are no longer in use and may
result in higher exposures than current generation machines.

Gold (2008) completed a comprehensive literature review of studies evaluating PCE exposures from a
variety of uses in the U.S. The most recent data for dry cleaning referenced in the article were from
studies completed between the years 1990 and 2002 (Gold et at.. 2008). The overall arithmetic means
from these studies for samples where the sampling time was greater than six hours were 11 ppm for
machine operators of dry-to-dry machines, 6.8 ppm for spotters, 1.3 ppm for pressers/seamstresses, and
7.4 ppm for counter clerks (Gold et at.. 2008). These data are higher than the 50th percentiles in EPA's
analysis; however, Gold (2008) only divides operator data between "transfer" and "dry-to-dry"
machines without further specification of machine types and does not differentiate non-operator
(spotters, pressers, counter clerks) exposure data between machine types. Therefore, machine operator
data may include second generation machines and data for non-operators may include employees at sites
using first or second generation machines.

Page 113 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Table 2-42. Summary of Worker Inhalation Exposure Monitoring Data for Dry Cleaning

Category

Scenario

8-hr
TWA
(ppm)

AC
(ppm)

A IK

(ppm)

I.AIK
(ppm)

Nil m her
of Data
Points

15-
Minute
TWA
(ppm)

Nil m her
of Data
Points

Worker Monitoring Data

Post-2006
NESHAP
Statistics51

High-End

20

6.5

5.2

2.7

21

94

9

Central
Tendency

3.6

1.2

0.9

0.3

33

Fourth and Fifth
Generation
Statistics13

High-End

5.6

1.9

1.5

0.8

114

899

6

Central
Tendency

1.0

0.3

0.2

9.16E-02

48

Occupational Non-User Monitoring Data

Post-2006
NESHAP
Statistics51

High-
End0

0.3

0.1

9.29E-02

4.77E-02

ld

No 15-minute
TWA data
available for
ONUs

Central
Tendency

C

0.3

0.1

8.18E-02

3.25E-02

Fourth and Fifth
Generation
Statistics13

High-End

0.1

4.09E-02

3.28E-02

1.68E-02

4

Central
Tendency

1.40E-02

4.65E-03

3.29E-03

1.31E-03

AC = Acute Concentration; ADC = Average Daily Concentration; and LADC = Lifetime Average Daily Concentration.
Equations and parameters for calculation of the AC, ADC, and LADC are described in Appendix B.
a Post-2006 NESHAP data are air samples collected from OSHA inspections or DoD and, based on the date of collection,
EPA assumed to be representative of the post-2006 mix of machine types as provided in the 2010 King County, WA survey
(Whittaker and Johanson. 20.1.1').

b Fourth and fifth generation data include only data where EPA could clearly identify the machine type in the study as fourth
or fifth generation. It does not include OSHA data, which are representative of a mix of machine generations but for which
machine types for individual samples could not be determined.

0 Only one data point was available for this scenario. However, different parameters are used for calculating high-end and
central tendency ADC and LADC. Therefore, a high-end and central tendency are presented based on the single data point.
d The single ONU data point comes from a sample taken on an inspector at a dry cleaning site. EPA assumes exposures to the
inspector would be similar to that of an ONU as inspectors are not expected to handle the chemical or operator dry cleaning
machines.

Source: (Defense Occupational and Environmental Health Readiness System - Industrial Hygiene. 20.1.8: OSHA. 20.1.7:
NIOSH. 2000a. b, 1999a. b, 1995)

2.11.3.3.2 Inhalation Exposure Assessment Results Using Modeling

Because there are multiple activities with potential PCE exposure at a dry cleaner, a multi-zone
modeling approach is used to account for PCE vapor generation from multiple sources. This model
framework was peer reviewed as part of the 2016 draft 1 -BP Risk Assessment (	lOljSe)- The

model has been updated to address public and peer review comments. The model also reflects additional
information that became available since 2016; specifically, several model input parameters have been

Page 114 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

refined. Figure 2-19 illustrates this multi-zone approach, which considers the following worker
activities:

•	Spot cleaning of stains on both dirty and clean garments: On receiving a garment, dry
cleaners inspect for stains or spots they can remove as much of as possible before cleaning the
garment in a dry cleaning machine. Spot cleaning may also occur after dry cleaning if the stains
or spots were not adequately removed. Spot cleaning occurs on a spotting board and can involve
the use of a spotting agent containing various solvents, such as PCE. Workers are exposed to
PCE when applying it via squeeze bottles, hand-held spray bottles, or even from spray guns
connected to pressurized tanks. Once applied, the worker may come into further contact with the
PCE if using a brush, spatula, pressurized air or steam, or their fingers to scrape or flush away
the stain (Young. 2012; NIOS 7a). For modeling, EPA assumed the near-field is a
rectangular volume covering the body of a worker.

•	Unloading garments from dry cleaning machines: At the end of each dry cleaning cycle,
workers manually open the machine door to retrieve cleaned garments. During this activity,
workers are exposed to PCE vapors remaining in the dry cleaning machine cylinder. For
modeling, EPA assumed that the near-field consists of a hemispherical area surrounding the
machine door, and that the entire cylinder volume of air containing PCE exchanges with the
workplace air, resulting in a "spike" in PCE concentration in the near-field, Cd, during each
unloading event. This concentration is directly proportional to the amount of residual PCE in the
cylinder when the door is opened. The near-field concentration then decays with time until the
next unloading event occurs.

•	Finishing and pressing: The cleaned garments taken out of the cylinder after each dry clean
cycle contain residual solvents and are not completely dried (Von Grote. 2003). The residual
solvents are continuously emitted into the workplace during pressing and finishing, where
workers manually place the cleaned garments on the pressing machine to be steamed and ironed.
EPA assumed any residual solvent is entirely evaporated during pressing, resulting in an increase
in the near-field PCE concentration during this activity. Workers are exposed to PCE vapors
while standing in vicinity of the press machine. For modeling, EPA assumed the near-field is a
rectangular volume covering the body of a worker.

Page 115 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Dry Cleaning
Machine

Vr

Qd^-'

Cn

Far-field (background)





Qff

lff







Figure 2-19. Illustration of the Dry Cleaning Multi-Zone Inhalation Exposure Model

As the figure shows, PCE vapor is generated in each of the three near-fields, resulting in worker
exposures at concentrations Cs, Cd, and Cf. The volume of each zone is denoted by Vs, Vd, and Vf. The
ventilation rate for the near-field zone (Qs, Qd, Qf) determines how quickly PCE dissipates into the far-
field (i.e., the facility space surrounding the near-fields), resulting in occupational non-user exposures to
PCE at a concentration Cff. Vff denotes the volume of the far-field space into which the PCE dissipates
out of the near-field. The ventilation rate for the surroundings, denoted by Qff, determines how quickly
PCE dissipates out of the surrounding space and into the outside air. Appendix I summarizes the
parameters and equations for the multi-zone model.

It should be noted that EPA did not identify information to estimate the use rate of PCE in spot cleaners;
however, IRTA (2007) and ERG (2005) indicate that the use of PCE in spot cleaners is minimal.
Specifically, IRTA (2007) state that only 150 gal of PCE-based spotting agents are used annually in
California (compared to 42,000 gal of TCE-based spotting agents). ERG (2005) stated that many PCE
spotting agents are categorized as oily type paint removers (OTPR), but that the majority of OTPR
spotting agents contain no PCE. Therefore, EPA set the use rate of PCE spotting agents to zero causing
the spotting zone of the model to become part of the far-field with exposure concentrations equivalent to
Cff.

The dry cleaning industry is characterized by a large number of small businesses, many are family-
owned and operated. EPA assumed small dry cleaners operate up to 12 hours a day and up to six days a
week. In addition, EPA assumed each facility has a single machine. The assumption of a single machine
per facility is supported by a recent industry study conducted in King County, Washington, where 96
percent of 151 respondents reported having only one machine at their facility. Four reported having two

Page 116 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

machines, and two reported having three machines (Whittaker and Joh an son. 2011). Based on the survey
results, this assumption is presumably representative of the majority of small dry cleaning shops.

For PCE, the model accounts for variation in the machine generations operated at each facility.
Specifically, the model uses a distribution to estimate the machine generation and then based on the
sampled machine generation in each iteration selects a distribution of machine cylinder concentrations
and residual solvent in clothing. The distribution of machine types is based on the 2010 survey of dry
cleaners in King County, WA, which estimated 7% were first or second generation, 26% of machines
were third generation or retrofitted second generation15, 61% were fourth or fifth generation, and 6%
were "other" (e.g., hydrocarbon or CO: machines) (Whittaker and Joh an son. ). Due to the limited
information on other machine types, the model only considers two scenarios: 1) facilities operating third
generation machines; and 2) facilities operating a fourth or fifth generation machine16. This is not
expected to introduce significant error in the exposure estimates as EPA expects the use of first and
second generation machines to be eliminated with the industry trending towards increasing usage of
fourth and fifth generation machines (see discussion in Section 2.11.3.3.1). Therefore, the 7% for these
machine types were assumed to be replaced by fourth or fifth generation resulting in 26% third
generation machines and 68% fourth or fifth generation machines. EPA then re-normalized the
distribution to consider only PCE machines resulting in a distribution of 28% third generation machines
and 72%) fourth or fifth generation machines.

The model estimates exposures for three types of workers within the modeled dry cleaning facility: 1) a
worker who performs spot cleaning; 2) a worker who unloads the dry cleaning machine and finishes and
presses the garments; and 3) an occupational non-user. However, the model for PCE assumes facilities
do not use PCE spot cleaning agents (discussed above in this section); therefore, spot cleaners are
exposed at concentrations equivalent to occupational non-users and are not assessed separately. Each
worker type is described in further detail below. EPA assumed each worker activity is performed over
the full 12-hour operating day.

•	EPA assumed spot cleaning occurs for a duration varying from two to five hours in the middle of
the twelve-hour work day. For PCE, the spot cleaning use rate is zero, so the worker is exposed
at the far-field concentration for the entire day. Spot cleaning can be performed for both dry
cleaned loads and for laundered loads.

•	EPA assumed a separate worker unloads the dry cleaning machine and finishes and presses the
garments. After each load, EPA assumed this worker spends five minutes unloading the machine,
during which he or she is exposed at the machine near-field concentration. After unloading, the
worker spends five minutes in the finishing near-field to prepare the garments. Then, the worker
spends another 20 minutes finishing and pressing the cleaned garments. During this 20-minute
period of finishing and pressing, the residual PCE solvent is off-gassed into the finishing near-
field. The amount of residual PCE solvent is estimated using measured data presented in von

15	For modeling purposes, retrofitted second generation machines are assumed to be equivalent to third generation machines.

16	The model treats fourth and fifth generation machines as equivalent as both are expected to reduce machine cylinder
concentrations to approximately 300 ppm (NIOSBL 1997b). The primary difference being that fifth generation machines have
an interlock preventing the machine door from being opened until the concentration is below 300 ppm whereas fourth
generation machines do not.

Page 117 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Grote (2003). These unloading and finishing activities are assumed to occur at regular intervals
throughout the twelve-hour day. The frequency of unloading and finishing depends on the
number of loads dry cleaned each day, which varies from 1 to 14, where 14 was the maximum
number of loads observed in the NIOSH (2010) and Blando (2010) studies. When this worker is
not unloading the dry cleaning machine or finishing and pressing garments, the worker is
exposed at the far-field concentration.

• EPA assumed one occupational non-user is exposed at the far-field concentration for twelve
hours a day. The occupational non-user could be the cashier, tailor, or launderer, who works at
the facility but does not perform dry cleaning activities.

Table 2-43 presents the Monte Carlo results with the Latin hypercube sampling method and 10,000
iterations. Statistics of the 12-hr TWA exposures (95th and 50th percentiles) are then calculated at the end
of the simulation after all iterations have completed. The AC, ADC, and LADC calculations are
integrated into the Monte Carlo simulation, such that the exposure frequency matches the model input
values for each iteration.

When comparing to the post-2006 NESHAP monitoring data results for workers, the model high-end is
higher than the monitoring data. This is likely because the model is meant to capture a wider range of
conditions than is likely captured in the monitoring data. The model central tendency for workers is
slightly less than half the central tendency for the post-2006 NESHAP monitoring data. This may be due
to the fact the majority of the post-2006 NESHAP data are from OSHA compliance inspections that are
often performed as a result of worker complaints and, therefore, may not necessarily be representative of
PCE concentrations encountered in the typical commercial dry cleaning establishment. Additionally, the
assumption that post-2006 NESHAP data is representative of the 2010 King County, WA survey results
may be inaccurate, and the data could actually represent sites with a higher frequency of third generation
machines, resulting in higher exposures. However, model results and monitoring data for the post-2006
NESHAP are within the same order of magnitude.

When comparing the model results to the fourth/fifth generation monitoring data results for workers, the
model high-end and central tendency are both an order of magnitude greater than the monitoring data.
This is expected as the model captures exposures from facilities with third and fourth/fifth generation
machines.

Page 118 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Table 2-43. Summary of Worker and Occupational Non-User Inhalation Exposure Modeling
Results for Dry Cleaning	

Scenario

12-hr TW A
(ppm)

AC
(ppm)

A IK

(ppm)

I.AIK
(ppm)

Workers Model Results

High-End

30

15

10

4.1

Central Tendency

1.4

0.7

0.5

0.2

Occupational Non-Users/Spot Cleaners Model Results

High-End

1.5

0.8

0.6

0.2

Central Tendency

0.1

5.43E-02

3.83E-02

1.44E-02

2.11.4 Water Release Assessment

2.11.4.1 Water Release Sources

The primary source of water releases from dry cleaning machines is wastewater from the water
separator. Water may be added to the system to remove water soluble impurities from the solvent or dry
sludge at the end of distillation (Ecb. 2005). It may also be present in the garments being dry cleaned
(Ecb. 2005). The refrigerated condenser used in third, fourth, and fifth generation machines condenses
both the PCE and any water in the air stream from the dry cleaning machine (	98). The

liquid stream is then fed to the water separator where the water is removed from the stream as waste and
PCE is recycled back to the system for reuse (	). Fourth and fifth generation machines

generate additional wastewater from the use of steam to regenerate carbon adsorbers used as secondary
vapor controls (	>8).

How facilities handle their produced separator water may be subject to state regulations. Under RCRA
regulations, produced water that contains at least 0.7 mg/L of PCE is a hazardous waste based on its
toxicity characteristic (U.S. EPA. 2019h). Various states may have regulations on permissible disposal
and treatment options for produced separator water containing PCE. For example, the Oregon
Department of Environmental Quality (DEQ) prohibits dry cleaners from disposing of their separator
water in the following manners, even if the separator water does not meet or exceed 0.7 mg/L PCE:
discharging to sewer, septic system, or state waters; using in a boiler; pouring on the ground; or
disposing in municipal trash (Oregon DEQ. 2018). The Oregon DEQ only allows the following
treatment and disposal methods for separator water: drumming the wastewater and shipping it offsite to
a hazardous waste facility; hard piping the separator water from the dry cleaning machine to an onsite
treatment unit; and manually transferring the separator water from the dry cleaning machine to an onsite
treatment unit (Oregon DEQ. 2018). Allowable onsite treatment units include secondary separators and
initial and secondary filters. Separator water treated to reduce PCE levels below 0.7 mg/L may be
discharged via evaporation to air (Oregon DEQ. 2018).

Best management practices published by Massachusetts also prohibits the discharge to sewer of
separator water that is hazardous waste but does allow the evaporation to air of the separator water as

Page 119 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

well as the drumming of separator water as hazardous waste for offsite disposal via a licensed treatment,
storage, and disposal facility (TSDF) (Massachusetts D.	).

Additional water releases of PCE may occur at sites using wet cleaning and hydrocarbon machines
from:

•	Residual PCE in clothing previously cleaned in a PCE dry cleaning machine and then washed in
the water or hydrocarbon machine;

•	Cross contamination at facilities that have both a water or hydrocarbon machine and a PCE
machine; and

•	PCE in spot cleaners used to pre-spot garments prior to cleaning in a water or hydrocarbon
machine (	>07; Morris and Wolf. 2005).

The extent to which these releases occur is unknown and therefore not included in this release
assessment. However, one study found up to 5.3 mg/L PCE in wet cleaning machine wash water and 1.1
mg/L PCE in wet cleaning machine rinse water at sites using both water machines and PCE machines;
0.48 mg/L PCE in the wet cleaning machine wash water from a site using water machines and PCE as a
spot cleaner; and up to 30 mg/L PCE in separator water from sites using hydrocarbon machines (the
source of the PCE at each of the studied facilities using hydrocarbon machines is not explicitly stated in
the study, but the authors state the same general sources as listed above) ( 2007; Morris and Wolf.
2005). The representativeness of these values for similar garment cleaning sites is unknown. EPA
expects spent water from wet cleaning machines is primarily discharged to sewer.

Given the variability in state regulations regarding the disposal practices of separator water and the
potential for PCE-contaminated wet cleaning machine water to be discharged to sewer, which is not
included in EPA's release assessment, EPA assesses the modeled produced separator water as
discharged to sewer (POTW). EPA expects this assumption will overestimate PCE releases to water
from dry cleaning machine separator water, but the release assessment underestimates PCE releases to
water from wet cleaning machines as these releases are not included. The overall directional bias of the
release assessment, accounting for both the overestimate and underestimate, is not known.

2.11.4.2 Water Release Assessment Results

To assess water releases from dry cleaners, EPA used data from the 2016 DMR (	).

EPA reviewed the reported SIC codes for each point source and assigned each point source to one of the
PCE conditions of use. However, the sites in the pollutant loading tool are not expected to contain all of
the dry cleaning sites in the U.S.; therefore, EPA supplemented the DMR data with modeled releases.
EPA considered industrial launderers and commercial dry cleaners separately for purposes of assessing
water releases.

In the 2016 DMR (U.S. EPA. 2016b). EPA identified eight sites that are likely industrial launderers
based on the reported SIC codes of 7212, 7216, and 721817. Based on the 2006 Dry Cleaning NESHAP
Economic Impact Analysis (	006a). there are an additional four industrial launderers that are

not in the 2016 DMR. These four sites may not be in DMR because they may have no water discharges
or because they discharge to sewer rather than surface water (sewer discharges not reported in DMR).

17 Seven of the eight sites reported one of these SIC codes, the other site did not report a SIC; rather, it was determined to be
an industrial launderer after review of the company's website.

Page 120 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Of the eight sites in the 2016 DMR, only two were identified as having non-zero discharges (IJ..S J_TA_
2016b). The results from the sites with non-zero releases are presented in Table 2-44. To calculate the
daily release, EPA averaged the annual release over the operating days of 289 days/yr (high-end release)
and 307 days/yr (central tendency release). The operating days are based on the distribution of operating
days used in the model discussed below, with the 50th percentile value being used to calculate the high-
end daily release and the 95th percentile value being used to calculate the central tendency daily release.

Table 2-44. Reported Wastewater Discharges of Perchloroethylene for Industrial Launderers in

2016 DMR

Site

Annual
Release
per Site
(kg/site-
vr)

Iligli-Knd
Release
Operating

Days
(davs/yr)

Central
Tendency
Release
Operating

Days
(days/yr)

Iligli-Knd
Release
(kg/site-
day)

Central
Tendency
Release
(kg/site-
day)

npdks

Code

Release
Media/
Treatment
Kacility
Type

Boise State
University,
Boise, ID

5.94E-02

289

307

2.05E-04

1.93E-04

IDG911006

Surface
Water

Unifirst,

Williamstown,

VT

1.37E-02

289

307

4.73E-05

4.45E-05

VT0000850

Surface
Water

Source: (U.S. EPA. 2016b)

In the 2016 DMR (U.S. EPA. 2016b). EPA identified four sites with non-zero discharges that are likely
commercial dry cleaners either based on reported SIC codes or review of company information available
online. It is unclear whether these sites are representative of typical commercial dry cleaning sites;
therefore, EPA used the Solvent Release in Water Discharge from Dry Cleaning Machines Model to
estimate releases from commercial dry cleaners.

The amount of wastewater generated from each site is dependent on the type of machine, the number of
dry cleaning machines at the site, the number of loads of garments cleaned per machine per day, the
weight of garments cleaned in each load, and the number of days per year the machine operates. To
account for variability in these parameters, EPA performed a Monte Carlo simulation with 100,000
iterations and the Latin Hypercube sampling method to model water releases from dry cleaning sites
using the Solvent Release in Water Discharge from Dry Cleaning Machines Model. A more detailed
description of the modeling approach count of 13 sites with sur is provided in Appendix J.

Based on data from a CARB survey of dry cleaners performed in 2003, the model assumes that the
volume of water released per pound of clothes cleaned is 0.0032 gal water/lb clothes for third generation
machines and 0.0037 gal water/lb clothes for fourth and fifth generation machines (Califora
Resources Board. 2006). The model uses the same machine type distribution as described for the Dry
Cleaning Multi-Zone Inhalation Exposure Model discussed in Section 2.11.3.3.2.

Page 121 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

The model assumes the load size ranges from 7 to 150 lb based on the King County survey (Whittaker
and Johanson. 2011) and the number of loads per day ranges from 1 to 14 based on observations from
NIOSH (2010) and Blando (2010). Based on survey data from CARB (2006) and Whittaker (2011). the
model assumes dry cleaning shops have between one and three machines. The model assumes that the
concentration of PCE in the wastewater stream is equal to the solubility of PCE in water, 206 mg/L. The
model results for both daily and annual releases per site and across all sites are presented in Table 2-45.
It should be noted that the distribution of release days/yr is taking into account when the annual release
is calculated in each iteration of the model; therefore, an exact value corresponding to the high-end and
central tendency annual release is not available. The values presented in the table are back-calculated by
dividing the estimated annual release by the daily release and rounding to the nearest whole number,
they are not necessarily representative of the 50th or 95th percentile operating days.

Table 2-45. Model Results for Perchloroethvlene Discharges to POTW from Dry Cleaning Sites

Scenario

Daily Release

per Site
(kg/site-day)

Annual Release
per Site
(kg/site-yr)

Annual
Release for All

Sites11
(kg/yr-all sites)

Release Days
(days/yr)

Release
Media/
Treatment
l-'acility Type

High-End

1.71E-03

0.5

6,310

288

POTW

Central Tendency

5.59E-04

0.2

2,057

287

POTW

a Releases for all sites calculated by multiplying per site releases by total number of commercial sites (12,822 commercial
sites).

For comparison results from the four commercial sites in the 2016 DMR are provided in Table 2-46.
Except for one site that reported an annual discharge of 2.76 kg, these discharges are comparable to the
annual PCE discharges to sewer estimated by the model.

Table 2-46. Summary of Direct Discharge Data for Commercial Dry Cleaning Reporters in the
2016 DMR

I'acility

2016 Reported Annual PCK
Discharge to Surface Water
(kg/site-year)

Chase Tower, Dallas,
TX

2.8

San Jacinto Tower,
Dallas, TX

3.05E-03

The Martin, Las
Vegas, NV

3.77E-02

The Stirling Club,
Las Vegas, NV

0.2

Average

0.7

Median

0.1

Source: (U.S. EPA. 20.1.6b')

Page 122 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

2.12 Adhesives, Sealants, Paints, and Coatings

2.12.1 Estimates of Number of Facilities

To determine the number of sites that use PCE-containing adhesives, sealants, paints, and coatings, EPA
considered 2014 NEI (U.S. EPA. 2016a). 2016TRI (	1), and 2016 DMR (

2016b) data. Sites in TRI and DMR do not differentiate between conditions of use; therefore, they have
been considered under other scenarios (e.g., OTVDs, processing aids, etc.) and are not considered again
here. In the 2014 NEI, EPA identified 60 sites reporting adhesive/sealant or paint/coating uses
(including one site reporting paint stripping) with 84 reports of spray applications, 4 reports of roll
coating applications, 5 reports of dip coating applications, 1 report of paint stripping, and 60 reports of
unspecified applications methods (U.	2016a)18. Of the 60 sites, 46 were identified as

paints/coatings uses, 11 were identified as adhesive/sealant uses, and 3 were identified as having both
coating and adhesive uses (	). It should be noted that this number may underestimate the

total number of sites using PCE-containing adhesives, sealants, paints, and coatings as NEI data only
covers specific industries which may not capture the entirety of industries using these products.
Additionally, NEI does not include operations that are classified as area sources because area sources are
reported at the county level and do not include site-specific information.

EPA did not identify data to determine the volume of PCE used in adhesives and coatings, but based on
market data, EPA expects no more than 3 to 10% of the national PCE production volume is used for
"miscellaneous" uses which includes coatings and adhesives (b JO I I; H\ -008) . EPA used 3%
of the national production volume, 4,412,190 kg/yr (National production volume = 324,240,744 lb x 3%
x 0.45 kg/lb = 4,412,190 kg/yr) as a bounding estimate for the volume of PCE used in coatings and
adhesives. EPA used 3% rather than 10% because the 3% value is more recent, and the miscellaneous
uses are expected to encompass other uses beyond adhesives and coatings and the 3% will limit the
overestimation from using a bounding estimate.

To estimate per site use rates of PCE, EPA averaged the 4,412,190 kg/yr for coatings and adhesives over
the total number of application lines in the 2014 NEI, resulting in an average of 28,651 kg
PCE/application line (4,412,190 kg/yr / (84 spray applications + 4 roll coating application lines + 5 dip
application lines + 1 paint stripping + 60 unspecified applications) = 4,412,190 kg/yr / 154 application
lines = 28,651 kg/yr). EPA then multiplied the average use rate per application line by the number of
application lines at the facility to get an annual use rate. It should be noted that these bounding estimates
likely overestimate the actual volume of PCE used in coatings and adhesives and the average annual use
rate at each facility as the 3% of the national production volume used to estimate these values is for
"miscellaneous uses" that go beyond just coatings and adhesives and there may be additional sites using
PCE-based coatings and adhesives.

Table 2-47 provides the number of application lines and total PCE use rate assessed for each site in NEI.
It should be noted that the use-rates in Table 2-47 are the use rates of PCE, not the coating or adhesive
product. The concentration of PCE in the products at each site is unknown; however, adhesive and
coating products identified in the Preliminary Information on Manufacturing, Processing, Distribution,
Use, and Disposal: Tetrachloroethylene (Perchloroethylene) (U.S. EPA. 2017c) had concentrations
reported on SDS's ranging from 0.1 to <100% for adhesives and 8.79 to <100% for coatings. The

18 Number of application methods is greater than the number of sites due to sites reporting multiple application methods.

Page 123 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

OECD ESD on the Use of Adhesives (OECD. 2015) does not have PCE-specific concentrations but
estimates organic solvent concentrations from 60 to 75% in adhesives. Similarly, the OECD ESD on
Coating Industry (Paints, Laquers and Varnishes) (OECD. 2009b) estimates organic solvent
concentrations from 30 to 80% in coatings.

Table 2-47. Perchloroethylene Use Rate at Coating and Adhesive Application Sites

She

Total Adhesive
Lines

Total Coaling
Lines

Total Lines

P( I] I se-Uale
(kg P( K/sile-
M)

3M -R&D Facility -
Maplewood Bldg 201,
Maplewood, MN

0

1

1

28,651

3P Processing, Wichita, KS

0

1

1

28,651

Accellent/Collegeville,
Collegeville, PA

0

1

1

28,651

Aerojet Rocketdyne, Inc., East
Camden, AR

6

0

6

171,904

Aerojet Rocketdyne, Inc., Rancho
Cordova, CA

0

1

1

28,651

Allen Industries, Inc.,
Greensboro, NC

0

1

1

28,651

Amphenol Corp - Aerospace
Operations, Sidney, NY

1

1

2

57,301

Aprotech Powertrain, Asheville,
NC

1

0

1

28,651

The Biltrite Corporation, Ripley,
MS

1

0

1

28,651

Brand FX Body Company -
Pocahontas, Pocahontas, IA

0

1

1

28,651

Brand FX Body Company - Swea
City, Swea City, IA

0

1

1

28,651

Britt Industries, Arlington
Heights, IL

0

1

1

28,651

Caddock Electronics Inc.,
Riverside, CAa

0

1

1

28,651

CF Martin & Co Inc./Upper
Nazareth, Nazareth, PA

0

1

1

28,651

Page 124 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Silo

Tolsil Adhesive
Lines

1 ol:il ( o;ilin«
Lines

Tolsil Lines

P( I] l se-Usite
(kg P( K/sile-
vr)

D.S. Brown Co, North Baltimore,
OH

2

1

3

85,952

Donaldson Co Inc, Stevens Point,
WI

0

1

1

28,651

Dow-Elco Inc., Montebello, CA

1

0

1

28,651

ExxonMobil Oil Corporation,
Torrance, CA

0

1

1

28,651

Fibermark, Brattleboro, VT

0

1

1

28,651

Forterra Pipe & Precast, Winter
Haven, FL

0

1

1

28,651

GKN Aerospace Chemtronics
Inc, El Cajon, CA

0

1

1

28,651

GKN Aerospace North America,
Inc. Berkeley, Hazelwood, MO

0

10

10

286,506

HH Brown Shoe Co/Cove Shoe
Martinsburg, Martinsburg, PA

1

0

1

28,651

Intrepid Powerboats, Inc., Largo,
FL

0

2

2

57,301

Irathane Systems Inc., Hibbing,
MN

0

2

2

57,301

Jayco Inc., Middlebury, IN

0

1

1

28,651

Jupiter Aluminum Corporation -
Beech Bottom Plant, Beech
Bottom, WV

0

1

1

28,651

Kansas City BPU - Quindaro,
Kansas City, KS

0

1

1

28,651

Keystone Recreational Vehicle
Company, Goshen, IN

0

1

1

28,651

Las Flores Canyon, Goleta, CA

0

1

1

28,651

Lord Corp/Cambridge Springs,
Cambridge Springs, PA

16

0

16

458,409

Lord Corp/Saegertown,
Saegertown, PA

1

0

1

28,651

Page 125 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Silo

Tolsil Adhesive
Lines

1 ol:il ( o;ilin«
Lines

Tolnl Lines

P( I] l se-Usite
(kg P( K/sile-
vr)

Lord Corporation, Bowling
Green, KY

8

0

8

229,205

Mastercraft Boat Company,
Vonore, TN

1

0

1

28,651

McNeilus Truck &
Manufacturing Inc., Dodge
Center, MN

0

42

42

1,203,325

Mfg Cherokee Pit, Tulsa, OK

0

1

1

28,651

Michelin Aircraft Tire Company,
Norwood, NC

0

1

1

28,651

Niacc-Avitech Technologies Inc.,
Clovis, CA

0

1

1

28,651

Northrop Grumman, Palmdale,
CA

0

1

1

28,651

Oharng Combined Support
Maintenance Shop, Columbus,
OH

0

2

2

57,301

Oil States Industries Special
Products Division, Arlington, TX

3

1

4

114,602

Olin Corp, East Alton, IL

0

1

1

28,651

Parker Hannifan Corporation,
Kings Mountain, NC

0

1

1

28,651

Portsmouth Naval Shipyard,
Kittery, ME

0

1

1

28,651

Precision Aerospace Corp,
Rancho Cucamonga, CA

0

1

1

28,651

R.D. Henry & Co., Wichita, KS

0

1

1

28,651

Rehau Inc., Cullman, AL

0

1

1

28,651

Ro-Lab Rubber Company Inc.,
Tracy, CA

1

0

1

28,651

Rotochopper Inc., Saint Martin,
MN

0

1

1

28,651

Safran Electrical & Power USA,
LLC, Sarasota, FL

0

1

1

28,651

Page 126 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

She

Tolsil Adhesive
Lines

Tolsil ( o;ilin«
Lines

Tolnl Lines

P( I] l se-Usite
(kg P( K/sile-
vr)

Spirit Aerosystems - Wichita,
Wichita, KS

0

1

1

28,651

SRC Of Lexington Inc.,
Lexington, KY

0

1

1

28,651

Timco, Dba Haeco Americas
Airframe Services, Greensboro,
NC

0

1

1

28,651

Tishomingo Acquisition LLC,
Dba Tbei, Tishomingo,
Tishomingo, MS

0

1

1

28,651

Toyo Automotive Parts (USA)
Inc., Franklin, KY

9

0

9

257,855

Trelleborg Vibracoustic Adhesive
Plant, Morganfield, KY

0

1

1

28,651

Trison Coatings, Inc., Lewisburg,
TN

0

1

1

28,651

Turbopower, LLC, Miami Lakes,
FL

0

1

1

28,651

Westmor Fluid Solutions LLC,
Columbus, MN

0

1

1

28,651

Zog Industries, Carpinteria, CA

0

1

1

28,651

a This site reports the sole paint stripping operation. For ease of presentation, paint stripping is counted in the "Total Coating

Lines" column.

Source: (U.S. EPA. 2016a)

2.12.2 Process Description

Based on products identified in Preliminary Information on Manufacturing, Processing, Distribution,
Use, and Disposal: Tetrachloroethylene (Perchloroethylene) (	1016a) and 2016 CDR reporting

(	2016d), PCE may be used in various adhesive, sealant, coating, paint, and paint stripper

products for industrial, commercial and consumer applications. Based on reporting in the 2014 NEI
typical application methods may include spray, roll, and dip applications (	). In the

2014 NEI (	1016a) there are 60 instances where the application method is not specified;

therefore, other applications methods (e.g., curtain, syringe/bead, roller/brush,
electrodeposition/electrocoating, and autodeposition) may also be used for these products.

The general process for adhesives and coatings include unloading liquid adhesives or coatings from
containers into the coating reservoir/application equipment, then applying the adhesive or coating to a
flat or three-dimensional substrate (OECD. 2015. 2009b). For adhesives substrates are then joined and

Page 127 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

allowed to cure with the volatile solvent (in this case PCE) evaporating during the curing stage (OECD.
2015). For solvent-based coatings, after application the substrates typically undergo a drying stage in
which the solvent evaporates from the coating (	>9b).

2.12.3 Exposure Assessment

2.12.3.1	Worker Activities

Worker activities may include unloading adhesive or coating products from containers into application
equipment, and, where used, manual application of the adhesive or coatings (e.g., use of spray guns or
brushes to apply product to substrate) (OECD. 2015). Workers may be exposed to PCE during the
application process if mists are generated such as during spray and roll applications (OECD. 2015).
Workers may also be exposed to PCE vapors that evaporate from the adhesive or coating as it is applied
or during the drying/curing process (	). EPA expects ONUs may be exposed to mists or

vapors that enter their breathing zone during routine work in areas where coating applications are
occurring.

2.12.3.2	Number of Potentially Exposed Workers

EPA estimated the number of workers and occupational non-users potentially exposed during use of
PCE-containing adhesives and coatings using Bureau of Labor Statistics' OES data (	)

and the U.S. Census" SUSB (U.S. Census Bureau. 2015) as well as the NAICS code reported by sites in
the 2014 NEI (U.S. EPA. 2016a). The method for estimating number of workers is detailed above in
Section 1.4.4 and Appendix A. These estimates were derived using industry- and occupation-specific
employment data from the BLS and U.S. Census. In the 2014 NEI, there were two sites with coating
operations that reported a NAICS code for which no Census data were available (U.S. EPA. 2016a). To
estimate the number of workers and ONUs at these sites, EPA used the average workers per site and
ONUs per site from the sites with known data. Table 2-48 provides the results of the number of worker
analysis for adhesives/sealants and Table 2-49 provides the results of the number of worker analysis for
paints/coatings19. There are approximately 410 workers and 160 ONUs potentially exposed during use
of adhesives/sealants and 1,900 workers and 1,100 ONUs potentially exposed during use of
paints/coatings.

19 Worker and ONU estimates for sites identified as having both adhesive and coating operations are included only in the
adhesives table.

Page 128 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Table 2-48. Estimated Number of Workers Potentially Exposed to Perchloroethylene During Use

of Adhesives and Sealants

NAICS
Code

Number of
Sites

Exposed
Workers
per Site11

Kxposed
Occupational
Non-l sers
per Site"

Total
Kxposed
Workers

Total
Kxposed
Occupational
Non-l sers

Total
Kxposed

316210

1

11

23

11

23

35

325998

2

14

5

28

9

37

326291

2

43

7

85

14

99

326299

3

27

4

82

13

96

332993

1

63

24

63

24

87

333132

1

21

10

21

10

31

334417

1

41

37

41

37

78

336390

1

45

13

45

13

58

336612

1

16

5

16

5

21

339113

1

20

6

20

6

27

Totalb

14

30

11

410

160

570

a Number of workers and occupational non-users per site are calculated by dividing the exposed number of workers or
occupational non-users by the number of establishments in the relevant NAICS codes. The workers/ONUs per site are then
multiplied by the number of sites within that NAICS to get the total exposed. The number of workers/ONUs per site is
rounded to the nearest integer.

b Totals have been rounded to two significant figures. Totals may not add exactly due to rounding.

Table 2-49. Estimated Number of Workers Potentially Exposed to Perchloroethylene During Use

of Paints and Coatings

NAICS
Code

Number of
Sites

Kxposed
Workers
per Site"

Kxposed
Occupational
Non-l sers
per Site"

Total
Kxposed
Workers

Total
Kxposed
Occupational
Non-l sers

Total
Kxposed

211111

1

2

4

2

4

6

221112

1

6

8

6

8

13

322130

1

120

18

120

18

139

324110

1

170

75

170

75

246

327390

1

11

2

11

2

13

331210

1

39

9

39

9

48

332812

4

7

2

29

7

35

332813

1

8

2

8

2

10

Page 129 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

NAICS
Code

Number of
Sites

Kxposed
Workers
per Site11

Kxposed
Omipsilionsil
Non-l sers
per Site"

lolsil
Kxposed
Workers

Tolsil
Kxposed
Omip;ilion;il
Non-l sers

ToUil
Kxposed

332994

1

11

4

11

4

15

332999



6

2

11

4

16

333120

1

23

11

23

11

34

333996

1

18

9

18

9

27

334220

1

17

18

17

18

35

334416

1

22

20

22

20

41

336211



33

4

133

18

150

336214



40

5

79

10

89

336390

1

45

13

45

13

58

336410

1

75

64

75

64

139

336411



184

155

551

465

1,016

336412

1

47

39

47

39

86

336413



41

35

123

104

227

336415

1

132

111

132

111

243

336611

1

61

19

61

19

80

336612

1

16

5

16

5

21

337110

1

3

2

3

2

6

337127

1

9

7

9

7

16

339920

1

9

2

9

2

11

339950

1

5

1

5

1

7

339992

1

7

2

7

2

9

339999

1

5

1

5

1

6

541710

2

1

9

2

19

21

Subtotal for
Known Data

44

41

24

1,790

1,073

2,863

No Data

2

38

21

76

43

118

Totalb

46

41

24

1,900

1,100

3,000

a Number of workers and occupational non-users per site are calculated by dividing the exposed number of workers or
occupational non-users by the number of establishments in the relevant NAICS codes. The workers/ONUs per site are then

Page 130 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

multiplied by the number of sites within that NAICS to get the total exposed. The number of workers/ONUs per site is
rounded to the nearest integer.

b Totals have been rounded to two significant figures. Totals may not add exactly due to rounding.

2.12.3.3 Occupational Exposure Results

EPA identified inhalation exposure monitoring data from a study at a single site in Poland using a PCE-
based adhesive, from three NIOSH investigations at three sites using PCE-based coatings, a study
submitted to EPA under TSCA for a truck plant using PCE-based coatings, and data provided to EPA
from DoD for spray coating processes (Defense Occupational and Environmental Health Readiness
System - Industrial Hygiene. 2018; Gromiec et at.. 2001. 's ^ ^ s , s - 5; Ford Motor >¦ s
NIOSH. 1981a). Due to the large variety in shop types that may use PCE-based adhesives and coatings,
it is unclear how representative these data are of a "typical" site using these products. However, EPA
does not have a model for estimating exposures from use of adhesives or paints/coatings; therefore, the
assessment is based on the identified monitoring data.

Gromiec (2002) studied chemicals and their air concentrations in a repair shop where rubber conveyor
belts were repaired at a brown coal mine in Poland. PCE was identified as a component of one of the
adhesives used to repair the convey belts (Gromiec et at.. 2002). The study collected a total of 13 PBZ
samples for employees in the repair shop and sample times were a minimum of 360 min (75% of the
working shift) (Gromiec et at.. 2002). The samples were collected from employees with the following
job titles: milling machine operators, assembler-vulcanizer, rolling machine operator, vulcanization
press operator, and roller and barrel vulcanizer (Gromiec et at.. 2002). Based on the job descriptions in
the report, only the assembler-vulcanizers are expected to handle the adhesive directly (Gromiec et at..
2002).

The study did not indicate the application method of the adhesive or the concentration of PCE in the
adhesive formulation; therefore, it is unknown how representative these data are of a "typical" PCE-
based adhesive formulation. The study did not provide discrete sample results; therefore, the high-end
exposure value is based on the max concentration of 0.81 ppm and the central tendency is based on the
mean concentration of 0.09 ppm reported in the study (Gromiec et at.. 2002). The study also did not
differentiate between worker and ONU exposures; therefore, EPA only presents a single set of exposure
results for the use of PCE-based adhesive. A summary of the inhalation exposure results for adhesives
can be found in Table 2-50.

Table 2-50. Summary of Inhalation Exposure Monitoring Data for Use of Perchloroethylene-
Based Adhesives

Scenario

8-hr TWA
(ppm)11

AC
(ppm)

A IK

(ppm)

I.AIK
(ppm)

Nilmher of
Data Points

High-End

0.8

0.3

0.2

9.50E-02

13

Central Tendency

8.85E-02

2.95E-02

2.02E-02

8.03E-03

AC = Acute Concentration; ADC = Average Daily Concentration; and LADC = Lifetime Average Daily Concentration.
Equations and parameters for calculation of the AC, ADC, and LADC are described in Appendix B.
a Exact sample times not given in study; however, study indicates that samples were taken for a minimum of 75% of the shift
(360 min). Therefore, EPA assumes that the results are representative of an 8-hr TWA exposure.

Source: (Gromiec et at. 20021

Page 131 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

The NIOSH studies (NIOS 3, 1986. 1981a) include the use of PCE-based paints and coatings
during construction of a nuclear power plant, at a specialty packaging products manufacturing site, and
an aluminum foundry. At the nuclear power plant, the PCE-based coating was applied as the second coat
of a three-step epoxy coating process (primer, second coat, and finish coat) (NIOSH. 1981a). Each
coating in the three-step process was a two-part epoxy coating with the two parts being mixed together
just prior to application (NIPS :1a). The coatings were applied primarily using an airless spray
gun; however, some applications were done via troweling (NIOSH. 1981a). Sample times ranged from
-2.5 to 4.5 hours; where sample times were less than eight hours, EPA converted to an 8-hr TWA
assuming exposure outside the sample time was zero.

At the specialty packaging products manufacturing site, PCE was identified as a component of one of
the coatings used in the coating of industrial-sized backing paper (NIOSH. 1993). In the process, the
backing paper was unraveled and conveyed through a coating system consisting of a coating tray,
application roller, and leveling rod (NIOSH. 1993). After coatings the paper was passed through an oven
for drying and curing and then re-rolled, cut to size and packaged for shipping (NIOSH. 1993). The
study collected samples from four workers working in the coating area (NIOSH. 1993). Sample times
ranged from -1.5 to 6.5 hours; where sample times were less than eight hours, EPA converted to an 8-hr
TWA assuming exposure outside the sample time was zero. It should be noted that two of the samples
measured concentrations between the limits of detection and the limit of quantitation (NIOS	).

At the aluminum foundry, PCE was identified as a component of a silver pigmented paint (NIOSH.
1986). The paint was applied via brushing or dipping to 10-15% of all the aluminum cores produced at
the site (NIOSH. 1986). The coating is used to prevent shrinkage of the aluminum as it cools in the mold
(NIOSH. 1986). The study collected two samples from workers applying the coating to the aluminum
cores (NIPS 6). Sample times ranged from 5 to 7 hours; where sample times were less than eight
hours, EPA converted to an 8-hr TWA assuming exposure outside the sample time was zero.

The study submitted to EPA under TSCA, measured exposure to workers at a truck plant during paint
mixing and pot spraying applications (Ford Motor Co. 1981). Paint applications occur in a boot with
ventilation to control emissions ( I Motor Co. 1981). The study collected three full-shift samples and
calculated a fourth full-shift TWA by combining four 15-min samples and one approximately 5-hour
sample collected on the same worker on the same day (Ford Motor Co. 1981). Sample times for the
three full-shift samples were not provided, but EPA assumed exposures concentrations were
representative of an 8-hr shift.

The data provided to EPA from DoD contained two samples collected in August 2007 and one sample
collected in June 2016 (Defense Pccupational and Environmental Health Readiness System - Industrial
Hygiene. 2018). All three samples were identified as for high-volume low-pressure spray applications
(Defense Pccupational and Environmental Health Readiness System - Industrial Hygiene. 2018). The
one data point from June 2016 was reported as having a sample time of zero (Defense Pccupational and
Environmental Health Readiness System - Industrial Hygie	). It is unclear what the result for this

data point represents; therefore, this data was not used in EPA's analysis. The other two data points had
a sample time of 15-min and 180 min (Defense Pccupational and Environmental Health Readiness
System - Industrial Hygiene. 2018). EPA converted the 180 min sample to an 8-hr TWA by assuming
zero exposure outside the sample time. The DoD data did not report a duration for the spray coating;

Page 132 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

however, EPA assumes the sample time is reflective of the duration of spray coating at the site and that
the worker will not perform other activities throughout the day that will result in PCE exposures.

A summary of the inhalation exposure results for coating applications can be found in Table 2-51. For
the 8-hr TWA, the 95th percentile of the data is presented as the high-end and the 50th percentile as the
central tendency. Due to the limited number of data points for the 15-minute TWA, the maximum is
presented as the high-end and the median is the central tendency. No data for ONUs were identified;
therefore, the table only includes results for workers. It should be noted that the PCE concentration in
the coatings used for each of the NIOSH studies were not provided; therefore, it is unclear how
representative these data are of a "typical" coating application.

Table 2-51. Summary of Inhalation Exposure Monitoring Data for Use of Perchloroethylene-
Based Paints/Coatings	

Scenario

8-hr
TWA
(ppni)

AC
(ppm)

A IK

(ppm)

I.AIK
(ppm)

Nil m her
of Data
Points

15-m i mile
TWA
(ppm)

Nil in her
of Data
Points

High-End

4.6

1.5

1.0

0.5

15

7.9

5

Central Tendency

0.2

7.78E-02

5.33E-02

2.12E-02

4.1

AC = Acute Concentration; ADC = Average Daily Concentration; and LADC = Lifetime Average Daily Concentration.
Equations and parameters for calculation of the AC, ADC, and LADC are described in Appendix B.

Source: (Defense Occupational and Environmental Health Readiness System - Industrial Hygiene. 20.1.8: NIOSH. .1.993.
.1.986: Ford Motor Co. 1.981: NIOSH. 1981a)

2.12.4 Water Release Assessment

2.12.4.1	Water Release Sources

The source of water releases at sites using PCE-based adhesives and coatings will vary depending on the
application methods and the control technologies used. The primary sources of water releases may
include: overspray losses from spray applications using water curtains to capture overspray and splatter
and mists generated during curtain and roll coating that are subsequently discharged to water (
2015. 2009b). Other potential sources include from the use of water to clean the containers and/or
equipment (U< 4 -"01 \ 2009b). However, for organic solvent-based products such as PCE-based
adhesives and coatings, EPA expects the majority of equipment and container cleaning to be performed
using organic solvents that are not discharged to water.

2.12.4.2	Water Release Assessment Results

EPA assessed bounding water release estimates using the PCE use-rates estimated in Section 2.12.1 and
loss fractions obtained from the ESD on Use of Adhesives (OE	) and the ESD on Coating

Industry (Paints, Laquers and Varnishes) (OECD. 2009b). Releases vary by the type of site using the
PCE-based coating or adhesive (i.e., automotive, aerospace, etc.), the amount of PCE-based coating or
adhesive used, and the application method of coating or adhesive.

The ESD on the Use of Adhesives (OECD. 2015) estimates the only application methods of adhesives
expected to result in discharges to water are roll and curtain applications. The ESD estimates transfer
efficiencies for roll and curtain applications may vary from 90 to 98% with splatter/mists generated

Page 133 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

during application disposed of to water, incineration, or landfill (OECD. 2015). For this bounding
estimate, EPA assumed a transfer efficiency of 90% (or loss of 10%) with all splatter/mists generated
discharged to POTW for sites indicating roll or curtain application methods (OB	). This

bounding estimate is likely an overestimate of actual releases. EPA also used this loss fraction to
estimate water discharges from all adhesive applications with unspecified application methods in the
2014 NEI (	1016a). This is based on the ESD indicating to use roll coating as the default

application method for all end-use industries (including unknown) except Motor and Non-Motor
Vehicle, Vehicle Parts, and Tire Manufacturing (Except Retreading) (OE<	). Based on the

NAICS codes reported in the 2014 NEI (	), none of sites with unspecified application

methods fall under the Motor and Non-Motor Vehicle, Vehicle Parts, and Tire Manufacturing (Except
Retreading) industry.

The ESD on Use of Adhesives (OECD. 2009b) estimates no releases of water for spray adhesive
applications as spray applications are expected to occur in spray booths or totally enclosed systems that
utilize dry filters to capture overspray and dry filters are expected to be disposed of to landfills or
incineration (OECD. 2009b). The ESD estimates releases from syringe/bead applications to be
negligible (OECD. 2009b). The ESD does not address releases from dip applications; however, EPA
does not expect discharges to water from dip applications as generation of mists/overspray is not
expected.

There were no roll coating applications reported for coating sites in the 2014 NEI (	»a);

however, there were 49 instances where the application method was unspecified. Based on information
in the ESD on the Coating Industry (OECD. 2009b). EPA assessed unspecified application methods as
spray applications. Water releases from dip coating and paint stripping are not expected; therefore, EPA
only water discharges from sites using spray application methods. Similar to adhesives, water releases
are not expected to occur from spray applications occurring in spray booths with dry filters; however,
unlike adhesives, the use of wet curtains to control overspray losses is expected in some coating industry
applications (OECD. 2009b). EPA assessed water releases from spray applications using water curtains
using the EPA/OPPT Automobile OEM Coating Overspray Loss Model which uses Equation 2-2 to
estimate releases to water:

Equation 2-2

AR = URx(l- TE) x McE x (1 - SrE)

Where:

AR

TE

McE

SrE

UR

Annual release of PCE from the application line (kg/yr)
fractional transfer efficiency of spray gun (unitless)
fractional spray mist capture efficiency (unitless)
fractional solid removal efficiency from captured mist (unitless)
use-rate of PCE (kg/application line-yr).

EPA estimated the type of spray booth used (dry filter or water curtain), values for TE, McE, and SrE
based on information from the ESD on the Coating Industry (OECD. 2009b). the draft GS on Use of
Spray Coatings in the Furniture Industry (U.S. EPA. 2004). the ESD on Automotive Refinishing
(	), and default values provided in the EPA/OPPT Automobile OEM Coating Overspray

Page 134 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Loss Model. Table 2-52 provides the list of values used for each industry. In general, the industry type
for each site was determined by the NAICS code reported in the 2014 NEI (I v H \ 2016a).

Table 2-52 provides values for six industry sectors (including an "other/unknown" sector) covered by
the various coatings ESDs and GSs. The table does not include values for metal packaging (i.e., metal
can), coil coating, and rail vehicle applications covered in the ESD on the Coatings Industry (

2009b) as the NAICS codes reported in the 2014 NEI (U.S. EPA. 2016a) were determined not to fall
into either category. The "Assessed Spray Booth Type" column indicates the type of spray booth used
for assessing water releases; however, in many instances, the use of both dry filters or water curtains
within an industry is possible. This column is not meant to indicate that this is the only booth type used;
rather, it represents the booth type used for this assessment based on available information of the
likelihood a particular booth type will be used within the industry. Where both dry filters and wet curtain
booths are equally likely or information to determine which is more likely was not available, EPA
assessed as a water curtain booth to determine a bounding estimate. For other/unknown industry sectors
(i.e., industry sectors not covered by an ESD or GS), EPA used the most protective values for water
curtains to generate a bounding water release estimate for these sites. It should be noted that several sites
in the 2014 NEI (	a) indicated the use of dry filters as the control approach; EPA

assessed zero discharges from these sites regardless of industry based on the use of dry filters.

Page 135 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Table 2-52. Parameters for Estimating Water Discharges from Spray Coating Applications

Industry
Type

Assessed
Spray
Booth
Type

Spray
Gun
Transfer
Efficiency
(%)

Mist
Capture
Efficiency
(%)

Solid
Removal
Efficiency
(%)

Source

Furniture
Manufacture

Dry
Filter

N/A - water releases not expected
for dry filter booths

(OECD. 2009b; U.S. EPA. 2004)

Automotive
Original
Equipment
Manufacture

Water
Curtain

20%

96%

90%

Default values for the EPA/OPPT
Automobile OEM Coating Overspray
Loss Model. The model estimates transfer
efficiencies based on a conventional spray
guna.

Automotive
Reftnishing

Dry
Filter

N/A - water releases not expected
for dry filter booths

(OECD. 2011a)

Marineb

Water
Curtain

65%

100%

95%

(OECD. 2009b)

Aerospace

Dry
filter0

N/A - water releases not expected
for dry filter booths

(OECD. 2009b)

Other/
Unknown

Water
Curtain

20%

96%

90%

Default values for the EPA/OPPT
Automobile OEM Coating Overspray
Loss Model. The model estimates transfer
efficiencies based on a conventional spray
gun.

a The ESD (OECD. 2009b) estimates water from coatings in the automotive OEM manufacture is captured and sent for
disposal with paint wastes; however, the ESD is based on information from the EU and it is unclear if this practice is
applicable to sites in the U.S. Therefore, EPA uses the more protective model values for a bounding estimate.
b The ESD (OECD. 2009b) estimates an overall transfer efficiency of 65% based on the prominent use of airless and air
assisted spray guns. The ESD (OECD. 2009b) estimates a solid removal efficiency of 90% but indicates that half of the
unremoved overspray is disposed of to land; therefore, EPA assessed a total solids removal efficiency of 95% to account for
both removed solids (90% of overspray) and unremoved solids released to land (5% of overspray).

0 The ESD (OECD. 2009b) indicates that due to the size, the majority of aerospace coating applications do not occur in spray
booths but in hangers. The ESD (OECD. 2009b) indicates that 80% of the overspray is expected to deposit on masking,
clothes, or the hanger floor and is subsequently disposed of in skips or washed to an interceptor tank where it is held until the
tank is emptied, with contents disposed of to incineration or landfill. The remaining 20% of overspray is expected to be
captured by fiberglass filters in air extraction units in the hanger. EPA assesses the use of fiber glass filters similar to dry
filter spray booths assuming no discharge to water.

EPA used the values in Table 2-52, the estimated use-rate per application line, and the number of
application lines per site to estimate the annual wastewater discharge for each coating site in the 2014
NEI. EPA then averaged annual discharges over 250 operating days/yr to estimate daily per site
discharges. Discharges from both adhesive and coating application sites are provided in Table 2-53. For
ease of presentation, Table 2-53 only includes sites that had non-zero discharges. The ESD on Use of
Adhesives (OECD. 2015) recommends assuming all aqueous discharges are sent to POTWs prior to
discharge to surface water. The ESD on the Coatings Industry (OECD. 2009b) does not indicate the

Page 136 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

prevalence of on-site treatment versus discharges to POTW; EPA assumes wastewater will be handled
similar to adhesive application sites and assesses discharges to POTW.

It should be noted that all models used to develop water discharges from coating and adhesive
applications represent estimates for the solids (i.e., non-volatile) portions of the coating or adhesive and
does not account for potential evaporation of volatiles from the mist prior to entering wastewater.
Therefore, these estimates likely overestimate actual wastewater discharges of PCE due to volatilization
(PCE vapor pressure is 18.5 mmHg at 25°C). This evaporation is difficult to estimate and is not
considered in this assessment. However, this overestimate is consistent with EPA's development of a
bounding estimate for these sites.

Table 2-53. Estimated Wastewater Discharges of Perchloroethylene from Coating and Adhesive
Application Sites	

Site

Silo Type

Nil m her
of Lines
with
Water
Releases

PC 1: l se-
Rate for
Lines with
Water
Releases
(kg/site-yr)

Annual
Release
(kg/site

-M)

Operating

Days
(davs/yr)

Daily
Release
(kg/site-
da v)

Release
Media

3M -R&D Facility
- Maplewood Bldg
201, Maplewood,
MN

Coating -
Other

1

28,651

2,200

250

8.8

POTW

3P Processing,
Wichita, KS

Coating -
Other

1

28,651

2,200

250

8.8

POTW

Accellent/
Collegeville,
Collegeville, PA

Coating -
Other

1

28,651

2,200

250

8.8

POTW

Aerojet Rocketdyne,
Inc., East Camden,
AR

Adhesive

6

171,904

17,190

250

69

POTW

Allen Industries, Inc.,
Greensboro, NC

Coating -
Other

1

28,651

2,200

250

8.8

POTW

Amphenol Corp -
Aerospace
Operations, Sidney,
NY

Coating -
Other

2

57,301

5,065

250

20

POTW

Brand FX Body
Company -
Pocahontas,
Pocahontas, IA

Coating -
Auto
OEM

1

28,651

2,200

250

8.8

POTW

Page 137 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Site

She Type

Nil in her
of Lines
with
\Y siler
Uelesises

PC 1: l se-
Usile lor
Lines willi
\\ siler
Uelesises
(kg/sile-vr)

A ii n usil
Uelesise
(kg/silo

-M)

Opersiling

Dsivs
(tlsivs/yr)

Dsiilv
Uelesise
(kg/silc-
clsiv)

Uelesise
Metlisi

Brand FX Body
Company - Swea
City, Swea City, IA

Coating -
Auto
OEM

1

28,651

2,200

250

8.8

POTW

CF Martin & Co
Inc./Upper Nazareth,
Nazareth, PA

Coating -
Other

1

28,651

2,200

250

8.8

POTW

D.S. Brown Co,
North Baltimore, OH

Coating -
Other

1

28,651

2,200

250

8.8

POTW

Donaldson Co Inc,
Stevens Point, WI

Coating -
Other

1

28,651

2,200

250

8.8

POTW

Dow-Elco Inc.,
Montebello, CA

Adhesive

1

28,651

2,865

250

11

POTW

ExxonMobil Oil
Corporation,
Torrance, CA

Coating -
Other

1

28,651

2,200

250

8.8

POTW

Fibermark,
Brattleboro, VT

Coating -
Other

1

28,651

2,200

250

8.8

POTW

Forterra Pipe &
Precast, Winter
Haven, FL

Coating -
Other

1

28,651

2,200

250

8.8

POTW

HH Brown Shoe
Co/Cove Shoe
Martinsburg,
Martinsburg, PA

Adhesive

1

28,651

2,865

250

11

POTW

Intrepid Powerboats,
Inc., Largo, FL

Coating -
Marine

2

57,301

1,003

250

4.0

POTW

Irathane Systems
Inc., Hibbing, MN

Coating -
Other

2

57,301

4,401

250

18

POTW

Jayco Inc.,
Middlebury, IN

Coating -
Other

1

28,651

2,200

250

8.8

POTW

Jupiter Aluminum
Corporation - Beech

Coating -
Other

1

28,651

2,200

250

8.8

POTW

Page 138 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Site

She Type

Nil in her
of Lines
with
\Y siler
Uelesises

PC 1: l se-
Usile lor
Lines willi
\\ siler
Uelesises
(kg/sile-vr)

A ii n usil
Uelesise
(kg/silo

-M)

Opersiling

Dsivs
(tlsivs/yr)

Dsiilv
Uelesise
(kg/silc-
clsiv)

Uelesise
Metlisi

Bottom Plant, Beech
Bottom, WV















Kansas City BPU -
Quindaro, Kansas
City, KS

Coating -
Other

1

28,651

2,200

250

8.8

POTW

Keystone

Recreational Vehicle
Company, Goshen,
IN

Coating -
Other

1

28,651

2,200

250

8.8

POTW

Las Flores Canyon,
Goleta, CA

Coating -
Other

1

28,651

2,200

250

8.8

POTW

Lord

Corp/S aegertown,
Saegertown, PA

Adhesive

1

28,651

2,865

250

11

POTW

Mastercraft Boat
Company, Vonore,
TN

Adhesive

1

28,651

2,865

250

11

POTW

McNeilus Truck &
Manufacturing Inc.,
Dodge Center, MN

Coating -
Auto
OEM

42

1,203,325

92,415

250

370

POTW

Michelin Aircraft
Tire Company,
Norwood, NC

Coating -
Other

1

28,651

2,200

250

8.8

POTW

Oharng Combined
Support Maintenance
Shop, Columbus, OH

Coating -
Other

2

57,301

4,401

250

18

POTW

Oil States Industries
Special Products
Division, Arlington,
TX

Coating -
Other

1

28,651

2,200

250

8.8

POTW

Olin Corp, East
Alton, IL

Coating -
Other

1

28,651

2,200

250

8.8

POTW

Page 139 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Site

She Type

Nil in her
of Lines
with
\Y siler
Uelesises

PC 1: l se-
Usile lor
Lines willi
\\ siler
Uelesises
(kg/sile-vr)

A ii n usil
Uelesise
(kg/silo

-M)

Opersiling

Dsivs
(tlsivs/yr)

Dsiilv
Uelesise
(kg/silc-
clsiv)

Uelesise
Metlisi

Parker Hannifan
Corporation, Kings
Mountain, NC

Coating -
Other

1

28,651

2,200

250

8.8

POTW

Portsmouth Naval
Shipyard, Kittery,
ME

Coating -
Marine

1

28,651

501

250

2.0

POTW

Rehau Inc., Cullman,
AL

Coating -
Auto
OEM

1

28,651

2,200

250

8.8

POTW

Rotochopper Inc.,
Saint Martin, MN

Coating -
Other

1

28,651

2,200

250

8.8

POTW

Safran Electrical &
Power USA, LLC,
Sarasota, FL

Coating -
Other

1

28,651

2,200

250

8.8

POTW

SRC Of Lexington
Inc., Lexington, KY

Coating -
Other

1

28,651

2,200

250

8.8

POTW

Tishomingo
Acquisition LLC,
Dba Tbei,
Tishomingo,
Tishomingo, MS

Coating -
Auto
OEM

1

28,651

2,200

250

8.8

POTW

Toyo Automotive
Parts (USA) Inc.,
Franklin, KY

Adhesive

4

114,602

11,460

250

46

POTW

Trelleborg
Vibracoustic
Adhesive Plant,
Morganfield, KY

Coating -
Other

1

28,651

2,200

250

8.8

POTW

Westmor Fluid
Solutions LLC,
Columbus, MN

Coating -
Other

1

28,651

2,200

250

8.8

POTW

Zog Industries,
Carpinteria, CA

Coating -
Other

1

28,651

2,200

250

8.8

POTW

Page 140 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

2.13 Maskant for Chemical Milling

2.13.1	Estimates of Number of Facilities

EPA estimated the number of sites using PCE as a maskant for chemical milling using information
obtained during meetings between EPA and industry stakeholders involved in the production, supply,
and use of chemical maskants. Data for estimates of the number of facilities is based on information
from AC Products (ACP) who supply over 99% of the solvent-based maskants sold in the U.S (AC
Products. 2017). According to AC Products (2017). a total of 679,000 gallons of maskants were sold in
North America in 2016 and 83% (by volume) were either solely or primarily PCE-based. ACP estimates
it sold 539,133 gallons of PCE-based maskants to approximately 71 U.S. customers using PCE-based
maskants with 6 of these customers using approximately 530,624 gallons (—98.4% by volume) of the
PCE-based maskants (AC Products. 2017).

EPA assumed the PCE-based maskant use for the six sites accounting for over 98% of the volume was
evenly distributed amongst the sites resulting in an average annual PCE-based maskant use of 88,437
gal/site-yr. EPA assumed the remaining 8,509 gal of maskant was evenly distributed amongst the other
65 sites resulting in an average annual PCE-based maskant use of 131 gal/site-yr. The concentration of
PCE in the maskants in unknown; therefore, the annual use rate of PCE per site could not be determined.

2.13.2	Process Description

Chemical maskants are elastomer-based coatings that are used to protect a substrate during exposure to a
chemical process (AC Products. 2017). They are used in chemical milling, plating, and anodizing
processes in the aerospace (military, commercial, and space), medical implants, and non-aerospace
military industries (AC Products. JO I ; l'-vh Met Inc.. JO I ).

Maskants are typically applied in dedicated coating application rooms via dipping of parts in a coating
tank or through automated airless spraying of coating onto the substrate (	ducts. 2017;

Ducommun Inc.. 2017; Spirit Aero Systems Inc.. 2017; Tech Met Inc.. 2017; Triumph Precision
Components JO I ; \\ eatherford Aerospace. 2017). The maskant coating is then cured, scribed, and
selectively removed in the desired locations to allow chemical milling (Spirit Aero Systems Inc.. 2017;
Tech Met Inc.. 2017; Weatherford Aerospace. 2017). The maskant forms a strong flexible film that can
withstand the chemical milling process such that only the exposed metal is milled (Spirit AeroSystems
Inc.. JO I ; l'-vh Met Inc.. JO I ; Weatherford Aerospace. 2017). Once the process is complete the
maskant can be peeled off the metal substrate (Weatherford Aerospace. 2017). According to AC
Products (2017). 95% (by volume) of the PCE-based maskants used in the U.S. are re-captured by the
customer and returned to the maskant manufacturer to make fresh maskant.

2.13.3	Exposure Assessment

2.13.3.1 Worker Activities

Information from stakeholder meetings and public comments indicate that in typical maskant application
processes the potential for exposure is low as the process is automated and performed in a dedicated
room (Ducommun Inc.. 2017; Spirit Aero Systems Inc.. JO I ; 1'^vh Met Inc.. JO I ). However, at least
one stakeholder indicated that employees may be exposed during maintenance operations (Spirit
Aero Systems Inc. J ). Specific maintenance activities were not described but may include adding
fresh maskant and handling of re-captured maskants.

Page 141 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

2.13.3.2 Number of Potentially Exposed Workers

EPA estimated the number of workers and occupational non-users potentially exposed during use of
PCE as a chemical maskant using Bureau of Labor Statistics' OES data (]i.S Jj]_S 1 ) and the U.S.
Census' SUSB (U.S. Census Bureau. ) as well as the primary NAICS and SIC code reported by
sites in the 2016 TRIO v ^ I \	2016 DMR (US. EPA. 2016b\ and/or the 2014 NEI (\ "

). The method for estimating number of workers is detailed above in Section 1.4.4 and
Appendix A. These estimates were derived using industry- and occupation-specific employment data
from the BLS and U.S. Census.

EPA identified 13 sites in the 2016 TRI that may use PCE-based maskants based on the NAICS code
reported, information submitted via public comments, and/or information obtained during stakeholder
meetings (Ducommun Inc.. 2017; Spirit Aero Systems Inc.. 2017; Tech Met Inc.. 2017; Triumph
Precision Compone^	, i v M \ », \\ iatherford Aerospace. JO I ). An additional 14 sites

were identified in the 2016 DMR as potential users of PCE-based maskants based on the reported SIC
codes and six sites were identified in the 2014 NEI based on reported emission unit information with
two of the sites in DMR and five of the sites in NEI being the same as sites identified in TRI (U.S. EPA.
2016a. b). The employment data from the U.S. Census SUSB and the Bureau of Labor Statistics" OES
data are based on NAICS code; therefore, SIC codes reported in the 2016 DMR had to be mapped to a
NAICS code to estimate the number of workers. A crosswalk of the SIC codes to the NAICS codes used
in the analysis are provided in Table 2-54.

Table 2-54. Crosswalk of Maskant SIC Codes in DMR to NAICS Codes

Sl( ( ode

Corresponding NAICS Code

3492 - Fluid Power Valves and Hose Fittings

332912 - Fluid Power Valve and Hose Fitting
Manufacturing

3721 - Aircraft

336411 - Aircraft Manufacturing

3724 - Aircraft Engines and Engine Parts

336412 - Aircraft Engine and Engine Parts
Manufacturing

3761 - Guided Missiles and Space Vehicles

336414 - Guided Missile and Space Vehicle
Manufacturing

3764 - Guided Missile and Space Vehicle
Propulsion Units and Propulsion Unit Parts

336415 - Guided Missile and Space Vehicle
Propulsion Unit and Propulsion Unit Parts
Manufacturing

9711 - National Security

928110-National Security

The reported NAICS codes (or NAICS code corresponding to a reported SIC code) were used to
estimate the number of workers and ONUs at each reporting site with three exceptions: 1) the site in the
2014 NEI that reported the NAICS code 311942 - Spice and Extract Manufacturing; 2) the sites
reporting the NAICS or SIC codes for National Security; and 3) one site in DMR that did not report a
SIC code. The site reporting the NAICS code 311942 in the 2014 NEI also reported their source
classification code as "Chemical Evaporation Surface Coating Operations Large Aircraft Other Not
Classified" and review of the company website indicated they specialize in the development,

Page 142 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

manufacture and application of products for the aerospace, automotive, marine and electronics
industries. Therefore, EPA assumes the reported NAICS code is an error and that the correct NAICS
code is 336411, Aircraft Manufacturing, which gives the most protective worker and ONU estimates of
the aircraft and aircraft parts manufacturing NAICS codes.

Employment data for NAICS code 928110 was not available from the U.S. Census Bureau or in the
Bureau of Labor Statistics. The sites reporting this NAICS or SIC code are either U.S. Air Force bases
or U.S. Naval bases; therefore, EPA assumes the activities at this site are typical of an aircraft or aircraft
parts manufacturer and uses worker and ONU estimates from the NAICS code 336411 to estimate the
number of workers at the site. Similarly, the one site without a SIC code in DMR was an airfield, which
EPA assumed to have activities similar of an aircraft or aircraft parts manufacturing and estimated
workers/ONUs using the NAICS code 336411.

The data from the 2016 TRI, 2016 DMR, and 2014 NEI only covers 28 unique sites; however, market
data from ACP indicates there are up to 71 sites using PCE-based maskants ( \(* Products. 201 ). To
estimate the number of workers and ONUs at the remaining sites EPA calculated the average number of
workers and ONUs per site from the 28 known sites. This resulted in 95 workers per site and 75 ONUs
per site at the unknown sites and a total of approximately 6,700 workers and 5,300 ONUs potentially
exposed during maskant uses of PCE (see Table 2-55).

Page 143 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Table 2-55. Estimated Number of Workers Potentially Exposed to Perchloroethylene During Use
of Chemical Maskants

NAICS Cotlc

Number of
Sites

Kxposed
Workers
per Site11

Kxposed
Occupaliona
1 Non-l sers
per She11

Total
Kxposed
Workers1'

Total
Kxposed
Occupaliona
1 Non-l sers1'

Total
Kxposed

332721

1

4

2

4

2

6

332812

1

7

2

7

2

9

332912

1

28

11

28

11

38

336214

1

40

5

40

5

45

336411

10

184

155

1,846

1,549

3,385

336412

4

47

39

186

157

343

336413

7

41

35

288

243

530

336414

1

372

314

372

314

686

336415

2

132

111

263

222

485

Subtotal for Known
Sites

28

108

89

3,024

2,504

5,529

Sites with
Unknown NAICS
or No Data

43

95

75

4,078

3,218

7,296

Total0

71

94

75

6,700

5,300

12,000

a Number of workers and occupational non-users per site are calculated by dividing the exposed number of workers or
occupational non-users by the number of establishments in the relevant NAICS codes. The workers/ONUs per site are then
multiplied by the number of sites within that NAICS to get the total exposed. The number of workers/ONUs per site is
rounded to the nearest integer. Number of workers and occupational non-users per site for sites with unknown NAICS codes
are calculated by averaging the values of the known sites.

b Total exposed workers and ONUs for sites with known NAICS are taken directly from the Bureau of Labor Statistics' OES
data the U.S. Census' SUSB. For sites with unknown NAICS codes the total workers and ONUs are estimated by multiplying
the workers and ONUs per site by the number of sites.

0 Totals have been rounded to two significant figures. Totals may not add exactly due to rounding.

2.13.3.3 Occupational Exposure Results

EPA identified inhalation exposure monitoring data from a single NIOSH investigation at an aircraft
parts manufacturing site using a dip coating application process for the maskants (NIOSH h>~'7Y The
NISOH report does not specify if PCE is the primary solvent in the maskant, the concentration of PCE
in the maskant, or the typical maskant use rates at the site. Due to the variety in both industry types and
typical per site maskant use rates and the uncertainty of the PCE concentration in the maskant, it is
unclear if these data are representative of a "typical" site. Additionally, these data were collected prior to
the promulgation of the Aerospace Manufacturing and Rework Facilities NESHAP which regulates the
emissions of hazardous air pollutants (HAPs) from various operation at aerospace facilities including
chemical milling. To the extent that this NESHAP reduces emissions of PCE into the workroom worker

Page 144 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

exposures may be lower than identified data. EPA does not have a model for estimating exposures from
maskant uses; therefore, the assessment is based on the identified monitoring data.

Table 2-56 summarizes the 8-hr and 4-hr TWA monitoring data for the use of PCE in maskants. The
data were obtained from a NIOSH HHE report from a site investigation in 1977 (NIOSH. 1977). The
95th percentile of the data is presented as the high-end and the 50th percentile as the central tendency.
The study collected data from multiple chemical mill operators at the site with sample times ranging
from approximately 2 to 7 hours (NIOSH. 1977). The report indicated that workers were rotated out of
the chemical mill area every four hours to minimize exposures CHIOS 7). EPA calculated 8-hr-
and 4-hr TWAs assuming exposure outside the sample time was zero.

Table 2-56 also includes a summary of 15-min TWA samples collected by the DoD between July 2013
and May 2017 during masking activities (Defense Occupational and Environmental Health Readiness
System - Industrial Hygiene. 2018). The DoD data contained nine samples that were measured below
the LOD (Defense Occupational and Environmental Health Readiness System - Industrial Hygiene.
2018). To estimate exposure concentrations for data below the LOD, EPA followed the Guidelines for
Statistical Analysis of Occupational Exposure Data (	») as discussed in Section 1.4.5.2.

The geometric standard deviation for the data was above 3.0; therefore, EPA used the to estimate the

exposure value as specified in the guidelines (	)).

Table 2-56. Summary of Worker Inhalation Exposure Monitoring Data for Chemical Maskants

Scenario

8-hr
TWA
(ppm)

AC
(ppm)

A IK

(ppm)

I.AIK
(ppm)

Nil m he
r of
Data
Points

4-hr
TWA
(ppm)

Nil in he
r of
Data
Points

15-
m i mile
TWA
(ppm)

N il in her
of Data
Points

High-End

2.1

0.7

0.5

0.2

24

3.2

9

28

20

Central Tendency

1.2

0.4

0.3

0.1

2.4

0.6

AC = Acute Concentration; ADC = Average Daily Concentration; and LADC = Lifetime Average Daily Concentration.

Equations and parameters for calculation of the AC, ADC, and LADC are described in Appendix B.

Source: (Defense Occupational and Environmental Health Readiness System - Industrial Hygiene. 20.1.8: NIOSH. 1977")

2.13.4 Water Release Assessment

2.13.4.1 Water Release Sources

According to AC Products (2017). 95% by volume of the PCI'-based maskants sold in the U.S. are
recaptured and returned to the maskant manufacturer for production of new maskant. Therefore, the
volume of PCE that may be released to any environmental media is expected to be no more than 5% of
the use volume. Based on information from public comments and stakeholder meetings and the volatility
of PCE, EPA expects the majority of the PCE that does not participate in the recapture process to be
released to air (Ducommim Inc.. 2017; Spirit AeroSystems Inc.. 2017; Tech Met Inc.. 2017; Triumph
Precision Components. 2017; Weatherford Aerospace. ^ ). However, there is some potential for PCE
to be released to water from the use of steam to regenerate carbon adsorbers used to control emissions in
the chemical milling area.

Page 145 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

2.13.4.2 Water Release Assessment Results

EPA identified 13 sites in the 2016 TRI (	) and 14 sites in the 2016 DMR (

2016b) that may be using PCE-based maskants. EPA assessed water releases from these sites using the
annual reported discharges from each site. In the 2016 TRI, only 2 of the 13 sites reported non-zero
discharges with both sites reporting discharges to POTW (U.S. EPA.! ). In the 2016 DMR, only 3
sites reported non-zero direct discharges to surface water (indirect discharges not reported in DMR) and
the remaining 11 sites reported zero direct discharges (	2016b).

To estimate the daily release, EPA used operating data from the 2014 NEI. In the 2014 NEI, there were
six sites reporting maskant operations with four reporting specifically for PCE and two reporting for
VOC only (U.S. EPA. 2016a). Each site provided operating hours per year for the masking operation.
EPA assumed eight hours per day of operation to calculate the number of operating days per year. If
assuming eight hours per day resulted in over 365 operating days per year; EPA assumed 24 operating
hours per day. EPA then mapped the operating days from NEI directly to sites in TRI and DMR where
possible and used the sites reported operating time to estimate daily releases. For sites that did not report
to NEI, EPA used the average operating time of 4,130 hr/yr and 24 hr/day to estimate the daily release
(	2016a). Release estimates for the five sites with water releases are presented in Table 2-57.

Page 146 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Table 2-57. Reported Wastewater Discharges of Perchloroethylene from Chemical Maskant Sites

Site

Annual
Release
(kg/day)

Operating

Days
(days/yr)

Daily
Release
(kg/year)11

NPDKS
Code

Release
Media/
Treatment
Kacility
Type

Source

Alliant Techsystems
Operations LLC, Elkton,
MDb

1.01E-03

172

5.86E-06

MD0000078

Surface
Water

(U.S.

EPA,
2016b)

Ducommun
Aerostructures Inc
Orange Facility, Orange,
CAb

0.5

172

2.64E-03

Not
available

POTW

(U.S.

EPA,
- )

GE Aviation, Lynn, MAb

0.1

172

8.62E-04

MA0003905

Surface
Water

(U.S.

EPA,
2016b)

McCanna Inc.,
Carpentersville, ILb

6.96E-02

172

4.05E-04

IL0071340

Surface
Water

(U.S.

EPA,
2016b)

Weatherford Aerospace
LLC, Weatherford, TXC

2.3

208

1.09E-02

Not
available

POTW

(U.S.

EPA,
- )

POTW = Publicly-Owned Treatment Works; WWT = Wastewater Treatment

a Daily releases are back-calculated from the annual release rate reported in the 2016 TRI or 2016 DMR and the operating
days.

b Operating days for these sites are based on the average operating time of 4,130 hr/yr and assuming 24 hr/day.

0 Operating days for this site is based on the sites reported operating time of 4,992 hr/yr in the 2014 NEI and assuming 24
hr/day.

Sources: (U.S. EPA. 2017d. 2016b)

It should be noted, that the majority of the sites identified from TRI and DMR as sites using PCE-based
maskants are based on reported NAICS and SIC codes, activities reported in TRI, and, where available,
information from public comments and stakeholder meetings. There is the potential that these sites
perform a different activity (e.g. metal degreasing) instead of or in addition to chemical milling
operations. Water releases in TRI and DMR are reported at the site level, not the operation level;
therefore, they are only considered under one expected condition of use to avoid double counting.

EPA did not identify any data to estimate releases from the other 44 sites that use PCE-based maskants.
However, sites that use PCE-based maskants are expected to be regulated by the Metal Finishing EG
(	2019c). As discussed in Section 2.5 for OTVDs, the Metal Finishing EG sets a discharge

limit for TTO concentration in wastewater stream not a PCE-speciftc limit. The Metal Finishing EG sets
a one-day maximum TTO discharge limit of 2.13 mg/L for BPT, BAT, PSES, NSPS, and PSNS.
Therefore, the concentration of PCE in wastewater streams using PCE-maskants is expected to be below
the TTO limit.

Page 147 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

2.14 Industrial Processing Aid

2.14.1	Estimates of Number of Facilities

To determine the number of sites that use PCE as a processing aid, EPA considered 2016 CDR (U.S.

), 2016 TRI (	), and 2016 DMR (	>016b). In the 2016 CDR, two

sites reported use of PCE as a processing aid in the industrial processing and use section (

2016d)20. Each site reported use as a processing aid in the petrochemical manufacturing industry and the
pesticide, fertilizer, and other agricultural chemical manufacturing (	i). Each site

reported the number sites using PCE as a processing aid as not known or reasonably ascertainable (U.S.

).

Based on the activities and NAICS codes reported in the 2016 TRI, EPA identified 64 facilities where
the primary condition of use is expected to be use of PCE as a processing aid (	). In the

2016 DMR data, there are 41 sites for which EPA expects the primary condition of use to be use of PCE
as a processing aid with seven sites being the same as TRI sites (	¥). NAICS and SIC

codes assumed to be using PCE as a processing aid include those related to petrochemical
manufacturing, agricultural product manufacturing (based on CDR reporting) and petroleum refineries
(for catalyst regeneration—see process description in Section 2.14.2). Based on the DMR and TRI data,
EPA assesses a total of 99 sites (64+41 = 105 sites - 7 duplicate sites = 98 sites) for the use of PCE as a
processing aid.

2.14.2	Process Description

According to the TRI Reporting Forms and Instructions (RFI) Guidance Document, a processing aid is a
"chemical that is added to a reaction mixture to aid in the manufacture or synthesis of another chemical
substance but is not intended to remain in or become part of the product or product mixture is otherwise
used as a chemical processing aid. Examples of such chemicals include, but are not limited to, process
solvents, catalysts, inhibitors, initiators, reaction terminators, and solution buffers" (	8c).

Additionally, processing aids are intended to improve the processing characteristics or the operation of
process equipment, but not intended to affect the function of a substance or article created (

2016c).

One processing aid use of PCE is for catalyst regeneration at petroleum refineries (American Fuel and
Petroleum Manufacturers. .«l , < Chemical. 2008). According to public comments from the
American Fuel and Petrochemical Manufacturers (2017). PCE is used in both the reforming and
isomerization processes at refineries. In the reforming process, PCE is added directly to a regenerator in
a Continuous Catalytic Regeneration reforming unit, and in the isomerization process, PCE is added to
the hydrocarbon feed (American Fuel and Petroleum Manufacturers. 2017). In both processes, PCE
provides chlorine ions to regenerate the catalysts and is consumed in the process (American Fuel and
Petroleum Manufacturers. 2017). Other specific processing aid uses of PCE were not identified;
however, EPA expects use as a process solvent to be amongst the major processing aid uses.

20 The industrial processing and use section of CDR is reported by manufacturers/importers of a chemical for the downstream
uses of the chemical. This includes processing and use activities at both the manufacture/import site and at customer sites.
Therefore, the total number of sites related to these reported uses may be equal to or greater than the number of CDR
reporting sites.

Page 148 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

2.14.3 Exposure Assessment

2.14.3.1	Worker Activities

At industrial facilities, workers are potentially exposed when unloading PCE from transport containers
into intermediate storage tanks and process vessels. Workers may be exposed via inhalation of vapor or
via dermal contact with liquids while connecting and disconnecting hoses and transfer lines. Once PCE
is unloaded into process vessels, it may be consumed in the process (e.g. when used for catalyst
regeneration) or be used until spent and sent for disposal.

ONUs are employees who work at the facilities that process and use PCE, but who do not directly
handle the material. ONUs may also be exposed to PCE but are expected to have lower inhalation
exposures and are not expected to have dermal exposures. ONUs for this condition of use may include
supervisors, managers, engineers, and other personnel in nearby production areas.

2.14.3.2	Number of Potentially Exposed Workers

EPA estimated the number of workers and occupational non-users potentially exposed during use of
PCE as a processing aid using Bureau of Labor Statistics' OES data (	2016) and the U.S.

Census' SUSB (U.S. Census Bureau. ) as well as the primary NAICS and SIC code reported by
each site in the 2016 TRI (	) or 2016 DMR (	b), respectively. The

employment data from the U.S. Census SUSB and the Bureau of Labor Statistics' OES data are based
on NAICS code; therefore, SIC codes reported in the 2016 DMR had to be mapped to a NAICS code to
estimate the number of workers. A crosswalk of the SIC codes to the NAICS codes used in the analysis
are provided in Table 2-58.

Table 2-58. Crosswalk of Processing Aid SIC Codes in DMR to NAICS Codes

Sl( ( ode

Corresponding NAICS Code

2865 - Cyclic Organic Crudes and Intermediates,
and Organic Dyes and Pigments

325194 - Cyclic Crude, Intermediate, and Gum
and Wood Chemical Manufacturing

2879 - Pesticides and Agricultural Chemicals, Not
Elsewhere Classified

325320 - Pesticide and Other Agricultural
Chemical Manufacturing

2911 - Petroleum Refining

324110 - Petroleum Refineries

2999 - Products of Petroleum and Coal, Not
Elsewhere Classified

324199 - All Other Petroleum and Coal Products
Manufacturing

Table 2-59 provides a summary of the reported NAICS codes (or NAICS mapped to the reported SIC
code), the number of sites reporting each NAICS code, and the estimated number of workers and ONUs
for each NAICS code as well as an overall total for use of PCE as a processing aid. It should be noted,
that in the 2016 DMR, two sites did not report a SIC code. To estimate the number of workers and
ONUs from these sites EPA calculated the average number of workers and ONUs per site from the other
known sites. There are approximately 14,000 workers and 6,000 ONUs potentially exposed during use
of PCE as a processing aid. The NAICS code 324110 for petroleum refineries has significantly more
workers and ONUs per site than other NAICS codes. This is likely due to the size of petroleum
refineries as compared to other chemical manufacturing industries. Refineries tend to be larger with
multiple process areas which may result in higher number of employees.

Page 149 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Table 2-59. Estimated Number of Workers Potentially Exposed to Perchloroethylene During Use
as a Processing Aid	

NAICS Cotlc

Number
of Sites

Kxposed
Workers
per Site"

Kxposed
Occupational
Non-l sers
per Site"

Total
Exposed
Workers

Total
Kxposed
Occupational
Non-l sers

Total
Kxposed

212393

1

24

6

24

6

30

324110

76

170

75

12,934

5,732

18,667

324199

3

17

8

52

23

75

325180

1

25

12

25

12

37

325199

7

34

16

239

113

352

325199

2

39

18

77

36

114

325311

1

17

5

17

5

23

325320

5

25

7

127

37

165

Unknown
NAICS

2

44

18

88

37

125

Totalb

99

140

61

14,000

6,000

20,000

a Number of workers and occupational non-users per site are calculated by dividing the exposed number of workers or
occupational non-users by the number of establishments in the relevant NAICS codes. The workers/ONUs per site are then
multiplied by the number of sites within that NAICS to get the total exposed. The number of workers/ONUs per site is
rounded to the nearest integer.

b Totals have been rounded to two significant figures. Totals may not add exactly due to rounding.

2.14.3.3 Occupational Exposure Results

EPA identified inhalation exposure monitoring data from four studies submitted to EPA under TSCA by
Dow Chemical (Dow Chem Co. 1983a. b, 1982. 1979). The exact function of PCE is each study is not
explicitly stated; however, the data was collected in the agricultural chemical production and
distribution, trichloroethylene production, and chloropyridines process areas. Based on CDR reporting,
PCE is used as a processing aid in agricultural chemical manufacturing; therefore, monitoring data
collected in the agricultural chemical production area is assessed as a processing aid use of PCE.
Similarly, chloropyridines are used as intermediates in both the pharmaceutical and agrochemical
industries (Scriven and Murugan. 2005). Both pharmaceutical and agrochemical industries are expected
to use PCE as a processing aid; therefore, monitoring data collected in the chloropyridine unit are also
assessed as a processing aid use. PCE can also be used as an inert material in trichloroethylene
production (Snedecor et at.. 2004). Use as an inert material would fall under processing aid uses;
therefore, monitoring data collected during trichloroethylene production is assessed as a processing aid
use.

Data were collected for a variety of workers in the process areas including operators, tank truck loading,
pipefitters, foreman, and technicians (Dow Chem Co. 1983a. b, 1982. 1979). Sample times ranged from
approximately 5.5 to 8 hours (Dow Chem Co. 1983a. h, rs82, 1979). Where sample times were less than

Page 150 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

eight hours, EPA calculated an 8-hr TWA assuming zero exposure outside the sample time. The data set
also included 22 data points for which the exact sample time was not provided; however, the submitted
studies indicate the results were calculated 8-hr TWAs; therefore, EPA included with the 8-hr TWAs
calculated from data with known sample times.

Table 2-60 presents a summary of the identified 8-hr TWA and 30-minute TWA monitoring data. For
the 8-hr TWA, the 95th percentile is presented as the high-end and the 50th percentile presented as the
central tendency. It should be noted that approximately 55% of the 8-hr TWA data were below the LOD.
To estimate exposure concentrations for these data, EPA followed the Guidelines for Statistical Analysis
of Occupational Exposure Data (U,	4b) as discussed in Section 1.4.5.2. The geometric

standard deviation for the data was above 3.0; therefore, EPA used the to estimate the exposure

value as specified in the guidelines (	|b). Because over 50% of the data are below the

LOD, calculating statistics from this data does present the potential to introduce biases into the results.
Estimation of exposure values for results below the LOD may over- or under-estimate actual exposure
thus skewing the calculated statistics higher or lower, respectively. The overall directional bias of the
exposure assessment, accounting for both the overestimate and underestimate, is not known.

For the 30-minute TWA, only two data point were available, one of which measured below the LOD.
Because only a single data point with a measured value was available, EPA could not calculate a
geometric standard deviation. Therefore, EPA presents two scenarios: 1) using the maximum as a
"higher value"; and 2) using the midpoint between the maximum and the LOD as a "midpoint" value.
These scenarios are plausible, but EPA cannot determine the statistical representativeness of the value.

Table 2-60. Summary of Worker inhalation Monitoring Data for Use of Perchloroethylene as a
Processing Aid	

Scenario

8-hr
TWA
(PPm)

AC
(ppm)

A IK

(ppm)

I.AIK
(ppm)

N il in her
of Data
Points

30-m ink*
TWA
(ppm)11

Nil in her
of Data
Points

High-End

1.2

0.4

0.3

0.1

89

2.2

2

Central Tendency

6.00E-02

2.00E-02

1.37E-02

5.44E-03

1.7

AC = Acute Concentration; ADC = Average Daily Concentration; and LADC = Lifetime Average Daily Concentration.
Equations and parameters for calculation of the AC, ADC, and LADC are described in Appendix B.
a Due to only two data points, one of which measured below the LOD, EPA presents two scenarios: 1) using the higher of the
two values; and 2) using the midpoint of the LOD and the maximum.

Source: (Dow Chem Co. 1983a. b, .1.982. 1979)

2.14.4 Water Release Assessment

2.14.4.1 Water Release Sources

Potential sources of water releases are expected to be similar to those described in Section 2.1.4.1 for
manufacturing and may include the following: equipment cleaning operations, aqueous wastes from
scrubbers/decanters, reaction water, process water from washing intermediate products, and trace water
settled in storage tanks (01	).

Page 151 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

2.14.4.2 Water Release Assessment Results

EPA assessed water releases using the annual discharge values reported to the 2016 TRI (

2017d) and the 2016 DMR (	b) by the 98 sites using PCE as a processing aid. In the 2016

TRI, 11 sites reported non-zero direct discharges to surface water and two of these 11 sites also reported
indirect discharges to POTW (	). All other sites in TRI reported zero direct or indirect

discharges (U.S. EPA. ). In the 2016 DMR, one sites reported a direct discharge to surface water
(indirect discharges not reported in DMR data) and the remainder reported zero indirect discharges (U.S.
).

To estimate the daily release, EPA assumed 300 days/yr of operation as given in the SpERC developed
by the European Solvent Industry Group for the manufacture of a substance (which includes use as a
process chemical or extraction agent) and averaged the annual release over the operating days (European
Solvents Industry Group. 2012). Table 2-61 summarizes the water releases from the 2016 TRI and DMR
for sites with non-zero discharges.

Table 2-61. Reported Wastewater Discharges of Perchloroethylene from Processing Aid Sites

Site

Annual
Release"
(kg/site-
vear)

Annual
Release

Days
(davs/vr)

Daily
Release
(kg/sit e-
dav)11

NPDKS ( ode

Release
Media/
Treatment
Facility Type

Source

Chevron Products
Co - Salt Lake
Refinery, Salt Lake
City, UT

1.74

300

0.01

UTG070261

Surface Water

(U.S. EPA.
2017d)

Chevron Products
Co Richmond
Refinery,
Richmond, CA

0.91

300

3.02E-03

CA0005134

Surface Water

(U.S. EPA.
2017d)

CHS McPherson
Refinery,
McPherson, KS

0.09

300

3.02E-04

KS0000337

Surface Water

(U.S. EPA.
2017d)

ExxonMobil Oil
Beaumont Refinery,
Beaumont, TX

7.26

300

0.02

Not available

Surface Water

(U.S. EPA.
2017d)

HollyFrontier El
Dorado Refining
LLC, El Dorado,
KS

0.91

300

3.02E-03

KS0000761

Surface Water

(U.S. EPA.
2017d)

Hunt Refining Co -
Tuscaloosa
Refinery,
Tuscaloosa, AL

4.01

300

0.01

AL0000973

Surface Water

(U.S. EPA.
2016b)

Page 152 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Site

Annual
Release11
(kg/si fo-
vea r)

Annual
Release

Days
(davs/vr)

Dailv
Release
(k«/sifo-
dav)11

NPDKS Code

Release
Media/
Treatment
Kacilitv Type

Source

Marathon
Petroleum Co LP,
Garyville, LA

2.72

300

0.01

LAU009485

Surface Water

ru.s. EPA.
2017d)

Occidental
Chemical Corp
Niagara Plant,
Niagara Falls, NY

25.85

300

0.09

NY0003336

Surface Water

ru.s. EPA.
2017d)

26.31

0.09

POTW

Tesoro Los Angeles
Refinery-Carson
Operations, Carson,
CA

0.45

300

1.51E-03

CA0000680

Surface Water

ru.s. EPA.
2017d)

108.0

0.36

POTW

The Dow Chemical
Co, Midland, MI

10.43

300

0.03

MI0000868

Surface Water

ru.s. EPA.
2017d)

Valero Refining Co
-Oklahoma Valero
Ardmore Refinery,
Ardmore, OK

2.27

300

0.01

OK0001295

Surface Water

ru.s. EPA.
2017d)

Valero Refining-
Texas LP Corpus
Christi West Plant,
Corpus Christi, TX

2.73

300

0.01

TX0063355

Surface Water

ru.s. EPA.
2017d)

POTW = Publicly-Owned Treatment Works; WWT = Wastewater Treatment

a Annual release amounts are based on the site reported values. Therefore, daily releases are calculated from the annual
release rate and assuming 300 days of operation per year.

Sources: (U.S. EPA. 2017d. 2016b)

2.15 Metalworking Fluids

2.15.1 Estimates of Number of Facilities

EPA did not identify information to estimate the number of facilities using metalworking fluids
containing PCE using information from CDR, TRI, or DMR or systematic review. However, sites using
metalworking fluids likely fall into similar NAICS codes as those identified for vapor degreasing/cold
cleaning. Therefore, it is possible that sites assessed under one of those conditions of use actually
perform metalworking activities rather than or in addition to degreasing. However, a HSIA (2008) report
estimated no more than 3% of the national PCE production volume is used for "miscellaneous" uses
(which includes metalworking fluids) compared to 7% for metal degreasing. Therefore, EPA expects the
majority of those sites to be performing degreasing activities with PCE rather than metalworking

Page 153 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

activities and they are not considered again here to avoid double counting21. It should be noted that only
a single PCE-based metalworking fluid product was identified (see process description in Section
2.15.2); therefore, the number of sites using PCE-containing metalworking fluids is expected to be
small.

2.15.2	Process Description

EPA identified one cutting fluid product in the Preliminary Information on Manufacturing, Processing,
Distribution, Use, and Disposal for PCE (	) that contains PCE. The safety data sheet

(SDS) and the company's product page indicate that PCE is present at <10 wt% in the formulation and
that the product's recommended use is an oil-based cutting and tapping fluid for use with copper, iron,
aluminum and magnesium materials (MSC Industrial Supply Inc.. JO I's; Winfield Brooks Company.
2014). Metalworking, cutting, and tapping fluids are all used in various metal shaping operations.

Cutting and tapping fluids are a subset of metalworking fluids that are used for the machining of internal
and external threads using cutting tools like taps and thread-mills (OE	). While some cutting

and tapping fluids may be used by consumers in a DIY setting, there is no indication that this product is
marketed solely to consumers, therefore, EPA assesses the industrial use of metalworking fluids in the
metal products and machinery (MP&M) industry. In general, industrial metal shaping operations include
machining, grinding, deformation, blasting, and other operations and may use different types of
metalworking fluids to provide cooling and lubrication and to assist in metal shaping and protect the part
being shaped from oxidation (	).

The OECD ESD on the Use of Metalworking Fluids (OECD. 2 ) provides a generic process
description of the industrial use of both water-based and straight oil metalworking fluids in the MP&M
industry. Based on the recommended use of "oil-based cutting and tapping fluid" listed in the SDS, EPA
assesses as a straight oil (Winfield Brooks Company. 2014). Metalworking fluids are typically received
in containers ranging from 5-gallon pails to bulk containers (	). Straight oils are transferred

directly into the trough of the metalworking machine without dilution (OECD. 2011c). The
metalworking fluids are pumped from the trough and usually sprayed directly on the part during metal
shaping (OECD, 2011). The fluid stays on the part and may drip dry before being rinsed or wiped clean.
Any remaining metalworking fluid is usually removed during a cleaning or degreasing operation
(OEC	).

2.15.3	Exposure Assessment

2.15.3.1 Worker Activities

Workers are expected to unload the metalworking fluid from containers; clean containers; dilute water-
based metalworking fluids; transfer fluids to the trough; performing metal shaping operations; rinse,
wipe, and/or transfer the completed part; change filters; transfer spent fluids; and clean equipment
( )•

ONUs include employees that work at the site where PCE is used in an industrial setting as a
metalworking fluid, but they typically do not directly handle the chemical and are therefore expected to

21 This statement is in reference to activities that involve PCE at each site. EPA expects that many sites may have both
metalworking and degreasing activities. The assumption is only that most of the sites use PCE as a degreasing solvent rather
than as a metalworking fluid, not whether metalworking activities (using non-PCE containing metalworking fluids) are
actually occurring at the site.

Page 154 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

have lower exposures. ONUs for metalworking fluids include supervisors, managers, and tradesmen that
may be in the processing area but do not perform tasks that result in the same level of exposures as
machinists.

Since PCE has a high vapor pressure (18.5 mmHg at 25°C), workers may be exposed to PCE when
handling liquid metalworking fluid, such as unloading, transferring, and disposing spent metalworking
fluids and cleaning machines and troughs. The greatest source of potential exposure is during metal
shaping operations. The high machine speeds can generate airborne mists of the metalworking fluids to
which workers can be exposed. Additionally, the high vapor pressure of PCE may lead to its evaporation
from the airborne mist droplets, potentially creating a fog of vapor and mist.

2.15.3.2	Number of Potentially Exposed Workers

The ESD on the Use of Metalworking Fluids cites a NIOSH study of 79 small machine shops, which
observed an average of 46 machinists per site ("OECD. , ). The ESD indicates that the "small" shops
refer to sites that machine a variety of products according to customer orders, rather than sites
manufacturing a large quantity of the same part (e.g., automobile part manufacturing) (	).

The ESD also cites an EPA effluent guideline development for the MP&M industry, which estimated a
single shift supervisor per shift, who may perform tasks such as transferring and diluting neat
metalworking fluids, disposing spent metalworking fluids, and cleaning the machines and troughs
( )•

Since the machinists perform the metal shaping operations, during which metalworking fluid mists are
generated, EPA assesses the machinists as workers, as they have the highest potential exposure. EPA
assessed the single shift supervisor per site as an ONU, as this employee is not expected to have as high
an exposure as the machinists. Assuming two shifts per day (hence two shift supervisors per day), EPA
assesses 46 workers and two ONUs per site (OE	). Although, per the ESD, it is possible the

shift supervisors may perform some tasks that may lead to direct handling of the metalworking fluid,
EPA assesses these shift supervisors as ONUs as their exposures are expected to be less than the
machinist exposures and EPA is assessing the machinists as workers, which yields a high worker-to-
ONU ratio of 23-to-l (	). The number of establishments that use PCE-based metalworking

fluids is unknown; therefore, EPA does not have data to estimate the total workers and ONUs exposed to
PCE from use of metalworking fluids.

2.15.3.3	Occupational Exposure Results

EPA did not identify any inhalation exposure monitoring data related to the use of PCE-based
metalworking fluids. Therefore, EPA assessed inhalation exposures using the ESD on the Use of
Metalworking Fluids (OE(	). The ESD estimates typical and high-end exposures for different

types of metalworking fluids. These estimates are provided in Table 2-62 and are based on a NIOSH
study of 79 small metalworking facilities (OECD. 2 ). The concentrations for these estimates are for
the solvent-extractable portion and do not include water contributions (01	). The "typical"

mist concentration is the geometric mean of the data and the "high-end" is the 90th percentile of the data
( )•

Page 155 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Table 2-62. ESP Exposure Estimates for Metalworking Fluids Based on Monitoring Data

Type of .Metalworking l luid

Typical Mist Concentration
(nig/in^)11

Iligli-Knd Mist Concentration
(nig/nr')1'

Conventional Soluble

0.19

0.87

Semi-Synthetic

0.20

0.88

Synthetic

0.24

1.10

Straight Oil

0.39

1.42

a The typical mist concentration is the geometric mean of the data (OECD. 2011c').
b The high-end mist concentration is the 90th percentile of the data (OECD. 2011c').
Source: (OECD. 20.1.1c')

The recommended use of the PCE-based metalworking fluid is an oil-based cutting and tapping fluid;
therefore, EPA assesses exposure to the PCE-based metalworking fluids using the straight oil mist
concentrations and the max concentration of PCE in the metalworking fluid. Straight oils are not diluted;
therefore, the concentration of PCE specified in the SDS (<10%) is equal to the concentration of PCE in
the mist. However, it should be noted that due to the evaporation of PCE from the metalworking fluid,
the actual concentration of PCE in the mist is expected to be less than the 10% estimated in the
metalworking fluid, resulting in an overestimate of exposure to PCE in the mist. Table 2-63 presents the
exposure estimates for the use of PCE-based metalworking fluids. The ESD estimates an exposure
duration of eight hours per day; therefore, results are presented as 8-hr TWA exposure values.

It should be noted that these estimates may underestimate exposures to PCE during use of metalworking
fluids as they do not account for exposure to PCE that evaporates from the mist droplets into the air.

This exposure is difficult to estimate and is not considered in this assessment. However, due to the
relatively low concentration of PCE in the metalworking fluid, the partial pressure may be low enough
such that evaporation of PCE from the mist is limited and this not a significant route of exposure.

Table 2-63. Summary of Exposure Results for Use of PCE in Metalworking Fluids Based on ESD
Estimates

Scenario

8-hr TWA
(ppm)11

AC
(ppm)

ADC
(ppm)

I.AIK
(ppm)

High-End

2.09E-02

6.98E-03

4.78E-03

2.45E-03

Central Tendency

5.75E-03

1.92E-03

1.31E-03

5.22E-04

AC = Acute Concentration; ADC = Average Daily Concentration; and LADC = Lifetime Average Daily Concentration.
Equations and parameters for calculation of the AC, ADC, and LADC are described in Appendix B.
a The PCE exposure concentrations are calculated by multiplying the straight oil mist concentrations in Table 2-62 by 10%
(the concentration of PCE in the metalworking fluid) and converting to ppm.

2.15.4 Water Release Assessment

2.15.4.1 Water Release Sources

The ESD states that water releases from use of straight oil metalworking fluids may come from disposal
of container residue and dragout losses from cleaning the part after shaping (i	). Facilities

Page 156 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

typically treat wastewater onsite due to stringent discharge limits to POTWs (QE	). Control

technologies used in onsite wastewater treatment in the MP&M industry include ultrafiltration, oil/water
separation, and chemical precipitation (	). Facilities that do not treat wastewater onsite

contract waste haulers to collect wastewater for off-site treatment (QE	).

2.15.4.2 Water Release Assessment Results

EPA assesses water release using TRI and DMR data. However, EPA cannot distinguish between sites
using metalworking fluids and sites using PCE in degreasers in TRI and DMR data; therefore, a single
set of water release for degreasing and metalworking fluid operations is presented in Section 2.5.4.2 for
OTVDs.

2.16 Wijpe Cleaning and Metal/Stone Polishes

2.16.1	Estimates of Number of Facilities

EPA did not identify information from EPA databases (e.g., CDR, TRI, or DMR) or in the results of the
systematic review process to estimate the number of sites using PCE for wipe cleaning and metal/stone
polishes. It is possible some sites using vapor degreasers or cold cleaners also use PCE for wipe
cleaning.

2.16.2	Process Description

PCE can be used as a solvent in non-aerosol degreasing and cleaning products. Non-aerosol cleaning
products typically involve dabbing or soaking a rag with cleaning solution and then using the rag to
wipe down surfaces or parts to remove contamination (	2014). The cleaning solvent is usually

applied in excess and allowed to air-dry (U.S. EPA. 2014). Parts may be cleaned in place or removed
from the service item for more thorough cleaning (U.S. EPA. 2014).

2.16.3	Exposure Assessment

2.16.3.1	Worker Activities

Workers are expected to be exposed to PCE vapors that evaporate from the PCE-soaked rag or the
solvent residue left behind on the substrate after wiping/polishing. Additional activities and use patterns
will vary depending on the specific site at which PCE product is being used.

2.16.3.2	Number of Potentially Exposed Workers

EPA did not identify information to estimate the number of workers or ONUs exposed to PCE during
use for wipe cleaning and metal/stone polishing. Wipe cleaning and metal/stone polishes can be used in
a large variety of industries that cannot be drilled down to a specific set of NAICS codes. Additionally,
EPA does not have information on market penetration to estimate number of workers even if a set of
"likely" NAICS codes were identified. Therefore, methodologies used in other conditions of use to
estimate workers cannot be used for this condition of use. It is possible some workers/ONUs at sites
using vapor degreasers or cold cleaners are also exposed to PCE from wipe cleaning activities.

2.16.3.3	Occupational Exposure Results

EPA identified inhalation exposure monitoring data from NIOSH investigations at two sites using PCE
for wipe cleaning. EPA did not identify exposure data specific to metal/stone polish applications;
therefore, these data were also used to assess the use of metal/stone polishes based on expected
similarities in the uses. Due to the large variety in shop types that may use PCE as a wipe cleaning

Page 157 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

solvent or metal/stone polish, it is unclear how representative these data are of a "typical" site. EPA does
not have a model for estimating exposures from wipe cleaning or metal/stone polish; therefore, the
assessment is based on the identified monitoring data. Table 2-64 summarizes 8-hr, 4-hr and 15-minute
TWA monitoring data for the use of PCE as a wipe cleaning solvent and metal/stone polish. Due to the
limited number of data points for workers 8-hr and 15-minute TWA results, the maximum of identified
data is presented as the high-end and the median is presented as the central tendency. There is only a
single 4-hr TWA data point for workers. Results based on a single value are plausible, but EPA cannot
determine the statistical representativeness of the value. For the ONU 8-hr TWA, the 95th percentile is
presented as the high-end and the 50th percentile as the central tendency. The ONU data included four
data points that are below the LOD. To estimate exposure concentrations for these data, EPA followed
the Guidelines for Statistical Analysis of Occupational Exposure Data (	2) as discussed

in Section 1.4.5.2. The geometric standard deviation for the data was above 3.0; therefore, EPA used the
to estimate the exposure value as specified in the guidelines (	|b).

The data were obtained from NIOSH HHEs conducted at a taxidermy shop and an air filter manufacturer
(NIOSH. 1983b. 1979). At the taxidermy shop, workers hand-rub a mixture of crushed corn cobs and
PCE on the animal fur to remove oils and fats that are deposited during the mounting process (NIOSH.
1979). The mixture is then removed from the fur by blowing it with air (NIOSH. 1979). The study notes
that the entire taxidermy process is done without adequate ventilation (NIOSH hr9). The sample times
ranged from approximately one hour to four hours; however, the study states that workers perform the
same task throughout the work shift and the exposures are representative of an eight-hour exposure
(NIOSH. 1979). Therefore, EPA consider these data as 8-hr TWA exposure values.

The air filter manufacture shop manufactures engine and machine filters for airplanes, trucks, railroads,
tank engines, and office machines (NIQS 3b)- Workers use rags dampened with solvent (primarily
PCE) to clean excess resins and adhesives from metal parts (NIOSH. 1983b). The study indicates that
workers are rotated between activities as an administrative control (NIQS 3b). The study collected
one three-hour from a worker and four 15-mintue samples. EPA calculated both 4-hr and 8-hr TWAs
from the three-hour sample assuming zero exposure outside the sample time.

Page 158 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Table 2-64. Summary of Worker Inhalation Monitoring Data for Use of Perchloroethylene as a
Wipe Cleaning Solvent and Metal/Stone Polishes	

Scenario

8-hr
TWA
(ppni)

AC
(ppni)

ADC
(ppm)

I.ADC

(ppm)

Nil m her
of Data
Points

4-hr
TWA
(ppm
)

Nil m be
r of
Data
Points

15-
m in ill
e

TWA

Number
of Data
Points

Worker Monitoring Data

High-End

228

76

52

27

4

9.5

1

103

4

Central
Tendency

132

44

30

12

66

Occupational Non-User Monitoring Data

High-End

23

7.7

5.3

2.7

6

No 4-hr or 15-minute data
identified for ONUs

Central
Tendency

2.18E-02

7.28E-03

4.98E-03

1.98E-03

AC = Acute Concentration; ADC = Average Daily Concentration; and LADC = Lifetime Average Daily Concentration.
Equations and parameters for calculation of the AC, ADC, and LADC are described in Appendix B.

Source: (NIOSH. 1983b. 1979)

2.16.4 Water Release Assessment

EPA does not expect releases of PCE to water from the use of PCE as a wipe cleaning solvent and
metal/stone polishes. Due to the volatility of PCE the majority of releases from the use of wipe cleaning
and metal/stone polish products will likely be to air as PCE evaporates from the rag/cloth used to apply
the solvent and the substrate surface. EPA expects any PCE residue that remains in the container or on
the rag/to be disposed of with shop trash that is either picked up by local waste management or by a
waste handler that disposes shop wastes as hazardous waste. There is a potential that PCE may drip from
the rag/cloth or the substrate surface onto shop floors or ground (for outdoor applications) and could
possibly end up in a floor drain (if the shop has one) or runoff into surface water or stormwater drains.
However, EPA expects the potential release to water from this to be minimal as there would be time for
PCE to evaporate before entering these pathways. This is consistent with estimates from the
International Association for Soaps, Detergents and Maintenance Products (AISE) SpERC for Wide
Dispersive Use of Cleaning and Maintenance Products, which estimates 100% of volatiles are released
to air (International Association for Soaps Detergents and Maintenance Products. 2012).

2.17 Other Spot Cleaning/Spot Removers (Including Carpet Cleaning)

2.17.1 Estimates of Number of Facilities

EPA did not identify information from EPA databases (e.g., CDR, TRI, or DMR) or in the results of the
systematic review process to estimate the number of sites using PCE in other spot cleaning/spot
removers.

Page 159 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

2.17.2	Process Description

EPA expects the use of spot cleaners/spot removers to involve spray applying PCE to the stained textile
(e.g., carpet) and then using brush or fingers to scrape away the stain. This condition of use includes
both professional carpet cleaning activities as well as spot cleaning activities at textile mills.

2.17.3	Exposure Assessment

2.17.3.1	Worker Activities

As previously described, workers are expected to spray PCE on to the stained textiles and then manually
scrape away the stain using a brush or fingers.

2.17.3.2	Number of Potentially Exposed Workers

EPA did not identify information from the systematic review process to estimate the total number of
workers and ONUs exposed from use of spot cleaners/spot removers. However, both the Fabric
Finishing GS (	la) and the ESD on the Use of Textile Dyes (OECD. 2017b) estimate three

to six workers exposed per site. It is unknown how many of those workers may be involved in the spot
cleaning process.

2.17.3.3	Occupational Exposure Results

EPA identified inhalation exposure monitoring data from a single NIOSH investigation at a garment
manufacturer. It is unclear how representative these data are of a "typical" spot cleaning/spot remover
scenario. The site had two spotting stations in the finishing department used to remove stains from
garments on an "as needed" basis (NIQSi >). The investigation collected three samples from
workers in the finishing department with sample times ranging from approximately 5.5 to 9 hours
(NIOSH. 1996). Where sample times were less than eight hours, EPA converted to an 8-hr TWA
assuming exposure outside the sample time was zero. Table 2-65 summarizes the 8-hr TWA monitoring
data for the use of PCE in spot cleaners/spot removers.

Page 160 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Table 2-65. Summary of Worker Inhalation Exposure Monitoring Data for Other Spot
Cleaning/Spot Removers (Including Carpet Cleaning)	

Scenario

8-hr TW A
(ppm)

AC

(ppm)

A IK

(ppm)

LAIK
(ppm)

Number of
Data Points

Worker Monitoring Data

Higher Valuea

0.2

7.69E-02

5.27E-02

2.70E-02

2

Midpoint Valuea

0.2

5.73E-02

3.92E-02

1.56E-02

Occupational Non-User Monitoring Data

High-Endb

3.00E-02

1.00E-02

6.85E-03

3.51E-03

1

Central Tendency13

3.00E-02

1.00E-02

6.85E-03

2.72E-03

AC = Acute Concentration; ADC = Average Daily Concentration; and LADC = Lifetime Average Daily Concentration.
Equations and parameters for calculation of the AC, ADC, and LADC are described in Appendix B.
a Due to only two data points identified for workers, EPA presents two scenarios: 1) using the higher of the two values; and
2) using the midpoint of the two values.

b Only one data point identified for ONUs. However, different parameters are used for calculating high-end and central
tendency ADC and LADC. Therefore, a high-end and central tendency are presented based on the single data point.

Source: fNIOSH. 19961

2,17,4 Water Release Assessment

EPA does not have information to estimate the releases to water from spot cleaners/spot removers. The
Fabric Finishing GS (	a) and ESD on the Use of Textile Dyes (OECD. 2017b) do not

address potential releases from spot cleaners. Due to the volatility of PCE, EPA expects the primary
release of PCE to be to air. However, there is the possibility that PCE deposits into water streams at
textile plants and is subsequently discharged other directly to water or indirectly to POTW or non-
POTW WWT.

2.18 Other Industrial Uses

2.18.1	Estimates of Number of Facilities

To determine the number of sites that use PCE for other industrial uses, EPA considered 2016 TRI (

), and 2016 DMR (	<) data. EPA identified 19 facilities in the 2016 TRI and

111 facilities in the 2016 DMR where EPA could not determine the condition of use or the condition of
use falls into an industrial use discussed in Section 2.18.2. Therefore, EPA assessed a total of 130 sites
for use of PCE in "other industrial uses".

2.18.2	Process Description

Based on information identified in EPA's preliminary data gathering and information obtained from TRI
and DMR, a variety of other industrial uses of PCE may exist. Based on information in the Use
Document, market profile, and NAICS/SIC codes reported in TRI and DMR, examples of these uses
include, but are not limited to, uses in textile processing, wood furniture manufacturing, foundry
applications, food manufacturing, and scientific research and development (	1017a. c, d,

2016b). EPA did not identify information on how PCE may be used at these facilities.

Page 161 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

2.18.3 Exposure Assessment

2.18.3.1	Worker Activities

Although information on worker activities at these sites was not identified, EPA expects workers to
perform activities similar to other industrial facilities. Therefore, workers may potentially be exposed
when unloading PCE from transport containers into intermediate storage tanks and process vessels.
Workers may be exposed via inhalation of vapor or via dermal contact with liquids while connecting and
disconnecting hoses and transfer lines.

ONUs are employees who work at the facilities that process and use PCE, but who do not directly
handle the material. ONUs may also be exposed to PCE but are expected to have lower inhalation
exposures and are not expected to have dermal exposures. ONUs for this condition of use may include
supervisors, managers, engineers, and other personnel in nearby production areas.

2.18.3.2	Number of Potentially Exposed Workers

EPA estimated the number of workers and occupational non-users potentially exposed during other
industrial uses of PCE using Bureau of Labor Statistics" OES data (U.S. BLS. 2016) and the U.S.

Census" SUSB (U.S. Census Bureau. ) as well as the primary NAICS and SIC code reported by
each site in the 2016TRI (	) or 2016 DMR (U.S. EPA. 2016b"). respectively. The

method for estimating number of workers is detailed above in Section 1.4.4 and Appendix A. These
estimates were derived using industry- and occupation-specific employment data from the BLS and U.S.
Census. The employment data from the U.S. Census SUSB and the Bureau of Labor Statistics' OES data
are based on NAICS code; therefore, SIC codes reported in the 2016 DMR had to be mapped to a
NAICS code to estimate the number of workers. A crosswalk of the SIC codes to the NAICS codes used
in the analysis are provided in Table 2-66. In the 2016 DMR there was one site that did not report a SIC
code but after review of the company's website, EPA determined that NAICS 311411 - Frozen Fruit,
Juice, and Vegetable Manufacturing was the most appropriate NAICS code to use for this site.

Table 2-66. Crosswalk of Other Industrial Use SIC Codes in DMR to NAICS Codes

SIC ( ode

Corresponding NAICS Code

1041 - Gold Ores

211120 - Gold Ore Mining

1221 - Bituminous Coal and Lignite
Surface Mining

212111 - Bituminous Coal and Lignite Surface Mining

1311 - Crude Petroleum and Natural Gas

211120 - Crude Petroleum Extraction

1423 - Crushed and Broken Granite

212313 - Crushed and Broken Granite Mining and
Quarrying

1429 - Crushed and Broken Stone, Not
Elsewhere Classified

212319 - Other Crushed and Broken Stone Mining and
Quarrying

1442 - Construction Sand and Gravel

212321 - Construction Sand and Gravel Mining

2026 - Fluid Milk

311511 - Fluid Milk Manufacturing

2033 - Canned Fruits, Vegetables,
Preserves, Jams, and Jellies

311421 - Fruit and Vegetable Canning

Page 162 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

SIC Cotlc

Corresponding NAICS Code

2046 - Wet Corn Milling

311221 - Wet Corn Milling

2066 - Chocolate and Cocoa Products

311351 - Chocolate and Confectionery Manufacturing
from Cacao Beans

2082 - Malt Beverages

312120 - Breweries

2087 - Flavoring Extracts and Flavoring
Syrups, Not Elsewhere Classified51

311900 - Other Food Manufacturing

2099 - Food Preparations, Not Elsewhere
Classified51

311900 - Other Food Manufacturing

2611 - Pulp Mills

322110-Pulp Mills

2672 - Coated and Laminated Paper, Not
Elsewhere Classified

322220 - Paper Bag and Coated and Treated Paper
Manufacturing

2679 - Converted Paper and Paperboard
Products, Not Elsewhere Classified

322299 - All Other Converted Paper Product
Manufacturing

2812 - Alkalies and Chlorine

325180 - Other Basic Inorganic Chemical Manufacturing

2822 - Synthetic Rubber (Vulcanizable
Elastomers)

325212 - Synthetic Rubber Manufacturing

2823 - Cellulosic Manmade Fibers

325220 - Artificial and Synthetic Fibers and Filaments
Manufacturing

2824 - Manmade Organic Fibers, Except
Cellulosic

325220 - Artificial and Synthetic Fibers and Filaments
Manufacturing

2833 - Medicinal Chemicals and Botanical
Products

325411 - Medicinal and Botanical Manufacturing

2836 - Biological Products, Except
Diagnostic Substances

325414 - Biological Product (except Diagnostic)
Manufacturing

2892 - Explosives

325920 - Explosives Manufacturing

3264 - Porcelain Electrical Supplies

327110 - Pottery, Ceramics, and Plumbing Fixture
Manufacturing

3297 -Nonclay Refractories

327120 - Clay Building Material and Refractories
Manufacturing

4911 - Electric Services'3

221100 - Electric Power Generation, Transmission and
Distribution

5171 - Petroleum Bulk stations and
Terminals

424710 - Petroleum Bulk Stations and Terminals

8731 - Commercial Physical and
Biological Research0

541700 - Scientific Research and Development Services

Page 163 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

SIC ( ode

Corresponding NAICS Code

8733 - Noncommercial Research
Organizations0

541700 - Scientific Research and Development Services

a The SIC codes 2087 and 2099 may map to any of the following NAICS codes: 311920, 311930, 311941, 311942, 311991,
or 311999. There is not enough information in the DMR data to determine the appropriate NAICS for each site; therefore,
EPA uses data for the 4-digit NAICS, 311900, rather than a specific 6-digit NAICS.

b The SIC code 4911 may map to any of the following NAICS codes: 221111, 221112, 221113, 221114, 221115, 221116,
221117, 221118, 221121, or 221122. There is not enough information in the DMR data to determine the appropriate NAICS
for each site; therefore, EPA uses data for the 4-digit NAICS, 221100, rather than a specific 6-digit NAICS.

0 The SIC codes 8731 and 8733 may map to any of the following NAICS codes: 541713, 541714, 541715, or 541720. There
is not enough information in the DMR data to determine the appropriate NAICS for each site; therefore, EPA uses data for
the 4-digit NAICS, 541700, rather than a specific 6-digit NAICS.

Table 2-67 provides a summary of the reported NAICS codes (or NAICS identified in the crosswalk),
the number of sites reporting each NAICS code, and the estimated number of workers and ONUs for
each NAICS code as well as an overall total for other industrial uses. It should be noted, that in the 2016
DMR, nine sites either did not report a SIC code or reported a SIC code for which no employment data
were available for the corresponding NAICS code. To estimate the number of workers and ONUs from
these sites EPA calculated the average number of workers and ONUs per site from the other known
sites. There are approximately 2,700 workers and 1,300 ONUs potentially exposed during other
industrial uses.

Table 2-67. Estimated Number of Workers Potentially Exposed to Perchloroethylene During
Other Industrial Uses

NAICS
Code

Number of
Sites

Kxposed
Workers
per Site11

Kxposed
Occupational
Non-l sers
per Site11

Total
Kxposed
Workers

Total
Kxposed
Occupational
Non-l sers

Total
Kxposed

212111

1

15

6

15

6

20

212221

3

29

11

86

33

119

212313

1

6

1

6

1

7

212319

1

5

1

5

1

7

212321

2

4

1

7

2

9

221100

32

5

7

169

230

399

311221

1

39

39

39

39

78

311351

1

9

2

9

2

11

311411

1

57

9

57

9

65

311421

1

22

3

22

3

25

311511

1

39

8

39

8

47

311900

2

16

3

32

7

39

Page 164 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

NAICS
Code

Number of
Sites

Kxposed
Workers
per Site11

Kxposed
Occupational
Non-l sers
per Site"

Total
Kxposed
Workers

Total
Kxposed
Occupational
Non-l sers

Total
Kxposed

312120

1

8

1

8

1

9

322110

1

100

15

100

15

116

322220

1

35

5

35

5

40

322299

1

19

2

19

2

22

324110

3

170

75

511

226

737

325110

4

64

30

255

120

375

325180

2

25

12

50

24

74

325199

3

39

18

116

55

170

325211

3

27

12

82

36

119

325212

2

25

11

49

22

71

325220

8

47

21

378

166

545

325411

2

24

15

49

30

79

325414

1

54

33

54

33

88

325612

1

17

4

17

4

20

325920

3

32

10

95

31

126

327110

1

13

2

13

2

16

327120

1

24

4

24

4

28

424710

30

1

0

43

5

48

541700

5

1

9

5

45

50

541712

1

1

10

1

10

12

Unknown
NAICS

9

30

12

274

108

382

Totalb

130

21

10

2,700

1,300

4,000

a Number of workers and occupational non-users per site are calculated by dividing the exposed number of workers or
occupational non-users by the number of establishments in the relevant NAICS codes. The workers/ONUs per site are then
multiplied by the number of sites within that NAICS to get the total exposed. The number of workers/ONUs per site is
rounded to the nearest integer.

b Totals have been rounded to two significant figures. Totals may not add exactly due to rounding.

2.18.3.3 Occupational Exposure Results

EPA did not identify any inhalation exposure monitoring data for the other industrial uses. Therefore,
EPA assessed inhalation exposures during these uses using the Tank Truck and Rail car Loading and

Page 165 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Unloading Release and Inhalation Exposure Model, assuming PCE is present at 100 percent
concentration when used.

The model estimates the potential concentration of PCE in air when it is unloaded or loaded at an
industrial facility. The model accounts for the displacement of saturated air containing the chemical of
interest as the container/truck is filled with liquid, emissions of saturated air containing the chemical of
interest that remains in the loading arm, transfer hose and related equipment, and emissions from
equipment leaks from processing units such as pumps, seals, and valves.

Based on comments from HSIA (2017). halogenated solvents, such as PCE, are expected to be delivered
in either tank trailers or tank cars. Therefore, EPA modeled the central tendency scenario as tank truck
loading/unloading. EPA modeled the high-end scenario as railcar loading/unloading since railcars are
larger and more likely to use longer transfer arms (and thus represent a higher exposure potential than
tank trucks). Based on requirements under the Clean Air Act, EPA expects the majority of industrial
sites to operate a vapor balancing system to minimize fugitive emissions when loading and unloading
tank trucks and rail cars. Therefore, EPA assumed emissions from displacement of saturated air as the
container is filled to be negligible.

For emissions of saturated air from loading arms and transfer hoses, EPA used engineering judgement to
estimate a hose volume of 2.0 gallons (based on a 2-in diameter and 12-ft long hose) for the central
tendency scenario. For the high-end scenario, EPA calculated the 95th percentile volume based on the
dimensions of several types of loading systems provided in an OPW Engineered Systems catalog (OPW
Engineered Systems. 2014) resulting in a volume of 17.7 gallons. For emissions from equipment leaks,
EPA used emission factors from EPA's Protocol for Equipment Leak Emission Estimates (

1995) for the synthetic organic chemical manufacturing industry and engineering judgement to
determine the number and types of equipment used in each scenario. EPA assumed the following
equipment are used:

•	Tank Truck Loading/Unloading:

o Liquid Service:

¦	Four valves (modeled as valves in light liquid service)

¦	One safety relief valve (modeled as valve in light liquid service)

¦	One bleed valve or sampling connection

¦	One hose connector
o Vapor Service:

¦	Three valves (modeled as valves in gas service)

¦	One pressure relief valve

¦	One bleed valve (modeled as a sampling connection)

¦	One hose connector

•	Railcar Loading/Unloading

o Liquid Service: EPA assumed, for the high-end scenario, two parallel liquid service lines,
each using the same equipment as assumed for tank trucks. Therefore, a total of:

¦	Eight valves (modeled as valves in light liquid service)

¦	Two safety relief valves (modeled as valve in light liquid service)

Page 166 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

¦	Two bleed valves or sampling connections

¦	Two transfer arm connectors

o Vapor Service: EPA assumed a single line in vapor service with the same equipment as
assumed for tank trucks.

¦	Three valves (modeled as valves in gas service)

¦	One pressure relief valve

¦	One bleed valve (modeled as a sampling connection)

¦	One transfer arm connector

Additional details of the model design and parameters is provided in Appendix E.

The model calculates both 8-hr TWA exposure concentrations, 1-hr TWA (high-end acute), and 30-min
TWA (central tendency acute) exposure concentrations. As shown in Table 2-13, the Tank Truck and
Railcar Loading and Unloading Release and Inhalation Exposure Model estimates a central tendency
and high-end exposure level of 0.01 ppm and 0.04 ppm as 8-hr TWA, respectively, during container
unloading activities. Note the model does not estimate exposure levels for ONUs for this activity; EPA
estimates that ONU exposures are lower than worker exposures, since ONUs do not typically directly
handle the chemical.

Table 2-68. Summary of Exposure Modeling Results for Other Industrial Uses of
Perchloroethylene	

Scenario

8-hr TW A
(ppm)

AC
(ppm)

A IK

(ppm)

I.AIK
(ppm)

Acute TWA"
(ppm)

High-End

3.60E-02

1.20E-02

8.21E-03

4.21E-03

0.3

Central Tendency

7.96E-03

2.65E-03

1.82E-03

7.22E-04

0.1

AC = Acute Concentration; ADC = Average Daily Concentration; and LADC = Lifetime Average Daily Concentration.
Equations and parameters for calculation of the AC, ADC, and LADC are described in Appendix B.
a High-end for Acute exposures is calculated as a 1-hr TWA and central tendency is calculated as a 30-min TWA.

2.18.4 Water Release Assessment

2.18.4.1	Water Release Sources

Specifics of the processes and potential sources of release for other industrial uses are unknown.
However, general potential sources of water releases in the chemical industry may include the
following: equipment cleaning operations, aqueous wastes from scrubbers/decanters, reaction water,
process water from washing intermediate products, and trace water settled in storage tanks (

).

2.18.4.2	Water Release Assessment Results

EPA assessed water releases using the annual discharge values reported to the 2016 TRI (

2017d) and the 2016 DMR (	b) by the 130 sites using PCE in other industrial uses. In the

2016 TRI, one site reported non-zero direct discharges to surface water and all the other sites reported
zero indirect or direct discharges (	). In the 2016 DMR, six sites reported a direct

Page 167 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

discharge to surface water (indirect discharges not reported in DMR data) and the remaining sites
reported zero direct discharges (	b).

To estimate the daily release, EPA assumed a 250 days/yr of operation and averaged the annual release
over the operating days. Table 2-69 summarizes the water releases from the 2016 TRI and DMR for
sites with non-zero discharges.

Table 2-69. Reported Wastewater Discharges of Perchloroethvlene from Other Industrial Uses

Site

Annual
Release11
(k«/sile-
vear)

Annual
Release

Days
(davs/vr)

Daily
Release
(k«/site-
dav)11

NPDKS ( ode

Release
Media/
Treatment
l-'acililv Type

Source

ExxonMobil Oil
Corp Joilet
Refinery,
Channahon, IL

1.2

250

4.72E-03

ILR10H432

Surface Water

(US. EPA.
2017d)

Natrium Plant, New
Martinsville, WV

7.9

250

3.15E-02

WV0004359

Surface Water

(US. EPA.
2016b")

Oxy Vinyls LP -
Deer Park PVC,
Deer Park, TX

78

250

0.3

TX0007412

Surface Water

(US. EPA.
2016b)

Princeton Plasma
Physics Lab (FF),
Princeton, NJ

0.1

250

5.30E-04

NJ0023922

Surface Water

(US. EPA.
2016b)

Tree Top Inc
Wenatchee Plant,
Wenatchee, WA

7.60E-03

250

3.04E-05

WA0051527

Surface Water

(US. EPA.
2016b)

Vesuvius USA
Corp Buffalo Plant,
Buffalo, NY

3.08E-02

250

1.23E-04

NY0030881

Surface Water

(US. EPA.
2016b)

William E. Warne
Power Plant, Los
Angeles County,
CA

2.82E-04

250

1.13E-06

CA0059188

Surface Water

(US. EPA.
2016b)

a Annual release amounts are based on the site reported values. Therefore, daily releases are calculated from the annual
release rate and assuming 250 days of operation per year.

Sources: (U.S. EPA. 2017d. 2016b)

2.19 Other Commercial Uses

2.19.1 Estimates of Number of Facilities

EPA did not identify information from EPA databases (e.g., CDR, TRI, or DMR) or in the results of the
systematic review process to estimate the number of sites using PCE for other commercial uses. EPA

Page 168 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

did identify seven facilities in the 2016 DMR where EPA could not determine the condition of use or the
condition of use falls into a commercial use discussed in Section 2.19.2. However, due to the large
variety of PCE-based products and uses of PCE, these seven sites are not expected to represent the
entirety of sites using PCE in other commercial applications.

2.19.2	Process Description

Based on information identified in EPA's preliminary data gathering and information obtained from
public comments, a variety of other commercial uses of PCE may exist. Examples of these uses include,
but are not limited to, metal (e.g., stainless steel) and stone polishes, inks and ink removal products,
photographic film applications, and mold cleaning, release, and protectant products. For many of these
uses PCE is expected to act similar to a cleaning solvent used to remove dirt or other contaminates from
substrates (e.g., metal polishes and ink removal products). However, in the photographic film industry,
PCE is used as a liquid-gate fluid to help protect scratching of optical negatives during filming (NIOSH.
1980a). Due to changes in technology (e.g., the use of digital equipment in place of traditional film), the
prevalence of use of PCE as a liquid-gate fluid is unknown.

2.19.3	Exposure Assessment

2.19.3.1	Worker Activities

The worker activity, use pattern, and associated exposure will vary for each condition of use. For
polishes, ink removal products, and mold release, EPA expects workers may be exposed to PCE vapors
that evaporate from the application material (rag, brush, etc.) or the substrate surface during use. For
inks, workers may be exposed to mists generated during the ink application process. For photographic
film, workers may be exposed to PCE that evaporates from the gating process.

2.19.3.2	Number of Potentially Exposed Workers

EPA has not identified information from the systematic review process on the number of sites and
potentially exposed workers associated with these uses. The use of PCE for these conditions of use is
expected to be minimal.

2.19.3.3	Occupational Exposure Results

EPA assessed exposure to these uses of PCE using data from identified studies. Table 2-70 summarizes
the 8-hr TWA and 15-min TWA data identified for these uses. For printing uses (includes uses of both
inks and ink removal products and commercial print shops), EPA identified data from six NIOSH
investigations at six printing facilities. Four of the printing sites investigated by NIOSH used PCE for
cleaning machines or printing plates, and 2 did not describe the function of PCE at the shop (NIOSH.
1994. 1984a. 1983a. 1982a. 1981c. 1980b). A total of 23 samples were collected at the sites with sample
times ranging from approximately 2 to 8.5 hours (citation for six NIOSH HHES used for printing
analysis). Where sample times were less than eight hours, EPA calculated the 8-hr TWA assuming
exposure outside the sample time was zero. For the 8-hr TWA, the 95th percentile is presented as the
high-end and the 50th percentile as the central tendency. There was a single 15-minute TWA sample.
Results based on a single value are plausible, but EPA cannot determine the statistical representativeness
of the value. There was single 8-hr TWA sample that measured below the LOD. To estimate exposure
concentrations for this data point, EPA followed the Guidelines for Statistical Analysis of Occupational
Exposure Data (	94b) as discussed in Section 1.4.5.2. The geometric standard deviation for

Page 169 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

the data was above 3.0; therefore, EPA used the —p to estimate the exposure value as specified in the
guidelines (	4b).

EPA also identified PCE exposure data at a U.S. photocopy shop using dry-process photocopiers
(Stefaniak et at.. 2000). The study collected three PBZ samples from workers at the photocopy shop
with sample times ranging from 7.5 to 8 hours (Stefaniak et at.. 2000). Where sample times were less
than eight hours, EPA calculated the 8-hr TWA assuming exposure outside the sample time was zero.
Only three data points were available; therefore, EPA presented the maximum as the high-end and the
median as the central tendency.

For photographic film uses, EPA identified a single NIOSH study that investigated PCE exposures at 14
optical film shops (NIOSH. 1980a). A total of 55 samples were collected at the sites with sample times
ranging from approximately two to eight hours (NIOSH. 1980a). Where sample times were less than
eight hours, EPA calculated the 8-hr TWA assuming exposure outside the sample time was zero. The
95th percentile Is presented as the high-end and the 50th percentile as the central tendency.

For mold release products, EPA did not identify any PBZ data; however, Gold et al. (Gold et at.. 2008)
completed a comprehensive literature review of studies evaluating PCE exposures from a variety of uses
in the U.S. that provided data from a single 1983 study that collected four area samples for the use of
PCE as a mold release agent. Sample times in the study were greater than six hours (Gold et al.. 2008).
EPA assessed exposures assuming area data are representative of exposures to personnel. However,
details of how/where the area samples were collected were not provided and it is unclear how
representative these area samples are of actual exposures. Discrete data were not available; therefore,
EPA presents the maximum as the high-end and the arithmetic mean as the central tendency.

There is a wide range of exposure results across the different commercial uses summarized in Table
2-70. This is likely due to the difference in how PCE is handled within each use and how easily PCE can
evaporate into the workers breathing zone.

Page 170 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Table 2-70. Summary of Exposure Monitoring Data for Other Commercial Uses of
Perchloroethylene	

Scenario

8-hr
TWA
(ppni)

AC
(ppni)

A IK

(ppni)

I.AIK
(ppm)

Nil in her
of Data
Points

15-mi mile
TWA
(ppm)

Nil in her
of Data
Points

Results for Printing Applications (Ink and Ink Removal Products)

High-End

5.9

2.0

1.4

0.7

23

0.2

1

Central Tendency

1.9

0.6

0.4

0.2

Results for Photocopying

High-End

5.00E-04

1.67E-04

1.14E-04

5.85E-05

3

No 15-minute data
identified for this use

Central Tendency

1.88E-04

6.25E-05

4.28E-05

1.70E-05

Results for Photographic Film Applications

High-End

56

19

13

6.6

62

117

40

Central Tendency

6.3

2.1

1.4

0.6

13

Results for Mold Release Products

High-End

0.2

6.67E-02

4.57E-02

2.34E-02

4

No 15-minute data
identified for this use

Central Tendency

0.1

3.33E-02

2.28E-02

9.07E-03

AC = Acute Concentration; ADC = Average Daily Concentration; and LADC = Lifetime Average Daily Concentration.
Equations and parameters for calculation of the AC, ADC, and LADC are described in Appendix B.

Source: (Gold et at 2008: Stefaniak et at 2000: NIOSH. .1.994. 1984a. 1983a. 1982a. 1981c. 1980a. b)

2.19.4 Water Release Assessment

2.19.4.1	Water Release Sources

Specifics of the processes and potential sources of release for these uses are unknown. Based on the
volatility of PCE, EPA expects the majority of PCE used for these applications to evaporate and be
released to air. EPA expects residuals in containers to be disposed of with general site trash that is either
picked up by local waste management or by a waste handler that disposes wastes as hazardous waste.

2.19.4.2	Water Release Assessment Results

Table 2-71 summarizes non-zero water releases from sites using PCE in other commercial uses reported
in the 2016 DMR (\ c< ! i1 \ 2016b). To estimate the daily release for the sites in Table 2-71, EPA
assumed 250 days/yr of operation and averaged the annual release over the operating days. These data
are not expected to capture the entirety of water releases from these uses; however, EPA does not have
information to estimate water releases from sites not reporting to DMR. Based on the SIC codes
reported in the DMR, the industries covered by these sites include special trade contractors, heavy
construction, and line-haul railroad operations.

Page 171 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Table 2-71. Reported Wastewater Discharges of Perchloroethylene from Other Commercial Uses
in the 2016 DMR

Site

Annual
Release"
(k«/sile-
vear)

Annual
Release

Days
(davs/vr)

Daily
Release
(k«/sile-
dav):l

NPDKS (ode

Release Media/

Treatment
l-'acililv Type

Union Station North
Wing Office Building,
Denver, CO

0.7

250

2.87E-03

COG315293

Surface Water

Confluence Park
Apartments, Denver,
CO

7.50E-02

250

3.00E-04

COG315339

Surface Water

Wynkoop Denver
LLCP St, Denver, CO

3.84E-02

250

1.54E-04

COG603115

Surface Water

100 Saint Paul, Denver
County, CO

1.07E-02

250

4.27E-05

COG315289

Surface Water

BPI-W estminster,
LLC(Owner)/Arcadi s
(Op), Denver, CO

8.60E-03

250

3.44E-05

COG315146

Surface Water

Safeway Inc, Denver,
CO

3.91E-03

250

1.56E-05

COG315260

Surface Water

Illinois Central
Railroad,

Thompsonville, IL

3.72E-03

250

1.31E-05

IL0070696

Surface Water

a Annual release amounts are based on the site reported values. Therefore, daily releases are calculated from the annual
release rate and assuming 250 days of operation per year.

Sources: (TJ.S. EPA. 20.1.6b')

2.20 Laboratory Chemicals

2.20.1	Estimates of Number of Facilities

EPA did not identify information from the systematic review process to estimate the number of sites
using PCE as a laboratory chemical.

2.20.2	Process Description

PCE is used in a variety of laboratory applications as a chemical reagent (Aerospace Industries
Association. 2017). Specific process descriptions for how PCE is used in lab applications is not known.
In general, PCE is expected to be received in small containers and used in small quantities on a lab
bench in a fume cupboard or hood. After use, waste PCE is collected and disposed or recycled. Figure
2-20 this general process.

Page 172 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Figure 2-20. General Laboratory Use Process Flow Diagram
2.20.3 Exposure Assessment

2.20.3.1 Worker Activities

Specific worker activities for using laboratory uses were not identified, but EPA expects that workers
may be potentially exposed to PCE in laboratories during multiple activities, including unloading of
PCE from the containers in which they were received, transferring PCE into laboratory equipment (i.e.,
beakers, flasks, other intermediate storage containers), dissolving substances into PCE or otherwise
preparing samples that contain PCE, analyzing these samples, and discarding the samples.

ONUs include employees that work at the sites where PCE 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. ONUs for this condition of use include supervisors, managers, and other employees
that may be in the laboratory 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 PCE.

2.20.3.1	Number of Potentially Exposed Workers

EPA did not identify information to estimate the total number of workers exposed to PCE at laboratory
facilities. However, EPA estimated the number of workers and ONUs per site using information from
the Bureau of Labor Statistics' OES data (	. 2016) and the U.S. Census' SUSB (U.S. Census

Bureau.: ). The method for estimating number of workers from the Bureau of Labor Statistics" OES
data and U.S. Census' SUSB data is detailed in Appendix A. These estimates were derived using
industry- and occupation-specific employment data from the BLS and U.S. Census.

EPA identified the NAICS code 541380, Testing Laboratories, as the code expected to include
laboratory chemical uses of PCE. Based on data from the BLS for this NAICS code and related SOC
codes, there are an average of one worker and nine ONUs per site, or a total of ten potentially exposed
workers and ONUs per site.

2.20.3.2	Occupational Exposure Results

EPA does not have data to assess worker exposures to PCE during laboratory use. However, due to the
expected safety practices when using chemicals in a laboratory setting, PCE is expected to be applied in
small amounts under a fume hood, thus reducing the potential for inhalation exposures.

Page 173 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

2.20.4 Water Release Assessment

2.20.4.1	Water Release Sources

The primary source of water releases at laboratories is expected to be from disposal of spent PCE
reagent. However, not all sites using PCE as a laboratory reagent are expected to dispose of PCE to
water. EPA expects some will collect PCE wastes with other hazardous lab materials to be collected and
disposed of by a waste contractor as hazardous waste.

2.20.4.2	Water Release Assessment Results

EPA did not identify information from the systematic review process on the number of laboratory sites
or the volume of PCE used in laboratory applications to estimate releases of PCE to water from
laboratory uses. The SpERC developed by the European Solvent Industry Group for laboratory reagents
(European solvents Industry Group. 2019b) estimates a 100% release scenario with 50% of the use
volume being released to municipal wastewater, sewer, or water course, and 50% released to air.
Therefore, no more than 50% of the use volume is expected to be released to water.

2.21 Waste Handling, Disposal, Treatment, and Recycling

2.21.1	Estimates of Number of Facilities

To determine the number of disposal, treatment, and recycling sites, EPA considered 2016 TRI (
), and 2016 DMR (	») data. Based on the activities and NAICS codes

reported in the 2016 TRI, EPA identified 38 facilities where the primary condition of use is expected to
be disposal or recycling of PCE-containing wastes (	). In the 2016 DMR data, there are

59 sites for which EPA expects the primary condition of use to be disposal/recycling of PCE wastes
based on the reported SIC codes and facility names, three of which are the same as sites identified in
TRI (U.S. EPA. 2016bY Based on the DMR and TRI data, EPA assesses a total of 94 sites (38+59 = 97
sites - 3 duplicate sites = 94 sites) for the disposal/recycling of PCE. NAICS codes used to identify
disposal/treatment sites are those related to waste disposal including any NAICS code under the 3-digit
NAICS code 562000, Waste Management and Remediation Services and NAICS codes excepted to
operate cement kilns (e.g., 327310, Cement Manufacturing) which are expected to burn various waste
products for fuel.

2.21.2	Process Description

Each of the conditions of use of PCE may generate waste streams of the chemical that are collected and
transported to third-party sites for disposal, treatment, or recycling. Industrial sites that treat or dispose
onsite wastes that they themselves generate are assessed in each condition of use assessment in Sections
2.1 through 2.20. Similarly, point source discharges of PCE to surface water are assessed in each
condition of use assessment in Sections 2.1 through 2.20 (point source discharges are exempt as solid
wastes under RCRA). Wastes of PCE that are generated during a condition of use and sent to a third-
party site for treatment, disposal, or recycling may include the following:

• Wastewater: PCE may be contained in wastewater discharged to POTW or other, non-public
treatment works for treatment. Industrial wastewater containing PCE discharged to a POTW may
be subject to EPA or authorized NPDES state pretreatment programs. The assessment of
wastewater discharges to POTWs and non-public treatment works of PCE is included in each of
the condition of use assessments in Sections 2.1 through 2.20.

Page 174 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

• Solid Wastes: Solid wastes are defined under RCRA as any material that is discarded by being:
abandoned; inherently waste-like; a discarded military munition; or recycled in certain ways
(certain instances of the generation and legitimate reclamation of secondary materials are
exempted as solid wastes under RCRA). Solid wastes may subsequently meet RCRA's definition
of hazardous waste by either being listed as a waste at 40 CFR §§ 261.30 to 261.35 or by
meeting waste-like characteristics as defined at 40 CFR §§ 261.20 to 261.24. Solid wastes that
are hazardous wastes are regulated under the more stringent requirements of Subtitle C of
RCRA, whereas non-hazardous solid wastes are regulated under the less stringent requirements
of Subtitle D of RCRA.

o Solid wastes containing PCE may be regulated as a hazardous waste under RCRA waste
codes D039 for wastes containing 0.7 mg/L or more of PCE (40 CFR 261.24), F001 for
spent halogenated solvents used in degreasing, (40 CFR 261.31), F002 for spent
halogenated solvents (40 CFR 261.31), and U210 for discarded commercial chemical
products, manufacturing chemical intermediates, off-specification commercial chemical
products, container residues, or spill residues. These wastes would be either incinerated
in a hazardous waste incinerator or disposed to a hazardous waste landfill.

• Wastes Exempted as Solid Wastes under RCRA: Certain conditions of use of PCE may generate
wastes of PCE that are exempted as solid wastes under 40 CFR § 261.4(a). For example, the
generation and legitimate reclamation of hazardous secondary materials of PCE may be exempt
as a solid waste.

2016 TRI data lists off-site transfers of PCE to land disposal, wastewater treatment, incineration, and
recycling facilities (	). About 32% of off-site transfers were incinerated, 1% sent to land

disposal, less than 1% sent to wastewater treatment, 66% is recycled off-site, and 1% is sent to other or
unknown off-site disposal/treatment (	). See Figure 2-21.

Page 175 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Hazardous Waste	Hazardous Waste

Generation	Transportation

Recycling



Treatment

Figure 2-21. Typical Waste Disposal Process

Source: (U.S. EPA. 2017b)

Municipal Waste Incineration

Municipal waste combustors (MWCs) that recover energy are generally located at large facilities
comprising an enclosed tipping floor and a deep waste storage pit. Typical large MWCs may range in
capacity from 250 to over 1,000 tons per day. At facilities of this scale, waste materials are not generally
handled directly by workers. Trucks may dump the waste directly into the pit, or waste may be tipped to
the floor and later pushed into the pit by a worker operating a front-end loader. A large grapple from an
overhead crane is used to grab waste from the pit and drop it into a hopper, where hydraulic rams feed
the material continuously into the combustion unit at a controlled rate. The crane operator also uses the
grapple to mix the waste within the pit, in order to provide a fuel consistent in composition and heating
value, and to pick out hazardous or problematic waste.

Facilities burning refuse-derived fuel (RDF) conduct on-site sorting, shredding, and inspection of the
waste prior to incineration to recover recyclables and remove hazardous waste or other unwanted
materials. Sorting is usually an automated process that uses mechanical separation methods, such as
trommel screens, disk screens, and magnetic separators. Once processed, the waste material may be
transferred to a storage pit, or it may be conveyed directly to the hopper for combustion.

Tipping floor operations may generate dust. Air from the enclosed tipping floor, however, is
continuously drawn into the combustion unit via one or more forced air fans to serve as the primary
combustion air and minimize odors. Dust and lint present in the air is typically captured in filters or
other cleaning devices in order to prevent the clogging of steam coils, which are used to heat the
combustion air and help dry higher-moisture inputs22.

Z2 J.B. Kitto, Eds., Steam: Its Generation and Use, 40th Edition. Babcock and Wilcox/American Boiler Manufacturers
Association, 1992.

Page 176 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Hazardous Waste Incineration

Commercial scale hazardous waste incinerators are generally two-chamber units, a rotary kiln followed
by an afterburner, that accept both solid and liquid waste. Liquid wastes are pumped through pipes and
are fed to the unit through nozzles that atomize the liquid for optimal combustion. Solids may be fed to
the kiln as loose solids gravity fed to a hopper, or in drums or containers using a conveyor23'24.

Incoming hazardous waste is usually received by truck or rail, and an inspection is required for all waste
received. Receiving areas for liquid waste generally consist of a docking area, pumphouse, and some
kind of storage facilities. For solids, conveyor devices are typically used to transport incoming waste.

Smaller scale units that burn municipal solid waste or hazardous waste (such as infectious and hazardous
waste incinerators at hospitals) may require more direct handling of the materials by facility personnel.
Units that are batch-loaded require the waste to be placed on the grate prior to operation and may
involve manually dumping waste from a container or shoveling waste from a container onto the grate.
See Figure 2-22 for a typical incineration process.

Emissions Stacx

Host Recovery

Combustion h- ~

Gas, TcmpctViturn
Rorh.ainn

Air Portion Coniro r-

Ash Hnnoi.ng

Scraobcr Wntrr nr
Ash Haod st>g

y

Disposal

Disposal

Figure 2-22.Typical Industrial Incineration Process

23	Environmental Technology Council's Hazardous Waste Resource Center; http://www.etc.org/advanced-technologies/high-
temperature-incineration.aspx

24	Incineration Services; Heritage; https://www.heritage-enviro.com/services/incineration/

Page 177 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Municipal Waste Landfill

Municipal solid waste landfills are discrete areas of land or excavated sites that receive household
wastes and other types of non-hazardous wastes (e.g. industrial and commercial solid wastes). Standards
and requirements for municipal waste landfills include location restrictions, composite liner
requirements, leachate collection and removal system, operating practices, groundwater monitoring
requirements, closure-and post-closure care requirements, corrective action provisions, and financial
assurance. Non-hazardous solid wastes are regulated under RCRA Subtitle D, but state may impose
more stringent requirements.

Municipal solid wastes may be first unloaded at waste transfer stations for temporary storage, prior to
being transported to the landfill or other treatment or disposal facilities.

Hazardous Waste Landfill

Hazardous waste landfills are excavated or engineered sites specifically designed for the final disposal
of non-liquid hazardous wastes. Design standards for these landfills require double liner, double leachate
collection and removal systems, leak detection system, run on, runoff and wind dispersal controls, and
construction quality assurance program25. There are also requirements for closure and post-closure, such
as the addition of a final cover over the landfill and continued monitoring and maintenance. These
standards and requirements prevent potential contamination of groundwater and nearby surface water
resources. Hazardous waste landfills are regulated under Part 264/265, Subpart N.

Solvent Recovery

Waste solvents are generated when it becomes contaminated with suspended and dissolved solids,
organics, water, or other substances. Waste solvents can be restored to a condition that permits reuse via
solvent reclamation/recycling. The recovery process involves an initial vapor recovery (e.g.,
condensation, adsorption and absorption) or mechanical separation (e.g., decanting, filtering, draining,
setline and centrifuging) step followed by distillation, purification and final packaging. Worker activities
are expected to be unloading of waste solvents and loading of reclaimed solvents. Figure 2-23 illustrates
a typical solvent recovery process flow diagram (U.S. EPA. 1980).

25 https://www.epa.gov/tiwperniittlng/hazardons-waste-management-facilitles-and-niiits

Page 178 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Figure 2-23. General Process Flow Diagram for Solvent Recovery Processes (U.S. EPA, 1980)
2.21.3 Exposure Assessment

2.21.3.1 Worker Activities

At waste disposal sites, workers are potentially exposed via dermal contact with waste containing PCE
or via inhalation of PCE vapor. Depending on the concentration of PCE in the waste stream, the route
and level of exposure may be similar to that associated with container unloading activities. See Section
2.3.3.3 for the assessment of worker exposure from chemical unloading activities.

Municipal Waste Incineration

At municipal waste incineration facilities, there may be one or more technicians present on the tipping
floor to oversee operations, direct trucks, inspect incoming waste, or perform other tasks as warranted by
individual facility practices. These workers may wear protective gear such as gloves, safety glasses, or
dust masks. Specific worker protocols are largely up to individual companies, although state or local
regulations may require certain worker safety standards be met. Federal operator training requirements
pertain more to the operation of the regulated combustion unit rather than operator health and safety.

Workers are potentially exposed via inhalation to vapors while working on the tipping floor. Potentially-
exposed workers include workers stationed on the tipping floor, including front-end loader and crane
operators, as well as truck drivers. The potential for dermal exposures is minimized by the use of trucks
and cranes to handle the wastes.

Page 179 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Hazardous Waste Incineration

More information is needed to determine the potential for worker exposures during hazardous waste
incineration and any requirements for personal protective equipment. There is likely a greater potential
for worker exposures for smaller scale incinerators that involve more direct handling of the wastes.

Municipal and Hazardous Waste Landfill

At landfills, typical worker activities may include operating refuse vehicles to weigh and unload the
waste materials, operating bulldozers to spread and compact wastes, and monitoring, inspecting, and
surveying and landfill site26.

2.21.3.2 Number of Potentially Exposed Workers

EPA estimated the number of workers and occupational non-users potentially exposed during
disposal/treatment of PCE using Bureau of Labor Statistics" OES data (U.S. BLS. 2016) and the U.S.
Census" SUSB (U.S. Census Bureau. ) as well as the primary NAICS and SIC code reported by
each site in the 2016 TR1 (	) or 2016 DMR (U.S. EPA. 2016b"). respectively. The

method for estimating number of workers is detailed above in Section 1.4.4 and Appendix A. These
estimates were derived using industry- and occupation-specific employment data from the BLS and U.S.
Census. The employment data from the U.S. Census SUSB and the Bureau of Labor Statistics' OES data
are based on NAICS code; therefore, SIC codes reported in the 2016 DMR had to be mapped to a
NAICS code to estimate the number of workers. A crosswalk of the SIC codes to the NAICS codes used
in the analysis are provided in Table 2-72. In the 2016 DMR there were 27 sites that either did not report
a SIC code or reported a SIC for which employment data were not available for the corresponding
NAICS code; for these sites, EPA used the average workers and ONUs per site calculated from the other
sites with known data.

26 http://www.calrecycle.ca.gov/SWfacilities/landfills/needfor/Operations.htm

Page 180 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Table 2-72. Crosswalk of Disposal SIC Codes in DMR to NAICS Codes

Sl( ( ode

Corresponding NAICS Code

3273 - Ready-Mixed Concrete

327320 - Ready-Mix Concrete Manufacturing

3295 - Minerals and Earths, Ground or Otherwise
Treated

327992 - Ground or Treated Mineral and Earth
Manufacturing

4953 - Refuse Systemsa

562200 - Waste Treatment and Disposal

4959 - Sanitary Services, Not Elsewhere
Classified13

562900 - Remediation and Other Waste
Management Services

7699 - Repair Shops and Related Services, Not
Elsewhere Classified0

562998 - All Other Miscellaneous Waste
Management Services

9511 - Air and Water Resource and Solid Waste
Management

924110 - Administration of Air and Water
Resource and Solid Waste Management Programs

a The SIC code 4953 may map to any of the following NAICS codes: 562211, 562212, 562213 or 562219. There is not
enough information in the DMR data to determine the appropriate NAICS code to use; therefore, EPA uses data for the 4-
digit NAICS, 562200, rather than a specific 6-digit NAICS.

b The SIC code 4959 may map to any of the following NAICS codes: 561710, 561790, 562910 or 562998. Based on the
condition of use for the site reporting this SIC code, EPA determined that the NAICS codes 592910 and 592998 most
accurately described the site. There is not enough information in the DMR data to determine which is more appropriate;
therefore, EPA uses data for the 4-digit NAICS, 562900, rather than a specific 6-digit NAICS.

0 The SIC code 7699 maps to several NAICS codes primarily related to repair services and not disposal services. After review
of the reporting company's website, this site was determined to be primarily engaged in disposal activities; therefore, EPA
determined the NAICS codes 562998 most accurately described the site.

Table 2-73 provides a summary of the reported NAICS codes (or NAICS identified in the crosswalk),
the number of sites reporting each NAICS code, and the estimated number of workers and ONUs for
each NAICS code as well as an overall total for disposal/treatment of PCE wastes. There are
approximately 1,600 workers and 700 ONUs potentially exposed during disposal/treatment of PCE
wastes.

Page 181 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Table 2-73. Estimated Number of Workers Potentially Exposed to Perchloroethylene During
Disposal/Treatment	

NAICS Code

Number
of Sites

Kxposed
Workers
per Site"

Kxposed
Occupational
Non-l sers
per Site11

Total
Kxposed
Workers1'

Total
Kxposed
Occupational
Non-l sers1'

Total
Kxposed

221112

1

6

8

6

8

13

324110

1

170

75

170

75

246

324191

2

20

9

40

18

58

325110

2

64

30

127

60

187

327310

8

22

3

174

27

201

327320

4

5

1

21

3

25

327992

2

17

3

34

7

41

562200

23

6

3

131

75

206

562211

19

9

5

171

98

269

562213

1

13

8

13

8

21

562219

1

3

2

3

2

4

562920

2

2

2

4

3

7

562998

1

1

1

1

1

3

Subtotal for Known
SIC/NAICS Data

67

13

6

897

384

1,281

Unknown or No Data

27

26

12

702

311

1,014

Total0

94

17

7

1,600

700

2,300

a Number of workers and occupational non-users per site are calculated by dividing the exposed number of workers or
occupational non-users by the number of establishments in the relevant NAICS codes. The workers/ONUs per site are then
multiplied by the number of sites within that NAICS to get the total exposed. The number of workers/ONUs per site is
rounded to the nearest integer. Number of workers and occupational non-users per site for sites with unknown NAICS codes
are calculated by averaging the values of the known sites.

b Total exposed workers and ONUs for sites with known NAICS are taken directly from the Bureau of Labor Statistics' OES
data the U.S. Census' SUSB. For sites with unknown NAICS codes the total workers and ONUs are estimated by multiplying
the workers and ONUs per site by the number of sites.

0 Totals have been rounded to two significant figures. Totals may not add exactly due to rounding.

2.21.3.3 Occupational Exposure Results

EPA did not identify any inhalation exposure monitoring data for disposal/recycling of PCE. See
Section 2.18.3.3 for the assessment of worker exposure from chemical unloading activities. EPA
assumes the exposure sources, routes, and exposure levels are similar to those at other industrial
facilities using PCE.

Page 182 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.
2.21.4 Water Release Assessment

2.21.4.1	Water Release Sources

Potential sources of water releases at disposal/recycling sites may include the following: aqueous wastes
from scrubbers/decanter, trace water settled in storage tanks, and process water generated during the
disposal/recycling process.

2.21.4.2	Water Release Assessment Results

EPA assessed water releases using the annual discharge values reported to the 2016 TRI (

2017d) and the 2016 DMR (	b) by the 94 disposal/treatment sites. In the 2016 TRI, four

sites reported non-zero indirect discharges to non-POTW WWT, four sites reported indirect discharges
to POTW, and all of the sites reported zero direct discharges to surface water. In the 2016 DMR, five
sites reported non-zero direct discharges to surface water (indirect discharges not reported in DMR data)
and the remaining sites reported no direct discharges.

To estimate the daily release, EPA assumed 250 days/yr of operation as and averaged the annual release
over the operating days. Table 2-74 summarizes the water releases from the 2016 TRI and DMR for
sites with non-zero discharges.

Table 2-74. Reported Wastewater Discharges of Perchloroethylene from Disposal/Treatment of
Perchloroethylene-Containing Wastes	

She

Annual
Release"
(kg/silc-
vear)

Annual
Release

Days
(davs/vr)

Daily
Release
(kg/sile-
da> ):l

NPDI SC ode

Release
Media/
Treatment
Kacililv Type

Source

Clean Harbors Deer
Park LLC, La Porte,
TX

87

250

0.3

TX0005941

Non-POTW
WWT

(US. EPA.
2017d)

Clean Harbors El
Dorado LLC, El
Dorado, AR

9.3

250

3.71E-02

AR0037800

Non-POTW
WWT

(US. EPA.
2017d)

Clean Harbors
Recycling Services
of Ohio LLC,
Hebron, OH

8.55E-03

250

3.42E-05

Not available

POTW

(US. EPA.
2017d)

Clean Water Of New
Yorklnc, Staten
Island, NY

0.9

250

3.77E-03

NY0200484

Surface Water

(US. EPA.
2016b)

Clifford G Higgins
Disposal Service Inc
SLF, Kingston, NJ

5.25E-02

250

2.10E-04

NJG160946

Surface Water

(US. EPA.
2016b)

Durez North
Tonawanda

1.48E-02

250

5.92E-05

NY0001198

Surface Water

(US. EPA.
2016b")

Page 183 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Site

Anniiiil

Release"
(kg/si fo-
vea r)

An mi:il
Release

Davs
(davs/vr)

Dailv
Release
(kg/sife-
dav)11

M»l)i:S( ode

Release
Media/
Treatment
l-'acilitv Type

Source

Occidental Chemical
Corporation, North
Tonawanda, NY













Heritage Thermal
Services, East
Liverpool, OH

9.07E-05

250

3.63E-07

OHO 107298

POTW

(US. EPA.
2017d)

Oiltanking Houston
Inc, Houston, TX

0.8

250

3.31E-03

TX0091855

Surface Water

(US. EPA.
2016b)

Pinewood Site
Custodial Trust,
Pinewood, SC

0.1

250

5.82E-04

SC0042170

Surface Water

(US. EPA.
2016b)

Safety-Kleen
Systems Inc,
Smithfield, KY

338

250

1.4

KY0098345

Non-POTW
WWT

(US. EPA.
2017d)

Safety-Kleen
Systems Inc, East
Chicago, IN

68

250

0.3

Not available

POTW

(US. EPA.
2017d)

Tier Environmental
LLC, Bedford, OH

30

250

0.1

Not available

POTW

(US. EPA.
2017d)

Tradebe Treatment
& Recycling LLC,
East Chicago, IN

1.4

250

5.44E-03

Not available

Non-POTW
WWT

(US. EPA.
2017d)

POTW = Publicly-Owned Treatment Works; WWT = Wastewater Treatment

a Annual release amounts are based on the site reported values. Therefore, daily releases are calculated from the annual
release rate and assuming 250 days of operation per year.

Sources: (U.S. EPA. 2017d. 2016b)

2.22 Other Department of Defense Uses

EPA reached out to the Department of Defense (DoD) for monitoring data for the first 10 chemical
substances that are the subject of the Agency's initial chemical risk evaluations. The DoD provided
monitoring data from its Defense Occupational and Environmental Health Readiness System - Industrial
Hygiene (DOEHRS-IH), which collects occupational and environmental health risk data from each
service branch. The DoD provided inhalation monitoring data for three branches of the military: the
Army, Air Force, and Navy (Defense Occupational and Environmental Health Readiness System -
Industrial Hygiene. 2018). These data are not distinguished among the three branches.

The following subsections provide an overview of the DOD data. It should be noted that where the
condition of use of the collected monitoring data could be clearly determined and fit into one of the

Page 184 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

conditions of use assessed in Sections 2.1 through 2.21; EPA included the data and relevant discussions
in the assessment of that condition of use. This section only serves to describe data that did not fit into
another previously identified condition of use.

2.22,1 Data Overview

The data provided by DoD included 49 data points for PCE from samples taken during a variety of
processes. Of these 49 data points, 41 were determined to fit under conditions of use assessed in
Sections 2.1 through 2.21 and are not discussed further here (Defense Occupational and Environmental
Health Readiness System - Industrial Hygiene. 2018). The remaining eight samples were collected
during one of the following processes:

1.	Oil Analysis;

2.	Water Pipe Repair;

3.	Conducting industrial hygiene surveys/Taking industrial hygiene samples;

4.	Cable End Molding; and

5.	Soldering/Desoldering (Defense Occupational and Environmental Health Readiness System -
Industrial Hygiene. 2018).

A summary of the personal breathing zone samples for these five DoD activities are summarized in
Table 2-75. EPA assumes all sample results indicated with a less than symbol were below the LOD.

Table 2-75. Summary of DoD Inhalation Monitoring Data Not Included in Assessments for Other
Conditions of Use

Process

Worker
Activity
Frequency

Process

Mill.
Sample

Max.
Sample

Nil m her

of
Samples

Sample
Duration
(mill)

Sample

Duration

Result

Result

Dale





(|i|im)

(ppin)



Oil Analysis

2-3
times/week

1-2 hours

4.14

6.61

2

15-64

February
25, 2008

Water Pipe Repair

2-3

times/month

Not
provided

-

<3.0

1

370

March 15,
2003

Conducting industrial
hygiene surveys/
Taking industrial
hygiene samples

2-3
times/week

Not
provided

<0.22

<0.98

3

15-134

September
30, 2006

Cable End Molding

Daily

Not
provided

-

<0.29

1

51

May 30,
2013

Soldering/Desoldering

Daily

Not
provided

-

<0.23

1

64

April 27,
2016

Source: (Defense Occupational and Environmental Health Readiness System - Industrial Hygiene. 2018)

Page 185 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

2.22.2	Process Description and Worker Activities

The DoD data did not provide specific process descriptions or worker activities beyond the process
name identified in Table 2-75 for these data.

2.22.3	Occupational Exposure Assessment

EPA did not assess exposures from the following processes as the sample times were less than 50% of
an 8-hr shift (assumed shift-time for these activities) and, therefore, may not be representative of actual
8-hr TWA exposures:

•	Conducting industrial hygiene surveys/Taking industrial hygiene samples;

•	Cable End Molding; and

•	Soldering/Desoldering.

EPA assessed exposures from the oil analysis and water pipe repair processes separately due to
differences in the frequency of activities. For the oil analysis process, EPA calculated an 8-hr TWA
exposure using the single sample result collected over 64 minutes. EPA believes this to be a reasonable
assumption as the process duration is specified as one to two hours. Therefore, EPA expects the sample
time to be representative of the time the worker spent handling PCE in the process with little potential
for exposure for the remainder of the shift27. The process duration for the water pipe repair process was
not provided; however, the sample time is sufficiently long (>6 hours) such that EPA assumes it is
representative of the duration that the worker handles PCE during the work-shift. Table 2-76
summarizes the results for both the oil analysis and water pipe repair processes.

Only one data point was available for the oil analysis. Results based on a single value are considered
plausible, but EPA cannot determine the statistical representativeness of the value. There was only one
data point available for the water pipe repair as well; however, it measured below the LOD. To estimate
values below the LOD, EPA referenced the Guidelines for Statistical Analysis of Occupational Exposure

Data (	i) which estimates the exposure value as if the geometric standard deviation

of the data is less than 3.0 and if the geometric standard deviation is 3.0 or greater (

1994b). However, there is only a single data point, so the geometric standard deviation is not statistically
meaningful. Therefore, EPA assesses the exposure as ranging from zero to the LOD (2.31 ppm) and
presents two scenarios: 1) using the LOD as a "higher value"; and 2) using half the LOD as a
"midpoint" value. These scenarios are plausible, but EPA cannot determine the statistical
representativeness of the value.

For the oil analysis, DoD reported the frequency as two to three times per week. Therefore, when
calculating the ADC and LADC, EPA adjusted the exposure frequency to reflect the expected number of
exposure days. For the high-end calculations, EPA used the maximum process frequency of three times
per week and for the central tendency calculations, EPA used the midpoint of the frequency, 2.5 times
per week. Assuming 50 weeks per year of exposure (standard EPA assumption allowing for two weeks
off), results in 150 exposure days/yr at the high-end and 125 exposure days at the central tendency.

27 The 15-min TWA for the oil analysis process was not used to calculate an 8-hr TWA exposure as it is not expected to be
representative of the duration the worker is handling PCE based on the reported process duration.

Page 186 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

For the water pipe repair, DoD reported the frequency as two to three times per month. Therefore, EPA
similarly adjusted the exposure frequency to reflect the expected number of exposure days when
calculating the ADC and LADC. For the high-end calculations, EPA used the maximum process
frequency of three times per month and for the central tendency (i.e., midpoint) calculations, EPA used
the midpoint of the frequency, 2.5 times per month. Assuming 12 months per year of exposure, results in
36 exposure days/yr at the high-end and 30 exposure days at the central tendency.

Table 2-76. Summary of Worker Inhalation Monitoring Data for Other DoD Uses of
Perchloroethylene	

Scenario

8-hr
TWA
(ppm)

AC
(ppm)

A IK

(ppm)

I.A IK'
(PP'ii)

Nil m her
of Data
Points

l-lir
TWA
(ppm)

N il in her
of Data
Points

15-
iii imile
TWA
(ppm)

N il in her
of Data
Points

Oil Analysis Results

High-Enda

0.9

0.3

0.1

6.19E-02

1

6.6

1

4.1

1

Central Tendency51

0.1

4.00E-02

Water Pipe Repair Results

Higher Value

2.3

0.8

7.60E-02

7.60E-02

1

No 1-hr or 15-minute data
identified for this use

Midpoint Value

1.2

0.4

3.17E-02

1.26E-02

AC = Acute Concentration; ADC = Average Daily Concentration; and LADC = Lifetime Average Daily Concentration.
Equations and parameters for calculation of the AC, ADC, and LADC are described in Appendix B.
a Only one data point identified for oil analysis. However, different parameters are used for calculating high-end and central
tendency ADC and LADC. Therefore, a high-end and central tendency are presented based on the single data point.

Source: (Defense Occupational and Environmental Health Readiness System - Industrial Hygiene. 2018)

2.22.4 Water Release Assessment

EPA did not identify data in systematic review to estimate wastewater discharges from DoD facilities
using PCE for conditions of use not included in Sections 2.1 through 2.21.

2.23 Dermal Exposure Assessment

Because PCE is a volatile liquid, the dermal absorption of PCE depends on the type and duration of
exposure. Where exposure is not occluded, only a fraction of PCE that comes into contact with the skin
will be absorbed as the chemical readily evaporates from the skin. However, dermal exposure may be
significant in cases of occluded exposure, repeated contacts, or dermal immersion. For example, work
activities with a high degree of splash potential may result in PCE liquids trapped inside the gloves,
inhibiting the evaporation of PCE and increasing the exposure duration.

To assess exposure, EPA used the Dermal Exposure to Volatile Liquids Model (see Equation 2-3) to
calculate the dermal retained dose for both non-occluded and occluded scenarios. The equation 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 and a "protection factor (PF)" to account
for glove use. Default PF values, which vary depending on the type of glove used and the presence of
employee training program, are shown in Table 2-77:

Page 187 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Equation 2-3

( Qu xfabs)

n 	 C v	ac-by v V v E"V

uexp	^ 1derm ^ 1 1

Where:

S is the surface area of contact (cm2)

Qu is the quantity remaining on the skin (mg/cm2-event)

Yderm is the weight fraction of the chemical of interest in the liquid (0 < Yderm < 1)

FT is the frequency of events (integer number per day)

fabs is the fraction of applied mass that is absorbed (Default for PCE: 0.13 for industrial facilities

and 0.19 for commercial facilities)

PF is the glove protection factor (Default: see Table 2-77)

Table 2-77. Glove Protection Factors for Different Dermal Protection Strategies

Dermal Protect ion Characteristics

Setting

Protection
Kactor. PV

a. No gloves used, or any glove / gauntlet without permeation data
and without employee training

Industrial and

Commercial

Uses

1

b. Gloves with available permeation data indicating that the
material of construction offers good protection for the substance

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

Table 2-78 presents the estimated dermal retained dose for workers in various exposure scenarios,
including what-if scenarios for glove use. The dose estimates assume one exposure event (applied dose)
per work day and that 13 to 19 percent28 of the applied dose is absorbed through the skin. The exposure
estimates are provided for each condition of use, where the conditions of us are "binned" based on the
maximum possible exposure concentration (Yderm) and the likely level of exposure. The exposure
concentration is determined based on EPA's review of currently available products and formulations
containing PCE:

• Bin 1: Bin 1 covers industrial uses that generally occur in closed systems. For these uses, dermal
exposure is likely limited to chemical loading/unloading activities (e.g. connecting hoses) and
taking quality control samples. EPA assesses the following glove use scenarios for Bin 1
conditions of use:

o No gloves used: Operators in these industrial uses, while working around closed-system
equipment, may not wear gloves or may wear gloves for abrasion protection or gripping
that are not chemical resistant.

28 The absorbed fraction (fabs) is a function of indoor air speed, which differs for industrial and commercial settings.

Page 188 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

o Gloves used with a protection factor of 5, 10, and 20: Operators may wear chemical-
resistant gloves when taking quality control samples or when connecting and
disconnecting hoses during loading/unloading activities. EPA assumes gloves may offer a
range of protection, depending on the type of glove and employee training provided.
Bin 2: Bin 2 covers industrial degreasing and chemical maskant uses, which are not closed
systems. For these uses, there is greater opportunity for dermal exposure during activities such as
charging and draining degreasing/milling equipment, drumming waste solvent, handling
recycled/re-captured maskants, and removing waste sludge. EPA assesses the following glove
use scenarios for Bin 2 conditions of use:

o No gloves used: Due to the variety of shop types in these uses the actual use of gloves is
uncertain. EPA assumes workers may not wear gloves or may wear gloves for abrasion
protection or gripping that are not chemical resistant during routine operations such as
adding and removing parts from degreasing equipment,
o Gloves used with a protection factor of 5, 10, and 20: Workers may wear chemical-
resistant gloves when charging and draining degreasing/milling equipment, drumming
waste solvent, handling recycled/re-captured maskants, and removing waste sludge. EPA
assumes gloves may offer a range of protection, depending on the type of glove and
employee training provided.

Bin 3: Bin 3 covers aerosol uses, where workers are likely to have direct dermal contact with
film applied to substrate and incidental deposition of aerosol to skin. EPA assesses the following
glove use scenarios for Bin 3 conditions of use:

o No gloves used: Actual use of gloves in this use is uncertain. EPA assumes workers may
not wear gloves or may wear gloves for abrasion protection or gripping that are not
chemical resistant during routine aerosol applications,
o Gloves used with a protection factor of 5 and 10: Workers may wear chemical-resistant
gloves when applying aerosol products. EPA assumes the commercial facilities in Bin 3
do not offer activity-specific training on donning and doffing gloves,
o Scenarios not assessed: EPA does not assess glove use with protection factors of 20 as
EPA assumes chemical-resistant gloves used in these industries would either not be
accompanied by training or be accompanied by basic employee training, but not activity-
specific training.

Bin 4: Bin 4 covers commercial activities of similar maximum concentration. Most of these uses
are uses at dry cleaners, and/or uses expected to have direct dermal contact with bulk liquids. At
dry cleaning shops, workers may be exposed to bulk liquids while charging and draining solvent
to/from machines, removing and disposing sludge, and maintaining equipment. Workers can also
be exposed to PCE used in spot cleaning products at the same shop. EPA assesses the following
glove use scenarios for Bin 4 conditions of use:

o No gloves used: Actual use of gloves in this use is uncertain. EPA assumes workers may

not wear gloves during routine operations (e.g., spot cleaning),
o Gloves used with a protection factor of 5 and 10: Workers may wear chemical-resistant
gloves when charging and draining solvent to/from machines, removing and disposing
sludge, and maintaining equipment. EPA assumes the commercial facilities in Bin 3 do
not offer activity-specific training on donning and doffing gloves,
o Scenarios not assessed: EPA does not assess glove use with protection factors of 20 as
EPA assumes chemical-resistant gloves used in these industries would either not be

Page 189 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

accompanied by training or be accompanied by basic employee training, but not activity-
specific training.

•	Bin 5: Bin 5 covers uses of metalworking fluids containing PCE. These product formulations are
expected to be used in industrial settings and workers may be exposed when unloading the
metalworking fluid from containers; transferring fluids to the trough; and performing metal
shaping operations. EPA assesses the following glove use scenarios for Bin 5 conditions of use:

o No gloves used: Actual use of gloves in this use is uncertain. EPA assumes workers may
not wear gloves during routine operations.

o Gloves used with a protection factor of 5, 10, and 20: Workers may wear chemical-
resistant gloves when unloading the metalworking fluid from containers; transferring
fluids to the trough; and performing metal shaping operations. EPA assumes gloves may
offer a range of protection, depending on the type of glove and employee training
provided.

•	Bin 6: Bin 6 covers uses of adhesives, sealants, paints, and coatings containing PCE. These
product formulations may have both industrial and commercial uses and workers may be
exposed when mixing coating/adhesive, charging products to application equipment (e.g., spray
guns, roll applicators, etc.), and cleaning application equipment. Other workers may also have
incidental contact with applied products during subsequent fabrication steps. EPA assesses the
following glove use scenarios for Bin 6 conditions of use:

o No gloves used: Actual use of gloves in this use is uncertain. EPA assumes workers may
not wear gloves or may wear gloves for abrasion protection or gripping that are not
chemical resistant during routine operations such as coating/adhesive applications and
fabrication steps.

o Gloves used with a protection factor of 5, 10, and 20 (industrial only): Workers may wear
gloves when mixing coating/adhesive, charging products to application equipment (e.g.,
spray guns, roll applicators, etc.), and cleaning application equipment. Coating/adhesive
applications may occur at both industrial and commercial facilities. EPA assumes that
commercial facilities in Bin 6 do not offer activity-specific training on donning and
doffing gloves, but that the industrial facilities may offer such training.

As shown in the table, the calculated retained dose is low for all non-occluded scenarios as PCE
evaporates quickly after exposure. Dermal exposure to liquid is not expected for occupational non-
users, as they do not directly handle PCE.

Page 190 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Table 2-78. Estimated Dermal Acute Retained Dose for Workers in All Conditions of Use









Dermal Kxposure (m«/k«-dav)

Condition of I se

liin

Max

^ (Ici lll

No (Joves
(PI- = 1)

Protective
(•loves

(Pr = 5)

Protective
(Jovcs (PI =
10)

Protective
(•loves
(Industrial uses.
PI- = 20)

Manufacture













Repackaging













Processing as a Reactant













Incorporation into Formulation, Mixture, or Reaction
Product

Bin 1

1.0

1.2 (CT)
3.5(HE)

0.2 (CT)
0.7 (HE)

0.1 (CT)
0.4 (HE)

5.89E-02 (CT)
0.2 (HE)

Industrial Processing Aid













Other Industrial Uses













Waste Handling, Disposal, Treatment, and Recycling













Batch Open-Top Vapor Degreasing













Batch Closed-Loop Vapor Degreasing













Conveyorized Vapor Degreasing

Bin 2

1.0

1.2(CT)

0.2 (CT)

0.1 (CT)

5.89E-02 (CT)

Web Degreasing

3 .5 (HE)

0.7 (HE)

0.4 (HE)

0.2 (HE)

Cold Cleaning













Maskant for Chemical Milling













Aerosol Degreasing and Aerosol Lubricants

Bin 3

0.98

1.8 (CT)
5 .3 (HE )

0.4 (CT)
1.1 (HE)

0.2 (CT)
0.5 (HE)

N/A


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.









Dcrniiil Kxposnre (mg/kg-riiiv)

Condition of I so

liin

Msix

^ (Icmi

No (Joves
(PI- = 1)

Protect ive
(•loves

(Pr = 5)

Protective
(Jovcs (PI =
10)

Protective
(•loves
(Inriiistrhil uses.
PI- = 20)

Dry Cleaning and Spot Cleaning













Wipe Cleaning

Bin 4

1.0

1.8 (CT)

0.4 (CT)

0.2 (CT)

N/A

Other Spot Cleaning/Spot Remover

5 .4 (HE)

1.1 (HE)

0.5 (HE)

Other Commercial Uses













Metalworking Fluids

Bin 5

0.10

0.1 (CT)
0.4 (HE)

2.35E-02 (CT)
7.06E-02 (HE)

1.18E-02 (CT)
3.53E-02 (HE)

5.89E-03 (CT)
1.77E-02 (HE)

Adhesives, Sealants, Paints, and Coatings (Industrial)

Bin 6

0.80

0.9 (CT)
2.8 (HE)

0.2 (CT)
0.6 (HE)

9.42E-02 (CT)
0.3 (HE)

4.71E-02 (CT)
0.1 (HE)

Adhesives, Sealants, Paints, and Coatings
(Commercial)

0.80

1.4 (CT)
4.3 (HE)

0.3 (CT)
0.9 (HE)

0.1 (CT)
0.4 (HE)

N/A

CT = Central Tendency; HE = High-End. Equations anc

parameters for calculation oi

" dermal exposures are described in Appendix K.

Page 192 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

3 Discussion of Uncertainties and Limitations

3.1	Variability

EPA addressed variability in 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 is
not known, assumptions are made to estimate the parameter distribution using available literature data.

3.2	Uncertainties and Limitations

Uncertainty is "the lack of knowledge about specific variables, parameters, models, or other factors" and
can be described qualitatively or quantitatively (	2001b). The following sections discuss

uncertainties in each of the assessed conditions of use scenarios.

3.2.1 Number of Workers

There are a number of uncertainties surrounding the estimated number of workers potentially exposed to
PCE, as outlined below. Most are unlikely to result in a systematic underestimate or overestimate but
could result in an inaccurate estimate.

CDR data are used to estimate the number of workers associated with manufacturing. There are inherent
limitations to the use of CDR data as they are reported by manufacturers and importers of PCE.
Manufacturers and importers are only required to report if they manufactured or imported PCE in excess
of 25,000 pounds at a single site during any calendar from 2012 to 2015; as such, CDR may not capture
all sites and workers associated with any given chemical. Second, the estimate is based on information
that is known or reasonably ascertainable to the submitter. CDR submitters (chemical manufacturers and
importers) do not always have accurate information on the number of potentially exposed workers at
downstream processing sites.

There are also uncertainties with BLS data, which are used to estimate the number of workers for the
remaining conditions of use. First, BLS' OES employment data for each industry/occupation
combination are only available at the 3-, 4-, or 5-digit NAICS level, rather than the full 6-digit NAICS
level. This lack of granularity 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, in reality, likely to use PCE
for the assessed conditions of use. 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 PCE exposure differs
from the overall distribution of workers in each NAICS, then this approach will result in inaccuracy.

Second, EPA's judgments about which industries (represented by NAICS codes) and occupations
(represented by SOC codes) are associated with the uses assessed in this report are based on EPA's
understanding of how PCE is used in each industry. Designations of which industries and occupations
have potential exposures is nevertheless subjective, and some industries/occupations with few exposures
might erroneously be included, or some industries/occupations with exposures might erroneously be
excluded. This would result in inaccuracy but would be unlikely to systematically either overestimate or
underestimate the count of exposed workers.


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

3.2.2	Analysis of Exposure Monitoring Data

To analyze the exposure data, EPA categorized individual PBZ data points as either "worker" or
"occupational non-user". The categorizations are based on descriptions of worker job activity as
provided in literature and EPA's judgment. In general, samples for employees that are expected to have
the highest exposure from direct handling of PCE are categorized as "worker" and samples for
employees that are expected to have lower exposure and do not directly handle PCE are categorized as
"occupational non-user".

Exposures for occupational non-users can vary substantially. Most data sources do not sufficiently
describe the proximity of these employees to the PCE exposure source. As such, exposure levels for the
"occupational non-user" category will have high variability depending on the specific work activity
performed. It is possible that some employees categorized as "occupational non-user" have exposures
similar to those in the "worker" category depending on their specific work activity pattern.

Some data sources may be inherently biased. For example, bias may be present if exposure monitoring
was conducted to address concerns regarding adverse human health effects reported following exposures
during use. Similarly, OSHA CEHD are obtained from OSHA inspections, which may be the result of
worker complaints, and may provide exposure results that are generally more conservative than the
industry average.

Some scenarios have limited exposure monitoring data in literature, if any. Where few data are
available, the assessed exposure levels are unlikely to be representative of worker exposure across the
entire job category or industry. In addition, exposure data for compliance safety and health officers may
not be representative of typical exposure levels for occupational non-users.

In cases where there was no exposure monitoring data, EPA may have used monitoring data from
similar conditions of use as surrogate. While these conditions of use have similar worker activities
contributing to exposures, it is unknown if the results will be fully representative of worker exposure
across different conditions of use.

Where the sample data set contains six or more data points, the 50th and 95th percentile exposure
concentrations were calculated from the sample to represent central tendency and high-end exposure
levels, using available data. The underlying distribution of the data, and the representativeness of the
available data, are not known. Where discrete data was not available, EPA used reported statistics (i.e.,
median, mean, 90th percentile, etc.). Since EPA could not verify these values, there is an added level of
uncertainty.

3.2.3	Near-Field/Far-Field Model Framework

The near-field/far-field approach is used as a framework to model inhalation exposure for many
conditions of use. The following describe uncertainties and simplifying assumptions generally
associated with this modeling approach:

• There is some degree of uncertainty associated with each model input parameter. In general, the
model inputs were determined based on review of available literature. Where the distribution of
the input parameter is known, a distribution is assigned to capture uncertainty in the Monte Carlo
analysis. Where the distribution is unknown, a uniform distribution is often used. The use of a

Page 194 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

uniform distribution will capture the low-end and high-end values but may not accurately reflect
actual distribution of the input parameters.

•	The model assumes the near-field and far-field are well mixed, such that each zone can be
approximated by a single, average concentration.

•	All emissions from the facility are assumed to enter the near-field zone. This assumption will
overestimate exposures and risks in facilities where some emissions do not enter the airspaces
relevant to worker exposure modeling.

•	The exposure models estimate airborne concentrations. Exposures are calculated by assuming
workers spend the entire activity duration in their respective exposure zones (i.e., the worker in
the near-field and the occupational non-user in the far-field). Since vapor degreasing and cold
cleaning involve automated processes, a worker may actually walk away from the near-field
during part of the process and return when it is time to unload the degreaser. As such, assuming
the worker is exposed at the near-field concentration for the entire activity duration may
overestimate exposure.

•	For certain PCE applications (e.g. vapor degreasing and cold cleaning), PCE vapor is assumed to
emit continuously while the equipment operates (i.e. constant vapor generation rate). Actual
vapor generation rate may vary with time. However, small time variability in vapor generation is
unlikely to have a large impact in the exposure estimates as exposures are calculated as a time-
weighted average.

•	The exposure models represent model workplace settings for each PCE condition of use. The
models have not been regressed or fitted with monitoring data.

Each subsequent section below discusses uncertainties associated with the individual model.

3.2.3.1 Tank Truck and Railcar Loading and Unloading Release and Inhalation
Exposure Model

For the other industrial uses and waste handling, disposal, treatment, and recycling conditions of use, the
Tank Truck and Railcar Loading and Unloading Release and Inhalation Exposure Model is used to
estimate the airborne concentration associated with generic chemical loading scenarios at industrial
facilities. Specific uncertainties associated with this model are described below:

•	After each loading event, the model assumes saturated air containing PCE that remains in the
transfer hose and/or loading arm is released to air. The model calculates the quantity of saturated
air using design dimensions of loading systems published in the OPW Engineered Systems
catalog and engineering judgment. These dimensions may not be representative of the whole
range of loading equipment used at industrial facilities handling PCE.

•	The model estimates fugitive emissions from equipment leaks using total organic compound
emission factors from EPA's Protocol for Equipment Leak Emission Estimates (

1995). and engineering judgement on the likely equipment type used for transfer (e.g. number of
valves, seals, lines, and connections). The applicability of these emission factors to PCE, and the
accuracy of EPA's assumption on equipment type are not known.

•	The model assumes the use of a vapor balance system to minimize fugitive emissions. Although
most industrial facilities are likely to use a vapor balance system when loading/unloading
volatile chemicals, EPA does not know whether these systems are used by all facilities that
potentially handle PCE.

Page 195 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

3.2.3.2	EPA AP-42 Loading Model and EPA Mass Balance Inhalation Model

For the incorporation into formulation for non-aerosol formulations assessment, the EPA AP-42 Loading
Model and the EPA Mass Balance Inhalation Model were used to estimate the airborne concentration
associated with loading of formulation into drums at industrial facilities. Specific uncertainties
associated with these models are described below:

•	The model assumes all formulated products are loaded into 55-gallon drums but does not
consider the potential for loading products into smaller containers instead of or in addition to
drums.

•	The model assumes that the process steps associated with drum loading occurs indoors, without
engineering controls, and in an open-system environment where vapors freely escape. In the absence of
industry-specific information, these assumptions provide for conservative estimates for exposures during
this operation. Actual exposures may be less due to various factors including closed-system loading and
unloading, the use of vapor recovery systems, or the automation of various process steps.

•	The model also does not consider exposure from unloading raw PCE from bulk containers (i.e.
tank trucks or railcars). Although EPA can estimate exposures during this unloading activity
using the Tank Truck and Railcar Loading and Unloading Release and Inhalation Exposure
Model, it is unclear if the same workers will perform both unloading and loading activities in the
same day. Therefore, it may not be accurate to combine estimates from each model to estimate a
total exposure. In the case where a worker is both unloading bulk containers and loading
products into drums on the same day, the overall error from not including exposures during
unloading in the results is expected to be small as the Tank Truck and Railcar Loading and
Unloading Release and Inhalation Exposure Model estimates an 8-hr TWA exposure of 0.01
ppm for tank truck unloading and an 8-hr TWA of 0.04 ppm for railcar unloading whereas the
model for drum loading estimates 8-hr TWAs ranging from 0.60 to 14.1 ppm.

•	The model does not account for other potential sources of exposure at industrial facilities, such
as sampling, equipment cleaning, and other process activities that can contribute to a worker's
overall 8-hr daily exposure. These model uncertainties could result in an underestimate of the
worker 8-hr exposure.

3.2.3.3	Vapor Decreasing and Cold Cleaning Models

The conveyorized vapor degreasing, web degreasing, and cold cleaning assessments use a near-field/far-
field approach to model worker exposure. In addition to the uncertainties described above, the vapor
degreasing and cold cleaning models have the following uncertainties:

•	To estimate vapor generation rate for each equipment type, EPA used a distribution of the
emission rates reported in the 2014 NEI for each degreasing/cold cleaning equipment type. NEI
only contains information on major sources not area sources. Therefore, the emission rate
distribution used in modeling may not be representative of degreasing/cold cleaning equipment
emission rates at area sources.

•	The emission rate for conveyorized vapor degreasing is based on equipment at a single site and
the emission rates for web degreasing are based on equipment from two sites. It is uncertain how
representative these data are of a "typical" site.

Page 196 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

•	EPA assumes workers and occupational non-users remove themselves from the contaminated
near- and far-field zones at the conclusion of the task, such that they are no longer exposed to
any residual PCE in air.

3.2.3.4	Brake Servicing Model

The aerosol degreasing assessment also uses a near-field/far-field approach to model worker exposure.

Specific uncertainties associated with the aerosol degreasing scenario are presented below:

•	The model references a CARB study (	300) on brake servicing to estimate use rate and
application frequency of the degreasing product. The brake servicing scenario may not be
representative of the use rates for other aerosol applications involving PCE.

•	The CARB study (CARB. 2000) presented 13 different aerosol degreasing formulations
containing PCE. For each Monte Carlo iteration, the model determines the PCE concentration in
product by selecting one of 13 possible formulations, assuming the distribution for each
formulation is equal to that found in a survey of brake cleaning shops in California. It is
uncertain if this distribution is representative of other geographic locations within the U.S.

•	Some of the aerosol formulations presented in the CARB study (	2000) were provided as
ranges. For each Monte Carlo iteration the model selects a PCE concentration within the range of
concentrations using a uniform distribution. In reality, the PCE concentration in the formulation
may be more consistent than the range provided.

3.2.3.5	Dry Cleaning Model

The multi-zone dry cleaning model also uses a near-field/far-field approach. Specific uncertainties

associated with the dry cleaning scenario are presented below:

•	The model assumes each facility only has one dry cleaning machine, cleaning one to fourteen
loads of garments per day. The number of machines is based on the 2010 King County, WA
survey (Whittaker and Johanson.: ) where 96 percent of 151 respondents reported having
only one machine at their facility. It is uncertain if this distribution is representative of other
geographic locations in the U.S. Larger facilities are likely to have more machines, which could
result in additional PCE exposures.

•	The model conservatively uses a hemispherical volume based on the dry cleaning machine door
diameter as the near-field for machine unloading. The small near-field volume results in a large
spike in concentration when the machine door is opened, where any residual PCE solvent is
assumed to be instantaneously released into the near-field. In reality, the residual solvent will
likely be released continuously over a period of time. In addition, the worker may move around
while unloading the garments, such that the worker's breathing zone will not always be next to
the machine door throughout the duration of this activity. Therefore, these assumptions may
result in an overestimate of worker exposure during machine unloading.

•	Many of the model input parameters were obtained from (Von Grote. 2003). which is a German
study. Aspects of the U.S. dry cleaning facilities may differ from German facilities. However, it
is not known whether the use of German data will under- or over-estimate exposure.

•	The model does not cover all potential worker activities at dry cleaners. For example, workers
could be exposed to PCE emitted due to equipment leaks, when re-filling PCE solvent into dry
cleaning machines, when interrupting a dry cleaning cycle, or when performing maintenance
activities (e.g., cleaning lint and button traps, raking out the still, changing solvent filter, and

Page 197 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

handling solvent waste) ("OSHA. 2005). However, there is a lack of information on these
activities in the literature, and the frequency of these activities is not well understood. The
likelihood of equipment leaks is dependent on whether the machines are properly maintained.
The frequency of solvent re-filling depends on a specific dry cleaner's workload and solvent
consumption rate, which is also affected by the presence of leaks. Based on observations
reported by NIOSH (2010) and Blandof2010). solvent charging is not performed every day. EPA
was unable to develop a modeling approach for these exposure activities due to the lack of
available information.

3,2,4 Modeled Dermal Exposures

The Dermal Exposure to Volatile Liquids Model used to estimate dermal exposure to PCE in
occupational settings. The model 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 and does not address variability in exposure duration and frequency.

Page 198 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

REFERENCES

Page 199 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Appendix A Approach for Estimating Number of Workers and
Occupational Non-Users

This appendix summarizes the methods that EPA used to estimate the number of workers who are
potentially exposed to PCE in each of its conditions of use. The method consists of the following steps:

1.	Identify the North American Industry Classification System (NAICS) codes for the industry
sectors associated with each condition of use.

2.	Estimate total employment by industry/occupation combination using the Bureau of Labor
Statistics" Occupational Employment Statistics (OES) data (	L 2016).

3.	Refine the OES estimates where they are not sufficiently granular by using the U.S. Census
Bureau ( ) Statistics of U.S. Businesses (SUSB) data on total employment by 6-digit NAICS.

4.	Estimate the percentage of employees likely to be using PCE instead of other chemicals (i.e., the
market penetration of PCE in the condition of use).

5.	Estimate the number of sites and number of potentially exposed employees per site.

6.	Estimate the number of potentially exposed employees within the condition of use.

Step 1: Identifying Affected NAICS Codes

As a first step, EPA identified NAICS industry codes associated with each condition of use. EPA
generally identified NAICS industry codes for a condition of use by:

•	Querying the	sus Bureau's NAICS Search tool using keywords associated with each
condition of use to identify NAICS codes with descriptions that match the condition of use.

•	Referencing EPA Generic Scenarios (GS's) and Organisation for Economic Co-operation and
Development (OECD) Emission Scenario Documents (ESDs) for a condition of use to identify
NAICS codes cited by the GS or ESD.

•	Reviewing Chemical Data Reporting (CDR) data for the chemical, identifying the industrial
sector codes reported for downstream industrial uses, and matching those industrial sector codes
to NAICS codes using Table D-2 provided in the CDR reporting instructions.

Each condition of use section in the main body of this report identifies the NAICS codes EPA identified
for the respective condition of use.

Step 2: Estimating Total Employment by Industry and Occupation

U.S. BLS (2016) OES data provide employment data for workers in specific industries and occupations.
The industries are classified by NAICS codes (identified previously), and occupations are classified by
Standard Occupational Classification (SOC) codes.

Among the relevant NAICS codes (identified previously), EPA reviewed the occupation description and
identified those occupations (SOC codes) where workers are potentially exposed to PCE. Table Apx
A-l shows the SOC codes EPA classified as occupations potentially exposed to PCE. These occupations
are classified into workers (W) and occupational non-users (O). All other SOC codes are assumed to
represent occupations where exposure is unlikely.

Page 200 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

TableApx A-l. SOCs with Worker and ONU Designations for All Conditions of Use Except Dry
Cleaning	

soc

Occupation

Designation

11-9020

Construction Managers

O

17-2000

Engineers

O

17-3000

Drafters, Engineering Technicians, and Mapping Technicians

0

19-2031

Chemists

0

19-4000

Life, Physical, and Social Science Technicians

0

47-1000

Supervisors of Construction and Extraction Workers

0

47-2000

Construction Trades Workers

w

49-1000

Supervisors of Installation, Maintenance, and Repair Workers

0

49-2000

Electrical and Electronic Equipment Mechanics, Installers, and Repairers

w

49-3000

Vehicle and Mobile Equipment Mechanics, Installers, and Repairers

w

49-9010

Control and Valve Installers and Repairers

w

49-9020

Heating, Air Conditioning, and Refrigeration Mechanics and Installers

w

49-9040

Industrial Machinery Installation, Repair, and Maintenance Workers

w

49-9060

Precision Instrument and Equipment Repairers

w

49-9070

Maintenance and Repair Workers, General

w

49-9090

Miscellaneous Installation, Maintenance, and Repair Workers

w

51-1000

Supervisors of Production Workers

0

51-2000

Assemblers and Fabricators

w

51-4020

Forming Machine Setters, Operators, and Tenders, Metal and Plastic

w

51-6010

Laundry and Dry-Cleaning Workers

w

51-6020

Pressers, Textile, Garment, and Related Materials

w

51-6030

Sewing Machine Operators

0

51-6040

Shoe and Leather Workers

0

51-6050

Tailors, Dressmakers, and Sewers

0

51-6090

Miscellaneous Textile, Apparel, and Furnishings Workers

0

51-8020

Stationary Engineers and Boiler Operators

w

51-8090

Miscellaneous Plant and System Operators

w

51-9000

Other Production Occupations

w

W = worker designation
O = ONU designation

Page 201 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

For dry cleaning facilities, due to the unique nature of work expected at these facilities and that different
workers may be expected to share among activities with higher exposure potential (e.g., unloading the
dry cleaning machine, pressing/finishing a dry cleaned load), EPA made different SOC code worker and
ONU assignments for this condition of use. Table Apx A-2 summarizes the SOC codes with worker and
ONU designations used for dry cleaning facilities.

Table Apx A-2. SOCs with Worker and ONU Designations for Dry Cleaning Facilities

SOC

Occupation

Designation

41-2000

Retail Sales Workers

O

49-9040

Industrial Machinery Installation, Repair, and Maintenance Workers

W

49-9070

Maintenance and Repair Workers, General

w

49-9090

Miscellaneous Installation, Maintenance, and Repair Workers

w

51-6010

Laundry and Dry-Cleaning Workers

w

51-6020

Pressers, Textile, Garment, and Related Materials

w

51-6030

Sewing Machine Operators

0

51-6040

Shoe and Leather Workers

0

51-6050

Tailors, Dressmakers, and Sewers

0

51-6090

Miscellaneous Textile, Apparel, and Furnishings Workers

0

W = worker designation
O = ONU designation

After identifying relevant NAICS and SOC codes, EPA used BLS data to determine total employment
by industry and by occupation based on the NAICS and SOC combinations. For example, there are
110,640 employees associated with 4-digit NAICS 8123 {Dry cleaning and Laundry Services) and SOC
51 -6010 (Laundry and Dry-Cleaning Workers).

Using a combination of NAICS and SOC codes to estimate total employment provides more accurate
estimates for the number of workers than using NAICS codes alone. Using only NAICS codes to
estimate number of workers typically result in an overestimate, because not all workers employed in that
industry sector will be exposed. However, in some cases, BLS only provide employment data at the 4-
digit or 5-digit NAICS level; therefore, further refinement of this approach may be needed (see next
step).

Step 3: Refining Employment Estimates to Account for lack of NAICS Granularity

The third step in EPA's methodology was to further refine the employment estimates by using total
employment data in the U.S. Census Bureau (2015) SUSB. In some cases, BLS OES's occupation-
specific data are only available at the 4-digit or 5-digit NAICS level, whereas the SUSB data are
available at the 6-digit level (but are not occupation-specific). Identifying specific 6-digit NAICS will
ensure that only industries with potential PCE exposure are included. As an example, OES data are

Page 202 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

available for the 4-digit NAICS 8123 Drycleaning and Laundry Services, which includes the following
6-digitNAICS:

•	NAICS 812310 Coin-Operated Laundries and Drycleaners;

•	NAICS 812320 Drycleaning and Laundry Services (except Coin-Operated);

•	NAICS 812331 Linen Supply; and

•	NAICS 812332 Industrial Launderers.

In this example, only NAICS 812320 is of interest. The Census data allow EPA to calculate employment
in the specific 6-digit NAICS of interest as a percentage of employment in the BLS 4-digit NAICS.

The 6-digitNAICS 812320 comprises 46 percent of total employment under the 4-digit NAICS 8123.
This percentage can be multiplied by the occupation-specific employment estimates given in the BLS
OES data to further refine our estimates of the number of employees with potential exposure.

Table_Apx A-3 illustrates this granularity adjustment for NAICS 812320.

Page 203 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

TableApx A-3. Estimated Number of Potentially Exposed Workers and ONUs under NAICS
812320

NAICS

SOC
( ode

SOC Description

Occupation
Designation

Km ploy mcnl
by SOC at

4-di«i(
NAICS level

% ol* Total
Km ploy mcnl

Kslimalcd
Kinploymenl
hy SOC at

6-digit
NAICS level

8123

41-2000

Retail Sales Workers

O

44,500

46.0%

20,459

8123

49-9040

Industrial Machinery
Installation, Repair, and
Maintenance Workers

W

1,790

46.0%

823

8123

49-9070

Maintenance and Repair
Workers, General

w

3,260

46.0%

1,499

8123

49-9090

Miscellaneous
Installation,

Maintenance, and Repair
Workers

w

1,080

46.0%

497

8123

51-6010

Laundry and Dry-
Cleaning Workers

w

110,640

46.0%

50,867

8123

51-6020

Pressers, Textile,
Garment, and Related
Materials

w

40,250

46.0%

18,505

8123

51-6030

Sewing Machine
Operators

0

1,660

46.0%

763

8123

51-6040

Shoe and Leather
Workers

0

Not Reported for this NAICS Code

8123

51-6050

Tailors, Dressmakers,
and Sewers

0

2,890

46.0%

1,329

8123

51-6090

Miscellaneous Textile,
Apparel, and
Furnishings Workers

0

0

46.0%

0

Total Potentially Exposed Employees

206,070



94,740

Total Workers





72,190

Total Occupational Non-Users





22,551

Note: numbers may not sum exactly due to rounding.
W = worker

O = occupational non-user

Source: (U.S. BLS. 20.1.6: U.S. Census Bureau. 20.1.5')

Page 204 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Step 4: Estimating the Percentage of Workers Using PCE Instead of Other Chemicals

In the final step, EPA accounted for the market share by applying a factor to the number of workers
determined in Step 3. This accounts for the fact that PCE may be only one of multiple chemicals used
for the applications of interest. EPA only identified market penetration data for a limited number of
conditions of use. In the absence of market penetration data for a given condition of use, EPA assumed
PCE may be used at up to all sites and by up to all workers calculated in this method as a bounding
estimate. This assumes a market penetration of 100%. Market penetration is discussed for each condition
of use in the main body of this report.

Step 5: Estimating the Number of Workers per Site

EPA calculated the number of workers and occupational non-users in each industry/occupation
combination using the formula below (granularity adjustment is only applicable where SOC data are not
available at the 6-digitNAICS level):

Number of Workers or ONUs in NAICS/SOC (Step 2) x Granularity Adjustment Percentage (Step 3) =
Number of Workers or ONUs in the Industry/Occupation Combination

EPA then estimated the total number of establishments by obtaining the number of establishments
reported in the U.S. Census Bureau (2015) SUSB data at the 6-digit NAICS level.

EPA then summed the number of workers and occupational non-users over all occupations within a
NAICS code and divided these sums by the number of establishments in the NAICS code to calculate
the average number of workers and occupational non-users per site.

Step 6: Estimating the Number of Workers and Sites for a Condition of Use

EPA estimated the number of workers and occupational non-users potentially exposed to PCE and the
number of sites that use PCE in a given condition of use through the following steps:

6. A. Obtaining the total number of establishments by:

i.	Obtaining the number of establishments from SUSB (U.S. Census Bureau. 2015) at the 6-
digit NAICS level (Step 5) for each NAICS code in the condition of use and summing
these values; or

ii.	Obtaining the number of establishments from the Toxics Release Inventory (TRI),
Discharge Monitoring Report (DMR) data, National Emissions Inventory (NEI), or
literature for the condition of use.

6.B. Estimating the number of establishments that use PCE by taking the total number of

establishments from Step 6. A and multiplying it by the market penetration factor from Step
4.

6.C. Estimating the number of workers and occupational non-users potentially exposed to PCE by
taking the number of establishments calculated in Step 6.B and multiplying it by the average
number of workers and occupational non-users per site from Step 5.

Page 205 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Appendix B Equations for Calculating Acute and Chronic (Non-
Cancer and Cancer) Inhalation Exposures

This report assesses PCE exposures to workers in occupational settings, presented as 8-hr time weighted
average (TWA). The 8-hr TWA exposures are then used to calculate acute exposure (AC), average daily
concentration (ADC) for chronic, non-cancer risks, and lifetime average daily concentration (LADC) for
chronic, cancer risks.

Acute workplace exposures are assumed to be equal to the contaminant concentration in air (8-hr TWA),
per EquationApx B-l.

EquationApx B-l

C x ED
AC =—	

AT

ri1 acute

Where:

AC = acute exposure concentration
C = contaminant concentration in air (TWA)

ED = exposure duration (hr/day)

ATacute = acute averaging time (hr)

ADC and LADC are used to estimate workplace exposures for non-cancer and cancer risks, respectively.
These exposures are estimated as follows:

Equation Apx B-2

Equation Apx B-3

Equation Apx B-4

CxEDxEFxWY

ADC or LADC =	—	—	

AT or ATC

day hr
AT = WYx 365 —x 24-

yr day

day hr
ATC = LTx 365 —x 24-

yr day

Where:

ADC	= Average daily concentration used for chronic non-cancer risk calculations

LADC = Lifetime average daily concentration used for chronic cancer risk calculations

ED	= Exposure duration (hr/day)

EF	= Exposure frequency (day/yr)

WY	= Working years per lifetime (yr)

AT	= Averaging time (hr) for chronic, non-cancer risk

ATc	= Averaging time (hr) for cancer risk

AWD	= Annual working days (day/yr)

f	= Fractional working days with exposure (unitless)

Page 206 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

LT = Lifetime years (yr) for cancer risk

The parameter values in Table Apx B-l are used to calculate each of the above acute or chronic
exposure estimates. Where exposure is calculated using probabilistic modeling, the AC, ADC, and
LADC calculations are integrated into the Monte Carlo simulation. Where multiple values are provided
for ED and EF, it indicates that EPA may have used different values for different conditions of use. The
rationale for these differences are described below in this section.

Table Apx B-l. Parameter Values for Calculating Inhalation Exposure Estimates

Parameter Name

Sv m hoi

Value

1 nil

Exposure Duration

ED

8 or 12

hr/day

Exposure Frequency

EF

250

258 (50th percentile) to 293 (95th
percentile) (dry cleaning only)
125 to 150 (DoD - oil analysis only)
30 to 36 (DoD - water pipe repair only)

days/yr

Working years

WY

31 (50th percentile)
40 (95th percentile)

years

Lifetime Years, cancer

LT

78

years

Averaging Time, non-
cancer

AT

271,560 (central tendency)51
350,400 (high-end)b

hr

Averaging Time, cancer

ATC

683,280

hr

a Calculated using the 50th percentile value for working years (WY)
b Calculated using the 95th percentile value for working years (WY)

Exposure Duration (ED)

EPA generally uses an exposure duration of 8 hours per day for averaging full-shift exposures with two
notable exceptions: manufacturing and results from the Dry Cleaning Multi-Zone Inhalation Exposure
Model. In the manufacturing data there were both 8-hr TWA and 12-hr TWA monitoring data. EPA
used an ED of 8 hours for the 8-hr TWA data and 12 hours for the 12-hr TWA data. For dry cleaning,
the monitoring data were generally 8-hr TWAs; therefore, EPA used an ED of 8 hours. However, EPA
assumes dry cleaners may operate up to 12 hours per day; therefore, when modeling dry cleaning
exposures using the multi-zone model, EPA modeled assuming an ED of 12 hours.

Page 207 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Exposure Frequency (EF)

EPA generally uses an exposure frequency of 250 days per year with two notable exceptions: dry
cleaning and DoD uses. EPA assumed dry cleaners may operate between five and six days per week and
50 to 52 weeks per year resulting in a range of 250 to 312 annual working days per year (AWD). Taking
into account fractional days exposed (f) resulted in an exposure frequency (EF) of 258 at the 50th
percentile and 293 at the 95th percentile. For the two DoD uses, information was provided indicating
process frequencies of two to three times per week (oil analysis) and two to three times per month (water
pipe repair). EPA used the maximum frequency for high-end estimates and the midpoint frequency for
central tendency estimates. For the oil analysis use this resulted in 125 days/yr at the central tendency
and 150 days/yr at the high-end. For the water pipe repair, this resulted in 30 days/yr at the central
tendency and 36 days/yr at the high-end.

EF is expressed as the number of days per year a worker is exposed to the chemical being assessed. In
some cases, it may be reasonable to assume a worker is exposed to the chemical on each working day. In
other cases, it may be more appropriate to estimate a worker's exposure to the chemical occurs during a
subset of the worker's annual working days. The relationship between exposure frequency and annual
working days can be described mathematically as follows:

EquationApx B-5

EF = fx AWD

Where:

EF = exposure frequency, the number of days per year a worker is exposed to the chemical
(day/yr)

f = fractional number of annual working days during which a worker is exposed to the
chemical (unitless)

AWD = annual working days, the number of days per year a worker works (day/yr)

U.S. BLS (2016) provides data on the total number of hours worked and total number of employees by
each industry NAICS code. These data are available from the 3- to 6-digit NAICS level (where 3-digit
NAICS are less granular and 6-digit NAICS are the most granular). Dividing the total, annual hours
worked by the number of employees yields the average number of hours worked per employee per year
for each NAICS.

EPA has identified approximately 140 NAICS codes applicable to the multiple conditions of use for the
ten chemicals undergoing risk evaluation. For each NAICS code of interest, EPA looked up the average
hours worked per employee per year at the most granular NAICS level available (i.e., 4-digit, 5-digit, or
6-digit). EPA converted the working hours per employee to working days per year per employee
assuming employees work an average of eight hours per day. The average number of days per year
worked, or AWD, ranges from 169 to 282 days per year, with a 50th percentile value of 250 days per
year. EPA repeated this analysis for all NAICS codes at the 4-digit level. The average AWD for all 4-
digit NAICS codes ranges from 111 to 282 days per year, with a 50th percentile value of 228 days per
year. 250 days per year is approximately the 75th percentile. In the absence of industry- and PCE-
specific data, EPA assumes the parameter/is equal to one for all conditions of use except dry cleaning.
Dry cleaning used a uniform distribution from 0.8 to 1 for f. The 0.8 value was derived from the

Page 208 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

observation that the weighted average of 200 day/yr worked (from BLS/Census) is 80% of the standard
assumption that a full-time worker works 250 day/yr. The maximum of 1 is appropriate as dry cleaners
may be family owned and operated and some workers may work as much as every operating day.

Working Years (WY)

EPA has developed a triangular distribution for working years. EPA has defined the parameters of the
triangular distribution as follows:

•	Minimum value: BLS CPS tenure data with current employer as a low-end estimate of the
number of lifetime working years: 10.4 years;

•	Mode value: The 50th percentile tenure data with all employers from SIPP as a mode value for
the number of lifetime working years: 36 years; and

•	Maximum value: The maximum average tenure data with all employers from SIPP as a high-end
estimate on the number of lifetime working years: 44 years.

This triangular distribution has a 50th percentile value of 31 years and a 95th percentile value of 40 years.
EPA uses these values for central tendency and high-end ADC and LADC calculations, respectively.

The U.S. BLS (2014) provides information on employee tenure with current employer obtained from the
Current Population Survey (CPS). CPS is a monthly sample survey of about 60,000 households that
provides information on the labor force status of the civilian non-institutional population age 16 and
over; CPS data are released every two years. The data are available by demographics and by generic
industry sectors but are not available by NAICS codes.

The U.S. Census Bureau (2019a) Survey of Income and Program Participation (SIPP) provides
information on lifetime tenure with all employers. SIPP is a household survey that collects data on
income, labor force participation, social program participation and eligibility, and general demographic
characteristics through a continuous series of national panel surveys of between 14,000 and 52,000
households (U.S. Census Bureau. 2019b). EPA analyzed the 2008 SIPP Panel Wave 1, a panel that
began in 2008 and covers the interview months of September 2008 through December 2008 (U.S.

Census Bureau. .Or \i, b). For this panel, lifetime tenure data are available by Census Industry Codes,
which can be cross-walked with NAICS codes.

SIPP data include fields for the industry in which each surveyed, employed individual works
(TJBIND1), worker age (TAGE), and years of work experience with all employers over the surveyed
individual's lifetime.29 Census household surveys use different industry codes than the NAICS codes
used in its firm surveys, so these were converted to NAICS using a published crosswalk (t; S Census
Bureau.: ). EPA calculated the average tenure for the following age groups: 1) workers age 50 and
older; 2) workers age 60 and older; and 3) workers of all ages employed at time of survey. EPA used

29 To calculate the number of years of work experience EPA took the difference between the year first worked
(TMAKMNYR) and the current data year (i.e., 2008). EPA then subtracted any intervening months when not working
(ETIMEOFF).

Page 209 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

tenure data for age group "50 and older" to determine the high-end lifetime working years, because the
sample size in this age group is often substantially higher than the sample size for age group "60 and
older". For some industries, the number of workers surveyed, or the sample size, was too small to
provide a reliable representation of the worker tenure in that industry. Therefore, EPA excluded data
where the sample size is less than five from our analysis.

TableApx B-2 summarizes the average tenure for workers age 50 and older from SIPP data. Although
the tenure may differ for any given industry sector, there is no significant variability between the 50th
and 95th percentile values of average tenure across manufacturing and non-manufacturing sectors.

Table Apx B-2. Overview of Average Worker Tenure from U.S. Census SIPP (Age Group 50+)

Industry Sectors

Average

Workii
50lh Percentile

lg Years
95"' Percentile

Maxim inn

All industry sectors relevant to the 10
chemicals undergoing risk evaluation

35.9

36

39

44

Manufacturing sectors (NAICS 31-33)

35.7

36

39

40

Non-manufacturing sectors (NAICS 42-81)

36.1

36

39

44

Source: (U.S. Census Bureau. 2019a")

Note: Industries where sample size is less than five are excluded from this analysis.

BLS CPS data provides the median years of tenure that wage and salary workers had been with their
current employer. Table Apx B-3 presents CPS data for all demographics (men and women) by age
group from 2008 to 2012. To estimate the low-end value on number of working years, EPA uses the
most recent (2014) CPS data for workers age 55 to 64 years, which indicates a median tenure of 10.4
years with their current employer. The use of this low-end value represents a scenario where workers are
only exposed to the chemical of interest for a portion of their lifetime working years, as they may
change jobs or move from one industry to another throughout their career.

Page 210 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Table Apx B-3. Median Years of Tenure with Current Em

Age

January 2008

January 2010

January 2012

January 2014

16 years and over

4.1

4.4

4.6

4.6

16 to 17 years

0.7

0.7

0.7

0.7

18 to 19 years

0.8

1.0

0.8

0.8

20 to 24 years

1.3

1.5

1.3

1.3

25 years and over

5.1

5.2

5.4

5.5

25 to 34 years

2.7

3.1

3.2

3.0

35 to 44 years

4.9

5.1

5.3

5.2

45 to 54 years

7.6

7.8

7.8

7.9

55 to 64 years

9.9

10.0

10.3

10.4

65 years and over

10.2

9.9

10.3

10.3

jloyer by Age Group

Source: (U.S. BLS. 20.1.4')

Lifetime Years (LT)

EPA assumes a lifetime of 78 years for all worker demographics.

Page 211 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Appendix C Sample Calculations for Calculating Acute and Chronic
(Non-Cancer and Cancer) Inhalation Exposures

Sample calculations for high-end and central tendency acute and chronic (non-cancer and cancer)
exposure concentrations for one condition of use, manufacturing, are demonstrated below. The
explanation of the equations and parameters used is provided in Appendix B.

C.l Example High-End AC, ADC, and LADC Calculations

Calculate AChe:

AChe —

Che x ED

ATi

acute

Calculate ADChe:

2.61 ppm x 8 hr/day
AChe =	24h^	= °"87 PPm

CHE x ED x EF xWY

adche —

hr

AT
days

2.61 ppm x 8-j— x 250—— x 40 years
day	year 7

ADChe =			r	= 0.60 ppm

„,r«ays hr	rr

40 years x 365—— x 24-j—
J	yr	day

Calculate LADChe:

LADChe —

CHE x ED XEF XWY
ATr

2.61 ppm x 8-^— x 250 ^a^5 x 40 years
, day year 7
LADChe =	-T—	= 0.31 ppm

78 years x 365—— x 24 hr/day
7	year	' 7

€.2 Example Central Tendency AC, ADC, and LADC Calculations

Calculate ACct:

ACct —

CCT x ED
~AT,

acute

0.03 ppm x 8 hrlday
AC"= 24 hr/day

Calculate ADCct:

ADCct =

CCT x ED XEF XWY

AT

Page 212 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

0.03 ppm x 8-^— x 250 ^a^S x 31 years
day	year 7

ADCct =	-j	r	= 0.01 ppm

„,raays hr	rr

31 years x 365—— x 24-j—
J	yr	day

Calculate LADCct:

Cct x ED x EF x WY

LADCct = —	—	

ATC

0.03 ppm x 8-7-— x 250 x 31 years

, „ _	day	year 7	„	,

LADCct =		= 2.95 x 10 3 ppm

78 years x 365 a^S x 24 hr/day
7	year	' 7

Page 213 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Appendix D Approach for Estimating Water Releases from
Manufacturing Sites Using Effluent Guidelines

This appendix presents a methodology for estimating water releases of PCE from manufacturing sites
using effluent guidelines (EGs). This method uses the maximum daily and maximum average monthly
concentrations allowed under the Organic Chemicals, Plastics and Synthetic Fibers (OCPSF) Effluent
Guidelines and Standards (	IT). EGs are national regulatory standards set forth by EPA for

wastewater discharges to surface water and municipal sewage treatment plants. The OCPSF EG applies
to facilities classified under the following SIC codes:

•	2821—Plastic Materials, Synthetic Resins, and Nonvulcanizable Elastomers;

•	2823—Cellulosic Man-Made Fibers;

•	2865—Cyclic Crudes and Intermediates, Dyes, and Organic Pigments; and

•	2869—Industrial Organic Chemicals, Not Elsewhere Classified.

Manufacturers of PCE would typically be classified under SIC code 2869; therefore, the requirements of
the OCPSF EG are assumed to apply to manufacturing sites. Subparts I, J, and K of the OCPSF EG set
limits for the concentration of PCE in wastewater effluent for industrial facilities that are direct
discharge point sources using end-of-pipe biological treatment, direct discharge point sources that do not
use end-of-pipe biological treatment, and indirect discharge point sources, respectively (U.S. EPA.
1987). Direct dischargers are facilities that discharge effluent directly to surface waters and indirect
dischargers are facilities that discharge effluent to publicly-owned treatment works (POTW). The
OCPSF limits for PCE in each of the Subparts are provided in TableApx D-l.

Table Apx D-l. Summary of OCPSF Effluent Guidelines for Perchloroethylene

OCPSI-" Subpart

.Maximum
for Any
One Day
(Hg/I.)

.Maximum
for Any
Monthly
Average
(ng/i<)

Basis

Subpart I - Direct Discharge Point Sources That
Use End-of-Pipe Biological Treatment

56

22

BAT effluent limitations
and NSPS

Subpart J - Direct Discharge Point Sources That
Do Not Use End-of-Pipe Biological Treatment

164

52

BAT effluent limitations
and NSPS

Subpart K - Indirect Discharge Point Sources

164

52

Pretreatment Standards
for Existing Sources
(PSES) and
Pretreatment Standards
for New Sources
(PSNS)

BAT = Best Available Technology Economically Achievable; NSPS = New Source Performance Standards; PSES =
Pretreatment Standards for Existing Sources; PSNS = Pretreatment Standards for New Sources.

Source: (TJ.S. EPA. 19871

Page 214 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

To estimate daily releases from the EG, EPA used EquationApx D-l to estimate daily releases and
EquationApx D-2 to estimate annual releases using the parameters in TableApx D-2. The prevalence
of end-of-pipe biological treatment is unknown; therefore, EPA used the discharge limits for direct
discharge point sources that do not use end-of-pipe biological treatment (Subpart J) and indirect
discharge point sources (Subpart K). EPA estimated a central tendency daily release using the limit for
the maximum monthly average (52 |_ig/L) from Subparts J and K, a high-end daily release using the limit
for the maximum for any one day (164 |_ig/L) from Subparts J and K, and an annual release using the
maximum monthly average from Subparts J and K.

Equation Apx D-l

DLxPWx PV
DR ~ 1,000,000,000 x OD

Equation Apx D-2

DLxPW x PV

AD = 	

1,000,000,000

Table Apx D-2. Default Parameters for Estimating Water Releases of Perchloroethylene from

Manufacturing Sites

Pa ram el er

Parameter Doseriplion

Default Value

I nil

DR

Daily release rate

Calculated from
equation

kg/site-day

DL

Discharge limita

Max Daily: 164
Average Daily: 52
Annual: 52

Hg/L

PW

Produced waterb

10

L/kg

PV

Annual PCE production volume

Site-specific

kg/site-yr

OD

Operating Days0

350

days/yr

AR

Annual release rate

Calculated from
equation

kg/site-yr

a Discharge limits are based on the maximum discharge limits allowed in the OCPSF EG, which correspond to the discharge
limits for direct discharge point sources with no biological end-of-pipe treatment (Subpart J) and indirect discharge points
sources (Subpart K) (citation for 40 C.F.R. 414). There is no "average" daily discharge limit set by the EGs; therefore, EPA
assumed that the average daily discharge concentration would be equal to the maximum monthly average discharge limit.
b The amount of produced water per kilogram of PCE produced is based on the SpERC developed by the European Solvent
Industry Group for the manufacture of a substance, which estimates 10 m3 of wastewater generated per metric ton of
substance produced and converted to 10 L/kg (European Solvents Industry Group. 2012).

0 Due to large throughput, manufacturing sites are assumed to operate seven days per week and 50 weeks per year with two
weeks per year for shutdown activities.

EPA did not identify PCE-specific information on the amount of wastewater produced per day. The
Specific Environmental Release Category (SpERC) developed by the European Solvent Industry Group
for the manufacture of a substance estimates 10 m3 of wastewater generated per metric ton of substance
produced (equivalent to 10 L water/kg of substance produced) (European Solvents Indu; 3up.

Page 215 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

2012). In lieu of PCE-specific information, EPA estimated wastewater flow using the SpERC specified
wastewater production volume and the annual PCE production rates for each facility. TableApx D-3
provides estimated daily production volume and wastewater flow for each facility that EPA used the EG
to assess water releases.

Table Apx D-3. Summary of Facility Perchloroethylene Production Volumes and Wastewater
Flow Rates

Nile

Annual
Production

Volume
(kg/site-vr)

Annual Operating
Days
(davs/yr)

Daily Production
Volume
(kg/site-dav)

Daily Wastewater
Mow
(1./site-day)

Axiall Corporation,
Westlake, LAa

20,743,859

350

59,268

592,682

Greenchem,

West Palm Beach,
FLa

20,743,859

350

59,268

592,682

Solvents &
Chemicals,
Pearland, TXb

59,626

350

170

1,704

Univar USA Inc,
Redmond, WAa

20,743,859

350

59,268

592,682

a The 2015 annual production volumes in the 2016 CDR for these sites was either claimed as CBI or withheld. EPA estimate
the production volume by subtracting known site production volumes from the national production volume and averaging the
result over all the sites with CBI or withheld production volumes and converting from pounds to kilograms.
b Annual production volume fortius site is based on the 2015 production volume reported in the 2016 CDR and converting
from pounds to kilograms.

EPA estimated both a maximum daily release and an average daily release using the OCPSF EG limits
for PCE for maximum on any one day and maximum for any monthly average, respectively. Prevalence
of end-of-pipe biological treatment at PCE manufacturing sites is unknown; therefore, EPA used limits
for direct discharges with no end-of-pipe biological treatment and indirect dischargers as conservative.
EPA estimated annual releases from the average daily release and assuming 350 days/yr of operation.

Example max daily, average daily, and annual water release calculations for PCE at manufacturing sites
based on the estimated production volume for Axiall Corporation (45,732,418 lbs/yr or 20,743,859
kg/yr)30:

day

Max DR =

164^n- x 10 t^- x 20,743,859 —
L kg	yr

1,000,000,000^ x 350

kg	yr

30 This estimated production volume is equal to the estimated production volume assessed for Greeenchem and Univar USA
Inc.

Page 216 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

52^x10-^x 20,743,859^

L kq yr	kg

Average DR =	-	-r— = 0.03

1,000,000,000^ x 350	day

kg	yr

52^x10-^x 20,743,859^	hn

L kq	yr	kg

AR =	-	7775	— = 10.79 —

i,ooo,ooo,ooo^|	yr

Page 217 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Appendix E Tank Truck and Railcar Loading and Unloading Release
and Inhalation Exposure Model Approach and Parameters

This appendix presents the modeling approach and model equations used in the Tank Truck and Railcar
Loading and Unloading Release and Inhalation Exposure Model. The model was developed through
review of relevant literature and consideration of existing EPA exposure models. The model approach is
a generic inhalation exposure assessment at industrial facilities that is applicable for any volatile
chemical with the following conditions of use:

•	Manufacture (loading of chemicals into containers);

•	Processing as a reactant/intermediate (unloading of chemicals);

•	Processing into formulation, mixture, or reaction products;

•	Import (repackaging); and

•	Other similar conditions of use at industrial facilities (e.g., industrial processing aid).

As an example, PCE at a manufacturing facility is expected to be packaged and loaded into a container
before distributing to another industrial processing or use site (e.g., formulation sites, sites using PCE as
an intermediate, and sites using PCE as a processing aid). At the industrial processing or use site, PCE is
then unloaded from the container into a process vessel before being incorporated into a mixture, used as
a chemical intermediate, or otherwise processed/used. For the model, EPA assumes PCE is unloaded
into tank trucks and railcars and transported and distributed in bulk. EPA also assumes the chemical is
handled as a pure substance (100 percent concentration).

Because PCE is volatile (vapor pressure above 0.01 torr at room temperature), fugitive emissions may
occur when PCE is loaded into or unloaded from a tank truck or railcar. Sources of these emissions
include:

•	Displacement of saturated air containing PCE as the container/truck is filled with liquid;

•	Emissions of saturated air containing PCE that remains in the loading arm, transfer hose, and
related equipment; and

•	Emissions from equipment leaks from processing units such as pumps, seals and valves.

These emissions result in subsequent exposure to workers involved in the transfer activity. The
following subsections address these emission sources.

E.l Displacement of Saturated Air Inside Tank Trucks and Railcars

For screening-level assessments, EPA typically uses the EPA/OAQPS AP-42 Loading Model to
conservatively assess exposure during container unloading activities (	). The model

estimates release to air from the displacement of air containing chemical vapor as a container/vessel is
filled with liquid (	). The model assumes the unloading activity displaces an air volume

equal to the size of the container, and that displaced air is either 50 percent or 100 percent saturated with
chemical vapor (	•).

Process units at facilities that manufacture PCE as a primary product; use PCE as a reactant or
manufacture PCE as a product or co-product; or are located at a plant that is a major source of hazardous

Page 218 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

air pollutants (HAPs) as defined in section 112(a) of the Clean Air Act are required to install and operate
a vapor capture system and control device (or vapor balancing system) for loading/unloading operations
(	). Therefore, EPA expects the majority of industrial facilities to use a vapor balance

system to minimize fugitive emissions when loading and unloading tank trucks and railcars. As such,
vapor losses from displacement of air is likely mitigated by the use of such systems. Actual fugitive
emissions are likely limited to any saturated vapor that remain in the hose, loading arm, or related
equipment after being disconnected from the truck or railcar. This emission source is addressed in the
next subsection.

E.2 Emissions of Saturated Air that Remain in Transfer Hoses/Loading
Arm

After loading is complete, transfer hoses and/or loading arms are disconnected from tank trucks and
railcars. Saturated air containing the chemical of interest that remains in transfer equipment may be
released to air, presenting a source of fugitive emissions. The quantity of PCE released will depend on
concentration in the vapor and the volume of vapor in the loading arm/hose/piping.

TableApx E-l presents the dimensions for several types of loading systems according to an OPW
Engineered Systems catalog (OPW Engineered Systems. 2014). OPW Engineered Systems (2014)
specializes in the engineering, designing, and manufacturing of systems for loading and unloading a
wide range of materials including petroleum products, liquefied gases, asphalt, solvents, and hazardous
and corrosive chemicals. These systems include loading systems, swivel joints, instrumentation, quick
and dry-disconnect systems, and safety breakaways. Based on the design dimensions, the table presents
the calculated total volume of loading arm/system and assumes the volume of vapor containing PCE
equals the volume of the loading arm/system.

Based on comments from HSIA (2017). halogenated solvents, such as PCE, are expected to be delivered
in either tank trailers or tank cars. Therefore, EPA modeled the central tendency scenario as tank truck
loading/unloading. EPA modeled the high-end scenario as railcar loading/unloading since railcars are
larger and more likely to use longer transfer arms (and thus represent a higher exposure potential than
tank trucks). To estimate the high-end transfer arm volume, EPA calculated the 95th percentile of the
OPW Engineered Systems loading arms volumetric data resulting in a high-end value of 17.7 gallons.
For the central tendency tank truck scenario, EPA assumed a 2-inch diameter, 12-ft long transfer hose.
This hose has a volume of 2.0 gallons.

Once the volume is known, the emission rate, Et (g/s), can be calculated as follows:

EquationApx E-l

_fx MW x 3,786.4 xVhxXxVP
tdisconnect X T X R X 3,600 X 760

Default values for Equation Apx E-l can be found in Table Apx E-2.

Page 219 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Table Apx E-l. Example Dimensions and Volume of Loading Arm/Transfer System

OPW Kngincercd Systems Transfer
Arm

1

Ar
2-inch

en«th o
in/Con ii

3-inch

f l.oadini
eel ion (ii

4-inch

'

)"

6-inch

Vo

2_
iiu'h

In me.

3-
inch

Vh (gs

4-
iiu'h

I)1'

6-
inoh

Unsupported Boom-Type Bottom Loader

149.875

158.5

165.25

191.75

2.0

4.9

9.0

23.5

"A" Frame Loader M-32-F

153.75

159.75

164.5

NA

2.1

4.9

8.9

NA

"A" Frame Hose Loader AFH-32-F

180.75

192.75

197.5

NA

2.5

5.9

10.7

NA

CWH Series Counterweighted Hose
Loader

NA

NA

309

NA

NA

NA

16.8

NA

Spring Balanced Hose Loader SRH-32-F

204.75

216.75

221.5

NA

2.8

6.6

12.0

NA

Spring Balanced Hose Loader LRH-32-F

NA

270

277.625

NA

NA

8.3

15.1

NA

Top Loading Single Arm Fixed Reach

201.75

207.75

212.5

NA

2.7

6.4

11.6

NA

Top Loading Scissor Type Arm

197.875

206.5

213.25

NA

2.7

6.3

11.6

NA

Supported Boom Arm B-32-F

327.375

335

341.5

NA

4.5

10.3

18.6

NA

Unsupported Boom Arm GT-32-F

215.875

224.5

231.25

NA

2.9

6.9

12.6

NA

Slide Skv\c Ann A-32F

279

292 5

305 125

\ A

3 X

9 0

16 6

\A

Nose without Trims for Arm

















Hose (EPA judgment)

120

—

—

—

1.6

—

—

—

a Total length includes length of piping, connections, and fittings.

b Calculated based on dimension of the transfer hose/connection, Vh = 7ir2L (converted from cubic inch to gallons).
Source: (OPW Engineered Systems. 20.1.4')

Page 220 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

TableApx E-2. Default Values for Calculating Emissions Rate of Perchloroethylene from
Transfer/Loading Arm	

Parameter

Parameter Description

Default Value

I nit

Et

Emission rate of chemical from transfer/loading system

Calculated from
model equation

g/s

f

Saturation factora

1

dimensionless

MW

Molecular weight of the chemical

165.833

g/mol

Vh

Volume of transfer hose

See Table Apx
E-l

gallons

r

Fill ratea

2 (tank truck)
1 (railcar)

containers/hr

tdisconnect

Time to disconnect hose/couplers (escape of saturated
vapor from disconnected hose or transfer arm into air)

0.25

hr

X

Vapor pressure correction factor

1

dimensionless

VP

Vapor pressure of the pure chemical

18.5

ton-

T

Temperature

298

IC

R

Universal gas constant

82.05

atm-
cm3/gmol-K

a Saturation factor and fill rate values are based on established EPA release and inhalation exposure assessment
methodologies (U.S. EPA. 2015b').

E.3 Emissions from Leaks

During loading/unloading activities, emissions may also occur from equipment leaks from valves,
pumps, and seals. Per EPA's Chapter 5: Petroleum Industry of AP-42 (U.S. EPA. 2015a) and EPA's
Protocol for Equipment Leak Emission Estimates (	>), the following equation can be used

to estimate emission rate El, calculated as the sum of average emissions from each process unit:

EquationApx E-2

Z	1,000

(^x^racxJV)x —

Parameters for calculating equipment leaks using Equation Apx E-2 can be found in Table Apx E-3.

Page 221 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

TableApx E-3. Parameters for Calculating Emission Rate of Perchloroethylene from Equipment
Leaks

I'arameler

Pa ra melor Doseripi ion

Default Value

I nil

El

Emission rate of chemical from equipment leaks

Calculated from
model equation

g/s

Fa

Applicable average emission factor for the
equipment type

See Table Apx
E-4

kg/hr-source

WFtoc

Average weight fraction of chemical in the stream

1

dimensionless

N

Number of pieces of equipment of the applicable
equipment type in the stream

See Table Apx
E-4

Source

To estimate emission leaks using this modeling approach, EPA modeled a central tendency loading rack
scenario using tank truck loading/unloading and a high-end loading rack scenario using railcar
loading/unloading as discussed in Appendix E.2. EPA used engineering judgment to estimate the type
and number of equipment associated with the loading rack in the immediate vicinity of the loading
operation. EPA assumes at least one worker will be near the loading rack during the entire duration of
the loading operation.

Table Apx E-4 presents the average emission factor for each equipment type, based on the synthetic
organic chemical manufacturing industry (SOCMI) emission factors as provided by EPA's 1995
Protocol	and the likely number of pieces of each equipment used for each chemical

loading/unloading activity, based on EPA's judgment. Note these emission factors are for emission rates
of total organic compound emission and are assumed to be applicable to PCE. In addition, these factors
are most valid for estimating emissions from a population of equipment and are not intended to be used
to estimate emissions for an individual piece of equipment over a short period of time.

Page 222 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Table Apx E-4. Default Values for Fa and N

Kqiiipinenl Typo

Service

SOCMI Emission
Kaclor. K\
(k«/lir-sourcc):l

N il in her of
Kqiiipinenl. N
(central Icii(Iciicy)

Number of
Kqiiipinenl. \
(high-end)

Valves

Gas
Light liquid
Heavy
liquid

0.00597
0.00403
0.00023

3 (gas)
5 (light liquid)

3 (gas)
10 (light liquid)

Pump seals'3

Light liquid
Heavy
liquid

0.0199
0.00862

—

—

Compressor seals

Gas

0.228

—

—

Pressure relief valves

Gas

0.104

1

1

Connectors

All

0.00183

2

3

Open-ended lines

All

0.0017

—

—

Sampling connections

All

0.015

2

3

a SOCMI average emission factors for total organic compounds from EPA's 1995 Protocol (U.S. EPA. 1995"). "Light liquid"
is defined as "material in a liquid state in which the sum of the concentration of individual constituents with a vapor pressure
over 0.3 kilopascals (kPa) at 20 °C is greater than or equal to 20 weight percent". "Heavy liquid" is defined as "not in
gas/vapor service or light liquid service." Since PCE has a vapor pressure of 18.5 mmHg (2.47 kPa) at 25 °C, EPA modeled
PCE liquid as a light liquid.

b The light liquid pump seal factor can be used to estimate the leak rate from agitator seals.

Source: (U.S. EPA. 1995)

EPA assumed the following equipment are used in loading racks for the loading/unloading of tank
trucks and railcars. FigureApx E-l illustrates an example tank truck and unloading rack equipment.

•	Tank Truck Loading/Unloading:

o Liquid Service:

¦	Four valves (modeled as valves in light liquid service)

¦	One safety relief valve (modeled as valve in light liquid service)

¦	One bleed valve or sampling connection

¦	One hose connector
o Vapor Service:

¦	Three valves (modeled as valves in gas service)

¦	One pressure relief valve

¦	One bleed valve (modeled as a sampling connection)

¦	One hose connector

•	Railcar Loading/Unloading

o Liquid Service: EPA assumed, for the high-end scenario, two parallel liquid service lines,
each using the same equipment as assumed for tank trucks. Therefore, a total of:

¦	Eight valves (modeled as valves in light liquid service)

¦	Two safety relief valves (modeled as valve in light liquid service)

Page 223 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

¦	Two bleed valves or sampling connections

¦	Two transfer arm connectors

o Vapor Service: EPA assumed a single line in vapor service with the same equipment as
assumed for tank trucks.

¦	Three valves (modeled as valves in gas service)

¦	One pressure relief valve

¦	One bleed valve (modeled as a sampling connection)

¦	One transfer arm connector



Vaporservice line

Liquid service line

w

FigureApx E-l. Illustration of Transfer Lines Used During Tank Truck Unloading and
Associated Equipment Assumed by EPA

E.4 Exposure Estimates

The vapor generation rate, G, or the total emission rate over time, can be calculated by aggregating

emissions from all sources:

•	During the transfer period, emissions are only due to leaks, with emission rate G = EL.

•	After transfer, during the disconnection of the hose(s), emissions are due to both leaks and
escape of saturated vapor from the hose/transfer arm with emission rate G = ET + EL.

The vapor generation rate can then be used with the EPA/OPPTMass Balance Inhalation Model to
estimate worker exposure during loading/unloading activities (U.S. EPA 2015b). Th q EPA/OPPT Mass
Balance Inhalation Model estimates the exposure concentration using Equation Apx E-3 and the default
parameters found in TableApx E-5 (U.S. EPA 2015b). TableApx E-6 presents exposure estimates for
PCE using this approach. These estimates assume one unloading/loading event per day and PCE is

Page 224 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

loaded/unloaded at 100% concentration. The loading operation occurs in an outdoor area with minimal
structure, with wind speeds of 9 mph (central tendency) or 5 mph (high-end).

EquationApx E-3

r

_ ^V

m — v

vm

TableApx E-5. Parameters for Calculating Exposure Concentration Using the EPA/OPPT Mass
Balance Model

Parameler

Pa ra in el er Desc riplio n

Default Value

In it

Cm

Mass concentration of chemical in air

Calculated from model equation

mg/m3

Cv

Volumetric concentration of
chemical in air

Calculated as the lesser of:

170.000XTXG 1,000,000 XXXVP

	or	

MWxQxk 760

ppm

T

Temperature of air

298

K

G

Vapor generation rate

El during transfer period
Et+El after transfer/during
disconnection of hose/transfer arm

g/s

MW

Molecular weight of the chemical

165.833

g/mol

Q

Outdoor ventilation rate

237,600 (central tendency)
26,400 x f 60 x ) (high-end)

V 5280/

ft3/min

vz

Air speed

440

ft/min

k

Mixing factor

0.5

dimensionless

X

Vapor pressure correction factor

1

dimensionless

VP

Vapor pressure of the pure chemical

18.5

torr

Vm

Molar volume

24.45 @ 25°C, 1 atm

L/mol

EPA also calculated acute and 8-hr TWA exposures as shown in Equation Apx E-4 and Equation Apx
E-5, respectively. The acute TWA exposure is the weighted average exposure during the entire exposure
duration per shift, accounting for the number of loading/unloading events per shift. The 8-hr TWA
exposure is the weighted average exposure during an entire 8-hr shift, assuming zero exposures during
the remainder of the shift. EPA assumed one container is loaded/unloaded per shift: one tank truck per
shift for the central tendency scenario and one railcar per shift for the high-end scenario.

Equation Apx E-4

( Cm(leak only) ^ Q^event ~ tdisconnect) \^m(leak and hose) ^ tdisconnect) ) ^ ^cont

Acute TWA = 				-	

"¦shift

Page 225 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

EquationApx E-5

(j-'m(leak only) ^ (.h-event ~ tdisconnect) {f'm{leak and hose) ^ ^disconnect)) ^ ^cont

8

= Airborne concentration (mass-based) due to leaks during unloading while hose
connected (mg/m3)

= Airborne concentration (mass-based) due to leaks and displaced air during hose

disconnection (mg/m3)

= Exposure duration of each loading/unloading event (hr/event); calculated as the
inverse of the fill rate, r: 0.5 hr/event for tank trucks and 1 hr/event for rail cars
= Exposure duration during the shift (hr/shift); calculated as hevent X Ncont¦ 0.5

hr/shift for tank trucks and 1 hr/shift for railcars
= Time duration to disconnect hoses/couplers (during which saturated vapor

escapes from hose into air) (hr/event)

= Number of containers loaded/unloaded per shift (event/shift); assumed one tank
truck per shift for central tendency scenario and one railcar per shift for high-end
scenario

TableApx E-6. Calculated Emission Rates and Resulting Exposures from the Tank Truck and
Railcar Loading and Unloading Release and Inhalation Exposure Model for Perchloroethylene

Scenario

Ki

(Jl/s)

Ki

(g/s)

Ki. +
Ki

(S/s)

Cm
(leaks
only)
(nig/m')

Cm
(leaks and
hose vapor)
(iiiK/iii-*)

Acule
TWA
(mg/iir")11

8-hr
TWA
(mg/nr')

Acute
TWA
(ppm)11

8-hr
TWA
(ppm)

Central
Tendency

0.049

0.001

0.050

0.85

0.88

0.86

0.054

0.13

0.01

High-End

0.059

0.012

0.071

1.85

2.24

1.95

0.24

0.29

0.04

a Acute TWA exposure is a 0.5-hr TWA exposure for the central tendency scenario and a 1-hr TWA exposure for the high-
end scenario.

Where:

Cm(leak only)

Cm(leak and hose)

hevent

hshift

tdisconnect

Ncont

Page 226 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Appendix F Drum Loading and Unloading Release and Inhalation
Exposure Model Approach and Parameters

This appendix presents the approach for central tendency and high-end inhalation exposure estimation
for the loading of formulated-products containing PCE into 55-gallon drums. This approach applies a
stochastic modeling approach to the EPA/OAQPS AP-42 Loading Model, which estimates air releases
during container loading and unloading, and the EPA/OPPTMass Balance Model, which estimates
inhalation exposures resulting from air releases (	).

This approach is intended to assess air releases and associated inhalation exposures associated with
indoor container loading scenarios at industrial and commercial facilities. 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, such as ventilation rate (Q), mixing factor (k), and vapor
saturation factor (f), are expected to vary from one facility to another. This approach addresses
variability for these parameters using a Monte Carlo simulation.

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 ventilation rate (Q), mixing factor (k), vapor saturation factor (f), and exposed working
years per lifetime (WY). 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 Industrial Edition, Version 7.0.0 (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
statistics were calculated directly in @Risk. The following subsections detail the model design equations
and parameters used for Inhalation exposure estimates.

F.l Model Design Equations

The EPA/OPPT Mass Balance Model includes the following equations for estimating mass
concentration of the chemical vapor in air (mg/m3):

EquationApx F-l

Cv x MW

Where:

Cm	= Mass concentration of chemical vapor in air (mg/m3)

Cv	= Volumetric concentration of chemical vapor in air (ppm)

Page 227 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

MW	= Molecular Weight of chemical (g/mol)

Vm	= Molar volume (L/mol)

EquationApx F-2

170,000 xTxG
r = 	

v MWxQxk

Where:



T

= Temperature (K)

G

= Average vapor generation rate (g/s)

MW

= Molecular weight of chemical (g/mol)

Q

= Ventilation rate (ftVmin)

K

= Mixing factor (Dimensionless)

The average vapor generation rate needed for the EPA/OPPTMass Balance Model is calculated from
the following EPA/OAQPS AP-42 Loading Model:

Equation Apx F-3

VP

f x MW x (3,785.4 xVc)xrxXx^
G ~	3,600 XT x R

Where:



G

= Average vapor generation rate (g/s)

f

= Saturation factor (Dimensionless)

MW

= Molecular weight of chemical (g/mol)

Vc

= Container volume (gallon)

r

= Container loading/unloading rate (number of containers/hr)

X

= Vapor pressure correction factor (Dimensionless), assumed equal to weight



fraction of component

VP

= Vapor pressure (at temperature, T) (mmHg)

T

= Temperature (K)

R

= Universal gas constant (atm-cm3/mol-K)

Mass concentration of the chemical vapor in air (Cm) calculated from Equation Apx F-l and the time
spent loading containers each day (trading) are then used in the following equation to estimate the 8-hr
TWA concentration used for estimating acute exposure concentrations (AC), average daily
concentrations (ADC) used for chronic non-cancer risk calculations (ADC) and lifetime average daily
concentration (LADC) used for chronic cancer risk calculations:

Equation Apx F-4

8-hrTWA = CmXtloadins
8 hr

To determine the amount of time spent each day loading containers, the model uses the following
equations:

Page 228 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

EquationApx F-5

t

N,

cd

loading

Where:

tloading

Ned
r

Loading duration (hrs)

Number of containers loaded per site per day (containers/site-day)
Container fill rate (drum/hr)

Equation Apx F-6

Ncd =

N,

cy_

OD

Where:

Ncy

OD

Number of containers per site per year (containers/site-yr)
Operating days (days/yr)

Equation Apx F-7

PV

Nrv =	

Vc_lb x wtfrac x Nsites

Where:
PV
Vc-lb
wtfrac

Nsites

Production volume for the condition of use (lb/yr)
Volume of container in pounds (lb/container)
weight fraction of PCE in the formulation (unitless)
Number of sites for the condition of use

F.2_ Model Parameters

Table Apx F-l summarizes the model parameters and their values for the EPA/OAQPS AP-42 Loading
Model and the EPA/OPPTMass Balance Model. Each parameter is discussed in detail the following the
subsections.

Page 229 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Table Apx F-l. Summary of Parameter Values and Distributions Used in the Inhalation Exposure Model

Input Parameter

Symbol

Unit

Constant

Model
Parameter
Values

Variable Model Parameter Values

Lower
Bound

Upper
Bound

Mode

Distribution

	line	

Rationale/Basis

EPA/OAQPS AP-42 Loading Model

Saturation factor

f

dimensionless

—

0.5

1.45

0.5

Triangular

See Section F.2.1

Molecular weight of the
chemical

MW

g/mol

165.833

—

—

—

—

Physical Property

Volume of container

Vc

gallons

55

—

—

—

—

See Section F.2.2

Fill rate

r

containers/hr

20

—

—

—

—

See Section F.2.3

Vapor pressure correction factor

X

dimensionless

—

Equal

to Xi

Equal

to Xi

Equal

to Xi

Uniform

See Section F.2.4

Vapor pressure of the pure
chemical

VP

ton-

18.5

—

—

—

—

Physical Property

Temperature

T

IC

298

—

—

—

—

Process Parameter

Universal gas constant

R

atm-cm3/gmol-
K

82.05

—

—

—

—

Physical Constant

Mol fraction of chemical

Xi

dimensionless

—

Equal

to
wtfrac

Equal

to
wtfrac

Equal

to
wtfrac

Uniform

See Section F.2.4

EPA/OPPT Mass Balance Inhalation Model

Ventilation rate (indoor)

Q

ft3/min

—

500

10,000

3,000

Triangular

See Section F.2.5

Mixing factor

k

dimensionless

—

0.1

1

0.5

Triangular

See Section F.2.6

Molar volume

Vm

L/mol

24.46

—

—

—

—

Physical Constant


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Input Parameter

Symbol

Unit

Constant

Model
Parameter
Values

Variable Model Parameter Values

Rationale/Basis

Lower
Bound

Upper
Bound

Mode

Distribution
Type

Other Parameters

Weight fraction of chemical in
dry cleaning and degreasing
formulation

wtfrac

dimensionless

1

—

—

—

—

See Section F.2.7

Weight fraction of chemical in
miscellaneous formulations

wtfrac

dimensionless

—

0.3

0.8

—

Uniform

See Section F.2.7

Lb per drum

Vc lb

lb/container

744.95

—

—

—

—

See Section F.2.2

Production volume for dry
cleaning

PV

lb/yr

32,424,074

—

—

—

—

See Section F.2.8

Production volume for
degreasing

PV

lb/yr

22,696,852

—

—

—

—

See Section F.2.8

Production volume for
miscellaneous formulations

PV

lb/yr

9,727,222

—

—

—

—

See Section F.2.8

Number of sites for dry cleaning
solvent formulation

Nsites

# of sites

5

—

—

—

—

See Section F.2.9

Number of sites for degreasing
solvent formulation

Nsites

# of sites

19

—

—

—

—

See Section F.2.9

Number of sites for
miscellaneous product
formulation

Nsites

# of sites

15

—

—

—

—

See Section F.2.9

Operating days

OD

day/yr

300

—

—

—

—

See Section F.2.10

Exposure Frequency

EF

day/yr

250

—

—

—

—

See Section F.2.11

Exposure Duration

ED

hr/day

8

—

—

—

—

See Section F.2.12

Page 231 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

F.2.1 Saturation Factor

The Chemical Engineering Branch Manual for the Preparation of Engineering Assessments, Volume 1
[CEB Manual] (	) indicates that during splash filling the saturation concentration was

reached or exceeded by misting with a maximum saturation factor of 1.45. The CEB manual indicates
that generation rate 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 requires least assumptions and 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 (	) for the EPA/OAQPS

AP-42 Loading Model for drums.

F.2.2 Volume of Container

EPA assumed formulated products were loaded into 55-gallon drums. It is possible that some formulated
products, such as coatings and adhesives, may be loaded into smaller containers (e.g., pails) for smaller
commercial and consumer applications; however, EPA does not have information to estimate the
volume packaged into drums versus smaller containers. Therefore, EPA assessed the entire throughput
as packaged into drums to give the most protective worker exposure estimates.

This value was then converted to lbs. using the density of PCE. This assumes that the density of the
formulated product will be similar to that of PCE. This may result in error when estimating number of
containers where the actual density of the formulation differs significantly from that of PCE.

F.2.3 Fill Rate

The ChemSTEER User Guide (	) provides a typical fill rate of 20 containers per hour

for 55-gallon drums.

F.2.4 Vapor Pressure Correction Factor and Mole Fraction

The ChemSTEER User Guide (	) assumes Raoulfs Law such that the vapor pressure

correction factor may be set equal to the mole fraction of PCE in the formulation. Due to the wide
variety of formulations PCE may be used in, the mole fraction of PCE in each product could not be
determined. Therefore, EPA assumed that the mole fraction is equal to the weight fraction of PCE in the
formulation (see Section F.2.7). This assumption is not expected to result in significant error in
formulations where the molecular weight (MW) of PCE is very similar to that of the other components.
However, if the MW of PCE differs significantly from the MW of the other components it may result in
error when estimating the vapor generation rate and corresponding exposure. If the MW of PCE is
significantly lower than the MW of other components then the mol fraction, correction factor and
resulting vapor generation and exposure will be overestimated. If the MW of PCE is significantly higher
than the other components the mol fraction, correction factor, and resulting vapor generation and
exposure will be underestimated.

F.2.5 Ventilation Rate

The CEB Manual (	) indicates general ventilation rates in industry range from 500 to

10,000 ftVmin, with a typical value of 3,000 ft3/min. The underlying distribution of this parameter is not
known; therefore, EPA assigned triangular distributions, since triangular distribution requires least
assumptions and is completely defined by range and mode of a parameter. EPA assumed the mode is
equal to the typical value provided by the CEB Manual (	i).


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

F.2.6 Mixing Factor

The CEB Manual (	) 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 triangular distributions, since triangular distribution requires least
assumptions and is completely defined by 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 Chem STEER User Guide (x ^ \ \\ b) for the EPA/OPPT Mass Balance Inhalation Model.

F.2.7 Weight Fraction of Chemical

The weight fraction of PCE in the product varies depending on specific product being formulated. For
formulation, EPA considered three types of formulations: 1) degreasing solvents, 2) dry cleaning
solvents, and 3) miscellaneous products. Miscellaneous products include coatings, adhesives,
metalworking fluids, and other niche use PCE-based products. These three categories were selected
based on the availability (or lack thereof) of market data. For example, market data from HSIA (2008)
estimated 7% of the production volume of PCE is used in degreasing, 10% is used in dry cleaning, and
3% is for "miscellaneous" uses. More specific market data for the third "miscellaneous" group were not
available; therefore, EPA could not divide the PCE production volume amongst the product types to
develop exposure estimates for each product type.

For degreasing and dry cleaning solvents EPA assumed the PCE weight fraction to be 100%. Typically,
the only materials expected to be added to degreasing and dry cleaning solvents are stabilizers used to
prevent decomposition during storage and use (European Chlorinated Solvents Association. 2011). PCE
generally requires less stabilizers than other chlorinated solvents with weight fractions of stabilizers
expected to be less than 0.5% in degreasing solvents, and less than 0.05% in dry cleaning solvents.
(European Chlorinated Solvents Association. 2011). Therefore, the assumption of 100% PCE in the
model is not expected to significantly impact exposure results.

For miscellaneous products, the concentration of PCE can vary greatly depending on the product being
formulated. For modeling purposes, EPA assessed used a uniform distribution of 30 to 80% PCE in the
formulated product based on the expected concentrations of solvents in organic solvent-borne coatings
estimated by the OECD ESD (QE 09b). This range was used as it is expected to encompass the
range of compositions for the majority of PCE-based products in this category (e.g., per the OECD ESD
(OECD. 2009a) typical organic solvent concentrations in adhesives is estimated to be between 60 to
75%) which falls within the range used in the model). While it is possible that some of the products
contain PCE concentrations outside this range, the error from this is expected to be small as, based on
the reported NAICS codes, 10 of the 15 formulation sites assessed in this category are either coatings
(including maskants) or adhesive formulation sites (see Section F.2.9).

F.2.8 Production Volume

HSIA (2008) estimated 7% of the production volume of PCE is used in degreasing, 10% is used in dry
cleaning, and 3% is for "miscellaneous" uses. More specific market data for the third "miscellaneous"
group were not available; therefore, EPA could not divide the PCE production volume amongst the
product types to develop exposure estimates for each product type. Based on the 2016 CDR (

Page 233 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

2016d), the national production volume of PCE in 2015 was 324,240,744 lbs. resulting in the following
PCE production volumes used in each category:

•	Degreasing - 22,696,852 lbs.;

•	Dry Cleaning - 32,424,074 lbs.; and

•	Miscellaneous - 9,727,222 lbs.

F.2.9 Number of Sites

Formulation sites were determined based on SIC codes reported by sites in the 2016 DMR (

2016b) and activities and NAICS codes reported by sites in the 2016 TRI (	i). This

resulted in a total of 39 formulation sites. Each site was than categorized as a degreasing solvent
formulation site, dry cleaning solvent formulation site, or a miscellaneous product formulation site for
use in modeling. Sites were categorizes based on reported NAICS codes (or SIC codes mapped to
NAICS codes) as follows:

•	Degreasing solvent formulation NAICS codes:

o 324110 - Petroleum Refineries31; and

o 325998 - All Other Miscellaneous Chemical Product and Preparation Manufacturing32.

•	Dry cleaning solvent formulation NAICS codes:

o 325611 - Soap and Other Detergent Manufacturing;
o 325612 - Polish and Other Sanitation Good Manufacturing; and
o 325613 - Surface Active Agent Manufacturing.

•	Miscellaneous formulation NAICS codes:

o All NAICS codes reported not listed above.

The categorization resulted in 19 formulation sites for degreasing solvents, 5 for dry cleaning solvents,
and 15 for miscellaneous products.

F.2.10 Operating Days

EPA assumed 300 days/yr of operation as given in the SpERC developed by the European Solvent
Industry Group for the formulation and (re)packing of substances and mixtures (European Solvents
Indus )up. 2019a). EPA uses 300 days per year rather than 350 (7 days/wk, 50 wks/yr) because it
is likely that formulation sites make multiple products not all of which will contain PCE. Drum loading
of PCE-based products is only expected to occur on days were PCE-containing products are produced.

F.2.11 Exposure Frequency

When calculating ADC and LADC, EPA uses an exposure frequency of 250 days/yr. This assumes
workers work five days per week 50 weeks per year, with two weeks off.

F.2.12 Exposure Duration

EPA assumes workers work a total of eight hours per day.

31	EPA does not typically expect petroleum refineries to formulate degreasing solvents; however, the one site reporting this
NAICS code to the 2016 TRI also reported as an importer to the 2016 CDR and reported its entire import volume as used on-
site and reported formulation of solvents for cleaning and degreasing.

32	This NAICS codes may also include sites manufacturing aerosol products; therefore, the total number of sites for
formulating degreasing solvents may be overestimated.

Page 234 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Page 235 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Appendix G Vapor Degreasing and Cold Cleaning Near-Field/Far-
Field Inhalation Exposure Models Approach and Parameters

This appendix presents the modeling approach and model equations used in the following models:

•	Conveyorized Degreasing Near-Field/Far-Field Inhalation Exposure Model;

•	Web Degreasing Near-Field/Far-Field Inhalation Exposure Model; and

•	Cold Cleaning Near-Field/Far-Field Inhalation Exposure Model.

The models were developed through review of the literature and consideration of existing EPA/OPPT
exposure models. These models use a near-field/far-field approach (	009). where a vapor

generation source located inside the near-field diffuses into the surrounding environment. Workers are
assumed to be exposed to PCE vapor concentrations in the near-field, while occupational non-users are
exposed at concentrations in the far-field.

The model uses the following parameters to estimate exposure concentrations in the near-field and far-
field:

•	Far-field size;

•	Near-field size;

•	Air exchange rate;

•	Indoor air speed;

•	Exposure duration;

•	Vapor generation rate; and

•	Operating hours per day.

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. 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 Industrial Edition,
Version 7.0.0. 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 the model at 100,000 iterations to capture the range of
possible input values (i.e., including values with low probability of occurrence).

Model results from the Monte Carlo simulation are presented as 95th and 50th percentile values. The
statistics were calculated directly in @Risk. The 95th percentile value was selected to represent high-end
exposure level, whereas the 50th percentile value was selected to represent typical exposure level. The
following subsections detail the model design equations and parameters for vapor degreasing and cold
cleaning models.

G.l Model Design Equations

FigureApx G-l through FigureApx G-3 illustrate the near-field/far-field modeling approach as it was
applied by EPA to each vapor degreasing and cold cleaning model. As the figures show, volatile PCE

Page 236 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

vapors evaporate into the near-field, resulting in worker exposures at a PCE concentration Cnf. The
concentration is directly proportional to the evaporation rate of PCE, G, into the near-field, whose
volume is denoted by Vnf. The ventilation rate for the near-field zone (Qnf) determines how quickly
PCE dissipates into the far-field, resulting in occupational non-user exposures to PCE at a concentration
Cff. Vff denotes the volume of the far-field space into which the PCE dissipates out of the near-field.
The ventilation rate for the surroundings, denoted by Qff, determines how quickly PCE dissipates out of
the surrounding space and into the outside air.

	Far-Field	

Figure Apx G-l. The Near-Field/Far-Field Model as Applied to the Cold Cleaning Near-

Field/Far-Field Inhalation Exposure Model

Page 237 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Far-Field

FigureApx G-2. The Near-Field/Far-Field Model as Applied to the Conveyorized Decreasing

Near-Field/Far-Field Inhalation Exposure Model

Far-Field

Figure Apx G-3. The Near-Field/Far-Field Model as Applied to the Web Degreasing Near-

Field/Far-Field Inhalation Exposure Model

The model design equations are presented below in Equation Apx G-l through Equation_Apx G-16.
Note the design equations are the same for each of the models discussed in this appendix.

Near-Field Mass Balance

Page 238 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

EquationApx G-l

Far-Field Mass Balance
Equation Apx G-2

Where:

Vnf Jt — CffQnf ~ CNFQNF + G

dt

V,

dC,

FF

FF'

dt

CnfQnf CffQnf CffQff

Vnf =

near-field volume;

Vff =

far-field volume;

Qnf =

near-field ventilation rate;

Qff =

far-field ventilation rate;

Cnf =

average near-field concentration;

Cff =

average far-field concentration;

G

average vapor generation rate; and

t

elapsed time.

Both of the previous equations can be solved for the time-varying concentrations in the near-field and
far-field as follows (	s09):

Equation Apx G-3
Equation Apx G-4

Where:

Equation Apx G-5

CNF = G(k1 + k2eXlt - k3eX2t)

C,

FF

= g(tt~

\Qff

+ k4eXlt — kseX2t

kt =

Equation Apx G-6

Equation Apx G-7

Equation Apx G-8

k7 =

ko =

(q»f + «J

QnfQff + ^-2^nf(.Qnf + Qff)
QnfQff^nf^i ~ ^2)

QnfQff + A.1Vnf(.Qnf + Qff)
QnfQff^nf(^i ~ ^2)

, MiVnf + Qnf\ ,

** = ( Q„ )"2

Page 239 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

EquationApx G-9

_ M2Vnf + Qnf\ j
5 V	/ 3

EquationA

= 0.5

EquationA

Az = 0.5

px G-10

(Qnf^ff + Vnf(.Qnf + Qff)
VNFVFF

Qnf^ff + VnfCQnf + Qff)\
VnfVFf	J

^ (QnfQff\
V V^jFVFF /

px G-ll

Qnf^ff + Vnf(Qnf + Qff)N

V	^/VF^FF	,



/Qnf^ff + Vnf(Qnf + Qff)\ _ . (QnfQff\
^/VF^FF	/ V ^/VF^FF '

EPA calculated the hourly TWA concentrations in the near-field and far-field using Equation Apx G-12
and Equation Apx G-13, respectively. Note that the numerator and denominator of Equation Apx G-12
and Equation Apx G-13 use two different sets of time parameters. The numerator is based on operating
times for the scenario (e.g., 13 hours for conveyorized degreasers, 24 hours for web degreasers, and 1 to
24 hours for cold cleaning, see Section G.2) while the denominator is fixed to an average time span,
t avg, of eight hours (since EPA is interested in calculating 8-hr TWA exposures). Mathematically, the
numerator and denominator must reflect the same amount of time. This is indeed the case since the
numerator assumes exposures are zero for any hours not within the operating time. Therefore,
mathematically speaking, both the numerator and the denominator reflect eight hours regardless of the
values selected for ti and h.

Equation Apx G-12

C,

JCNFdt JG(/cx + k2eXlt — k3eX2t)dt

NFTWA — t

J0 dt

Vavg

(fci t

k7eXlt2 k3eX2t2
A1	X2

G[k-[t2 +	 —



bavg

Equation Apx G-13

C,

St' Crrdt ^ c U-+ kte^'- kse^')

dt

FF TWA — t

' f*avg dt

Vavg

(-

\Q,

k4eXlt2 _ kseX2t2\ _ I tx k4eXltl _ kseX2t^
QfF ^1	^2 / WFF ^1	^2 >

G TT~ +

Lavg

Page 240 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

To calculate the mass transfer to and from the near-field, the free surface area, FSA, is defined to be the
surface area through which mass transfer can occur. Note that the FSA is not equal to the surface area of
the entire near-field. EPA defined the near-field zone to be a rectangular box resting on the floor;
therefore, no mass transfer can occur through the near-field box's floor. FSA is calculated in
EquationApx G-14, below:

EquationApx G-14

FSA = 2{LnfHnf) + 2 (WnfHnf) + (LnfWnf)

Where: Lnf, Wnf, and Hnf are the length, width, and height of the near-field, respectively. The near-
field ventilation rate, Qnf, is calculated in Equation Apx G-15 from the near-field indoor wind speed,
vnf, and FSA, assuming half of FSA is available for mass transfer into the near-field and half of FSA is
available for mass transfer out of the near-field:

Equation Apx G-15

1

Qnf — — vnfFSA

The far-field volume, Vff, and the air exchange rate, AER, is used to calculate the far-field ventilation
rate, Qff, as given by Equation Apx G-16:

Equation Apx G-16

Qff = R

Using the model inputs described in Section G.2, EPA estimated PCE inhalation exposures for workers
in the near-field and for occupational non-users in the far-field. EPA then conducted the Monte Carlo
simulations using @Risk (Version 7.0.0). The simulations applied 100,000 iterations and the Latin
Hypercube sampling method for each model.

G.2 Model Parameters

TableApx G-l through TableApx G-3 summarize the model parameters and their values for each of
the models discussed in this Appendix. Each parameter is discussed in detail in the following
subsections.

Page 241 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

TableApx G-l. Summary of Parameter Values and Distributions Used in the Conveyorized Degreasing Near-Field/Far-Field

Input Parameter

Symbol

Unit

Constant Model
Parameter Values

Variable Model Parameter Values

Comments

Value

Basis

Lower
Bound

Upper
Bound

Mode

Distribution
Type

Far-field volume

Vff

ft3

—

—

10,594

70,629

17,657

Triangular

See Section G.2.1

Air exchange rate

AER

hr"1

—

—

2

20

3.5

Triangular

See Section G.2.2

Near-field indoor
wind speed

VNF

ft/hr

—

—

—

23,882

—

Lognormal

See Section G.2.3

cm/s

—

—

—

202.2

—

Lognormal

Near-field length

Lnf

ft

10

—

—

—

—

Constant Value

See Section G.2.4

Near-field width

Wnf

ft

10

—

—

—

—

Constant Value

Near-field height

Hnf

ft

6

—

—

—

—

Constant Value

Starting time

tl

hr

0

—

—

—

—

Constant Value

Constant

Exposure Duration

t2

hr

8

—

—

—

—

Constant Value

See Section G.2.5

Averaging Time

tavg

hr

8

—

—

—

—

Constant Value

See Section G.2.6

Vapor generation
rate

G

mg/hr

—

—

1.85E+06

1.85E+06

—

Discrete

See Section G.2.7

lb/hr

—

—

4.083

4.083

—

Discrete

Operating hours per
day

OH

hr/day

13

—

—

—

—

Discrete

See Section G.2.8


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

TableApx G-2. Summary of Parameter Values and Distributions Used in the Web Degreasing Near-Field/Far-Field Inhalation
Exposure Model 					

Input Parameter

Symbol

Unit

Constant Model
Parameter Values

Variable Model Parameter Values

Comments

Value

Basis

Lower
Bound

Upper
Bound

Mode

Distribution
Type

Far-field volume

Vff

ft3

—

—

10,594

70,629

17,657

Triangular

See Section G.2.1

Air exchange rate

AER

hr"1

—

—

2

20

3.5

Triangular

See Section G.2.2

Near-field indoor
wind speed

VNF

ft/hr

—

—

—

23,882

—

Lognormal

See Section G.2.3

cm/s

—

—

—

202.2

—

Lognormal

Near-field length

Lnf

ft

10

—

—

—

—

Constant Value

See Section G.2.4

Near-field width

Wnf

ft

10

—

—

—

—

Constant Value

Near-field height

Hnf

ft

6

—

—

—

—

Constant Value

Starting time

tl

hr

0

—

—

—

—

Constant Value

Constant.

Exposure Duration

t2

hr

8

—

—

—

—

Constant Value

See Section G.2.5

Averaging Time

tavg

hr

8

—

—

—

—

Constant Value

See Section G.2.6

Vapor generation
rate

G

mg/hr

—

—

9.09E+03

2.24E+04

—

Discrete

See Section G.2.7

lb/hr

—

—

0.020

0.049

—

Discrete

Operating hours per
day

OH

hr/day

24

—

—

—

—

Discrete

See Section G.2.8

Page 243 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

TableApx G-3. Summary of Parameter Values and Distributions Used in the Cold Cleaning Near-Field/Far-Field Inhalation
Exposure Model 					

Input Parameter

Symbol

Unit

Constant Model
Parameter Values

Variable Model Parameter Values

Comments

Value

Basis

Lower
Bound

Upper
Bound

Mode

Distribution
Type

Far-field volume

Vff

ft3

—

—

10,594

70,629

17,657

Triangular

See Section G.2.1

Air exchange rate

AER

hr"1

—

—

2

20

3.5

Triangular

See Section G.2.2

Near-field indoor
wind speed

VNF

ft/hr

—

—

—

23,882

—

Lognormal

See Section G.2.3

cm/s

—

—

—

202.2

—

Lognormal

Near-field length

Lnf

ft

10

—

—

—

—

Constant
Value

See Section G.2.4

Near-field width

Wnf

ft

10

—

—

—

—

Constant
Value

Near-field height

Hnf

ft

6

—

—

—

—

Constant
Value

Starting time

tl

hr

0

—

—

—

—

Constant
Value

Constant.

Exposure Duration

t2

hr

—

—

1

8

—

Discrete

See Section G.2.5

Averaging Time

tavg

hr

8

—

—

—

—

Constant
Value

See Section G.2.6

Vapor generation
rate

G

mg/hr

—

—

5.13E-02

5.63E+04

—

Discrete

See Section G.2.7

lb/hr

—

—

1.13E-07

0.12

—

Discrete

Operating hours per
day

OH

hr/day

—

—

1

24

—

Discrete

See Section G.2.8

Page 244 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

G.2.1 Far-Field Volume

EPA used the same far-field volume distribution for each of the models discussed. The far-field volume
is based on information obtained from von Grote (2003)) that indicated volumes at German metal
degreasing facilities can vary from 300 to several thousand cubic meters. They noted that smaller
volumes are more typical and assumed 400 and 600 m3 (14,126 and 21,189 ft3) in their exposure models
(Von Grote. 2003). These are the highest and lowest values EPA identified in the literature; therefore,
EPA assumes a triangular distribution bound from 300 m3 (10,594 ft3) to 2,000 m3 (70,629 ft3) with a
mode of 500 m3 (the midpoint of 400 and 600 m3) (17,657 ft3).

G.2.2 Air Exchange Rate

EPA used the same air exchange rate distribution for each of the models discussed. The air exchange
rate is based on data from Hellweg (2009) and information received from a peer reviewer during the
development of the 2014 TSCA Work Plan Chemical Risk Assessment Trichloroethylene: Degreasing,
Spot Cleaning and Arts & Crafts Uses (SCG. 2013). Hellweg (2009) reported that average air exchange
rates for occupational settings using mechanical ventilation systems vary from 3 to 20 hr"1. The risk
assessment peer reviewer comments indicated that values around 2 to 5 hr"1 are likely (SCG 2013). in
agreement with the low end reported by Hellweg (2009). Therefore, EPA used a triangular distribution
with the mode equal to 3.5 hr"1, the midpoint of the range provided by the risk assessment peer reviewer
(3.5 is the midpoint of the range 2 to 5 hr"1), with a minimum of 2 hr"1, per the risk assessment peer
reviewer (SC	) and a maximum of 20 hr"1 per Hellweg (2009).

G.2.3 Near-Field Indoor Air Speed

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 separate
distributions for these industrial and commercial settings and used the industrial distribution for facilities
performing vapor degreasing and/or cold cleaning.

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.
However, a mean air speed (averaged over a work area) is the required input for the model.


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

G.2.4 Near-Field Volume

EPA assumed a near-field of constant dimensions of 10 ft x 10 ft x 6 ft resulting in a total volume of 600

ft3.

G.2.5 Exposure Duration

EPA assumed the maximum exposure duration for each model is equal to the entire work-shift (eight
hours). Therefore, if the degreaser/cold cleaning machine operating time was greater than eight hours,
then exposure duration was set equal to eight hours. If the operating time was less than eight hours, then
exposure duration was set equal to the degreaser/cold cleaning machine operating time (see Section
G.2.8 for discussion of operating hours).

G.2.6 Averaging Time

EPA was interested in estimating 8-hr TWAs for use in risk calculations; therefore, a constant averaging
time of eight hours was used for each of the models.

G.2.7 Vapor Generation Rate

For the vapor generation rate from each machine type (conveyorized, web, and cold), EPA used a
discrete distribution based on the annual unit emission rates reported in the 2014 NEI (U.S. EPA
2016a). Annual unit emission rates were converted to hourly unit emission rates by dividing the annual
reported emissions by the reported annual operating hours (see Section G.2.8). Reported annual
emissions in NEI without accompanying reported annual operating hours were not included in the
analysis. Emission rates reported as zero were also excluded as it is unclear if this is before or after
vapor controls used by the site and if the vapor controls used would control emissions into the work area
(thus reducing exposure) or only control emissions to the environment (which would not affect worker
exposures). TableApx G-4 summarizes the data available in the 2014 NEI for the relevant machine
types.

Table Apx G-4. Summary of Perchloroethylene Vapor Degreasing and Cold Cleaning Data from

the 2014 NEI





Units with Zero
Emissions

Units without

Units Used
in Analysis

Unit Type

Total Units

Accompanying
Operating Hours

Conveyorized Degreasers

1

0

0

1

Web Degreasers

10

0

0

10

Cold Cleaning Machines

34

6

2

26

Source: (U.S. EPA. 2016a)

Table Apx G-5 through Table Apx G-7 summarize the distribution of hourly unit emissions for each
machine type calculated from the annual emission in the 2014 NEI. It should be noted that the emission
rate for convey orized degreasing is based on a single unit emission rate and it is unclear how
representative this emission rate is of a "typical" conveyorized degreaser.

Page 246 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

TableApx G-5. Distribution of Perchloroethylene Conveyorized Degreasing Unit Emissions

Count
of Units

Unit Emissions
(lb/unit-hr)

Fractional
Probability

1

4.08

1.0000

TableApx G-6. Distribution of Perchloroethylene Web Degreasing Unit Emissions

Count
of Units

Unit Emissions
(lb/unit-hr)

Fractional
Probability

1

0.0495

0.1000

1

0.0495

0.1000

1

0.0495

0.1000

1

0.0495

0.1000

1

0.0330

0.1000

1

0.0330

0.1000

4

0.0200

0.4000

Page 247 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

TableApx G-7. Distribution of Perchloroethylene Cold Cleaning Unit

Count
of Units

Unit Emissions
(lb/unit-hr)

Fractional
Probability

1

0.124

0.0385

1

0.085

0.0385

1

0.022

0.0385

1

1.17E-02

0.0385

1

4.02E-03

0.0385

1

8.03E-04

0.0385

1

4.01E-04

0.0385

1

2.67E-04

0.0385

1

2.66E-04

0.0385

1

2.30E-04

0.0385

1

2.01E-04

0.0385

1

2.01E-04

0.0385

1

1.34E-04

0.0385

1

9.13E-05

0.0385

1

9.13E-05

0.0385

1

9.13E-05

0.0385

1

9.13E-05

0.0385

1

9.13E-05

0.0385

1

2.77E-05

0.0385

1

2.28E-05

0.0385

1

2.17E-05

0.0385

1

1.83E-05

0.0385

1

1.49E-06

0.0385

1

2.98E-07

0.0385

1

2.98E-07

0.0385

1

1.13E-07

0.0385

Emissions

G.2.8 Operating Hours

For the operating hours of each machine type (conveyorized, web, and cold), EPA used a discrete
distribution based on the daily operating hours reported in the 2014 NEI. It should be noted that not all
units had an accompanying reported daily operating hours; therefore, the distribution for the operating
hours per day is based on a subset of the reported units. Table Apx G-8 through Table Apx G-10
summarize the distribution of operating hours per day for each machine type. It should be noted that the
operating hours for conveyorized degreasers is based on a single unit operating time and it is unclear
how representative this is of a "typical" conveyorized degreaser.

Page 248 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

TableApx G-8. Distribution of Perchloroethylene Conveyorized Degreasing Operating Hours

Count of
Occurrences

Operating
Hours
(hr/day)

Fractional
Probability

1

13

1.0000

TableApx G-9. Distribution of Perchloroethylene Web Degreasing Operating Hours

Count of
Occurrences

Operating
Hours
(hr/day)

Fractional
Probability

7

24

1.0000

Table Apx G-10. Distribu

ion of Perchloroethylene Cold Cleaning Operating Hours



Operating



Count of

Hours

Fractional

Occurrences

(hr/day)

Probability

19

24

0.7037

7

8

0.2593

1

1

0.0370

Page 249 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Appendix H Brake Servicing Near-Field/Far-Field Inhalation
Exposure Model Approach and Parameters

This appendix presents the modeling approach and model equations used in the Brake Servicing Near-
Field/Far-Field Inhalation Exposure Model. The model was developed through review of the literature
and consideration of existing EPA/OPPT exposure models. This model uses a near-field/far-field
approach (	009). where an aerosol application located inside the near-field generates a mist of

droplets, and indoor air movements lead to the convection of the droplets between the near-field and far-
field. Workers are assumed to be exposed to PCE droplet concentrations in the near-field, while
occupational non-users are exposed at concentrations in the far-field.

The model uses the following parameters to estimate exposure concentrations in the near-field and far-
field:

•	Far-field size;

•	Near-field size;

•	Air exchange rate;

•	Indoor air speed;

•	Concentration of PCE in the aerosol formulation;

•	Amount of degreaser used per brake j ob;

•	Number of degreaser applications per brake job;

•	Time duration of brake j ob;

•	Operating hours per week; and

•	Number of j obs per work shift.

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. 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 Industrial Edition,
Version 7.0.0. 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 the model at 100,000 iterations to capture the range of
possible input values (i.e., including values with low probability of occurrence).

Model results from the Monte Carlo simulation are presented as 95th and 50th percentile values. The
statistics were calculated directly in @Risk. The 95th percentile value was selected to represent high-end
exposure level, whereas the 50th percentile value was selected to represent central tendency exposure
level. The following subsections detail the model design equations and parameters for the brake
servicing model.

H.l Model Design Equations

In brake servicing, the vehicle is raised on an automobile lift to a comfortable working height to allow
the worker (mechanic) to remove the wheel and access the brake system. Brake servicing can include
inspections, adjustments, brake pad replacements, and rotor resurfacing. These service types often

Page 250 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

involve disassembly, replacement or repair, and reassembly of the brake system. Automotive brake
cleaners are used to remove oil, grease, brake fluid, brake pad dust, or dirt. Mechanics may occasionally
use brake cleaners, engine degreasers, carburetor cleaners, and general purpose degreasers
interchangeably (CARB, 2000). Automotive brake cleaners can come in aerosol or liquid form (CARB.
2000): this model estimates exposures from aerosol brake cleaners (degreasers).

FigureApx H-l illustrates the near-field/far-field modeling approach as it was applied by EPA to brake
servicing using an aerosol degreaser. The application of the aerosol degreaser immediately generates a
mist of droplets in the near-field, resulting in worker exposures at a PCE concentration Cnf. The
concentration is directly proportional to the amount of aerosol degreaser applied by the worker, who is
standing in the near-field-zone (i.e., the working zone). The volume of this zone is denoted by Vnf. The
ventilation rate for the near-field zone (Qnf) determines how quickly PCE dissipates into the far-field
(i.e., the facility space surrounding the near-field), resulting in occupational bystander exposures to PCE
at a concentration Cff. Vff denotes the volume of the far-field space into which the PCE dissipates out
of the near-field. The ventilation rate for the surroundings, denoted by Qff, determines how quickly PCE
dissipates out of the surrounding space and into the outside air.

—~

Figure Apx H-l. The Near-Field/Far-Field Model as Applied to the Brake Servicing Near-

Field/Far-Field Inhalation Exposure Model

In brake servicing using an aerosol degreaser, aerosol degreaser droplets enter the near-field in non-
steady "bursts," where each burst results in a sudden rise in the near-field concentration. The near-field
and far-field concentrations then decay with time until the next burst causes a new rise in near-field
concentration. Based on site data from automotive maintenance and repair shops obtained by CARB
(CARB. 2000) for brake cleaning activities and as explained in Appendix H.2.5 and H.2.9, the model
assumes a worker will perform an average of 11 applications of the degreaser product per brake job with
five minutes between each application and that a worker may perform one to four brake jobs per day
each taking one hour to complete. EPA modeled two scenarios: one where the brake jobs occurred back-

Page 251 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

to-back and one where brake jobs occurred one hour apart. In both scenarios, EPA assumed the worker
does not perform a brake job, and does not use the aerosol degreaser, during the first hour of the day.

EPA denoted the top of each five-minute period for each hour of the day (e.g., 8:00 am, 8:05 am, 8:10
am, etc.) as tm,n. Here, m has the values of 0, 1, 2, 3, 4, 5, 6, and 7 to indicate the top of each hour of the
day (e.g., 8 am, 9 am, etc.) and n has the values of 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, and 11 to indicate the top
of each five-minute period within the hour. No aerosol degreaser is used, and no exposures occur, during
the first hour of the day, to,o to to,n (e.g., 8 am to 9 am). Then, in both scenarios, the worker begins the
first brake job during the second hour, ti,o (e.g., 9 am to 10 am). The worker applies the aerosol
degreaser at the top of the second 5-minute period and each subsequent 5-minute period during the hour-
long brake job (e.g., 9:05 am, 9:10 am,... 9:55 am). In the first scenario, the brake jobs are performed
back-to-back, if performing more than one brake job on the given day. Therefore, the second brake job
begins at the top of the third hour (e.g., 10 am), and the worker applies the aerosol degreaser at the top
of the second 5-minute period and each subsequent 5-minute period (e.g., 10:05 am, 10:10 am,... 10:55
am). In the second scenario, the brake jobs are performed every other hour, if performing more than one
brake job on the given day. Therefore, the second brake job begins at the top of the fourth hour (e.g., 11
am), and the worker applies the aerosol degreaser at the top of the second 5-minute period and each
subsequent 5-minute period (e.g., 11:05 am, 11:10 am,... 11:55 am).

In the first scenario, after the worker performs the last brake job, the workers and occupational non-users
(ONUs) continue to be exposed as the airborne concentrations decay during the final three to six hours
until the end of the day (e.g., 4 pm). In the second scenario, after the worker performs each brake job,
the workers and ONUs continue to be exposed as the airborne concentrations decay during the time in
which no brake jobs are occurring and then again when the next brake job is initiated. In both scenarios,
the workers and ONUs are no longer exposed once they leave work.

Based on data from CARB (	:000\ EPA assumes each brake job requires one 14.4-oz can of

aerosol brake cleaner as described in further detail below. The model determines the application rate of
PCE using the weight fraction of PCE in the aerosol product. EPA uses a uniform distribution of weight
fractions for PCE based on facility data for the aerosol products in use (CARB. 2000).

The model design equations are presented below in EquationApx H-l through EquationApx H-14.

Near-Field Mass Balance
Equation Apx H-l

Far-Field Mass Balance
Equation Apx H-2

dCFF

^FF~dt~ = ^NF®NF ~ ^FF^NF ~ CffQff

Where:

Vnf = near-field volume;

Vff = far-field volume;

Qnf = near-field ventilation rate;

Page 252 of 316

V,

dC,

NF

NF

dt

CffQnf CnfQnf


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Qff = far-field ventilation rate;

Cnf = average near-field concentration;

Cff = average far-field concentration; and
t	= elapsed time.

Solving EquationApx H-l and EquationApx H-2 in terms of the time-varying concentrations in the
near-field and far-field yields Equation Apx H-3 and Equation Apx H-4, which EPA applied to each of
the 12 five-minute increments during each hour of the day. For each five-minute increment, EPA
calculated the initial near-field concentration at the top of the period (tm,n), accounting for both the burst
of PCE from the degreaser application (if the five-minute increment is during a brake job) and the
residual near-field concentration remaining after the previous five-minute increment (tm,n-i; except
during the first hour and tm,o of the first brake job, in which case there would be no residual PCE from a
previous application). The initial far-field concentration is equal to the residual far-field concentration
remaining after the previous five-minute increment. EPA then calculated the decayed concentration in
the near-field and far-field at the end of the five-minute period, just before the degreaser application at
the top of the next period (tm,n+i). EPA then calculated a 5-minute TWA exposure for the near-field and
far-field, representative of the worker's and ONUs' exposures to the airborne concentrations during each
five-minute increment using Equation Apx H-13 and Equation Apx H-14. The k coefficients
(Equation Apx H-5 through Equation Apx H-8) are a function of the initial near-field and far-field
concentrations, and therefore are re-calculated at the top of each five-minute period. In the equations
below, where the subscript "m, n-1" is used, if the value of n-1 is less than zero, the value at "m-1, 11"
is used and where the subscript "m, n+1" is used, if the value of n+1 is greater than 11, the value at
"m+1, 0" is used.

Equation Apx H-3
Equation Apx H-4
Equation Apx H-5

/Cl £ 	

1>Lm,n

Cnf t j.1 = (kit eXlt+ k2t e^)

iyir>Lm,n+1 v	z>Lm,n J

CpF t j.1 =(^3t eXlt — k4t eX2t)

rr>Lm,n+1 v 3,im,n	^>Lm,n J

QnF (CFF 0(f?n,ri) — Cnf,o(tm,n)) — A.2VNFCNF,o(j;m,n)

Lm,n	— ^2)

Equation Apx H-6

QnF	— ^FF,0	+ ^l^NF^NF.oiSm.n)

2,tm'n	Vnf(A1 — a2)

Equation Apx H-7

(.QnF + AjVnf)(,QnF (CFF,o(tm,n) ~ ^NF.oi^m.n)) ~ ^-2^NF^NF,0
3,tm'n	Qnf^nf(Ai — ^2)

Page 253 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

EquationApx H-8

(.Qnf + ^2^nf){Qnf (cWF0(tmn) — CFF 0(tmnj^ + \1 VNFCNF 0(tmn))

4'tm,n	Qnf^nf(.^-i ~ ^2)

EquationA

Xx = 0.5

EquationA

X2 = 0.5

px H-9

_ /QnF^FF + Vnf(QnF + Qff)\ I/Qnf^ff + Vnf(Qnf + Qff)\ _ a (QnfQff\

\ ^NF^FF ) J \ VnF^FF J ^ ^NF^FF '

px H-10

(Qnf^ff + Vnf(Qnf + Qff)\ /Qnf^ff + Vnf(Qnf + Qff)\ _ . (QnfQff\
\	^NF^FF	) J\	VnfVfF	) \VNFVFF)

Equation Apx H-ll

(	0, m = 0

CNF,o{pm,n) — j —f1,000——+ CWF(tm n_1) , n > 0 /or all m where brake job

V NF ^	9 '

Equation Apx H-12

r	0, m = 0

FF,o\tm,n) — {CFF(trriin^1), for all n where m > 0

occurs

Equation Apx H-13

C,

rk	k	\ (k	k	N

Um,n-l	j 2>,-m,n—l ^t-, J _ / Um.n-l	| 2-tm,n-l

Ai	^2/1 ^1	^2

NF, 5-min TWA, tr,

^2 ^1

Equation Apx H-14

^1	^2 J \ ^ 1	/

CfF, 5-min TWA, tm „ —	7 7

r2 rl

After calculating all near-field/far-field 5-minute TWA exposures (i.e., CWF 5-minTWA,tmn and

Cpp 5-min twa, tmn) f°r each five-minute period of the work day, EPA calculated the near-field/far-field

8-hour TWA concentration and 1-hour TWA concentrations following the equations below:

Equation Apx H-15

n	2jTO=0 Hr! = o[C/VF,5-min TWA,tmn X 0.0833 hr\

NF, 8-hr TWA =	!

Page 254 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

EquationApx H-16

2m=0 2n=o[^FF,5-minTWA,tm/n X 0.0833 hr\

CNF, 8-hr TWA ~	Q~hr~

Equation Apx H-17

n	_ UriiofC/VF.S-min TWA,tmn X 0.0833 kr\

CNF, 1-hr TWA =

Equation Apx H-18

n	_ hh=o[CFF,5-minTWA,tmtn X 0.0833 hr\

CFF, 1-hr TWA =	Thr^

EPA calculated rolling 1-hour TWA's throughout the workday and the model reports the maximum
calculated 1-hour TWA.

To calculate the mass transfer to and from the near-field, the free surface area (FSA) is defined to be the
surface area through which mass transfer can occur. The FSA is not equal to the surface area of the
entire near-field. EPA defined the near-field zone to be a hemisphere with its major axis oriented
vertically, against the vehicle, and aligned through the center of the wheel (see FigureApx H-l). The
top half of the circular cross-section rests against, and is blocked by, the vehicle and is not available for
mass transfer. The FSA is calculated as the entire surface area of the hemisphere's curved surface and
half of the hemisphere's circular surface per EquationApx H-19, below:

Equation Apx H-19

FSA = (2 X ^uRnf) + (2 x nR

If.

Where: Rnf is the radius of the near-field

The near-field ventilation rate, Qnf, is calculated in Equation Apx H-20 from the indoor wind speed,
vnf, and FSA, assuming half of the FSA is available for mass transfer into the near-field and half of the
FSA is available for mass transfer out of the near-field:

Equation Apx H-20

1

Qnf — — vnfFSA

The far-field volume, Vff, and the air exchange rate, AER, is used to calculate the far-field ventilation
rate, Qff, as given by Equation Apx H-21:

Equation Apx H-21

Qff = VppAE R

Using the model inputs described in Appendix H.2, EPA estimated PCE inhalation exposures for
workers in the near-field and for occupational non-users in the far-field. EPA then conducted the Monte

Page 255 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Carlo simulations using @Risk (Version 7.0.0). The simulations applied 100,000 iterations and the Latin
Hypercube sampling method.

H.2 Model Parameters

Table Apx H-l summarizes the model parameters and their values for the Brake Servicing Near-
Field/Far-Field Inhalation Exposure Model. Each parameter is discussed in detail in the following
subsections.

Page 256 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

TableApx H-l. Summary of Parameter Values and Distributions Used in the Brake Servicing Near-Field/Far-Field Inhalation
Exposure Model					

Input
Parameter

Symbol

Unit

Constant Model
Parameter Values

Variable Model Parameter Values

Comments

Value

Basis

Lower
Bound

Upper
Bound

Mode

Distribution
Type

Far-field
volume

Vff

m3

—

—

206

70,679

3,769

Triangular

Distribution based on data collected
bv CARB (CARB. 2000V

Air

exchange
rate

AER

hr"1

—

—

1

20

3.5

Triangular

Demou (2009) identifies tvoical
AERs of 1 hr"1 and 3 to 20 hr"1 for
occupational settings without and
with mechanical ventilation
systems, respectively. Hellweg
(2009) identifies average AERs for
occupational settings utilizing
mechanical ventilation systems to
be between 3 and 20 hr"1. Golsteijn
(2014) indicates a characteristic
AER of 4 hr"1. Peer reviewers of
EPA's 2013 TCE draft risk
assessment commented that values
around 2 to 5 hr"1 may be more
likelv (SCG, 2013), in agreement
with Golsteiin (2014). A triangular
distribution is used with the mode
equal to the midpoint of the range
provided by the peer reviewer (3.5
is the midpoint of the range 2 to 5
hr"1).

Near-field
indoor
wind speed

VNF

ft/hr

—

—

0

23,882

—

Lognormal

Lognormal distribution fit to
commercial-type workplace data
from Baldwin (1998).

cm/s

—

—

0

202.2

—

Lognormal

Near-field
radius

Rnf

m

1.5

—

—

—

—

Constant
Value

Constant.

Starting
time for

tl

hr

0

—

—

—

—

Constant
Value

Constant.


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Input





Constant Model
Parameter Values

Variable Model Parameter Values



Symbol

Unit









Comments

Parameter

Value

Basis

Lower
Bound

Upper
Bound

Mode

Distribution
Type

each



















application
period



















End time
for each
application
period

t2

hr

0.0833

—

—

—

—

Constant
Value

Assumes aerosol degreaser is
applied in 5-minute increments
during brake job.

Averaging
Time

tavg

hr

8

—

—

—

—

Constant
Value

Constant.



















Discrete distribution of PCE-based



















aerosol product formulations based



















on survey results from CARB

PCE
weight

wtfrac

wt frac





0.20

0.99



Discrete

(CARB, 2000). Where the weight
fraction of PCE in the formulation

fraction

















was given as a range, EPA assumed
a uniform distribution within the
reported range for the PCE
concentration in the product.

Degreaser
Used per
Brake Job

wd

oz/job

14.4

—

—

—

—

Constant
Value

Based on data from CARB (CARB,
2000).

Number of
Applicatio
ns per Job

Na

Application
s/job

11

—

—

—

—

Constant
Value

Calculated from the average of the
number of applications per brake
and number of brakes per job.

Amount



















Used per

Applicatio

n

Amt

g PCE/
application

—

—

7.4

36.7

—

Calculated

Calculated from wtfrac, Wd, and

Na.

Page 258 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Input
Parameter

Symbol

Unit

Constant Model
Parameter Values

Variable Model Parameter Values

Comments

Value

Basis

Lower
Bound

Upper
Bound

Mode

Distribution
Type

Operating
hours per
week

OHpW

hr/week

—

—

40

122.5

—

Lognormal

Lognormal distribution fit to the
operating hours per week observed
in CARB (CARB, 2000) site visits.

Number of
Brake Jobs
per Work
Shift

Nj

jobs/site-
shift

—

—

1

4

—

—

Calculated from the average
number of brake jobs per site per
year, OHpW, and assuming 52
operating weeks per year and 8
hours per work shift.

Page 259 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

H.2.1 Far-Field Volume

The far-field volume is based on information obtained from CARB (GARB. 2000) from site visits of
137 automotive maintenance and repair shops in California. CARB (2000) indicated that shop volumes
at the visited sites ranged from 200 to 70,679 m3 with an average shop volume of 3,769 m3. Based on
this data EPA assumed a triangular distribution bound from 200 m3 to 70,679 m3 with a mode of 3,769
m3 (the average of the data from CARB (2000)).

CARB measured the physical dimensions of the portion of the facility where brake service work was
performed at the visited facilities. CARB did not consider other areas of the facility, such as customer
waiting areas and adjacent storage rooms, if they were separated by a normally closed door. If the door
was normally open, then CARB did consider those areas as part of the measured portion where brake
servicing emissions could occur (CARB. 2000) . CARB's methodology for measuring the physical
dimensions of the visited facilities provides the appropriate physical dimensions needed to represent the
far-field volume in EPA's model. Therefore, CARB's reported facility volume data are appropriate for
EPA's modeling purposes.

H.2.2 Air Exchange Rate
The air exchange rale ( \l R) is hased on data from Demon ( ). I lei I u eg ( ). Golsteijn ( ).

and information received from a peer reviewer during the development of the 2014 TSCA Work Plan
Chemical Risk Assessment Trichloroethylene: Degreasing, Spot Cleaning and Arts & Crafts Uses (SCO.
2013). Demou (2009) identifies typical AERs of 1 hr"1 and 3 to 20 hr"1 for occupational settings without
and with mechanical ventilation systems, respectively. Similarly, Hellweg (2009) identifies average
AERs for occupational settings using mechanical ventilation systems to vary from 3 to 20 hr"1. Golsteijn
(2014) indicates a characteristic AER of 4 hr"1. The risk assessment peer reviewer comments indicated
that values around 2 to 5 hr"1 are likely (SCG. 2013). in agreement with Golsteijn (2014) and the low
end reported by Demou (2009) and Hellweg (2009). Therefore, EPA used a triangular distribution with
the mode equal to 3.5 hr"1, the midpoint of the range provided by the risk assessment peer reviewer (3.5
is the midpoint of the range 2 to 5 hr"1), with a minimum of 1 hr"1, per Demou (2009) and a maximum of
20 hr"1 per Demou (2009) and Hellweg (2009).

H.2.3 Near-Field Indoor Air Speed

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 separate
distributions for these industrial and commercial settings and used the commercial distribution for
facilities performing aerosol degreasing.

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 commercial facilities to a lognormal distribution with the
following parameter values: mean of 10.853 cm/s and standard deviation of 7.883 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


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

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.
However, a mean air speed (averaged over a work area) is the required input for the model.

H.2.4 Near-Field Volume

EPA defined the near-field zone to be a hemisphere with its major axis oriented vertically, against the
vehicle, and aligned through the center of the wheel (see FigureApx H-l). The near-field volume is
calculated per EquationApx H-22. EPA defined a near-field radius (Rnf) of 1.5 meters, approximately
4.9 feet, as an estimate of the working height of the wheel, as measured from the floor to the center of
the wheel.

Equation Apx H-22

1 4

VNF = 2 X g

H.2.5 Application Time

EPA assumed an average of 11 brake cleaner applications per brake job (see Appendix H.2.9). CARB
observed, from their site visits, that the visited facilities did not perform more than one brake job in any
given hour (	00). Therefore, EPA assumed a brake job takes one hour to perform. Using an

assumed average of 11 brake cleaner applications per brake job and one hour to perform a brake job,
EPA calculates an average brake cleaner application frequency of once every five minutes (0.0833 hr).
EPA models an average brake job of having no brake cleaner application during its first five minutes
and then one brake cleaner application per each subsequent 5-minute period during the one-hour brake
job.

H.2.6 Averaging Time

EPA was interested in estimating 8-hr TWAs for use in risk calculations; therefore, a constant averaging
time of eight hours was used.

H.2.7 Perchloroethylene Weight Fraction

CARB (2000) collected information on PCE concentrations from safety data sheets (SDS) of PCE-based
aerosol products used at 54 automotive maintenance and repair facilities. EPA used a discrete
distribution to model the PCE weight fraction based on the number of occurrences of each formulation
type. In some instances, the concentration of PCE was reported as a range. For these formulation types,
EPA used a uniform distribution to model the PCE weight fraction within the formulation type.

Table Apx H-2 provides a summary of the reported PCE weight fractions in the SDS's and the number
of occurrences of each formulation type, and the fractional probability of each formulation type.

Page 261 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Table Apx H-2. Summary of Perchloroethylene-

3ased Aerosol Degreaser Formulations

Formulation Type ID
Assigned by EPA for
Use in the Model

Perchloroethylene
Weight Fraction

Number of
Occurrences

Fractional Probability

1000

0.65-0.94

29

0.5370

1100

0.99

2

0.0370

1200

0.90

1

0.0185

1300

0.70-0.94

2

0.0370

1400

0.25-0.85

1

0.0185

1500

0.90-0.99

2

0.0370

1600

0.65-0.75

1

0.0185

1700

0.89

1

0.0185

1800

0.60-0.99

2

0.0370

1900

0.20-0.50

1

0.0185

2000

0.55

9

0.1667

2100

0.85

1

0.0185

2200

0.98

2

0.0370

Total

54

1.0000

H.2.8 Volume of Degreaser Used per Brake Job

CARB (2000) assumed that brake jobs require 14.4 oz of aerosol product. EPA did not identify other
information to estimate the volume of aerosol product per job; therefore, EPA used a constant volume of
14.4 oz per brake job based on CARB (2000).

H.2.9 Number of Applications per Brake Job

Workers typically apply the brake cleaner before, during, and after brake disassembly. Workers may
also apply the brake cleaner after brake reassembly as a final cleaning process (CARB. 2000).

Therefore, EPA assumed a worker applies a brake cleaner three or four times per wheel. Since a brake
job can be performed on either one axle or two axles (CARB. 2000). EPA assumed a brake job may
involve either two or four wheels. Therefore, the number of brake cleaner (aerosol degreaser)
applications per brake job can range from six (3 applications/brake x 2 brakes) to 16 (4
applications/brake x 4 brakes). EPA assumed a constant number of applications per brake job based on
the midpoint of this range of 11 applications per brake job.

H.2.10 Amount of Perchloroethylene Used per Application

EPA calculated the amount of perchloroethylene used per application using Equation Apx H-23. The
calculated mass of perchloroethylene used per application ranges from 7.4 to 36.7 grams.

Page 262 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

EquationApx H-23

Wd x wtfrac x 28.3495^-

Amt =	—

Na

Where:

Amt	= Amount of PCE used per application (g/application);

Wd	= Weight of degreaser used per brake j ob (oz/j ob);

Wtfrac	= Weight fraction of PCE in aerosol degreaser (unitless); and

Na	= Number of degreaser applications per brake j ob (applications/j ob).

II.2.11 Operating Hours per Week

CARB (2000) collected weekly operating hour data for 54 automotive maintenance and repair facilities.
The surveyed facilities included service stations (fuel retail stations), general automotive shops, car
dealerships, brake repair shops, and vehicle fleet maintenance facilities. The weekly operating hours of
the surveyed facilities ranged from 40 to 122.5 hr/week. EPA fit a lognormal distribution to the surveyed
weekly operating hour data. The resulting lognormal distribution has a mean of 16.943 and standard
deviation of 13.813, which set the shape of the lognormal distribution. EPA shifted the distribution to
the right such that its minimum value is 40 hr/week and set a truncation of 122.5 hr/week (the truncation
is set as 82.5 hr/week relative to the left shift of 40 hr/week).

H.2.12 Number of Brake Jobs per Work Shift

CARB (2000) visited 137 automotive maintenance and repair shops and collected data on the number of
brake jobs performed annually at each facility. CARB calculated an average of 936 brake jobs
performed per facility per year. EPA calculated the number of brake jobs per work shift using the
average number of jobs per site per year, the operating hours per week, and assuming 52 weeks of
operation per year and eight hours per work shift using Equation Apx H-24 and rounding to the nearest
integer. The calculated number of brake jobs per work shift ranges from one to four.

Equation Apx H-24

site-year shift

I =	1	

J r„weeks	...

52	x OHpW

yr	r

Where:

Nj	= Number of brake jobs per work shift (jobs/site-shift); and

OHpW = Operating hours per week (hr/week).

Page 263 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Appendix I Dry Cleaning Multi-Zone Inhalation Exposure Model
Approach and Parameters

This appendix presents the modeling approach and model equations used in the Dry Cleaning Multi-
Zone Inhalation Exposure Model. The model was developed through review of relevant literature and
consideration of existing EPA exposure models. This model uses a near-field/far-field approach (
2009). where a vapor generation source located inside the near-field diffuses into the surrounding
environment. Workers are assumed to be exposed to PCE vapor concentrations in the near-field, while
occupational non-users are exposed at concentrations in the far-field. Because there are multiple
activities with potential PCE exposure at a dry cleaner, a multi-zone modeling approach is used to
account for PCE vapor generation from multiple sources. The model considers the following three
worker activities:

•	Spot cleaning of stains on both dirty and clean garments: On receiving a garment, dry
cleaners inspect for stains or spots they can remove as much of as possible before cleaning the
garment in a dry cleaning machine. Spot cleaning may also occur after dry cleaning if the stains
or spots were not adequately removed. Spot cleaning occurs on a spotting board and can involve
the use of a spotting agent containing various solvents, such as PCE. Workers are exposed to
PCE when applying it via squeeze bottles, hand-held spray bottles, or even from spray guns
connected to pressurized tanks. Once applied, the worker may come into further contact with the
PCE if using a brush, spatula, pressurized air or steam, or their fingers to scrape or flush away
the stain (Young. 2012; NIOS 7a). For modeling, EPA assumed the near-field is a
rectangular volume covering the body of a worker.

•	Unloading garments from dry cleaning machines: At the end of each dry cleaning cycle, dry
cleaning workers manually open the machine door to retrieve cleaned garments. During this
activity, workers are exposed to PCE vapors remaining in the dry cleaning machine cylinder. For
modeling, EPA assumed that the near-field consists of a hemispherical area surrounding the
machine door, and that the entire cylinder volume of air containing PCE exchanges with the
workplace air, resulting in a "spike" in PCE concentration in the near-field, Cd, during each
unloading event. This concentration is directly proportional to the amount of residual PCE in the
cylinder when the door is opened. The near-field concentration then decays with time until the
next unloading event occurs.

•	Finishing and pressing: The cleaned garments taken out of the cylinder after each dry clean
cycle contain residual solvents and are not completely dried (Von Grote. 2003). The residual
solvents are continuously emitted into the workplace during pressing and finishing, where
workers manually place the cleaned garments on the pressing machine to be steamed and ironed.
EPA assumed any residual solvent is entirely evaporated during pressing, resulting in an increase
in the near-field PCE concentration during this activity. Workers are exposed to PCE vapors
while standing in vicinity of the press machine. Because this activity is typically performed
while standing, EPA assumed the near-field to be a rectangular volume covering the upper body
of the worker.

The model uses the following parameters to estimate exposure concentrations in the near-field and far-
field:

Page 264 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

•	Far-field size;

•	Near-field size;

•	Air exchange rate;

•	Indoor air speed;

•	Exposure duration;

•	Concentration of solvent in the drum after the dry cleaning cycle;

•	Residual solvent adhered to garments after dry cleaning;

•	Spot cleaning use rate; and

•	Operating hours per day.

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. 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 Professional Edition,
Version 7.0.0. 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 the model at 10,000 iterations to capture the range of
possible input values (i.e., including values with low probability of occurrence). Note: this is fewer
iterations than used for the near-field/far-field models described in other appendices as the multi-zone
model takes significantly longer to run and 10,000 iterations allowed the simulation to be complete in a
reasonable amount of time while still capturing the variability of each parameter.

Model results from the Monte Carlo simulation are presented as 95th and 50th percentile values. The
statistics were calculated directly in @Risk33. The 95th percentile value was selected to represent high-
end exposure level, whereas the 50th percentile value was selected to represent central tendency
exposure level. The following subsections detail the model design equations and parameters for the dry
cleaning model.

1.1 Model Design Equations

Figure Apx 1-1 illustrates the near-field/far-field modeling approach as it was applied by EPA to the
Dry Cleaning Multi-Zone Inhalation Exposure Model. As the figure shows, PCE vapor is generated in
each of the three near-fields, resulting in worker exposures at concentrations Cs, Cd, and Cf. The
volume of each zone is denoted by Vs, Vd, and Vf. The ventilation rate for the near-field zone (Qs, Qd,
Qf) determines how quickly PCE dissipates into the far-field (i.e., the facility space surrounding the
near-fields), resulting in occupational non-user exposures to PCE at a concentration Cff. Vff denotes the
volume of the far-field space into which the PCE dissipates out of the near-field. The ventilation rate for
the surroundings, denoted by Qff, determines how quickly PCE dissipates out of the surrounding space
and into the outside air.

33 @Risk; Palisade; https://www.palisade.com/risk/

Page 265 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Dry Cleaning
Machine

x VD /
\ ^ u ~
Qd^-'
Cn

n



Sp

lea

3 ° VI

_ r-J-



QS

era

c* vs

Qf



31







00'



3;



era

L

CpVf

Far-field (background)

Cc

Q,

-FF

FF

FigureApx 1-1. Illustration of the Dry Cleaning Multi-Zone Inhalation Exposure Model

The model design equations are presented below in EquationApx 1-1 through EquationApx 1-15.

Near-Field Mass Balance for Spot Cleaning (Multi-Zone)

EquationApx 1-1

dCs

^s~dt = ^FF®S ~ ^sQs + Gs

Near-Field Mass Balance for Finishing (Multi-Zone)

EquationApx 1-2

dCF

^F~dt = ^FF®F ~ ^fQf + GF

Near-Field Mass Balance for Dry Cleaning Machine (Multi-Zone)

EquationApx 1-3

dCD

Vd ~rr = CffQd ~ cdqd

dt

Far-Field Mass Balance
EquationApx 1-4

Where:

Vs

d C

Vff ^ = CsQs + CFQF + CdQd — CppQs — CFFQF — CFFQD — CFFQFF

= near-field volume for spot cleaning;

Page 266 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Vf

near-field volume for finishing;

Vd =

near-field volume for unloading dry cleaning machine;

Vff =

far-field volume;

Qs =

near-field ventilation rate for spot cleaning;

Qf

near-field ventilation rate for finishing;

Qd

near-field ventilation rate for dry cleaning machine;

Qff =

far-field ventilation rate;

Cs

average near-field concentration for spot cleaning;

CF

average near-field concentration for finishing;

Cd =

average near-field concentration for dry cleaning machine;

Cff =

average far-field concentration;

Gs

average vapor generation rate for spot cleaning;

G,

average vapor generation rate for finishing; and

t

elapsed time.

To calculate the mass transfer to and from the near-field, the free surface area, FSA, is defined to be the
surface area through which mass transfer can occur. Note that the FSA may not be equal to the surface
area of the entire near-field.

For spot-cleaning, EPA defined the near-field zone to be a rectangular box resting on the floor;
therefore, no mass transfer can occur through the near-field box's floor. FSA is calculated using
EquationApx 1-5:

EquationApx 1-5

FSAs = 2{LSHS) + 2(WSHS) + (.LSWS)

For finishing, EPA defined the near-field zone to be a rectangular box resting on the floor; therefore, no
mass transfer can occur through the near-field box's floor. FSA is calculated using Equation Apx 1-6:

EquationApx 1-6

FSAf = 2 (LnfHnf) + 2 (WnfHnf) + (LnfWnf)

For dry cleaning, EPA defined the near-field zone to be a hemispheric area projecting from the door of
the dry cleaning machine, calculated as Equation Apx 1-7:

EquationApx 1-7

FSAd = 2nrp

Where:

FSAs =

free surface area for spot cleaning;

FSAf =

free surface area for finishing;

FSAd =

free surface area for dry cleaning machine;

Ls =

near-field length for spot cleaning;

HS

near-field height for spot cleaning;

Ws =

near-field width for spot cleaning;

Lf =

near-field length for finishing;

Page 267 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Hf = near-field height for finishing;

Wf = near-field width for finishing; and

td = radius of the dry cleaning machine door opening.

The near-field ventilation rates, Qs, Qd, and Qf are calculated from the near-field indoor wind speed,
vnf, and FSA, using EquationApx 1-8 through EquationApx I-10, assuming half of FSA is available
for mass transfer into the near-field and half of FSA is available for mass transfer out of the near-field.
The near-field indoor wind speed is assumed to be the same across all three near fields:

EquationApx 1-8

Qs = 2 vnfFSAs

EquationApx 1-9

1

Qf — 2 vnfF$Af

EquationApx 1-10

Qd = 2 VnfFSAd

The far-field volume, Vff, and the air exchange rate, AER, is used to calculate the far-field ventilation
rate, Qff, as given by Equation Apx 1-11:

EquationApx 1-11

Qff = VppAE R

The model results in the following four, coupled ordinary differential equations (ODEs) given in
Equation Apx 1-12 through Equation Apx 1-15:

EquationApx 1-12

EquationApx 1-13

EquationApx 1-14

dCs _ Qs Qs r Gs

~7T ~ 7TLs +7T + 7T
at Vs Vs Vs

dCF Qp Qp Gp

—- =	Cp +—Cpp + —

dt VF F VF FF VF

dCD _ Qd Qd

~7T ~ ~ ~w~ + 7T~

dt VD VD

EquationApx 1-15

dCpp _ Qs n , Qf n , Qd n Qs + Qf + Qd + Qff n

~ Tt ls + t7 + Tt ld	77	lff

Clt V FF	VpF	VpF	VpF

Page 268 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

When solving coupled ODEs, it is common to transform the equations into a standard mathematical
format. This standard mathematical format allows one to more easily identify appropriate solution
methodologies from standard mathematical references. EPA transformed these four ODEs into the
following format in EquationApx 1-16 through EquationApx 1-19:

EquationApx 1-16

y'i = anyi + 0-iaJa + 9i

EquationApx 1-17

y'i = a22y2 + a24y4 + g2

EquationApx 1-18

y'i = a33y3 + «34y4

EquationApx 1-19

Va ~ O-AlJl + aA2j2 + aA3y3 + aAAjA

Where:

And:

Cs — yi CF — y2 CD — y3 CFF — y4

Qs+Qf+Qd+Qff
vFF

— Cl44

Page 269 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

These ordinary differential equations can be solved using a numerical integration method. EPA used the
fourth-order Runge-Kutta method (RK4). RK4 numerically integrates a system of coupled ordinary
differential equations from time step n to n+1 with a constant time step size of h using the following
equations (shown for generic variables yi, y2, y3, and y4 as a function of t).

EquationApx 1-20

dyx

~^~ = fiit.yv y2,y 3, y4)

EquationApx 1-21

dy2

-^- = f2(t,yvy2, y3, yd

EquationApx 1-22

dy3

-fa=h(t,y\,y2,y*,y4)

EquationApx 1-23

dy4

~^~ = fiit, yvy 2, y 3, y4)

Where, for each ODE j = 1, 2, 3, 4 (where 1 = spot cleaning, 2 = finishing, 3 = dry cleaning machine,
and 4 = far field):

EquationApx 1-24

k{ = fj(t,y1,y2,y3,y4)

EquationApx 1-25

111	11

= fjtf + 2h,yi + 2k^h,y2 + 2k*k,y3 + 2k*k,y4 + 2k^

EquationApx 1-26

111	11

ki = fjtf + 2h,yi + 2kzh,y2 + 2kzh,y3 + 2k*h'y4 + 2^)

EquationApx 1-27

= fj(t + h,y1+ k\h,y2 + k\h,y3 + fc|/i,y4 +

EquationApx 1-28

1

yf+1 = yj1 +-h(k( + 2 kJ2 + 2k}3 + k{)

RK4 is an explicit integration method, meaning it solves for the dependent variables at step n+1
explicitly using the dependent variables at step n. RK4 is a fourth-order method, which means the local
truncation error at a single integration step is on the order of h5, while the total global error is on the

Page 270 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

order of h4.

The choice of step size h is such to allow a successful integration of the system of differential equations.
If parameter values are chosen such that the differential equation coefficients (the a terms in
EquationApx 1-16 through EquationApx 1-19) are sufficiently large, the differential equations may
become stiff. Stiff differential equations would require sufficiently small time step sizes to allow their
integration. Stiffness can be difficult to predict. If stiffness is encountered, meaning if the solution
diverges to unrealistic values, such as infinity, the step size should be reduced to see if that allows for
successful integration.

Exposure Estimate Equations

The dry cleaning industry is characterized by a large number of small businesses, many are family-
owned and operated. EPA assumed small dry cleaners operate up to 12 hours a day and up to six days a
week. In addition, EPA assumed each facility has a single machine. The assumption of a single machine
per facility is supported by a recent dry cleaning industry study conducted in King County, Washington,
where 96 percent of 151 respondents reported having only one machine at their facility. Four reported
having two machines, and two reported having three machines (Whittaker and Joh an son. 2011). Based
on the survey results, this assumption is presumably representative of the majority of small dry cleaning
shops.

The model accounts for variation in the machine generations operated at each facility. Specifically, the
model uses a distribution to estimate the machine generation and then based on the sampled machine
generation in each iteration selects a distribution of machine cylinder concentrations and residual solvent
in clothing. The distribution of machine types is based on the 2010 survey of dry cleaners in King
County, WA, which estimated 7% were first or second generation, 26% of machines were third
generation or retrofitted second generation34, 61% were fourth or fifth generation, and 6% were "other"
(e.g., hydrocarbon or CO: machines) (Whittaker and Johanson. 2011). Due to the limited information on
other machine types, the model only considers two scenarios: 1) facilities operating third generation
machines; and 2) facilities operating a fourth or fifth generation machine35. These assumptions are not
expected to introduce significant error in the exposure estimates as, based on bans on first and second
generation machines in the 1993 and 2006 Perchloroethylene NESHAPs for Dry Cleaning Facilities,
EPA expects the use of PCE in first and second generation machines to be eliminated (
2006a). Additionally, based on several survey results and projections (presented in Table Apx I-1), EPA
expects the industry to be trending towards increasing usage of fourth and fifth generation machines.
Therefore, the 7% of facilities reporting using first- and second-generation machines were assumed to be
replaced by fourth or fifth generation resulting in 26% third generation machines and 68% fourth or fifth
generation machines. The model only considers exposure at facilities using PCE; therefore, EPA re-
normalized the distribution to consider only PCE machines resulting in a distribution of 28% third
generation machines and 72% fourth or fifth generation machines.

34	For modeling purposes, retrofitted second generation machines are assumed to be equivalent to third generation machines.

35	The model treats fourth and fifth generation machines as equivalent as both are expected to reduce machine cylinder
concentrations to approximately 300 ppm (CDC, 1997). The primary difference being that fifth generation machines have an
interlock preventing the machine door from being opened until the concentration is below 300 ppm whereas fourth
generation machines do not.

Page 271 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Table Apx 1-1. Summary of Survey Responses for Dry Cleaning Machine Generations

Machine Type

Percent of Survey Respondents or Projected Facilities

2000 HSIA
Survey (ERG,
2005)

2003 CA Survey
(California Air
Resources Board,
2006)

2006 Projection
(ERG. 2005)

2010 King Cunty

WA Survey
(Whittaker and
Johanson, 2011)

1st Generation

1.4%

1%

1%

1%

2nd Generation

3%

—

1%

6%

2nd Generation
Retrofitted

—

2%

—

3%

3rd Generation

65%

62%

37%

23%

4th Generation

31%

28%

61%

28%

5th Generation

—

—

—

33%

Other

—

2%

—

6%

Total

100%

95%

100%

100%

EPA assessed three types of workers within the modeled dry cleaning facility: 1) a worker who performs
spot cleaning; 2) a worker who unloads the dry cleaning machine and finishes and presses the garments;
and 3) an occupational non-user. Each worker type is described in further detail below. EPA assumed
each worker activity is performed over the full 12-hour operating day.

•	EPA assumed spot cleaning occurs for a duration varying from two to five hours in the middle of
the 12-hour day. The worker is exposed at the spot cleaning near-field concentration during this
time, and at the far-field concentration for the remainder of the day. Spot cleaning can be
performed for both dry cleaned loads and for laundered loads.

•	EPA assumed a separate worker unloads the dry cleaning machine and finishes and presses the
garments. After each load, EPA assumed this worker spends five minutes unloading the machine,
during which he or she is exposed at the machine near-field concentration. After unloading, the
worker spends five minutes in the finishing near-field to prepare the garments. Then, the worker
spends another 20 minutes finishing and pressing the cleaned garments. During this 20-minute
period of finishing and pressing, the residual PCE solvent is off-gassed into the finishing near-
field. The amount of residual PCE solvent is estimated using measured data presented in von
Grote (2003). These unloading and finishing activities are assumed to occur at regular intervals
throughout the twelve-hour day. The frequency of unloading and finishing depends on the
number of loads dry cleaned each day, which varies from 1 to 14, where 14 was the maximum
number of loads observed in the NIOSH (2010) and Blando (2010) studies. When this worker is
not unloading the dry cleaning machine or finishing and pressing garments, the worker is
exposed at the far-field concentration.

•	EPA assumed one occupational non-user is exposed at the far-field concentration for 12 hours a
day. The occupational non-user could be the cashier, tailor, or launderer, who works at the
facility but does not perform dry cleaning activities.

Page 272 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Using the model inputs described in Section 1.2, EPA estimated PCE inhalation exposures for workers
performing spot cleaning, workers unloading the dry cleaning machine and performing finishing and
pressing activities, and for occupational non-users in the far-field. EPA then conducted the Monte Carlo
simulations using @Risk (Version 7.0.0). The simulations applied 10,000 iterations and the Latin
Hypercube sampling method for each model.

1,2 Model Parameters

Table Apx 1-2 summarizes the model parameters and their values for the Dry Cleaning Multi-Zone
Inhalation Exposure Model. Each parameter is discussed in detail in the following subsections.

Page 273 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Table Apx 1-2. Summary

Exposure Model

Input Parameter

Symbol

Unit

Constant Model
Parameter Values

Variable Model Parameter Values

Notes/Comments

Value

Basis

Lower
Bound

Upper
Bound

Mode

Distribution
Type

Facili

ty Parameters

Facility Height

Fh

ft

12

Median

—

—

—

—

See Section 1.2.1.1

Facility Floor Area

Farea

ft2

—

—

500

20,000

—

Beta

See Section 1.2.1.1

Far-field volume

Vff

ft3

—

—

6,000

240,000

—

—

See Section 1.2.1.1

Air exchange rate

AER

hr"1

—

—

1

19

3.5

Triangular

See Section 1.2.1.2

Near-field indoor wind
speed

VNF

ft/hr

—

—

—

202.2

—

Lognormal

See Section 1.2.1.3

cm/s

—

—

—

23,882

—

Lognormal

Dry Cleaning Machine Paramet

ers

Machine Door Diameter

D

ft

2.083

—

—

—

—

—

See Section 1.2.2.1

Number of Loads per Day

LD

loads/day

—

—

1

14

—

Uniform

See Section 1.2.2.2

Load Time

LT

hr/load

0.5

—

—

—

—

—

See Section 1.2.2.3

3rd Generation Machine
Cylinder PCE
Concentration

Cc 3RD

ppm

—

—

2,000

8,600

—

Uniform

See Section 1.2.2.4

4th Generation Machine
Cylinder PCE
Concentration

Cc_4TH

ppm

—

—

240

360

—

Uniform

See Section 1.2.2.4

Cylinder Volume

Vc

m3

—

—

0.24

0.64

—

Uniform

See Section 1.2.2.5

Starting time

ti

hr

0

—

—

—

—

—

Constant value.

Exposure Duration

t2

hr

0.083

—

—

—

—

—

See Section 1.2.2.6

Finishing anc

Pressing Paramet

ers

Near-field length

Lnf

ft

10

—

—

—

—

—

See Section 1.2.3.1

Near-field width

Wnf

ft

10

—

—

—

—

—

Near-field height

Hnf

ft

6

—

—

—

—

—

3rd Generation Machine
Residual Solvent

Rsolvent 3RD

g/kg

—

—

0.26

3.75

—

Discrete

See Section 1.2.3.2

4th Generation Machine
Residual Solvent

Rsolvent 4TFI

g/kg

—

—

0.12

1.26

—

Discrete

See Section 1.2.3.2

Load Size

LS

lb/load

30

—

—

—

—

—

See Section 1.2.3.3

Exposure Duration

t3

hr

0.33

—

—

—

—

—

See Section 1.2.3.4


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Input Parameter

Symbol

Unit

Constant Model
Parameter Values

Variable Model Parameter Values

Notes/Comments

Value

Basis

Lower
Bound

Upper
Bound

Mode

Distribution
Type

Spot Cleaning Parameters

Near-field length

Lnf

ft

10

—

—

—

—

—

See Section 1.2.4.1

Near-field width

Wnf

ft

10

—

—

—

—

—

Near-field height

Hnf

ft

6

—

—

—

—

—

Use Rate

UR

gal/yr

0

—

—

—

—

—

See Section 1.2.4.2

Exposure Duration

U

hr

—

—

2

5

—

Uniform

See Section 1.2.4.3

Other Paramei

ters

Operating hours per day

OH

hr

12

—

—

—

—

—

See Section 1.2.5.1

Operating days

OD

days/yr

—

—

249

313

300

Triangular

See Section 1.2.5.2

Fractional days of exposure

f

unitless

—

—

0.8

1.0

—

Uniform

See Section 1.2.5.3

Page 275 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

1.2.1 Facility Parameters

1.2.1.1 Far-Field Volume

EPA calculated the far-field volume by setting a distribution for the facility floor area and multiplying
the floor area by a facility height of 12 ft (median value per California Air Resources Board (CARB)
(2006) study) as discussed in more detail below.

The 2006 CARB California Dry Cleaning Industry Technical Assessment Report (California Air
Resources Board. 2006) and the Local Hazardous Waste Management Program in King County A
Profile of the Dry Cleaning Industry in King County, Washington (Whittaker and Johanson. 2011)
provide survey data on dry cleaning facility floor area. The CARB (California Air Resources Board.
2006) study also provides survey data on facility height. Using survey results from both studies, EPA
composed the following distribution of floor area. To calculate facility volume, EPA used the median
facility height from the CARB (California Air Resources Board. 2006) study. The facility height
distribution in the CARB (California Air Resources Board. 2006) study has a low level of variability, so
the median height value of 12 ft presents a simple but reasonable approach to calculate facility volume
combined with the floor area distribution.

Table Apx 1-3. Composite Distribution of Dry Cleaning Facility Floor Areas

Floor

Percentile



Area

(as

Source

Value (ft2)

fraction)



20,000

1

(Whittaker and Johanson,
2011)

3,000

0.96

(Whittaker and Johanson,
2011)

2,000

0.84

(Whittaker and Johanson,
2011)

1,600

0.5

(California Air Resources
Board, 2006)

1,100

0.1

(California Air Resources
Board, 2006)

500

0

(California Air Resources
Board, 2006)

EPA fit a beta function to this distribution with parameters: ai = 6.655, a.2 = 108.22, min = 500 ft2, max
= 20,000 ft2

1.2.1.2 Air Exchange Rate

von Grote et al. (2006)von Grote (2006) indicated typical air exchange rates (AERs) of 5 to 19 hr"1 for
dry cleaning facilities in Germany. Klein (1994) indicated AERs of 1 to 19 hr"1, with a mean of 8 hr"1 for
dry cleaning facilities in Germany. During the 2013 peer review of EPA's 2013 draft risk assessment of
TCE, a peer reviewer indicated that air exchange rate values around 2 to 5 hr"1 are likely (SCG. 2013). in
agreement with the low end of the ranges reported by von Grote (2006). and Klein (1994). A triangular
distribution is used with the mode equal to the midpoint of the range provided by the peer reviewer (3.5
is the midpoint of the range 2 to 5 hr"1). The minimum and maximum of the distribution are 1 and 19 hr"
1, respectively.


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

1.2.1.3 Near-Field Indoor Air Speed

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 categorizing 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 commercial distribution
for dry cleaners.

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 air speed surveys representative of commercial facilities were fit to a lognormal distribution with
the following parameter values: mean of 10.853 cm/s and standard deviation of 7.883 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.
However, a mean air speed (averaged over a work area) is the required input for the model.

1.2.2 Dry Cleaning Machine Parameters

1.2.2.1	Machine Door Diameter

EPA determined an approximate door diameter of 25 inches by reviewing images of several 4th
generation PCE machine models manufactured by Bowe and Firbimatic.

1.2.2.2	Number of Loads per Day

EPA used a uniform distribution for the number of loads per day ranging from 1 to 14 based on
observations from NIOSH (2010) and Blando (2010).

1.2.2.3	Load Time

EPA estimates that dry cleaning loads using PCE have an average cycle duration of 30 minutes (0.5
hours). This estimate is consistent with von Grote (2003). which estimated total cleaning and finishing
batch times of between 45 to 65 minutes for machines equivalent to U.S. 3rd generation machines and
between 50 to 70 minutes for machines equivalent to U.S. 4th generation machines, von Grote (2003)
further estimated that between one-fourth and one-third of the total cleaning and finishing batch time is
spent finishing/pressing (see Section 1.2.3.4). EPA assumed a total cleaning and finishing batch time of
60 minutes with the following breakdown:

• The finish/pressing duration is 20 minutes (see Section 1.2.3.4);

Page 277 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

•	The time to unload the garments from the machine is 5 minutes based on engineering judgment;

•	The time to prepare the garments for finishing/pressing is 5 minutes based on engineering
judgment; and

•	The machine cycle duration is 30 minutes based on the total cleaning and finishing batch time of
60 minutes minus the above task durations.

1.2.2.4	Machine Cylinder Concentration

EPA used two different distributions for machine cylinder concentration depending on the machine type
being modeled (third or fourth generation). For third generation machines, EPA used a uniform
distribution from 2,000 to 8,600 ppm to estimate the machine cylinder concentration after a dry cleaning
cycle. ERG (2005) indicated that the use of refrigerated condensers (the vapor control system used in
third generation machines) can reduce PCE concentrations in the drum to between 2,000 and 8,600 ppm.

For fourth generation machines, EPA used a uniform distribution from 240 to 360 ppm to estimate the
machine cylinder concentration after a dry cleaning cycle. NIOSH (1997a) indicated that the use of
refrigerated condensers and carbon adsorbers in fourth generation machines can reduce the PCE
concentration in the drum below 300 ppm after the cycle is complete. EPA used a uniform distribution
of 300 ppm +/- 20% to account for variability and uncertainty in the residual concentration.

1.2.2.5	Cylinder Volume

EPA assessed the cylinder volume using a uniform distribution of 0.24 to 0.64 m3 based on data from
von Grote (2003). von Grote (2003) presented the five most common machine sizes used in Germany
based on a 2002 survey with sizes ranging from 0.24 to 0.64 m3. EPA did not have data on the machine
sizes or distributions used in the U.S. Therefore, EPA modeled the cylinder volume using the range
provided by von Grote (2003) and assuming a uniform distribution of machine sizes.

1.2.2.6	Exposure Duration

EPA assumes it takes the worker five minutes to unload the dry cleaning machine.

1.2.3 Finishing and Pressing Parameters

1.2.3.1	Near-Field Volume

EPA assumed a near-field of constant dimensions of 10 ft x 10 ft x 6 ft resulting in a total volume of 600
ft3.

1.2.3.2	Residual Solvent

EPA used two different distributions for the amount of residual solvent that adheres to garments after the
dry cleaning cycle depending on the machine type being modeled (third or fourth generation). The
distributions for both machine types are based on data from von Grote (2003) who estimated residual
solvent for both normal loads and "off-the-peg" loads, von Grote (2003) defines "off-the-peg" loads as
loads with suits and jackets with shoulder pads and estimates that approximately 20% of all loads
cleaned are off-the-peg with the remaining 80% being normal loads. For third generation machines, von
Grote (2003) presents data estimating 0.26 g residual solvent/kg clothes for normal loads and 3.75 g
residual solvent/kg clothes for off-the peg loads. It should be noted that von Grote (2003) uses different
definitions of machines generations than used in the U.S. The fourth-generation machines in von Grote
(2003) are defined as non-vented dry-to-dry machines with refrigerated condensers which corresponds

Page 278 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

to third-generation machines in the U.S. Therefore, EPA used data for fourth-generation machines in
von Grote (2003) to model U.S. third-generation machines.

von Grote (2003) does not have a machine generation corresponding to fourth-generation machines in
the U.S. von Grote (2003) fourth-generation machines correspond to U.S. third-generation machines and
von Grote (2003) fifth-generation machines correspond to U.S. fifth-generation machines (machines
with refrigerated condensers, carbon adsorbers, and interlocks on the door). However, the only
difference between U.S. fourth- and fifth-generation machines is the presences of interlocks on the door
to prevent workers from opening prior to the solvent concentration dropping below 300 ppm. As
discussed in Section 1.2.2.4, fourth-generation machines are also expected to reduce cylinder
concentrations after a cycle to 300 ppm. Therefore, EPA expects residual solvent for fourth-generation
machines to be similar to fifth-generation machines and uses residual solvent data from von Grote
(2003) for fifth-generation machines in the estimates for fourth-generation machines, von Grote (2003)
presents data estimating 0.12 g residual solvent/kg clothes for normal loads and 1.26 g residual
solvent/kg clothes for off-the peg loads for fifth-generation machines. EPA assumes a discrete
distribution for both third- and fourth-generation estimates assuming 80% of loads are normal loads and
20% are off-the-peg von Grote (2003).

1.2.3.3 Load Size

The CARB (California Air Resources Board. 2006) and King County (Whittaker and Johanson. 2011)
studies provide machine capacities, and the King County study also provides data on actual size of loads
used by dry cleaners. EPA used the King County study data on actual load sizes to build a distribution.

TableApx 1-4 summarizes the survey results for respondents' primary (if facility has more than one
machine) or only machine. The study reports a maximum reported load of 150 lb, a minimum reported
load of 7 lb, and a median reported load of 30 lb for the primary machine (Whittaker and Johanson.
2011).

Table Apx 1-4. Survey Responses of Actual Pounds Washed per Load for Primary Machine (if

Actual Pounds

Results for Primary Machine

of Clothes

Number of

Percent of

Washed

Respondents

Respondents

1 - 10

4

3

11-20

36

25

21-30

76

53

31-40

16

11

41-50

6

4

51+

6

4

Total

144

100

EPA used these survey results to build a distribution to describe the actual wash loads per machine, as
summarized in Table Apx 1-5. To build this distribution, EPA set the following:

Page 279 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

•	The maximum, median, and minimum were set as 150 lb, 30 lb, and 7 lb, respectively, as stated
in the King County survey report (Whittaker and Johanson. 2011).

•	The 96th percentile was set at 50 lb as the high-end of the bin "41 to 50 lb". Per TableApx 1-4,
4% of respondents reported greater than 50 lb; therefore, 96% of facilities reported 50 lb or less.

•	The 28th percentile was set at 20 lb as the high-end of the bin "11 to 20 lb". Per Table Apx 1-4,
28% of respondents reported 20 lb or less.

EPA then determined the best-fit distribution using the software @Risk.

Table Apx 1-5. Distribution of Actual Load Sizes from 2010 King County Survey

Actual Load Washed
(lb)

Percentile
(as fraction)

150

1

50

0.96

30

0.5

20

0.28

7

0

Source: (Whittaker and Johanson. 2011)

EPA fit a beta distribution to this distribution with parameters: ai = 2.3927, 0.2 = 12.201, min = 7 lb, max
= 150 lb. The root-mean squared (RMS) error is 0.0365, Figure Apx 1-2 illustrates this fit.

Page 280 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Fit Comparison for Load Size	nil ii'll'lll

RbHMaG«ieral(2.39Z7.12.201,7,150) Minimum ^^^7.00	7.00

49-6 Maximum	150.00	150.00

Mean	31.68	30.45

Mode	N/A	22.81

Median	30.00	28.23

StdDev	19.37	13.41

Skewness	2.5076	0.8640

[xurtosis	13.5139	3.7151

LeftX	9.3	9.3

LeftP	5.0%	0.7%

Right X	49.6	49.6

Right P	95.0%	90.6%

Dif. X	40.24	40.24

|pif. P	90.0%	89.9%

[1%	7.46	9.71

5%	9.32	12.70

10%	11.64	15.03

15%	13.96	16.94

20%	16.29	18.66

25%	18.61	20.28

30%	20.91	21.86

35%	23.18	23.42

40%	25.45	24.98

45%	27.73	26.58

50%	30.00	28.23

55%	32.17	29.96

60%	34.35	31.78

65%	36.52	33.75

70%	38.70	35.90

75%	40.87	38.32

80%	43.04	41.11

85%	45.22	44.50

90%	47.39	48.94

95%	49.57	55.81

99%	125.00	69.31

FigureApx 1-2. Fit Comparison of Beta Cumulative Distribution Function to Load Size Survey

Results

1.2.3.4 Exposure Duration

EPA assumed workers take 20 minutes to press and finish each load. This estimate is consistent with
von Grote (2003). which estimated that residual solvent will evaporate continuously over a period of
approximately between one-fourth and one-third of the total time to clean and finish a single load of
garments, von Grote (2003) estimated total cleaning and finishing batch times of between 45 to 65
minutes for machines equivalent to U.S. 3rd generation machines and between 50 to 70 minutes for
machines equivalent to U.S. 4th generation machines. This yields an overall range of finishing/pressing
times of approximately 11 to 23 minutes.

1.2.4 Spot Cleaning Parameters

1.2.4.1	Near-Field Volume

EPA assumed a near-field of constant dimensions of 10 ft x 10 ft x 6 ft resulting in a total volume of 600

ft3.

1.2.4.2	Spot Cleaning Use Rate

EPA did not identify information to estimate the use rate of PCE in spot cleaners; however, IRTA
(2007) and ERG (2005) indicate that the use of PCE in spot cleaners is minimal. Specifically, IRTA
(2007) state that only 150 gal of PCE-based spotting agents are used annually in California (compared to
42,000 gal of TCE-based spotting agents). ERG (2005) stated that many PCE spotting agents are
categorized as oily type paint removers (OTPR), but that the majority of OTPR spotting agents contain

Page 281 of 316

5.0%

90.0%

5.0%

0.7%

89.9%

9.4%







/

















h

/























@RISK for Excel
Palisade Corporation









I

/















L



















3 C
r

) c

J "3

) c
v£

> c
j a

) c
J c

>	c

>	fN

) C
1

) o
vC

4 tH


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

no PCE. Therefore, EPA set the use rate of PCE spotting agents to zero. This results in the spot cleaning
near-field of the model to become part of the far-field with exposure concentrations equivalent to Cff.

1.2.4.3 Exposure Duration

IR.TA (2007) used data collected from dry cleaners to develop two model PCE-based dry cleaners: a
small and large dry cleaner. The authors estimated the small dry cleaner spends 2.46 hr/day spotting and
the large dry cleaner spends 5 hr/day spotting. EPA models the spot cleaning duration as a uniform
distribution varying from 2 to 5 hr/day.

1.2.5 Other Parameters

1.2.5.1	Operating Hours

EPA assumed a typical dry cleaner operates 12 hours per day based on engineering judgment.

1.2.5.2	Operating Days per Year

EPA modeled the operating days per year using a triangular distribution from 250 to 312 days per year
with a mode of 300 days per year36. The low-end operating days per year is based on the assumption that
at a minimum the dry cleaner operates five days per week and 50 weeks per year. The mode of 300 days
per year is based on an assumption that most dry cleaners will operate six days per week and 50 weeks
per year. The high-end value is based on the assumption that the dry cleaner would operate at most six
days per week and 52 weeks per year, assuming the dry cleaner is open year-round.

1.2.5.3	Fractional Number of Operating Days that a Worker Works

To account for lower exposure frequencies and part-time workers, EPA defines a fractional days of
exposure as a uniform distribution ranging from 0.8 to 1.0. EPA expects a worker's annual working days
may be less than the operating days based on BLS/Census data that showed the weighted average
worked hours per year and per worker in the dry cleaning sector is approximately 1,600 (i.e., 200 day/yr
at 8 hr/day) which falls outside the range of operating days per year used in the model (250 to 312
day/yr with mode of 300 day/yr).

The low end of the range, 0.8, was derived from the observation that the weighted average of 200 day/yr
worked (from BLS/Census) is 80% of the standard assumption that a full-time worker works 250 day/yr.
The maximum of 1.0 is appropriate as dry cleaners may be family owned and operated and some
workers may work as much as every operating day. EPA defines the exposure frequency as the number
of operating days (250 to 312 day/yr) multiplied by the fractional days of exposure (0.8 to 1.0).

36 For modeling purposes, the minimum value was set to 249 days per year and the maximum to 313 days per year; however,
these values have a probability of zero; therefore, the true range is from 250 to 312 days per year.

Page 282 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Appendix J Solvent Releases in Water Discharge from Dry Cleaning
Machines Model Approach and Parameters

This appendix presents the modeling approach and model equations used in the Solvent Release in
Water Discharge from Dry Cleaning Machines Model. This model estimates the PCE contained in the
produced separator water from the PCE dry cleaning machine. FigureApx J-l illustrates an example
process flow diagram of a 5th generation dry cleaning machine (NIOS	). This diagram

illustrates two sets of controls on the vapor loop: a refrigerated condenser and a carbon adsorber. The
refrigerated condenser condenses PCE from the vapor prior to its return to the machine chamber. The
refrigerated condenser also incidentally condenses any water vapor present in the chamber vapor. The
mixed water/PCE condensate drops down into a water separator. The water separator then separates the
heavier PCE from the lighter water, returning the PCE to the solvent tank. However, some PCE may
remain in the water phase at a concentration up to its solubility. The water is discharged, disposed, or
treated, depending on state regulations. Fourth generation machines will have an almost identical
process as 5th generation machines except they do not have the monitor shown in Figure Apx J-l
(NIOSH. 1997b). Third generation machines will also be similar to the process shown in Figure Apx
J-l, except they do not use a monitor in the machine to control residual solvent levels in the chamber
and do not use carbon adsorbers as secondary vapor controls (NIOSH. 1997b).

CDNTRMINRTED

Figure Apx J-l. Process Flow Diagram of a 5th Generation Dry Cleaning Machine (NIOSH.

1997a)

The model is based on the EPA/OPPT Water Saturation Loss Model, which assumes that water
contacted with the chemical becomes saturated with the chemical and remains saturated at the time of
disposal (	). The EPA/OPPT Water Saturation Model uses Equation Apx J-l and

Page 283 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Equation_Apx J-2 to calculate annual and daily amount of chemical released per site, respectively (U.S.
).

EquationApx J-l

WS x CF x Amt

DR = 1000

Equation Apx 3-2

AR = DR X OD

Where:

DR = Daily release rate (kg/site-day)

WS = Water solubility of the chemical (mg/L)

CF = Correction factor (unitless)37

Amt = An amount of water in which the chemical reaches saturation (kg/site-day)
AR = Annual release rate (kg/site-yr)

OD = Operating Days (days/yr)

This model uses the same basic principles as used in EPA/OPPT Water Saturation Model; however,
instead of the default value for "Amt" it uses several parameters and distributions to estimate high-end
and central tendency daily and annual release estimates including:



volume of produced water per pound of clothes cleaned;

• load size:

number of loads per day;
number of machines per site; and
operating days.

To account for parameter distributions, EPA used a Monte Carlo simulation with 100,000 iterations. The
following subsections describe EPA's modeling approach for estimating PCE water releases at dry
cleaning sites, including supporting rationale, calculations, and input parameters.

J.l_ Model Design Equations

The daily and annual release volumes of PCE from produced separator water are calculated using
Equation Apx J-3 and Equation Apx J-4.

Equation Apx J-3

WS x PW x LS x LD x NM

DR =	w	

2.20462 -P-

kg

Equation Apx J-4

AR = DR X OD

37 A correction factor that may be used to account for: 1) multiples of an amount (Amt) of water (e.g.,
a known or estimated correction of the water solubility of the chemical; and/ or 3) other corrections.

Page 284 of 316

multiple washings); 2)


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Where:

DR	= Daily release (kg/site-day)

WS	= PCE solubility in water (lb PCE/gal water)

PW	= Produced water (gal water/lb clothes)

LS	= Load size (lb clothes/load)

LD	= Number of loads per day (loads/machine-day)

NM = Number of machines per site (machines/site)

AR	= Annual release (kg/site-day)

OD	= Operating days per year (days/yr)

J.2 Model Parameters

TableApx J-l summarizes the model parameters and their values.

TableApx J-l. Summary of Parameter Values and Distributions for the Solvent Release in Water
Discharge from Dry Cleaning Machines Model	

Input
Parameter

Symbol

1 nil

Constant
Parameters

Value

\ ariahk

1 .ower
Hound

Pa ram el

I pper
Hound

l'IS

Mode

Distribution
Type

PCE Solubility
in Water

WS

lb PCE/gal

0.0017

—

—

—

Constant
Value

Produced Water

PW

gal water/
lb clothes

—

0.0032

0.0037

—

Discrete

Load Size

LS

lb clothes/
load

—

7

150

—

Beta

Number of
Loads per Day

LD

loads/day

—

1

14

—

Uniform

Number of
Machines

NM

machines/
site

—

1

3

1

Discrete

Operating Days

OD

days/yr

—

250

312

300

Triangular

J.2.1 Solubility in Water

The Problem Formulation of the Risk Evaluation for Perchloroethylene (Ethene, 1,1,2,2-Tetrachloro)
(	2018b) identifies a PCE solubility in water at 25 °C of 206 mg/L. This is converted to

0.0017 lb/gal. This parameter is kept at a constant value in the model.

J.2.2 Produced Water

The CARB California Dry Cleaning Industry Technical Assessment Report surveyed dry cleaning
facilities in California in 2003 (California Air Resources Board. 2006). The survey results of PCE
facilities found an average produced separator water of 141 gal per year for primary machines
(equivalent of 3rd generation machines) and 191 gal per year for secondary machines (equivalent of 4th

Page 285 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

and 5th generation machines) (California Air Resources Bo; 36). The survey also found PCE
facilities clean an average of 44,000 lb of clothes per year for primary machines and 52,000 lb of clothes
per year for secondary machines (California Air Resources Bo; 36).

EPA calculated produced separator water emission factors by dividing the average annual produced
separator water volume by the average annual clothes cleaned for each machine type. EPA calculated
the following emission factors:

•	Primary machine (3rd generation): 141 gal/yr / 44,000 lb clothes/yr = 0.00320 gal water/lb
clothes

•	Secondary machine (4th and 5th generation): 191 gal/yr / 52,000 lb clothes/yr = 0.00367 gal
water/lb clothes

EPA defined the distribution of the produced separator water emission factor as the distribution of
machine types. Using data from the King County survey results (Whittaker and Joh an son. 2011). EPA
built a distribution of current market shares of 3rd generation (including converted 2nd generation)
machines and 4th and 5th generation machines.

Since EPA expects the use of first- and second-generation machines to be eliminated, the 7% for these
machine types were assumed to be replaced by fourth or fifth generation machines to give the most
conservative water release estimate. EPA then re-normalized the distribution to consider only PCE
machines resulting in 28% of facilities using third generation machines and 72% using fourth or fifth
generation machines. TableApx J-2 summarizes the 2010 King County survey results and TableApx
J-3 shows the discrete distribution used for produced water.

Table Apx J-2. Distribution of Machine Types Based on 2010 King County Survey Results

PCK Machine Type

Percent of Respondents
Reporting this as their
Machine Type

Normalized Percentage of PCK
Machine Types Accounting for
3"' and 4M,/5lh (>eneration Only

1st Generation

1%

—

2nd Generation

6%

—

2nd Generation retrofitted

3%

28%

3rd Generation

23%

4th Generation

28%

72%

5th Generation

33%

Other (non-PCE machines)

6%

—

Total

100%

100%

Source: (Whittaker and Johanson. 20.1.1')

Page 286 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

TableApx J-3. Distribution of Produced Separator Water Emission Factors by Machine Type
Used in Model

Amount of W ater per lb
Clothing W ashed
(gal/lb)

Probability of
Value

0.00320

0.28

0.00367

0.72

J.2.3 Load Size

The CARB (California Air Resources Board. 2006) and King County (Whittaker and J oh an son. 2011)
studies provide machine capacities, and the King County study also provides data on actual size of loads
used by dry cleaners. EPA used the King County study data on actual load sizes to build a distribution.

Table Apx J-4 summarizes the survey results for respondents' primary (if facility has more than one
machine) or only machine. The study reports a maximum reported load of 150 lb, a minimum reported
load of 7 lb, and a median reported load of 30 lb for the primary machine (Whittaker and Johanson.

201 n.

Table Apx J-4. Survey Responses of Actual Pounds Washed per Load for Primary Machine (if
more than one machine) from 2010 King County Survey	

Actual Pounds
of Clothes
Washed

Results for Pri
Number of Respondents

marv Machine
Percent of Respondents

1 - 10

4

3

11-20

36

25

21-30

76

53

31-40

16

11

41-50

6

4

51+

6

4

Total

144

100

Source: (Whittaker and Johanson. 20.1.1')

EPA used these survey results to build a distribution to describe the actual wash loads per machine, as
summarized in Table Apx J-5. To build this distribution, EPA set the following:

•	The maximum, median, and minimum were set as 150 lb, 30 lb, and 7 lb, respectively, as stated
in the King County survey report (Whittaker and Johanson. 2011).

•	The 96th percentile was set at 50 lb as the high-end of the bin "41 to 50 lb". Per Table Apx J-4,
4% of respondents reported greater than 50 lb; therefore, 96% of facilities reported 50 lb or less.

•	The 28th percentile was set at 20 lb as the high-end of the bin "11 to 20 lb". Per Table Apx J-4,
28% of respondents reported 20 lb or less.

Page 287 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

EPA then determined the best-fit distribution using the software @Risk38.

TableApx J-5. Distribution of Actual Load Sizes from 2010 King County Survey

Actual Load
W ashed (lb)

Percentile
(as fraction)

150

1

50

0.96

30

0.5

20

0.28

7

0

Source: (Whittaker and Johanson. 20.1. D

EPA fit a beta distribution to this distribution with parameters: al = 2.3927, a2 = 12.201, min = 7 lb, max
= 150 lb. The root-mean squared (RMS) error is 0.0365. Figure Apx J-2 illustrates this fit.

38 @Risk; Palisade; https://www.palisade.com/risk/

Page 288 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Fit Comparison for Load Size

RiskBetaGeneral(2.3927,12.201,7,150)

5.0%

90.0%

5.0%

0.7%

89.9%

9.4%







/

















h

/























@RISK for Excel
Palisade Corporation









I,

/















i.



















3 C
C

) c

J "3

) c
v£

> c
j a

) c
J c

>	c

>	fN

) C
! m

) o
vC

1 tH

¦

Input 1

BetaGeneral

Minimum

7.00

7.00

Maximum

150.00

150.00

Mean

31.68

30.45

Mode

N/A

22.81

Median

30.00

28.23

Std Dev

19.37

13.41

Skewness

2.5076

0.8640

Kurtosis

13.5139

3.7151

LeftX

9.3

9.3

LeftP

5.0%

0.7%

Right X

49.6

49.6

Right P

95.0%

90.6%

Dif. X

40.24

40.24

Dif. P

90.0%

89.9%

1%

7.46

9.71

5%

9.32

12.70

10%

11.64

15.03

15%

13.96

16.94

20%

16.29

18.66

25%

18.61

20.28

30%

20.91

21.86

35%

23.18

23.42

40%

25.45

24.98

45%

27.73

26.58

50%

30.00

28.23

55%

32.17

29.96

60%

34.35

31.78

65%

36.52

33.75

70%

38.70

35.90

75%

40.87

38.32

80%

43.04

41.11

85%

45.22

44.50

90%

47.39

48.94

95%

49.57

55.81

99%

125.00

69.31

FigureApx J-2. Fit Comparison of Beta Cumulative Distribution Function to Load Size Survey

Results

J.2.4 Number of Loads per Day

EPA used a uniform distribution for the number of loads per day ranging from 1 to 14 based on
observations from NIOSH (2010) and Blando (2010).

J.2.5 Number of Machines per Site

Based on survey data from C ARB (California Air Resources Board. 2006) and Whittaker (2011). the
model assumes dry cleaning shops have between one and three machines.

Table Apx J-6 summarizes the survey results for number of machines per facility from both the 2003
CARB survey (California Air Resources Board. 2006) and the 2010 King County survey (Whittaker and
Johanson. 2011). The results of the two surveys are similar. The CARB survey includes industrial
facilities: <1% of respondents identified as industrial facilities; 96% identified as plant/retail; 3%
identified as other; and a total of <1.5% identified as government, nonprofit, and hotel/motel (California
Air Resources Board. 2006). Since the CARB survey includes industrial facilities as respondents, EPA
used the CARB survey results for number of machines per facility.

Page 289 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Table A

)x J-6. Survey Results of Number of Machines per Facility



Percent of
Respondents in
( AUIJ(

Board, 2006)
Survey"

Percent of
Respondents in King
County (

)

Survey1'

.Number of Machines

Percent of shops with 1 machine

92%

96%

Percent of shops with 2 machines

8%

2.7%

Percent of shops with >2 machines

<1%

1.3%

a The 2003 CARB survey had 1,634 respondents.

b The 2010 King County survey had 151 respondents respond to this survey question.

Source: (Whittaker and Johanson. 20.1.1; California Air Resources Board. 2006)

TableApx J-7 summarizes the distribution of number of machines per facility used in the model based
on the results of the CARB survey. The probabilities were normalized to sum to 100%. A maximum of 3
machines per facility was used based on the maximum number of machines reported in the 2010 King
County Survey (Whittaker and Johanson. 2011).

Table Apx J-7. Distribution of Number of Machines per Facility Used in the Model

Number of
Machines per
l-'acility

Probability of
Value

1

0.91

2

0.08

3

0.01

J.2.6 Operating Days per Year

EPA used a triangular distribution of operating days per year defined as the following:

•	Minimum value: 250 days/yr; consistent with operating 5 days/week and 50 weeks/yr.

•	Mode value: 300 days/yr; consistent with operating 6 days/week and 50 weeks/yr.

•	Maximum value: 312 days/yr; consistent with operating 6 days/week and 52 weeks/yr.

The triangular distribution is of discrete values only, as the number of operating days is a whole number
of days. EPA calculated the probability of each value of operating days using EquationApx J-5:

EquationApx J-5

P(x)

2(x — a)
(b — a) (c — a)

2 (b — x)
(b — a) (b — c)

for a < x < c

for c < x < b

Page 290 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Where:

P(x)

Probability of operating days value
Value of operating days
Minimum value
Maximum value
Mode value

x
a

b

c

Since only discrete values are used in the distribution, the bounds of the distribution are set as 249 and
313 to ensure the sampled minimum and maximum are 250 and 312, respectively. Per Equation Apx
J-5, p(x) is equal to zero when x equals a or b. Therefore, setting a equal to 249 and b equal to 313
ensures 250 and 312 are the true bounds of the sample results.

The assumed distribution of operating days is supported by observed data. The 2003 CARB survey
(California Air Resources Board. 2006) found that, of the 1,634 respondents, 100% of facilities are open
at least Monday through Friday. Approximately 96% of facilities are open on Saturday but closed on
Sunday, and approximately 4% of facilities are open on Sunday but closed on Saturday. Therefore, sites
are not expected to operate fewer than five days per week or greater than six days per week.

Page 291 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Appendix K Dermal Exposure to Volatile Liquids Model Approach
and Parameters

This method was developed through review of relevant literature and consideration of existing exposure
models, such as EPA/OPPT models and the European Centre for Ecotoxicology and Toxicology of
Chemicals Targeted Risk Assessment (ECETOC TRA).

K.1^ Incorporating the Effects of Evaporation

K.1.1 Modification of EPA/OPPT Models

Current EPA dermal models do not incorporate the evaporation of material from the dermis. The dermal
potential dose rate, Dexp (mg/day), is calculated as (	2015b):

EquationApx K-l

Dexp — S X Qu x Yderm x FT

Where:

S is the surface area of contact (cm2)

Qu is the quantity remaining on the skin (mg/cm2-event)

Yderm is the weight fraction of the chemical of interest in the liquid (0 < Yderm < 1)

FT is the frequency of events (integer number per day).

Here Qu does not represent the quantity remaining after evaporation, but represents the quantity
remaining after the bulk liquid has fallen from the hand that cannot be removed by wiping the skin (e.g.,
the film that remains on the skin).

One way to account for evaporation of a volatile solvent would be to add a multiplicative factor to the
EPA/OPPT model to represent the proportion of chemical that remains on the skin after evaporation,/abs

(0 s

Kasting (2006) developed a diffusion model to describe the absorption of volatile compounds applied to
the skin. As of part of the model, Kasting (2006)define a ratio of the liquid evaporation to absorption, %.
They derive the following definition of x (which is dimensionless) at steady-state:

Page 292 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

EquationApx K-3

P MW3A

-3. .0.78 vy

X = 3.4 X 10 u n 7,

j/O./oc
Aoct

Where:

u is the air velocity (m/s)

Koct is the octanol:water partition coefficient

MW is the molecular weight

Sw is the water solubility (|ag/cm3)

Pvp is the vapor pressure (torr)

Chemicals for which %» 1 will largely evaporate from the skin surface, while chemicals for which x
« 1 will be largely absorbed; x = 1 represents a balance between evaporation and absorption.

Equation Apx K-3 is applicable to chemicals having a log octanol/water partition coefficient less than
or equal to three (log Kow < 3)39. The equations that describe the fraction of the initial mass that is
absorbed (or evaporated) are rather complex (Equations 20 and 21 of Kasting (2006) but can be solved.

K.2.1 Small Doses (Case 1: Mo< Msat)

In the small dose scenario, the initial dose (Mo) is less than that required to saturate the upper layers of
the stratum corneum (Mo < Msat), and the chemical is assumed to evaporate from the skin surface at a
rate proportional to its local concentration.

For this scenario, Frasch (2012) calculated the fraction of applied mass that is absorbed, based on the
infinite limit of time (i.e. infinite amount of time available for absorption after exposure):

Equation Apx K-4

, ™abs(co) 2+ fx

Tabs - Mq "2 + 2/

Where:

mabs is the mass absorbed
Mo is the initial mass applied

/is the relative depth of penetration in the stratum corneum (f= 0.1 can be assumed)
X is as previously defined

Note the simple algebraic solution in Equation Apx K-4 provides a theoretical framework for the total
mass that is systemically absorbed after exposure to a small finite dose (mass/area) of chemical, which
depends on the relative rates of evaporation, permeation, and the initial load. At "infinite time", the
applied dose is either absorbed or evaporated (Frasch. 2012). The finite dose is a good model for splash-
type exposure in the workplace (Frasch andBunge. 2015).

The fraction of the applied mass that evaporates is simply the complement of that absorbed:

39 For simplification, Kasting (2006') does not consider the resistance of viable tissue layers underlying the stratum corneum,
and the analysis is applicable to hydrophilic-to-moderately lipophilic chemicals. For small molecules, this limitation is
equivalent to restricting the analysis to compounds where Log Kow < 3.

Page 293 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

EquationApx K-5

mevap(co)	2 x~fX

Mq ~ 'abs ~ 2 + 2/

Where:

mevap is the mass evaporated

The fraction absorbed can also be represented as a function of dimensionless time x (Dt/h2), as shown in
EquationApx K-6:

Equation Apx K-6

00	? 2

, mabs	1	-A 2t^ ( X +K. \ (cosjl-f) ln-CQ5ln\

tabs~M0-	6 n )\x2+Xn2+X) A f'K )

where the eigenvalues An are the positive roots of the equation:

Equation Apx K-7

An ¦ cot (ln) + x = 0

Equation Apx K-6 and Equation Apx K-7 must be solved analytically. It should be noted that the
dimensionless time x is not a representation of exposure duration for a work activity; rather, it represents
the amount of time available for absorption after the initial exposure dose is applied. Since most dermal
risk assessments are typically more concerned with the quantity absorbed, rather than the time course of
absorption, the simple algebraic solution is recommended over the analytical solution.

K.2.2 Large Doses (Case 2: Mo > Msat)

For large doses (Mo > Msat), the chemical saturates the upper layers of the stratum corneum, and any
remaining amount forms a residual layer (or pool) on top of the skin. The pool acts as a reservoir to
replenish the top layers of the membrane as the chemical permeates into the lower layer. In this case,
absorption and evaporation approach steady-state values as the dose is increased, similar to an infinite
dose scenario.

The steady-state fraction absorbed can be approximated by Equation Apx K-8:

Equation Apx K-8

1

fabsi.00) — _j_

Table Apx K-l presents the estimated absorbed fraction calculated using the steady-state approximation
for large doses (Equation Apx K-8) for PCE.

Table Apx K-l. Estimated Fraction Evaporated ant

Absorbed (^abs) using Equation Apx K-8



Chemical Name

I'erchloroelhvlene



Page 294 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

CASRN

127-18-4

Molecular Formula

CiCU

Molecular Weight
(g/mol)

165.833

Pvp (torr)

18.5

Universal gas constant, R

0.0821

(L*atm/K*mol)

Temperature, T (K)

303

Log Kow

3.4

Koct

2511.9

Sw (g/L)

0.206

Sw (|ig/cm3)

206

Industrial Setting

u (m/s)a

0.1674

Evaporative Flux, y

6.95

Fraction Evaporated

0.87

Fraction Absorbed

0.13

Commercial Setting

u (m/s)a

0.0878

Evaporative Flux, y

4.20

Fraction Evaporated

0.81

Fraction Absorbed

0.19

a EPA used air speeds from Baldwin (1998): the 50th percentile of industrial occupational environments of 16.74 cm/s is used
for industrial settings and the 50th percentile of commercial occupational enviromnents of 8.78 cm/s is used for commercial
settings.

K.3 Comparison of/abS to Experimental Values for 1-BP

Sections K.2 presents theoretical frameworks for estimating the fraction of volatile chemical absorbed in
finite dose, infinite dose, and transient exposure scenarios. It is unclear whether these frameworks have
been validated against measured data for the specific chemicals of current OPPT interest. Where
available, experimental studies and actual measurements of absorbed dose are preferred over theoretical
calculations.

In a 2011 study, Frasch (2011 )tested dermal absorption characteristics of 1-BP. For the finite dose
scenario, Frasch (2011) determined that unoccluded exposure resulted in less than 0.2 percent of applied
1-BP dose penetrated the skin - a value substantially lower than the theoretical ~6 percent absorbed
estimated using EquationApx K-8. While this discrepancy is unexplained, the Frasch (201 1)study
recognized the large standard deviation of certain experimental results, and the difficulty of spreading a
small, rapidly evaporating dose of 1-BP evenly over the skin surface. Frasch (2011) also raised the
possibility that 1-BP may dehydrate the stratum corneum, thereby decreasing the skin permeability after
initial exposure.

Page 295 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

K.4 Potential for Occlusion

Gloves can prevent the evaporation of volatile chemicals from the skin, resulting in occlusion.

Chemicals trapped in the glove may be broadly distributed over the skin (increasing S in EquationApx
K-l), or if not distributed within the glove, the chemical mass concentration on the skin at the site of
contamination may be maintained for prolonged periods of time (increasing Qu in Equation Apx K-l).
Conceptually, occlusion is similar to the "infinite dose" study design used in in vitro and ex vivo dermal
penetration studies, in which the dermis is exposed to a large, continuous reservoir of chemical.
The impact of occlusion on dermal uptake is complex: continuous contact with the chemical may
degrade skin tissues, increasing the rate of uptake, but continuous contact may also saturate the skin,
slowing uptake (Danelk et at.. ). These phenomena are dependent upon the chemical, the vehicle
and environmental conditions. It is probably not feasible to incorporate these sources of variability in a
screening-level population model of dermal exposure without chemical-specific studies.

Existing EPA/OPPT dermal models (Equation Apx K-l) could theoretically be modified to account for
the increased surface area and/or increased chemical mass in the glove. This could be achieved through
a multiplicative variable (such as used in Equation Apx K-2 to account for evaporative loss) or a change
in the default values of S and/or Qu. It may be reasonable to assume that the surface area of hand in
contact with the chemical, S, is the area of the whole hand owing to the distribution of chemical within
the glove. Since Qu reflects the film that remains on the skin (and cannot be wiped off), a larger value
should be used to reflect that the liquid volume is trapped in the glove, rather than falling from the hand.
Alternatively, the product S x Qu (cm2 x mg/cm2-event) could be replaced by a single variable
representing the mass of chemical that deposits inside the glove per event, M (mg/event):

Equation Apx K-9

Dexp — My. Yderm x FT

Garrod (2001) surveyed contamination by in volatile components of non-agricultural pesticide products
inside gloves across different job tasks and found that protective gloves were nearly always
contaminated inside. While the study does not describe the exact mechanism in which the contamination
occurs (e.g. via the cuff, permeation, or penetration through imperfections in glove materials), it
quantified inner glove exposure as "amount of product per unit time", with a median value of 1.36 mg
product per minute, a 75th percentile value of 4.21 mg/min, and a 95th percentile value of 71.9 mg/min. It
is possible to use these values to calculate the value of M, i.e. mass of chemical that deposits inside the
glove, if the work activity duration is known.

Assuming an activity duration of one hour, the 50th and 95th percentile values translate to 81.6 mg and
4,314 mg of inner glove exposure. While these values may be used as default for M in Equation Apx
K-9, EPA notes the significant difference between the 50th and 95th percentile deposition, with the 95th
percentile value being two times more conservative than the defaults for the EPA/OPPT 2-Hand Dermal
Exposure Model (where the product S x Qu is 2,247 mg/event). Given the significant variability in inner
glove exposure and lack of information on the specific mechanism in which the inner glove
contamination occurs, EPA addresses the occlusion scenario in combination with other glove
contamination and permeation factors through the use of a protection factor, as described in the next
section.

Page 296 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

EPA does not expect occlusion scenarios to be a reasonable occurrence for all conditions of use.
Specifically, occlusion is not expected at sites using chemicals in closed systems where the only
potential for dermal exposure is during the connecting/disconnecting of hoses used for
unloading/loading of bulk containers (e.g., tank trucks or rail cars) or while collecting quality control
samples including manufacturing sites, repackaging sites, sites processing the chemical as a reactant,
formulation sites, and other similar industrial sites. Occlusion is also not expected to occur at highly
controlled sites, such as electronics and pharmaceuticals manufacturing sites, where, due to purity
requirements, the use of engineering controls is expected to limit potential dermal exposures. EPA also
does not expect occlusion at sites where contact with bulk liquid chemical is not expected such as
aerosol degreasing sites where workers are only expected to handle the aerosol cans containing the
chemical and not the actual bulk liquid chemical.

EPA expects occlusion to be a reasonable occurrence at sites where workers may come in contact with
bulk liquid chemical and handle the chemical in open systems. This includes conditions of use such as
vapor degreasing, cold cleaning, and dry cleaning where workers are expected to handle bulk chemical
during cleanout of spent solvent and addition of fresh solvent to equipment. Similarly, occlusion may
occur at coating or adhesive application sites when workers replenish application equipment with liquid
coatings or adhesives.

K.5 Incorporating Glove Protection

Data about the frequency of effective glove use - that is, the proper use of effective gloves - is very
limited in industrial settings. Initial literature review suggests that there is unlikely to be sufficient data
to justify a specific probability distribution for effective glove use for a chemical or industry. Instead,
the impact of effective glove use should be explored by considering different percentages of
effectiveness (e.g., 25% vs. 50% effectiveness).

Gloves only offer barrier protection until the chemical breaks through the glove material. Using a
conceptual model, Cherrie (Cherrie et at.. 2004) proposed a glove workplace protection factor - the ratio
of estimated uptake through the hands without gloves to the estimated uptake though the hands while
wearing gloves: this protection factor is driven by flux, and thus varies with time. The ECETOC TRA
model represents the protection factor of gloves as a fixed, assigned protection factor equal to 5, 10, or
20 (Marquart et	). Where, similar to the APR for respiratory protection, the inverse of the

protection factor is the fraction of the chemical that penetrates the glove.

The protection afforded by gloves can be incorporated into the EPA/OPPT model (EquationApx K-l)
by modification of Qu with a protection factor, PF (unitless, PF > 1):

Equation Apx K-10

Dexp = S x x Yderm x FT

Given the limited state of knowledge about the protection afforded by gloves in the workplace, it is
reasonable to utilize the PF values of the ECETOC TRA model (Marquart et at.. ), rather than
attempt to derive new values. Table Apx K-2 presents the PF values from ECETOC TRA model

Page 297 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

(version 3). In the exposure data used to evaluate the ECETOC TRA model, Marquart (2017) reported
that the observed glove protection factor was 34, compared to PF values of 5 or 10 used in the model.

TableApx K-2. Exposure Control Efficiencies and Protection Factors for Different Dermal
Protection Strategies from ECETOC TRA v3	

Dermal Protection Characteristics

Affected I ser
(¦roup

Indicated
Efficiency
(%)

Protection
l-'actor.
PI-

a. Any glove / gauntlet without permeation data and
without employee training

Both industrial and
professional users

0

1

b. Gloves with available permeation data indicating that
the material of construction offers good protection for the
substance

80

5

c. Chemically resistant gloves (i.e., as b above) with
"basic" employee training

90

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 users
only

95

20

K.6 Proposed Dermal Dose Equation

Accounting for all parameters above, the proposed, overall equation for estimating dermal exposure is:
EquationApx K-ll

( Qu xfabs)

n 	 C v	uusj y	j-,j,

uexp	pp	^ 1derm ^ 1 1

EPA presents exposure estimates for the following deterministic dermal exposure scenarios:

•	Dermal exposure without the use of protective gloves (Equation Apx K-l 1, PF = 1)

•	Dermal exposure with the use of protective gloves (Equation Apx K-l 1, PF = 5)

•	Dermal exposure with the use of protective gloves and employee training (EquationApx K-ll,
PF = 20 for industrial users and PF = 10 for professional users)

•	Dermal exposure with occlusion (Equation Apx K-9)

EPA assumes the following parameter values for EquationApx K-ll in addition to the parameter
values presented in Table Apx K-l:

•	S, the surface area of contact: 535 cm2 (central tendency) and 1,070 cm2 (high-end), representing
the total surface area of both hands.

Page 298 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

•	Qu, the quantity remaining on the skin: 1.4 mg/cm2-event (central tendency) and 2.1 mg/cm2-
event (high-end). These are the midpoint value and high-end of range default value, respectively,
used in the EPA/OPPT dermal contact with liquids models (	3).

•	Yderm, the weight fraction of the chemical of interest in the liquid: EPA will assess a unique value
of this parameter for each occupational scenario or group of similar occupational scenarios.

•	FT, the frequency of events: 1 event per day. EquationApx K-l 1 shows a linear relationship
between FT and Dexp; however, this fails to account for time between contact events. Since the
chemical simultaneously evaporates from and absorbs into the skin, the dermal exposure is a
function of both the number of contact events per day and the time between contact events. EPA
did not identify information on how many contact events may occur and the time between
contact events. Therefore, EPA assumes a single contact event per day for estimating dermal
exposures.

For Equation Apx K-9, EPA assumes the quantity of liquid occluded underneath the glove (M) is equal
to the product of the entire surface area of contact (S = 1,070 cm2) and the assumed quantity of liquid
remaining on the skin (Qu = 2.1 mg/cm2-event), which is equal to 2,247 mg/event. See discussion in
Section K.4.

K.7 Equations for Calculating Acute and Chronic (Non-Cancer and
Cancer) Dermal Dose

Equation Apx K-l 1 estimates dermal potential dose rates (mg/day) to workers in occupational settings.
The potential dose rates are then used to calculate acute retained doses (ARD), and chronic retained
doses (CRD) for non-cancer and cancer risks.

Acute retained doses are calculated using Equation Apx K-12.

Equation Apx K-12

Where:

ARD	= acute retained dose (mg/kg-day)

Dexp	= dermal potential dose rate (mg/kg)

BW	= body weight (kg)

CRD is used to estimate exposures for non-cancer and cancer risks. CRD is calculated as follows:

Equation Apx K-13

Dexv x EF x WY
CRD = p

BW x (AT or ATC)
Equation Apx K-l4

day

AT = WY x 365——
yr

Page 299 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

EquationApx K-15

day

ATC = LTx 365

yr

Where:

CRD	= Chronic retained dose used for chronic non-cancer or cancer risk calculations

EF	= Exposure frequency (day/yr)

WY	= Working years per lifetime (yr)

AT	= Averaging time (day) for chronic, non-cancer risk

ATc	= Averaging time (day) for cancer risk

LT	= Lifetime years (yr) for cancer risk

Table Apx K-3 summarizes the default parameter values used to calculate each of the above acute or
chronic exposure estimates. Where multiple values are provided for EF, it indicates that EPA may have
used different values for different conditions of use. The rationales for these differences are described
below in this section.

Table_Apx K-3. Parameter Values for Calculating Dermal Dose Estimates

Parameter Name

Sv m hoi

Value

1 nil

Exposure Frequency

EF

250

258 (50th percentile) to 293 (95th
percentile) (dry cleaning only)
125 to 150 (DoD - oil analysis only)
30 to 36 (DoD - water pipe repair only)

days/yr

Working years

WY

31 (50th percentile)
40 (95th percentile)

years

Lifetime Years, cancer

LT

78

years

Body Weight

BW

80

kg

Averaging Time, non-
cancer

AT

11,315 (central tendency)51
14,600 (high-end)b

day

Averaging Time, cancer

ATC

28,470

day

a Calculated using the 50th percentile value for working years (WY)
b Calculated using the 95th percentile value for working years (WY)

Page 300 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Exposure Frequency (EF)

EPA generally uses an exposure frequency of 250 days per year with two notable exceptions: dry
cleaning and DoD uses. EPA assumed dry cleaners may operate between five and six days per week and
50 to 52 weeks per year resulting in a range of 250 to 312 annual working days per year (AWD). Taking
into account fractional days exposed (f) resulted in an exposure frequency (EF) of 258 at the 50th
percentile and 293 at the 95th percentile. For the two DoD uses, information was provided indicating
process frequencies of two to three times per week (oil analysis) and two to three times per month (water
pipe repair). EPA used the maximum frequency for high-end estimates and the midpoint frequency for
central tendency estimates. For the oil analysis use this resulted in 125 days/yr at the central tendency
and 150 days/yr at the high-end. For the water pipe repair, this resulted in 30 days/yr at the central
tendency and 36 days/yr at the high-end.

EF is expressed as the number of days per year a worker is exposed to the chemical being assessed. In
some cases, it may be reasonable to assume a worker is exposed to the chemical on each working day. In
other cases, it may be more appropriate to estimate a worker's exposure to the chemical occurs during a
subset of the worker's annual working days. The relationship between exposure frequency and annual
working days can be described mathematically as follows:

EquationApx K-16

EF = fx AWD

Where:

EF = exposure frequency, the number of days per year a worker is exposed to the chemical
(day/yr)

f = fractional number of annual working days during which a worker is exposed to the
chemical (unitless)

AWD = annual working days, the number of days per year a worker works (day/yr)

U.S. BLS (2016) provides data on the total number of hours worked and total number of employees by
each industry NAICS code. These data are available from the 3- to 6-digit NAICS level (where 3-digit
NAICS are less granular and 6-digit NAICS are the most granular). Dividing the total, annual hours
worked by the number of employees yields the average number of hours worked per employee per year
for each NAICS.

EPA has identified approximately 140 NAICS codes applicable to the multiple conditions of use for the
ten chemicals undergoing risk evaluation. For each NAICS code of interest, EPA looked up the average
hours worked per employee per year at the most granular NAICS level available (i.e., 4-digit, 5-digit, or
6-digit). EPA converted the working hours per employee to working days per year per employee
assuming employees work an average of eight hours per day. The average number of days per year
worked, or AWD, ranges from 169 to 282 days per year, with a 50th percentile value of 250 days per
year. EPA repeated this analysis for all NAICS codes at the 4-digit level. The average AWD for all 4-
digit NAICS codes ranges from 111 to 282 days per year, with a 50th percentile value of 228 days per
year. 250 days per year is approximately the 75th percentile. In the absence of industry- and PCE-
specific data, EPA assumes the parameter/is equal to one for all conditions of use except dry cleaning.
Dry cleaning used a uniform distribution from 0.8 to 1 for f. The 0.8 value was derived from the

Page 301 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

observation that the weighted average of 200 day/yr worked (from BLS/Census) is 80% of the standard
assumption that a full-time worker works 250 day/yr. The maximum of 1 is appropriate as dry cleaners
may be family owned and operated and some workers may work as much as every operating day.

Working Years (WY)

EPA has developed a triangular distribution for working years. EPA has defined the parameters of the
triangular distribution as follows:

•	Minimum value: BLS CPS tenure data with current employer as a low-end estimate of the
number of lifetime working years: 10.4 years;

•	Mode value: The 50th percentile tenure data with all employers from SIPP as a mode value for
the number of lifetime working years: 36 years; and

•	Maximum value: The maximum average tenure data with all employers from SIPP as a high-end
estimate on the number of lifetime working years: 44 years.

This triangular distribution has a 50th percentile value of 31 years and a 95th percentile value of 40 years.
EPA uses these values for central tendency and high-end ADC and LADC calculations, respectively.

The U.S. BLS (2014) provides information on employee tenure with current employer obtained from the
Current Population Survey (CPS). CPS is a monthly sample survey of about 60,000 households that
provides information on the labor force status of the civilian non-institutional population age 16 and
over; CPS data are released every two years. The data are available by demographics and by generic
industry sectors but are not available by NAICS codes.

The U.S. Census Bureau (2019a) Survey of Income and Program Participation (SIPP) provides
information on lifetime tenure with all employers. SIPP is a household survey that collects data on
income, labor force participation, social program participation and eligibility, and general demographic
characteristics through a continuous series of national panel surveys of between 14,000 and 52,000
households (U.S. Census Bureau. 2019b). EPA analyzed the 2008 SIPP Panel Wave 1, a panel that
began in 2008 and covers the interview months of September 2008 through December 2008 (U.S.

Census Bureau. .Or \i, b). For this panel, lifetime tenure data are available by Census Industry Codes,
which can be cross-walked with NAICS codes.

SIPP data include fields for the industry in which each surveyed, employed individual works
(TJBIND1), worker age (TAGE), and years of work experience with all employers over the surveyed
individual's lifetime.40 Census household surveys use different industry codes than the NAICS codes
used in its firm surveys, so these were converted to NAICS using a published crosswalk (t; S Census
Bureau.: ). EPA calculated the average tenure for the following age groups: 1) workers age 50 and
older; 2) workers age 60 and older; and 3) workers of all ages employed at time of survey. EPA used

40 To calculate the number of years of work experience EPA took the difference between the year first worked
(TMAKMNYR) and the current data year (i.e., 2008). EPA then subtracted any intervening months when not working
(ETIMEOFF).

Page 302 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

tenure data for age group "50 and older" to determine the high-end lifetime working years, because the
sample size in this age group is often substantially higher than the sample size for age group "60 and
older". For some industries, the number of workers surveyed, or the sample size, was too small to
provide a reliable representation of the worker tenure in that industry. Therefore, EPA excluded data
where the sample size is less than five from our analysis.

TableApx K-4 summarizes the average tenure for workers age 50 and older from SIPP data. Although
the tenure may differ for any given industry sector, there is no significant variability between the 50th
and 95th percentile values of average tenure across manufacturing and non-manufacturing sectors.

Table Apx K-4. Overview of Average Worker Tenure from U.S. Census SIPP (Age Group 50+)

Industry Sectors

Average

Workii
50,h Percentile

lg Years
95"' Percentile

Maxim inn

All industry sectors relevant to the 10
chemicals undergoing risk evaluation

35.9

36

39

44

Manufacturing sectors (NAICS 31-33)

35.7

36

39

40

Non-manufacturing sectors (NAICS 42-81)

36.1

36

39

44

Source: (U.S. Census Bureau. 2019a")

Note: Industries where sample size is less than five are excluded from this analysis.

BLS CPS data provides the median years of tenure that wage and salary workers had been with their
current employer. Table Apx K-5 presents CPS data for all demographics (men and women) by age
group from 2008 to 2012. To estimate the low-end value on number of working years, EPA uses the
most recent (2014) CPS data for workers age 55 to 64 years, which indicates a median tenure of 10.4
years with their current employer. The use of this low-end value represents a scenario where workers are
only exposed to the chemical of interest for a portion of their lifetime working years, as they may
change jobs or move from one industry to another throughout their career.

Page 303 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Table Apx K-5. Median Years of Tenure with Current Employer by Age Group

Age

January 2008

January 2010

January 2012

January 2014

16 years and over

4.1

4.4

4.6

4.6

16 to 17 years

0.7

0.7

0.7

0.7

18 to 19 years

0.8

1.0

0.8

0.8

20 to 24 years

1.3

1.5

1.3

1.3

25 years and over

5.1

5.2

5.4

5.5

25 to 34 years

2.7

3.1

3.2

3.0

35 to 44 years

4.9

5.1

5.3

5.2

45 to 54 years

7.6

7.8

7.8

7.9

55 to 64 years

9.9

10.0

10.3

10.4

65 years and over

10.2

9.9

10.3

10.3

Source: (U.S. BLS. 20.1.4')

Lifetime Years (LT)

EPA assumes a lifetime of 78 years for all worker demographics.

Body Weight (BW)

EPA assumes a body weight of 80 kg for all worker demographics.

AC Products. (2017). Maskants and their use in aerospace: Regulatory compliance of the industry.
(EPA-HQ-OPPT-2016-0732-0077). Washington, D.C.: AC Products.
https://www.regulations.gov/document?D=EPA~HQ~OPPT~~2Q16~0732~0077

Aerospace Industries Association. (2017). Re: Posting EPA-HQ-OPPT-2016-0732 tetrachloroethylene
(perchloroethylene) (CASRN 127-18-4). (EPA-HQ-OPPT-2016-0732-0015). Washington, D.C.:
Aerospace Industries Association (AIA). https://www.regulations.gov/document?D=EPA-HQ-
QPPT-2016-0732-0015

AIHA. (2009). Mathematical models for estimating occupational exposure to chemicals. In CB Keil; CE
Simmons; TR Anthony (Eds.), (2nd ed.). Fairfax, VA: AIHA Press. https://online-

aiha.org/amsssa/ecssashop.show product detail	Je=detail&p product serno=889

American Fuel and Petroleum Manufacturers. (2017). Re: EPA's "Risk evaluation scoping under TSCA
for ten chemical substances; Reopening of comment period; Tetrachloroethylene (also known
as perchloroethylene)". (EPA-HQ-OPPT-2016-0732-0018). Washington, D.C.: American Fuel &
Petroleum Manufacturers (AFPM). https://www.regulations.gov/document?D=EPA-HQ-OPPT-

Page 304 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

ATSDR. (2014). Toxicological profile for tetrachloroethylene (Draft for public comment). Atlanta, GA:
US Department of Health and Human Services, Public Health Service.

http://www.atsdr.cdc.gov/ToxProfiles/tp.asp?id=265&tid=48
Baldv in iH Mjynard, AD. (1998). A survey of wind speed in indoor workplaces. Ann Occup Hyg 42:

303-313. http://dx.doi.org/10.1016/50003-4878(98)00031-3
Blando, JD; Schill, DP: De La Cruz. MP: Zhang. L; Zhane. J. (2010). Preliminary study of propyl bromide
exposure among New Jersey dry cleaners as a result of a pending ban on perchloroethylene. J
Air Waste Manag Assoc 60: 1049-1056. http://dx.doi.ore/10.3155/1047-3289.60.9.1049
Brodkin, CA; Daniell. W; Checkoway, H; Echeven so i * 'ohnson, J: Wang. K; Sohaey, R; Green. D;
Redlich, C; Gretch. D. (1995). Hepatic ultrasonic changes in workers exposed to
perchloroethylene. Occup Environ Med 52: 679-685.

California Air Resources Board. (2006). California Dry Cleaning Industry Technical Assessment Report.
Stationary Source Division, Emissions Assessment Branch.
https://www.arb.ca.gov/toxics/dryclean/finaldrvcleantechreport.pdf
CARB. (2000). Initial statement of reasons for the proposed airborne toxic control measure for

emissions of chlorinated toxic air contaminants from automotive maintenance and repair
activities.

Cherrie, JW; Semple, S; Brouwer. D. (2004). Gloves and Dermal Exposure to Chemicals: Proposals for
Evaluating Workplace Effectiveness. Ann Occup Hyg 48: 607-615.
http://dx.doi.org/10.1093/annhyg/meh060
Cosgrove, H; Hygiene. I. (1994). Perchloroethylene Survey, Radiator Specialty Company. (EPA-HQ-
OPPT-2016-0732-0027). Charlotte, NC: Cosgrove Health & Hygiene Inc.
https://www.regulations.gov/document?D=EPA-HQ-OPPT-2016-0732-0027
Dancik. Y; Bigliardi, PL: Bigliardi-Qi, Me. (2015). What happens in the skin? Integrating skin permeation
kinetics into studies of developmental and reproductive toxicity following topical exposure.
Re prod Toxicol 58: 252-281. http://dx.doi.Org/10.1016/i.reprotox.2015.10.001
Defense Occupational and Environmental Health Readiness System - Industrial Hygiene. (2018). Email
between DOD and EPA: RE: [Non-DoD Source] Update: DoD exposure data for EPA risk
evaluation - EPA request for additional information [Personal Communication]. Washington,
D.C.: U.S. Department of Defense.

Demou, E; Hellweg, S: Wilson. MP: Hammond. SK; McKone, TE. (2009). Evaluating indoor exposure
modeling alternatives for LCA: A case study in the vehicle repair industry. Environ Sci Technol
43: 5804-5810. http://dx.doi.org/10.1021/es803551v
DLI; NCA. (2017). Public comment on tetrachloroethylene. TSCA review and scoping. (EPA-HQ-OPPT-

2016-0732). https://www.regulations.gov/document?D=EPA-HQ-OPPT-2016-0732-0007
Dow Che (1979). Evaluation of work exposures in ag production and distribution department
(apd2) operations, pittsburg, for 1978 with cover letter. (OTS: OTS0206690; 8EHQ Num: NA;
DCN: 878214806; TSCATS RefID: 25878; CIS: NA). Dow Chem Co.

Dow ChemCo. (1982). CHLOR-PYRIDINES -1981 INDUSTRIAL HYGIENE SURVEY (SANITIZED). (OTS:

OTS0515873; 8EHQ Num: NA; DCN: 86-870002349; TSCATS RefID: 309318; CIS: NA).
Dow ChemCo. (1983a). 1982 INDUSTRIAL HYGIENE MONITORING - CHLOROPYRIDINES (SANITIZED).
(OTS: OTS0515889; 8EHQ Num: NA; DCN: 86-870002365; TSCATS RefID: 309350; CIS: NA).

Page 305 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Dow Chem Co, (1983b). Chemical exposure evaluation - Trichloroethylene production plant (sanitized).

(EPA/OTS; Doc #86-870002355). Dow Chem Co.

Dow Chem Co. (1983c). RESULTS OF 1982 PERSONNEL MONITORING - PERCHLOROETHYLENE/CARBON
TETRACHLORIDE PRODUCTION AREA. (OTS: OTS0515871; 8EHQ Num: NA; DCN: 86-870002347;
TSCATS RefID: 309314; CIS: NA).

Dow Chem Co. (1984). INDUSTRIAL HYGIENE SURVEYS DURING 1983 AT THE EASTERN DIVISION

MARINE TERMINAL AT JOLIET, ILLINOIS (SANITIZED). (OTS: OTS0515882; 8EHQ Num: NA; DCN:
86-870002358; TSCATS RefID: 309336; CIS: NA).

Dow Chemical. (1983). 1982 industrial hygiene survey environmental operations department [TSCA
Submission], (OTS0515872. 86-870002348. TSCATS/309316).
https://ntrl.ntis.eov/NTRL/dashboard/searchResults/titleDetail/OTSQ515872.xhtml
Dow Chemical. (2008). Product safety assessment: Perchloroethylene. http://N/A

(2017). HSIA Support Ducommun. (EPA-HQ-OPPT-2016-0732-0027). Washington, D.C.:
Ducommun Inc. https://www. regulations.gov/document?D=EPA~HQ~OPPT-2016~0732~0027
Durkee, J. (2014). Cleaning with solvents: Methods and machinery. Oxford, UK: Elsevier Inc.

https://www.sciencedirect.com/book/9780323225205/cleaning~with~solvents-methods-and~
machinery

Ecb. (2005). European Union risk assessment report: Tetrachloroethylene. Part 1 - Environment.
(EINECS No: 204-825-9). United Kingdom: European Commission - Joint Research Centre
Institute for Health and Consumer Protection European Chemicals Bureau.

Eisenberg. J: Ramsey, J. (2010). Health hazard evaluation report no. HETA 2008-0175-3111, Evaluation
of 1-Bromopropane use in four New Jersey commercial dry cleaning facilities. (HETA 2008-0175-
3111). Cincinnati, OH: National Institute for Occupational Safety and Health.

EPA. US: Consulting. I. (2004). The U.S. solvent cleaning industry and the transition to non ozone
depleting substances, https://www.epa.gov/sites/production/files/2014~

11/docume nts/epasolventma rketreport.pdf
ERG. (2005). [Letter from Eric Goehl and Jennifer O'Neil, Eastern Research group, Inc, to Dry Cleaning
Docket, Subject: Background information document] [Personal Communication].
http://www3.epa.gov/airtoxics/dryperc/ll-14-05background.pdf
European Chlorinated Solvents Association. (2011). Health profile on perchloroethylene.

http://www.chlorinated-solvents.eu/index.php/sustainability/hazard~risk-assessments/health~
profiles

European Solvents Industry Group. (2012). SPERC fact sheet: Manufacture of substance - industrial
(solvent-borne). Brussels, Belgium: European Solvents Industry Group (ESIG).
https://www.esig.org/reach~ges/environment/

European Solvents Industry Group. (2019a). Industrial - solvent-borne (formulation and (re)packaging

of substances and mixtures), https://www.esig.org/reach~ges/environment/

European solvents Industry Group. (2019b). SPERC factsheet - Laboratory Reagents - Professional
(solvent-borne). Brussels, Belgium: European solvents Industry Group.
https://www.esig.org/reach~ges/environment/

Ford Motor Co. (1981). Industrial hygiene survey - spray booths, oil house, roll weld, bonderite deck,
trimline. (OTS: OTS0206239; 8EHQ Num: NA; DCN: 878210810; TSCATS RefID: 17580; CIS: NA).

Page 306 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Frasch, FH, (2012). Dermal Absorption of Finite doses of Volatile Compounds. J Pharm Sci 101: 2616-

2619. http://dx.doi.org/10.1080/15287394
Frasch. HF; Bunge, AL. (2015). The transient dermal exposure II: post-exposure absorption and
evaporation of volatile compounds. J Pharm Sci 104:1499-1507.
http://dx.doi.org/10.1002/ips.24334
Frasch. HF: Dotson, GS: Barbero, AM. (2011). In vitro human epidermal penetration of 1-
bromopropane. J Toxicol Environ Health A 74:1249-1260.
http://dx.doi.ore/10.1080/15287394.2011.595666
Garrod, AN: Phillips. AM: Pemberton, JA. (2001). Potential exposure of hands inside protective gloves—
a summary of data from non-agricultural pesticide surveys. Ann Occup Hyg 45: 55-60.
http://dx.doi.ore/10.1016/S0003~4878(00)00013~2
Gilbert. D; Goyer, M: Lyman, W; Magif, G; Walker. P; Wallace. D; Wechsler, A: Yee, J. (1982). An

exposure and risk assessment for tetrachloroethylene. (EPA-440/4-85-015). Washington, DC:
U.S. Environmental Protection Agency, Office of Water Regulations and Standards.
http://nepis.epa.gov/Exe/ZyPURL.cgi?Dockey=2000LLQH.txt
Gold. LS: De Roos. AJ; Waters. M: Stewart. P. (2008). Systematic literature review of uses and levels of
occupational exposure to tetrachloroethylene [Review]. J Occup Environ Hyg 5: 807-839.
http://dx.doi.ore/10.1080/15459620802510866
Golsteijn. L; Huizer. D; Hauck. M: van Zelm, R: Huiibregts, MA. (2014). Including exposure variability in
the life cycle impact assessment of indoor chemical emissions: the case of metal degreasing.
Environ Int 71: 36-45. http://dx.doi.Org/10.1016/i.envint.2014.06.003
liec. JP; Wesolowski, W; Brzeznicki, S: Wroblewska-Jakubowska, K; Kucharska, M. (2002).

Occupational exposure to rubber vulcanization products during repair of rubber conveyor belts
in a brown coal mine. J Environ Monit 4:1054-1059. http://dx.doi.ore/10.1039/b209207g
Hellweg, S: Demou, E; Bruzzi, R; Meiier, A: Rosenbaum, RK; Huiibregts. MA: McKone, TE. (2009).

Integrating human indoor air pollutant exposure within Life Cycle Impact Assessment [Review].
Environ SciTechnol 43: 1670-1679. http://dx.doi.org/lQ.1021/es< OK I <.

Honeywell. (2017). [Personal communication between H. Jason Ay res, Honeywell and Tyler Lloyd, U.S.

EPA, regarding tetrachloroethylene] [Personal Communication].

HSIA. (2008). Perchloroethylene - White Paper. Arlington, VA.

http://www.nttworldwide.com/docs/percwp2008.pdf
HSIA. (2017). Public comment on tetrachloroethylene. TSCA Review and Scoping. (EPA-HQ-OPPT-2016-

0732-0027). https://www.regulations.gov/document?D=EPA~HQ~OPPT~2016~0732~0027
HSIA. (2018). Comment letter of Halogenated Solvents Industry Alliance, Inc. (HSIA) regarding Docket
ID: EPA-HQ-OPPT-2016-0732-0097 [Personal Communication].
https://www. regu lations.gov/docume nt?D=EPA~HQ~OPPT~2Q16~0732~0097
Hydrocarbon Online. (1999). Geon, Oxychem Complete Vinyl Joint Venture.

https://www.hyd rocarbononline.com/doc/geon~oxvchem~complete~vinyl~ioint~venture~0001
International Association for Soaps Detergents and Maintenance Products. (2012). AISE SPERC fact

sheet - wide dispersive use of cleaning and maintenance products. International Association for
Soaps Detergents and Maintenance Products, https://www.aise.eu/our~activities/regulatorv-
context/reach/environmental-exposure-assessment.aspx

Page 307 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

1RTA, (2007). Spotting chemicals: Alternatives to perchloroethylene and trichloroethylene in the textile
cleaning industry. Prepared for: Cal/EPA's Department of Toxic Substances Control and U.S.
Environmental Protection Agency Region IX.

http://www.irta.us/reports/DTSC%20Spottine%20Chemical%20for%20Web.pdf
Kanegsbere. B: Kaneesbere. E. (2011). Handbook for critical cleaning, cleaning agents and systems (2nd

ed.). Boca Raton, FL: CRC Press.

Kastine. BG: Miller, MA. (2006). Kinetics of finite dose absorption through skin 2: Volatile compounds. J

Pharm Sci 95: 268-280. http://dx.doi.ore/10-1002/ips.20497
Klein, P; Kurz, J. (1994). [Reduction of Solvent Concentrations in Surroundings of Dry-Cleaning Shops].

Bonningheim, Germany: Hohenstein Physiological Institute on Clothing.

Ludwie, HR; Meister, MV: Roberts. PR: Cox. C. (1983). Worker exposure to perchloroethylene in the
commercial dry cleaning industry. Am Ind Hyg Assoc J 44: 600-605.
http://dx.doi.ore/10.1080/15298668391405391
Marauart, H; Franken, R: Goede, H; Fransman, W; Schinkel, J. (2017). Validation of the dermal
exposure model in ECETOCTRA. Ann Work Expo Health 61: 854-871.
http://dx.doi.ore/10.1093/annweh/wxx059
Massachusetts DEP. (2015). Dry Cleaners Environmental Certification Workbook. Massachusetts, USA:

Massachusetts DEP. https://www.mass.eov/files/documents/2016/08/xh/dcwkbk.pdf
Materna, BL. (1985). Occupational exposure to perchloroethylene in the dry cleaning industry. AIHA J
46: 268-273.

Morford, RG. (2017). Comment submitted by Richard G. Morford, General Counsel, Enviro Tech

International, Inc. https://www.reeulations.eov/document?D=EPA-HQ-OPPT-2C
Morris. M: Wolf. K. (2005). Evaluation of New and Emerging Technologies forTextile Cleaning. Institute
for Research and Technical Assistance (IRTA).

http://wsppn.ore/pdf/irta/Emereine Technoloeies Textile %20Clea,pdf
MSG Industrial Supply Inc. (2019). AlumTap: 1 Pt Can Cutting & Tapping Fluid.

https://www, mscdirect. com/prod uct/details/00265025?rltem=00265025
Newmoa. (2001). Pollution prevention technology profile - Closed loop vapor degreasing. Boston, MA.

http://www.newmoa.ore/prevention/p2tech/ProfileVaporDeereasine.pdf
NIQSH. (1977). Health Hazard Evaluation Determination, Report No. HHE-77-63-449, McDonnell

Aircraft Company, St. Louis, Missouri (pp. 77-63). (NIOSH/00076128). Hervin, RL; Stroman, R;
Belanger, P; Ruhe, R; Collins, C; Dyches, T.

NIQSH. (1979). Health Hazard Evaluation Determination Report No. HHE-78-95-596, Jonas Brothers

Taxidermy Co., Denver, Colorado (pp. 78-95). (NIOSH/00091563). Gunter, BJ; Lybarger, JA.
NIQSH. (1980a). Health hazard evaluation report no. HHE 79-42-685, Motion Picture Screen

Cartoonists, Local 841, New York, New York. (HHE 79-42-685). Cincinnati, OH: National Institute
for Occupational Safety and Health.

NIQSH. (1980b). Health Hazard Evaluation report no. HHE 80-18-691, Looart Press Incorporate,

Colorado Springs,Colorado. (HHE 80-18-691). Washington, DC: U.S. Department of Health and
Human Services.

NIQSH. (1981a). Health Hazard Evaluation Report 80-154-1027: Bechtel Power Corporation. (HHE 80-
154-1027). NIOSH. https://www.cdc.eov/niosh/hhe/reports/pdfs/80~154~

1027.pdf? id=10.26616/NIQSHHHE801541027

Page 308 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

NIOSH. (1981b). Health Hazard Evaluation Report 80-201-816: Peterson/Puritan Company. (HE 80-201-
816). NIOSH. https://www.cdc.gov/niosh/hhe/reports/pdfs/80~201~
816.pdf?id=10.26616/NIQSHHHE80201816

NIOSH. (1981c). Health hazard evaluation report no. HETA 81-105-831, Labels West, Inc., Redmond,

Washington. (HETA 81-105-831). Cincinnati, OH: National Institute for Occupational Safety and
Health.

NIOSH. (1982a). Health hazard evaluation report no. HETA 81-310-1039, King-Smith Printing Company,
Detroit, Michigan. (HETA 81-310-1039). Cincinnati, OH: National Institute for Occupational
Safety and Health.

NIOSH. (1982b). Health hazard evaluation report no. HETA 82-040-119, Synthes Ltd. (USA), Monument,
Colorado. (HETA 82-040-119). Cincinnati, OH: National Institute for Occupational Safety and
Health.

NIOSH. (1983a). Health Hazard Evaluation Report No. HETA-83-266-1391, McCourt Label Company,
Bradford, Pennsylvania. (NIOSH/00137711). NIOSH.

NIOSH. (1983b). Health Hazard Evaulation Report HETA 81-409-1290: The Donaldson Company, Inc.
(HETA 81-409-1290). NIOSH. https://www.cdc.gov/niosh/hhe/reports/pdfs/81~409~
1290. pdf?id=10.26616/NIQSHHETA814091290

NIOSH. (1984a). Health Hazard Evaluation Report HETA 83-425-1500: Westview Press. (HETA 83-425-
1500). NIOSH. https://www.cdc.gov/niosh/hhe/reports/pdfs/1983~0425~1500.pdf

NIOSH. (1984b). Health hazard evaluation report no. HETA 82-075-1545, Pratt & Whitney Aircraft,

West Palm Beach, Florida. (HETA 82-075-1545). Cincinnati, OH: National Institute for
Occupational Safety and Health.

NIOSH. (1986). Health Hazard Evaluation Report No. HETA-85-482-86-116-1730, Winters Industry
Foundry, Canton, Ohio (pp. 85-482). (NIOSH/00166571). Stephenson, RL; Albrecht, WN.

NIOSH. (1993). Health hazard evaluation report no. HETA 91-004-2316, Daubert Coated Products, Inc.,
Dixon, Illinois. (HETA 91-004-2316). Cincinnati, OH: National Institute for Occupational Safety
and Health.

NIOSH. (1994). Health hazard evaluation report no. HETA 93-0351-2413, Goodwill Industires of

America, Inc. Bethesda, Maryland. (HETA 93-0351-2413). Cincinnati, OH: National Institute for
Occupational Safety and Health.

NIOSH. (1995). In-depth survey report: control of perchloroethylene exposures in commercial dry

cleaners at Brown's Cleaners, Sant Monica, California. (ECTB 201-16a). The National Institute for
Occupational Safety and Health.

NIOSH. (1996). Health hazard evaluation report No. HETA 96-0135-2612, Eagle Knitting Mills, Inc.,

Shawano, Wisconsin. Cincinnati, OH: U.S. National Institute for Occupational Safety and Health.

NIOSH. (1997a). Control of health and safety hazards in commercial drycleaners: chemical exposures,
fire hazards, and ergonomic risk factors. (DHHS (NIOSH) Publication Number 97-150). Atlanta,
GA. http://www.cdc.gov/niosh/docs; )]_

NIOSH. (1997b). Hazard control: Control of exposure to perchloroethylene in commercial drycleaning
(machine design) (HC 18). (97-156).

https://www.cdc.gov/niosh/docs/hazardcontrol/pdfs/hcl8.pdf?id=10.26616/NIOSHPUB97156

Page 309 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

NIOSH. (1999a). Evaluation of Eight Dry Cleaning Shops with State-of-the-Art Control Equipment.
Report on Task 1. Perchloroethylene in Dry Cleaning Shops. NIOSH.
https://ntrl.ntis.eov/NTRL/dashboard/searchResults/titleDetail/PB99168890.xhtml
NIOSH. (1999b). In-depth survey report evaluation of control technology for perchlorethylene in dry
cleaning shops. (ECTB 240-15). Cincinnati, OH: NIOSH.
https://www.cdc.eov/niosh/surveyreports/pdfs/240-15.pdf
NIOSH. (2000a). In-depth survey report evaluation of control technology for perchlorethylene in dry
cleaning shops. (ECTB 240-12). Cincinnati, OH: NIOSH.
https://www.cdc.eov/niosh/surveyreports/pdfs/240-12.pdf
NIOSH. (2000b). In-depth survey report: Comparision of perchloroethylene exposures before and after
the installation of local exhaust ventilation at a commercial dry cleaners at drycleaning plus.
(ECTB 240-13). CDC. http://www.cdc.eov/niosh/nioshtic-2/20000629.html
NIOSH. (2001). Evaluation of Solvent Exposures from the Degreaser. Trilthic Inc., IN. (HETA 2000-0233-
2845). NIOSH Publishing Office: National Institute of Occupational Safety and Health.
http://www.cdc.eov/niosh/hhe/reports/pdfs/2000-0233-2845.pdf
NIOSH. (2002a). In-depth survey report: control of perchloroethylene (PCE) in vapor degreasing

operations, site #1. (EPHB 256-19b). Cincinnati, Ohio: National Institute for Occupational Safety
and Health (NIOSH).

NIOSH. (2002b). In-depth survey report: Control of perchloroethylene (PCE) in vapor degreasing

operations, site #2. (EPHB 256-16b). CDC. https://www.cdc.eov/niosh/surveyreports/pdfs/256-
16b.pdf

NIOSH. (2002c). In-depth survey report: control of perchloroethylene (PCE) in vapor degreasing

operations, site #4. (EPHB 256-18b). Cincinnati, Ohio: National Institute for Occupational Safety
and Health (NIOSH).

NIOSH. (2002d). In-depth survey report: Control of perchloroethylene exposure (PCE) in vapor
degreasing operations, site #3. (EPHB 256-17b). CDC.
https://www.cdc.eov/niosh/survevreports/pdfe	df

NIOSH. (2003). Respirator Usage in Private Sector Firms [Technical Report]. Washington D.C.: United
States Department of Labor, Bureau of Labor Statistics and National Institute for Occupational
Safety and Health, https://www.cdc.eov/niosh/docs/respsurv/

NTP. (2014). 13th Report on carcinogens [NTP] (13 ed.). (PB2015500015). Research Triangle Park, NC:

U.S. Department of Health and Human Services, Public Health Service.

OECD. (2004). Emission scenario document on lubricants and lubricant additives. (JT00174617). Paris,

France. http://www.olis.oecd.org/olis/2004doc.nsf/LinkTo/env-im-mono(2004)21
OECD. (2009a). Emission scenario document on adhesive formulation. (JT03263583). Paris, France.
OECD. (2009b). Emission scenario documents on coating industry (paints, lacquers and varnishes).

(JT03267833). Paris, France.

OECD. (2010). Scoping Document for Emission Scenario Document on Manufacturing and Use of

Printing Inks. Organization for Economic Co-operation and Development (OECD).

OECD. (2011a). 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.

Page 310 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

http://www.oecd.org/officialdocuments/publicdisplavdocumentpdf/?cote=env/im/mono(2004
)22/revl&doclanguage=en
OECD. (2011b). Emission scenario document on the chemical industry [Website].

http://www.oecd.org/env/ehs/risk-assessment/emissionscenariodocuments.htm
OECD. (2011c). Emission scenario document on the use of metalworking fluids. (JT03304938).

Organization for Economic Cooperation and Development.

OECD. (2015). Emission scenario document on use of adhesives. In Series on Emission Scenario
Documents No 34. (Number 34). Paris, France.

http://www.oecd.org/officialdocuments/publicdisplavdocumentpdf/?cote=ENV/JM/MONO(2Q
Stdoclanguage=en

Of r (2017a). Draft ESD on Vapor Degreasing - Internal EPA document. Organization for Economic Co-
operation and Development (OECD).

OECD. (2017b). Emission Scenario Document (ESD) on the use of textile dyes.

http://www.oecd.org/chemicalsafety/risk-assessment/emissionscenariodocuments.htm
OPW Engineered Systems. (2014). Loading Systems Catalog. OPW Engineered Systems: A Dover
Company. ES-LS-6/15-2M; Uploaded November 18, 2014. OPW Engineered Systems.
http://www.opwglobal.com/docs/libraries/sales~literature/chemical~industrial/engineered~
systems/brochures/opw-es-loading systems cat.pdf?sfvrsn=6
Oregon DEQ. (2018). Dry cleaners: wastewater management. Portland, Oregon, USA.

https://www.oregon.gov/deq/FilterDocs/ENGWastewaterManagement.pdf
OSHA. (2005). Reducing worker exposure to perchloroethylene (PERC) in dry cleaning. (OSHA 3253-

05N). Washington, DC: U.S. Department of Labor, Occupational Safety & Health Administration.
https://www.osha.gov/dsg/guidance/perc.html
OSHA. (2017). Chemical Exposure Health Data (CEHD) provided by OSHA to EPA. U.S. Occupational

Safety and Health Administration.

Rudnick, M. (2017). Comment submitted by Michelle Rudnick, Senior Manager, Regulatory Affairs, CRC
Industries, Inc., Part 5. https://www.regulations.gov/document?D=EPA~HQ~OPPT~2
0

SCG. (2013). Final peer review comments for the OPPT trichloroethylene (TCE) draft risk assessment
[Website], https://www.epa.gov/sites/production/files/2017~

06/documents/tce consolidated peer review comments September 5 2013.pdf
Scriven, 11 \ s Murugan. R. (2005). Pyridine and Pyridine Derivatives.

http://dx.doi.org/10.1002/0471238961.16251809190318Q9.a01.pub2
Seiji, K; Inoue. 0: Jin. C; Liu. YT; Cai. SX; Ohashi, M: Watanabe, T; Nakatsuka, H; Kawai. T; ikeda, M.

(1989). Dose-excretion relationship in tetrachloroethylene-exposed workers and the effect of
tetrachloroethylene co-exposure on trichloroethylene metabolism. Am J Ind Med 16: 675-684.
http://dx.doi.org/10.1002/aiim.47001606Q7
Seiji, \ !ih C; Watanabe, T; Nakatsuka. H; Ikeda, M. (1990). Sister chromatid exchanges in peripheral
lymphocytes of workers exposed to benzene, trichloroethylene, or tetrachloroethylene, with
reference to smoking habits. Int Arch Occup Environ Health 62: 171-176.
http://dx.doi.org/10.1007/BF0Q383594
Smart. BE: Fernandez. RE. (2000). Kirk-Othmer encyclopedia of chemical technology

Page 311 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Fluorinated aliphatic compounds [Encyclopedia]. Hoboken, NJ: John Wiley and Sons, Inc.

http://dx.doi.ore/10.1002/0471238961.0612211519130118.a01
Snedeco ckman, JC; Mertens, JA. (2004). Chloroethylenes and chloroethanes.

Sol ^ h; Robins, TG; Sampaio, C. (1990). Perchloroethylene exposure assessment among dry cleaning

workers. AIHA J 51: 566-574.

Spirit AeroSystems Inc. (2017). Perchloroethylene usage at Spirit AeroSystems, Inc. (EPA-HQ-OPPT-
2016-0732-0077). Washington, D.C.: Spirit AeroSystems Inc.
https://www.regulations.gov/document?D=EPA~HQ~OPPT~~2Q16~0732~0077
Stefaniak. AB: Breysse, PN; Murray, MPM; Rooney, BC: Schaefer. J. (2000). An evaluation of employee
exposure to volatile organic compounds in three photocopy centers. Environ Res 83:162-173.
http://dx.doi.org/10.1006/enrs.2000.4Q61
Tech Met Inc. (2017). Tech Met letter to HSIA. (EPA-HQ-OPPT-2016-0732-0027). Washington, D.C.:

Tech Met Inc. https://www.regulations.gov/document?D=EPA-HQ-OPPT-2016-0732-0027
Tomer. A: Kane. J. (2015). The great port mismatch. U.S. goods trade and international transportation.
The Global Cities Initiative. A joint project of Brookings and JPMorgon Chase.
https://www.brookings.edu/wp-content/uploads/2015/06/brgkssrvygcifreightnetworks.pdf
Triumph Precision Components. (2017). Perchloroethylene presentation. (EPA-HQ-OPPT-2016-0732-
0077). Washington, D.C.: Triumph Precision Components.
https://www.regulations.gov/document?D=EPA-HQ-OPPT-2016-0732-0077
U.S. BLS. (2014). Employee Tenure News Release.

http://www.bls.gov/news.release/archives/tenure 09182014.htm
U.S. BLS. (2016). May 2016 Occupational Employment and Wage Estimates: National Industry-Specific

Estimates [Website], http://www.bls.gov/oes/tables.htm
U.S. Census Bureau. (2015). Statistics of U.S. Businesses (SUSB).

https://www.census.gov/data/tables/2015/econ/susb/2015-susb-annual.html
U.S. Census Bureau. (2019a). Survey of Income and Program Participation data [Website].

https://www.census.gov/programs-surveys/sipp/data/datasets/2008-panel/wave-l.html
U.S. Census Bureau. (2019b). Survey of Income and Program Participation: SIPP introduction and
history. Washington, DC. https://www.census.gov/programs-surveys/sipp/about/sipp-
introduction-historv.html

U.S. EPA. (1977). Control of volatile organic emissions from solvent metal cleaning [EPA Report]. (EPA-
450/2-77-022). Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of Air
and Waste Management, Office of Air Quality Planning and Standards.

U.S. EPA. (1980). AP-42. Compilation of air pollutant emission factors

Chapter 4.7: Waste Solvent Reclamation (5th ed.). Research Triangle Park, NC: Office of Air and
Radiation, Office of Air Quality and Planning Standards.
https://www3.epa.gov/ttn/chief/ap42/ch04/index.html
U.S. EPA. (1981). AP-42. Compilation of air pollutant emission factors. Chapter 4.6: Solvent degreasing.

Washington, DC. http://www3.epa.gov/ttn/chief/ap42/ch04/final/c4s06.pdf
U.S. EPA. (1985). Occupational exposure and environmental release assessment of
tetrachloroethylene. Office of Pesticides and Toxic Substances.

Page 312 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

U.S. EPA. (1987). Organic chemicals and plastics and synthetic fibers category effluent limitations

guidelines, pretreatment standards, and new source performance standards. Final rule. Fed Reg
52: 42522-42584.

U.S. EPA. (1989). Locating and estimating air emissions from sources of perchloroethylene and
trichloroethylene. (EPA-450/2-89-013). Research Triangle Park, NC: U.S. EPA.
https://www3.epa.gov/ttn/chief/le/perc.pdf
U.S. EPA. (1991a). 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.

U.S. EPA. (1991b). Dry cleaning facilities - Draft background information for proposed standards. (EPA
450/3-91-020a). Research Triangle Park,NC: U.S. Environmental Protection Agency, Office of Air
Quality Planning and Standards.

https://nepis.epa.gov/Exe/ZvPDF.cgi/000Q2H9E.PDF?Dockev=00002H9E.PDF
U.S. EPA. (1992). Guidelines for exposure assessment. Federal Register 57(104):22888-22938 [EPA
Report], (EPA/600/Z-92/001). Washington, DC.
http://cfpub.epa.gov/ncea/cfm/recordisplay.cfm?deid=15263
U.S. EPA. (1994a). Fabric finishing - generic scenario for estimating occupational exposures and

environmental releases -final.

U.S. EPA. (1994b). Guidelines for Statistical Analysis of Occupational Exposure Data: Final. United States

Environmental Protection Agency :: U.S. EPA.

U.S. EPA. (1994c). National Emission Standards for Hazardous Air Pollutants from the Synthetic Organic
Chemical Manufacturing Industry for Process Vents, Storage Vessels, Transfer Operations, and
Wastewater. (40 C.F.R. §§ 63.110 and 63.126-130). Washington, DC.

U.S. EPA. (1995). Protocol for Equipment Leak Emission Estimates. (EPA-453/R-95-017). Research
Triangle Park, NC: Office of Air and Radiation, Office of Air Quality and Planning Standards.
https://www3.epa.gov/ttn/chief/efdocs/eauiplks.pdf
U.S. EPA. (1997). Solvent Cleaning. Volume III, Chapter 6. pp. 6.2.1. Washington, DC.

http://www3.epa.gov/ttnchiel/eiip/techreport/volume03/iii06fin.pdf
U.S. EPA. (1998). Cleaner technologies substitutes assessment for professional fabricare processes.

(EPA 744-B-98-001). Washington, DC: U.S. Environmental Protection Agency, Office of Pollution
Prevention and Toxics.

https://nepis.epa.gov/Exe/Z¥PDF.cgi/2000120G.PDF?Docke¥=2000120G.PDF
U.S. EPA. (2001a). Guide to industrial assessments for pollution prevention and energy efficiency [EPA
Report]. (EPA/625/R-99/003). Cincinnati, OH: Office of Research and Development, National
Risk Management Research Laboratory, Center for Environmental Research Information.
U.S. EPA. (2001b). Risk assessment guidance for superfund: Volume III - Part A, Process for conducting
probabilistic risk assessment [EPA Report]. (EPA 540-R-02-002). Washington, DC: U.S.
Environmental Protection Agency, Office of Emergency and Remedial Response.

U.S. EPA. (2004). Draft GS on spray coatings in furniture - internal EPA document.

U.S. EPA. (2006a). Economic impact analysis of the perchloroethylene dry cleaning residual risk
standard (pp. 1-19). (EPA 452/R-06-005). Research Triangle Park,NC: U.S. Environmental
Protection Agency, Office of Air Quality Planning and Standards, Health and Environmental
Impacts Division. https://nepis.epa.gov/Exe/ZvPDF.cgi/P100QFLJ.PDF?Dockey= P100QFLJ.PDF

Page 313 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

U.S. EPA. (2006b). Risk assessment for the halogenated solvent cleaning source category [EPA Report].
(EPA Contract No. 68-D-01-052). Research Triangle Park, NC: U.S. Environmental Protection
Agency, Office of Air Quality Planning and Standards.
https://www.regulations.gov/document?D=EPA~HQ~OAR~2002~0009~0Q22
U.S. EPA. (2014). Degreasing with TCE in commercial facilities: Protecting workers [EPA Report].

Washington, DC: U.S. Environmental Protection Agency, Office of Pollution Prevention and
Toxics.

U.S. EPA. (2015a). AP-42: Compilation of Air Pollutant Emissions Factors. Fifth Edition, Volume I;

Chapter 5: Petroleum Industry. Research Triangle Park, NC: Office of Air and Radiation, Office of
Air Quality and Planning Standards, https://www3.epa.eov/ttn/chief/ap42/ch05/index.html
U.S. EPA. (2015b). ChemSTEER user guide - Chemical screening tool for exposures and environmental
releases. Washington, D.C. https://www.epa.gov/sites/production/files/2015~
05/documents/user guide.pdf
U.S. EPA. (2016c). Instructions for reporting 2016 TSCA chemical data reporting. (EPA/600/R-09/052F).
Washington, DC: U.S. Environmental Protection Agency, Office of Pollution Prevention and
Toxics, https://www.epa.gov/chemical~data~reporting/instructions~reporting~2016~tsca~
chemical-data-reportine

U.S. EPA. (2016d). 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
U.S. EPA. (2016e). TSCA work plan chemical risk assessment: Peer review draft 1-bromopropane: (n-
Propyl bromide) spray adhesives, dry cleaning, and degreasing uses CASRN: 106-94-5 [EPA
Report], (EPA 740-R1-5001). Washington, DC.

https://www.epa.gov/sites/production/files/2016~03/documents/l~
bp report and appendices final.pdf
U.S. EPA. (2017a). EPA Use and Market Profile. (EPA-HQ-OPPT-2016-0732-0058).

U.S. EPA. (2017b). Learn the Basics of Hazardous Waste, https://www.epa.gov/hw/learn~basics~
hazardous-waste

U.S. EPA. (2017c). Preliminary Information on Manufacturing, Processing, Distribution, Use, and
Disposal: Tetrachloroethylene (Perchloroethylene) [Comment].

https://www.regulations.gov/document?D=EPA-HQ~OPPT-2Q'U 0 2-0003
U.S. EPA. (2018a). 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/production/files/2018~

06/documents/final application of sr in tsca 05-31-18.pdf
U.S. EPA. (2018b). Problem formulation of the risk evaluation for perchloroethylene (ethene, 1,1,2,2-
tetrachloro). (EPA-740-R1-7017). Washington, DC: Office of Chemical Safety and Pollution
Prevention, United States Environmental Protection Agency.
https://www.epa.gov/sites/production/files/2018~

06/documents/perc problem formulation 5~31~2018v3.pdf
U.S. EPA. (2018c). TRI Reporting Forms and Instructions (RFI) Guidance Document.
https://ofmpub.epa.gov/apex/guideme ext/f?p=guideme ext:41:0::

Page 314 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

U.S. EPA. (2019a). 40 CFR 413: Electroplating Point Source Category. (40 CFR Part 413). Washington,
D.C. https://www.ecfr.gov/cgi~bin/text~

idx?SID=5c5al9d4dd?29dble53fb9ca47el6706&mc=true&node=pt40.31.413&rgn=div5
U.S. EPA. (2019b). 40 CFR 414: Organic chemicals, plastics, and synthetic fibers. (40 CFR Part 414).
Washington, D.C. https://www.ecfr.gov/cgi~bin/text~

idx?SlD=5c5a!9d4dd?29dble53fb9c	>&mc=true&node=pt40.31.414&rgn=div5

U.S. EPA. (2019c). 40 CFR 433: Metal finishing point source company. (40 CFR Part 433). Washington,
D.C. https://www.ecfr.gov/cgi~bin/text~

ic	jf42f8141c0887e5f4&mc=true&node=pt40.32.433&rgn=div5

U.S. EPA. (2019d). 40 CFR 465: Coil coating point source category. (40 CFR Part 465). Washington, D.C.
https://www.ecfr.gov/cgi~bin/text~

idx?SlD=117c2452100fl78f42f8141c0887e5f4&mc=true&node=pt40.32.465&rgn=div5
U.S. EPA. (2019e). 40 CFR 469: Electrical and electronic components point source category. (40 CFR
Part 469). Washington, D.C. https://www.ecfr.gov/cgi~bin/text~
idx?SlD=117c2452100fl78f42f8141c0887e5f4&mc=true&node=pt40.32.469&rgn=div5
U.S. EPA. (2019f). Aluminum forming point source category. (40 CFR Part 467). Washington, D.C.
https://www.ecfr.gov/cgi~bin/text~

idx?SlD=117c2452100fl78f42f8141c0887e5f4&mc=true&node=pt40.32.467&rgn=div5
U.S. EPA. (2019g). Iron and steel manufacturing point source category. (40 CFR Part 420). Washington,
D.C. https://www.ecfr.gov/cgi~bin/text~

idx?SlD=5c5al9d4dd?29dble53fb9c >706&mc=true&node=pt40.31.420&rgn=div5
U.S. EPA. (2019h). Toxicity characterisic. (40 CFR § 261.24(b)). Washington, D.C.
https://www.ecfr.gov/cgi~bin/text~

idx?SlD=15aal9372ec20f4bd9e3bfe68134df39&mc=true&node=pt40.28.261&rgn=div5#se40.2
8.261 12

Von Grote, J. (2003) Occupational Exposure Assessment in Metal Degreasing and Dry Cleaning -
Influences of Technology Innovation and Legislation. A dissertation submitted to the Swiss
Federal Institute of Technology Zurich for the degree of Doctor of Natural Sciences. (Swiss
Federal Institute of Technology Zurich, Retrieved from https://www.research-
col lection, ethz.ch/handle/20.500.11850/116460
von Grote, J; Huriimann, C; Scheringer, M: Hungerbuhler, K. (2006). Assessing occupational exposure to
perchloroethylene in dry cleaning. J Occup Environ Hyg 3: 606-619.
http://dx.doi.org/10.1080/1545962060Q912173
Vulcan Chemicals. (1992). INDUSTRIAL HYGIENE STUDY OF

PERCHLOROETHYLENE/METHYLCHLOROFORM BLENDED AEROSOL BRAKE CLEANERS (FINAL
REPORT) WITH COVER LETTER DATED 031292. (OTS: OTS0535416; 8EHQ Num: NA; DCN: 86-
920000858; TSCATS RefID: 422422; CIS: NA).

Vulcan Chemicals. (1993). INDUSTRIAL HYGIENE STUDY OF METHYLENE

CHLORIDE/PERCHLOROETHYLENE/METHYLCHLOROFORM BLENDED AEROSOL BRAKE CLEANERS.
(OTS: OTS0556634; 8EHQ Num: NA; DCN: 86940000038; TSCATS RefID: NA; CIS: 86940000038).
Vulcan Chemicals. (1994a). INDUSTRIAL HYGIENE SIMULATIONS FOR AEROSOL BRAKE CLEANERS -

POTENTIAL OCCUPATIONAL EXPOSURES TO PERCHLOROETHYLENE, WITH COVER LETTER DATED
05/04/94. (OTS: OTS0557466; 8EHQ Num: NA; DCN: NA; TSCATS RefID: 442936; CIS: NA).

Page 315 of 316


-------
PEER REVIEW DRAFT. DO NOT CITE OR QUOTE.

Vulcan Chemicals. (1994b). Task Report- Cold Cleaning Field Tests of Perchloroethylene / Alcohol

Blends Vickers Electromechanical, Wichita, KS. (OTS: OTS0556807; 8EHQ Num: NA; DCN: 86-
940000212; TSCATS RefID: NA; CIS: 86940000212).

Weatherford Aerospace. (2017). Perchloroethylene Weatherford presentation. (EPA-HQ-OPPT-2016-
0732-0077). Washington, D.C. https://www.regulations.gov/document?D=EPA-HQ-QPPT-2016-
0732-0077

Whittaker, SG; Johanson, CA. (2011). A profile of the dry cleaning industry in King County, Washington:
Final report. (LHWMP 0048). Seattle, WA: Local Hazardous Waste Management Program in King
County.

http://www.hazwastehelp.org/publications/publications detail.aspx?DoclD=Oh73%2fQilg9Q%
3d

Winfield Brooks Company. (2014). Safety Data Sheet: Original Formula Alumtap.

http://wwwl.mscdirect.com/MSDS/MSDS00002/00265025-2016042Q.PDF
Young. ML. (2012). Pre-spotting step toward better cleaning [Website].

https://americandrvcleaner.com/articles/pre-spotting-step-toward-better-cleaning

Page 316 of 316


-------