EPA 560/6-76-006
ASSESSMENT OF WASTEWATER MANAGEMENT,
TREATMENT TECHNOLOGY, AND ASSOCIATED COSTS
FOR ABATEMENT OF PCBs CONCENTRATIONS IN
INDUSTRIAL EFFFLUENTS
TASK II
FEBRUARY 3,1976
FINAL REPORT
s^° '^
U.S. ENVIRONMENTAL PROTECTION AGENCY
OFFICE OF TOXIC SUBSTANCES
WASHINGTON, D.C. 20460
-------
EPA 560/6-76-006
ASSESSMENT OF WASTEWATER MANAGEMENT, TREATMENT
TECHNOLOGY, AND ASSOCIATED COSTS FOR ABATEMENT
OF PCBs CONCENTRATIONS IN INDUSTRIAL EFFLUENTS
Task II
Contract No. 68-01-3259
EPA Project Officer: Thomas Kopp
For
Environmental Protection Agency
Office of Toxic Substances
4th and M Streets, S.W.
Washington, D.C. 20460
February 3, 1976
-------
REVIEW NOTICE
This report has been reviewed by the Office of Toxic Substances, EPA,
and approved for publication. Approval does not signify that the contents
necessarily reflect the views and policies of the Enviroranental Protection Agency,,
nor does mention of trade names or commercial products constitute endorsement
or recommendation for use.
-------
ABSTRACT
This document presents the findings of a study of available wastewater
managemsnt and treatment technology for the purpose of determining toxic pol-
lutant effluents concentrations and daily load achievable in three industrial
categories: polychlorinated biphenyls (PCBs) manufacturing; capacitor manu-
facturing; and transformer manufacturing.
All plants in the above categories have PCB discharges to either water-
ways or sewage treatment plants, under normal operating conditions. All
plants have discharges to storm sewers or directly to waterways under heavy
rainfall conditions.
Extensive survey of wastewater treatment technologies and cooperative
laboratory work with several suppliers of treatment equipment and research
facilities has confirmed that carbon adsorption technology is the best current
candidate for successful removal of PCBs from the wastewaters. As an alterna-
tive uv-ozonation was considered. This technology is still in the research
stage; however, it offers potential of complete destruction of PCBs all the
way to CO.,, water and HCl.
^
Another adsorbent technology now in the development stage, AMBERLITE
polymeric adsorbents, has demonstrated a PCBs removal efficiency roughly
equivalent to carbon during laboratory tests. Further testing is needed with
this adsorbent to accurately assess its potentiality.
For scrap oils and burnable solid wastes generated at these plants, high
temperature, controlled incineration offers a straightforward method of de-
struction, whereas scientific landfilling appears to be the best suited mode
of disposal for nonburnable contaminated solids.
Zero discharge objectives can be best achieved by eliminating discharge
streams and developing recycle systems. All non-contact cooling water would
be segregated, cooled, and recycled. All other wastewater streams would be
pretreated. The portion of the pretreated water which would be used in
the plant would be treated with carbon, while the excess water would be
-------
incinerated in a specially designed system which would allcw for energy
recovery.
Supporting data, rationale for the selection of above recommended treat-
ment technologies and associated costs are contained in this report.
11.
-------
TABLE OF CONTENTS
SECTION I - INTRODUCTION
1.0 OBJECTIVE OF THIS STUDY 1
2.0 SCOPE OF THIS STUDY
2.1 Industry Characterization 2
2.2 Control and Treatment Technologies for PCBs Wastes . . 2
2.3 Capital and Operating Costs for Selected Treatment
Technologies 3
3.0 REPORT CONTENT 3
SECTION II - SUMMARY AND CONCLUSIONS 4
SECTION III - RECOMMENDATIONS 15
SECTION IV - WATER USE AND WASTE CHARACTERIZATION 18
1.0 INTRODUCTION 18
2.0 SPECIFIC WATER USES 18
3.0 INDUSTRY AND PROCESS WASTE CHARACTERIZATION 19
3.1 Manufacturing Process - Polychlorinated Biphenyls (PCBs) 20
3.1.1 Process Description 20
3.1.1.1 PCB Production and Usage 22
3.1.2 Raw Wastes 24
3.1.3 Plant Water Usage 24
3.1.4 Wastewater Treatment and Housekeeping 27
3.1.4.1 Treatment Facility for the Effluent from
Sauget Complex 31
3.1.5 Plant Effluents 31
3.2 Askarel Capacitor Manufacturing Industry 32
3.2.1 Askarel Capacitor Manufacturing Plants 34
3.2.1.1 Askarel Handling. . 36
3.2.1.2 Process Description 37
3.2.1.3 Raw Wastes 42
3.2.1.4 Water Use 42
in.
-------
TABLE OF CONTENTS (Can't)
SECTION IV (Con't)
3.2.1.5 Wastewater Treatment .... 46
3.2.1.6 Effluent Composition 52
3.3 Askarel Transformer Manufacturing Industry 52
3.3.1 Transformer Manufacturing Plants 59
3.3.1.1 Askarel Handling 61
3.3.1.2 Process Description 62
3.3.1.2.1 Assembly and Askarel Filling
Procedure for the Distribution
and Power Transformers. ... 62
3.3.1.3 Raw Wastes 68
3.3.1.4 Water Use 69
3.3.1.5 Wastewater Treatment 71
3.3.1.6 Effluent Composition 75
SECTION V - SELECTION OF POLLUTANT PARAMETERS 79
1.0 INTRODUCTION 79
2.0 SIGNIFICANCE AND RATIONALE FOR SELECTION OF POLLUTANT
PARAMETERS 80
2.1 Polychlorinated Biphenyls 80
2.2 Chemical Oxygen Demand (COD) 81
2.3 Fats, Oils and Greases 82
2.4 Suspended Solids 82
2.5 Dissolved Solids 83
2.6 Other Constituents 83
SECTION VI - WASTEWATER TREATMENT TECHNOLOGIES 84
1.0 INTRODUCTION 84
1.1 Similarities and Contrasts Between PCBs Wastes and Con-
trol Practices in their Production, and in their Use
as Dielectrics 84
1.2 Summary of Waste Management Problem Areas 85
1.2.1 Waste Liquid PCBs and Contaminated Scrap Oil . . 85
1.2.2 PCBs in Wastewaters 86
1.2.3 PCBs Contaminated Solid Wastes 86
1.2.4 Air Emissions of PCBs 87
IV.
-------
TABLE OF CONTENTS (Con't)
SECTION VI (Con't) Page
1.3 Summary of Current PCBs Waste Control Practices .... 87
1.3.1 Control of Waste Liquid PCBs and Contaminated
Scrap Oils 87
1.3.2 Control of PCBs in Wastewaters 88
1.3.3 Control of Solid Wastes Contaminated with PCBs . 89
1.3.4 Control of Air Emissions of PCBs 91
2.0 CANDIDATE PCBs WASTE TREATMENT TECHNOLOGIES CONSIDERED ... 92
2.1 Treatment of Waste Liquid PCBs and Contaminated Scrap
Oils 93
2.1.1 Incineration 93
2.1.2 Sanitary or Scientific Landfill 95
2.2 Treatment of Wastewaters Containing PCBs 95
2.2.1 Carbon Adsorption 95
2.2.1.1 PCBs Adsorption Testing by Carborundum
Company 97
2.2.1.2 PCBs Adsorption Testing by ICI-US ... 98
2.2.1.3 PCBs Adsorption Testing by Calgoh Corp. 102
2.2.1.3.1 Adsorption Treatment of the
Wastewater 105
2.2.1.3.2 Reactivation of the Granular
Carbon 106
2.2.1.3.3 Carbon Treatment 107
2.2.1.3.4 Materials of Construction . . 107
2.2.1.4 Carbon Regeneration Alternatives - Wet
Catalytic Oxidation 108
2.2.1.5 Further Applications Data 109
2.2.2 Ultraviolet-Assisted Ozonation 109
2.2.2.1 Molecular Responses to Ultraviolet
Region Energy Ill
2.2.2.2 Photodegradation of PCBs 112
2.2.2.3 Experimental Factors in UV-Assisted
Ozone Oxidation of PCBs 113
2.2.2.4 Destruction of PCBs and Refractory
Organics at Houston Research, Inc.. . 114
2.2.2.4.1 PCBs Destruction Data .... 114
2.2.2.4.2 Operating Data Obtained from
Refractory Organics Tests . 114
v.
-------
TABLE OF CONTENTS (Con't)
SECTION VI (Con't) Page
2.2.2.5 Destruction of PCBs and Refractory
Organics at Westgate Research Corp. . . 116
2.2.2.5.1 PCBs Destruction Data 116
2.2.2.5.2 Pilot Scale Tests of Refractory
Organics Decomposition . . .122
2.2.2.6 Laboratory Test Results from AiResearch
Corp 127
2.2.2.7 Garments on UV-Ozone Tests 127
2.2.3 Non-Carbon Adsorbents for PCBs 127
2.2.3.1 The Amberlite XAD Series of Macroreticular
Resins 128
2.2.3.1.1 PCBs Adsorption Testing .... 128
2.2.3.1.2 Process Concept for Resin Ad-
sorption of PCBs 129
2.3 Treatment of PCBs - Contaminated Solid Wastes 130
2.3.1 Incineration 130
2.3.2 Sanitary Landfill 132
2.4 Treatment of Air Emissions 132
2.4.1 Condensation Methods 132
2.4.2 Granular Adsorption Methods 132
2.4.3 Catalytic Oxidation of Organics in Evaporated
Effluents 132
2.5 The Potential for Zero Discharge 133
3.0 RATIONALE AND SELECTIONS OF CURRENTLY RECOMMENDED WASTE TREAT-
MENT METHODS 134
3.1 Incineration Recommended for Liquid PCBs and Scrap Oils . 134
3.2 Carbon Adsorption and UV-Assisted Ozonation Recoitmended
for PCBs in Wastewater 135
3.3 Incineration and Landfill Recommended for Contaminated
Solids 136
3.4 Dry Carbon Filter Adsorption Recommended for Control of
Air Emissions 136
SECTION VII - COST OF TREATMENT AND CONTROL TECHNOLOGIES 138
1.0 SUMMARY 138
1.1 Plant to Plant Cost Variations 140
VI.
-------
TABLE OF CONTENTS (Con't)
SECTION VII - Con't Page
1.2 Activated Carbon Terminal Treatment System 142
1.3 Costs Based on Volume Flow 143
2.0 COST REFERENCES AND RATIONALE 143
2.1 Interest Costs and Equity Financing Charges 143
2.2 Time Basis for Costs 144
2.3 Useful Service Life 144
2.4 Depreciation 144
2.5 Capital Costs 144
2.6 Annual Capital Costs 144
2.7 Land Costs 144
2.8 Operating Expenses 145
2.9 Rationale for Incineration Costs 145
2.10 Definition of Levels of Treatment 148
2.10.1 Effect of Varying Effluent PCB Concentrations . . 148
2.10.2 Zero Discharge Definition 149
3.0 PCB WASTEWATER TREATMENT PLANT DESIGN PARAMETERS AND COSTS . . 150
3.1 Introduction 150
3.2 Pretreatment 154
3.2.1 Equalization and Separation Basin 154
3.2.1.1 Equalization Basin Capital and Operating
Cost 156
3.2.1.2 Equalization Basin Costs 156
3.2.2 Multimedia Filtration 156
3.2.2.1 Filtration Costs 160
3.3 Activated Carbon Treatment System for Removal of PCBs
From Wastewater 160
3.3.1 Activated Carbon System Design 167
3.3.1.1 Regeneration of Spent Carbon 175
3.3.1.2 Economic Analysis of Carbon Treatment of
PCB Contaminated Wastewater 178
3.3.2 UV-Ozonation - Potential Alternative 180
3.3.2.1 Description of System Components and
Rationale for Cost Evaluation 180
3.3.2.2 Capital Costs 188
3.3.2.3 Operating Costs 188
3.3.2.4 Total Treatment System Cost for Zero
Discharge System 200
vn.
-------
TABLE OF CONTENTS (Con't)
SECTION VII (Con't)
3.3.2.5 Total Wastewater Treatment Plant Costs
Using Activated Carbon or Ultraviolet
Light Catalyzed Ozonation 200
3.3.3 Cost of Implementing Carbon Adsorption Treatment
for Selected Plants 205
3.3.4 Zero Discharge 212
3.3.4.1 Description of Dry Cooling Tower . . . 212
3.3.4.2 Design Criteria 213
3.3.4.3 Capital Costs 213
3.3.4.4 Operating Costs 213
3.3.4.5 Cost of "Zero Discharge" for Selected
Plants 218
3.3.5 Comparison of "Zero Discharge" with Discharge to
Surface Waters 218
3.4 Scrap Oil Incineration 222
3.5 Solid Waste Incineration • 223
3.6 Total Industry Treatment Costs 224
APPENDIX A - Plant 106's Position Statement on Dow's XFS-4169 Capacitor
Dielectric Liquid
APPENDIX B - PCB Adsorption Testing by XAD-4 Resin
APPENDIX C - Description of Macroreticular Resins from Rohn and Haas Co.
.APPENDIX D - Non-Carbon Adsorption and Other Research Stage PCB Treatment
Technologies
APPENDIX E - Energy Requirement for Various Treatment Options
Vlll.
-------
LIST OF TABLES
SECTION II
1. Sunmary of Waste Loads 5
2. Maximum PCB Industry Treatment Costs 12
SECTION IV
3.1.1.1-1 PCB Manufacture and Sales 23
3.1.1.1-2 Approximate Molecular Composition of Aroclors .... 25
3.2.1-1 U.S. Capacitor Manufacturing Industry Using PCBs ... 35
3.2.1.3-1 Non-Product PCB Discharges 43
3.2.1.3-2 Quantity of Waste Loads 44
3.2.1.6-1 Range of Flow Rates & PCB Concentration in Effluents
from Capacitor Manufacturing Plants 53
3.2.1.6-2 Comparison of Discharges 54
3.2.1.6-3 Intake Water PCB Concentration 55
3.3.1-1 U.S. Transformer Manufacturing Industry Using PCBs . . 60
3.3.1.3-1 Non-Product PCB Discharges 70
3.3.1.6-1 PCB Concentration in Effluents from Transformer
Manufacturing Plants 76
3.3.1.6-2 Influent and Effluent Compositions of Plant 103 ... 77
SECTION VI
2.2.1.1-1 Carborundum Co. Tests of PCBs (Aroclor 1254) Removal
from Water by an Experimental Carbon 99
2.2.1.2-1 ICI-US of PCBs (Aroclor 1254) Removal from Water by
Two Types of Commercial Carbons 101
2.2.1.3-1 Results of Calgon Corp. Laboratory Isotherm Tests for
Carbon Removal of PCBs 103
2.2.2.5.1-1 UV Ozonolysis Destruction of Typical Capacitor and
Transformer PCBs at Westgate Research 121
2.2.2.5.2-3 Simulated Two-Stage, Continuous UV-Ozonation of a 5
Component Mix at Westgate Research Corp 126
ix.
-------
LIST OF TABLES (Con't)
SECTION VII
3.2.1-1 Flew Equalization Basin Sizing and Costs , 158
3.2.2-1 Total Installed Pretreatment System Cost 161
3.2.2-2 Annual Operating Cost for Pretreatment for Removal of
PCS from Wastewater , . 162
3.3-1 Carbon System Feed Characteristics 166
3.3.1-1 Carbon System Design Criteria 168
3.3.1-2 Carbon Adsorption System Sizing and Cbsts 170
3.3.1-3 Activated Carbon Adsorption System Annual Operating
Costs 171
3.3.1-4 Typical Activated Carbon Specifications 174
3.3.2.2-1 Capital Cost Ozone - U.V. System 189
3.3.2.2-2 Ozone Generation and Dissolution Cost Data - Capital
Costs 197
3.3.2.3-1 Ozone Generation and Dissolution Cost Data - Power Cost 199
3.3.2.3-2 U.V. Light Power and System Maintenance Cost 202
3.3.2.4-1 Total Installed System Cost for U.V.-Ozone System . . 203
3.3.2.4-2 Annual Operating Cost for U.V. Ozone System 204
3.3.2.5-1 Total Installed Plant Cost with Activated Carbon ... 206
3.3.2.5-2 Total Annual Operating Cost for PCB Removal from Waste-
water with Activated Carbon 207
3.3.3-1 Costs for Treatment of all Flows and Discharging to
Receiving Stream with Carbon Final Treatment .... 210
3.3.3-2 Costs for Enclosed Recirculation and Cooling of Non-Contact
Cooling Water and Treatment of all other Flows
Using Carbon as Final Treatment 211
3.3.4.4-1 Incinerator Annual Operating Cost for Zero Discharge
Method 216
3.3.4.4-2 Annual Operating Cost for Totally Enclosed Cooling
Water Loops 217
3.3.4.5-1 Costs for Zero Discharge ....... 219
3.3.4.5-2 Costs for Zero Discharge 220
3.3.4.5-3 Costs for Zero Discharge 221
x.
-------
LIST OF TABLES (Can't)
SECTION VII (Con't) Page
3.6-1 Maximum Capacitor Industry Cost 225
3.6-2 Maximum Transformer Industry Cost 225
3.6-3 Cost for Monsanto Treatment 225
3.6-4 Maximum PCB Industry Treatment Cost 226
3.6T5 Maximum PCB Industry Treatment Cost Neglecting
Rainwater Runoff
XI.
-------
LIST OF FIGURES
SECTION IV
3.1.1-1A
3.1.1-1B
3.1.2-1
3.1.4-1
3.2.1.2-1
3.2.1.2-1
3.3.1.2.1-1
3.3.1.2.1-2
3.3.1.2.1-3
3.3.1.5-1
SECTION VI
2.2.1.3-1
2.2.2.4.1-1
2.2.2.4.1-2
2.2.2.4.2-1
2.2.2.4.2-2
2.2.2.5.2-1
2.2.2.5.2-2
2.2.3.1.2-1
Preparation of Crude Chlorinated Biphenyls -
Monsanto Krummrich Plant 21
Distillation of Crude Products - Monsanto's Krummrich
Plant 21
Non-Product PCB Discharges at Monsanto's Krummrich
Plant
26
Process Flow Diagram of the John Zink Incinerator at
Monsanto's Krumtnrich Plant 29
Generalized Flow Diagram for the Manufacturing of
Large Capacitors 38
Generalized Flow Diagram for the Manufacturing of
Snail Capacitors 39
Transformer Filling with Vapor Phase Predrying of
Interiors 63
Transformer Filling with Oven or Vacuum Chamber Pre-
drying of Transformer Internals 66
Transformer Filling Operation with Oven Predrying of
Assembled Hardware 67
Process Flow Diagram for Thermal Oxidizer Incinerator
at Plant 103 73
Equilibrium Carbon Adsorption of PCBs from Water at
Low Concentrations (Calgon Data) 104
Lab Scale Apparatus for Reaction and Mass Transfer
Studies at Houston Research, Inc 115
Aroclor 1254 Destruction by UV-Assisted Ozonation . . 117
Ozone Oxidation of Acetic Acid, Effect of UV and
Temperature. 118
Ozone/UV Oxidation of Acetic Acid; Effect of Increased
Radiation Input 119
The Effect of UV Path Length on TOC Destruction . . . 124
Schematic of Bench Reaction System at Westgate
Research Corp 125
PCBs Removal Process Concept Flow Sheet by Rohm and
Haas Company 131
Xll.
-------
LIST OF FIGURES (Con't)
SECTION VII Page
2.7-1 Required Treatment Plant Area for Removal of PCB
from Wastewater 146
3.1-1 Flow Diagram of Treatment System for Removal of PCB
from Wastewater 153
3.2.1-1 Flow Equalization Basin Installed Capital Cost . . . 157
3.2.2-1 Pretreatrtent System Capital Cost 163
3.2.2-2 Annual Cost for Pretreatment 164
3.3.1-1 Activated Carbon System Installed Cost 172
3.3.1.1-1 Carbon Regeneration Cost vs Usage 176
3.3.1.2-1 Activated Carbon Wastewater Treatment System Total
Annual Cost Based on Different Contaminant
Parameters 179
3.3.1.2-2 Cost per 1000 Gallon for PCB Removal from Wastewater
with Carbon 181
3.3.2.1-1 Proposed Reactor Configuration - Elevation 183
3.3.2.1-2 Plot of Houston Research Test Data for Destruction of
PCB Using Ultraviolet Light Catalyzed Ozonation . . 186
3.3.2.2-1 UV-Ozone System Installed Cost
Reactor Residence 110 Min 193
3.3.2.2-2 UV-Ozone System Installed Cost
Reactor Residence 210 Min 194
3.3.2.2-3 UV-Ozone System Intalled Cost
Reactor Residence 265 Min 195
3.3.2.2-4 UV-Ozone System Installed Cost
Reactor Residence 365 Min 196
3.3.2.2-5 Ozone Production and Dissolution Capital Costs . . . 198
3.3.2.2-1 Ozone Generation and Dissolution Power Costs .... 201
3.3.2.5-1 Carbon System - Total Installed Treatment Plant
Cost for Removal of PCB from Wastewater 208
3.3.2.5-2 Total Annual Treatment Cost for Removal of PCB from
Wastewater Using Carbon as Final Treatment .... 209
3.3.2.3-1 Incineration Cost as Used for Zero Discharge .... 214
3.3.4.3-2 Cost for Totally Enclosed Cooling System 215
Xlll.
-------
ACKNCWLEDGEMENTS
This report was prepared by the staff of Versar Inc., Springfield, Virginia
\d.th the aid of the staff of Clark, Dietz and Associates Engineering, Inc., Urbana,
Illinois. Dr. Robert L. Durfee of Versar was the Program Manager. Major Contributions
were made by Mrs. Gayaneh Contos and Dr. E. E. Hackman, III of Versar; and Mr. Roger
Crawford, Mr. Charles Huether and Dr. Ken Price of Clark, Dietz.
The considerable aid furnished by personnel of the Environmental Protection
Agency, Hazardous Materials Evaulation Branch and Office of Toxic Substances is
acknowledged. Mr. Tom Kopp of OTS served as Project Officer and Mr. Ralph Holtje
provided guidance for the project effort.
Appreciation is also extended to the Electric Industries Association, The
National Electrical Manufacturers Association and the individual companies that
cooperated in this effort.
xiv.
-------
SECTION I
INTRODUCTION
1.0 OBJECTIVE OF THIS STUDY
The major objective of this study was to gather/ evaluate and interpret
specific kinds of information about industrial PCBs wastes in production of
PCB and in electrical component manufacturing. The information was to pro-
vide the basis for establishment of enforceable PCBs effluent standards for
industrial discharges. It was also to be useable in determinations concern-
ing allowances for variances with pollution control regulations.
The specific need for the study was rooted in the conclusion of the ad-
judicatory hearing following publication of the proposed effluent standards
by EPA on December 29, 1973. The conclusion at that time was that the pro-
posed standards were based on inadequate data.
Thus there developed a high priority need for an assessment of wastewater
management and treatment technology in these selected industries, the capital
and operating costs to be borne in practice, the effluent concentrations and
daily waste load of PCBs achievable, and the premise developing technology
holds for future improvements in discharge and cost reductions.
2.0 SCOPE OF THIS STUDY
The overall scope was divided into three parts:
• Industry Characterization, to obtain the basic data and insights
to be used in placing the following tasks in perspective.
* Investigations of Control and Treatment Technologies, to provide
the technical bounds in establishing the allowable PCBs discharges,
and give the rationale used in reccmmending technologies for
practical current applications.
• Capital and Operating Costs^ for Selected Treatment Designs, to
provide the economic factors for use in decisions involving cost
and benefit analyses.
-------
2.1 Industry Characterization
This portion was to completely describe, with the aid of flow sheets,
the PCBs production method, and the methods of manufacture of transformers arrl
capacitors which contain PCBs. The intent was to cover every aspect of PCBs
handling, storage, and loss as wastes in both production and manufacture.
Data on quantities, character and concentrations of all discharges or losses
to the environment were to be determined using available information provided
by the plant personnel during on-site visits and by reasonable engineering
extrapolations and estimates. Potential methods for reduction of discharges
were to be noted, and costs already accumulated by the plants in reducing
PCBs discharges were to be collected and interpreted.
2.2 Control and Treatment Technologies for PCBs Wastes
This portion of the study required assembly, examination to deter-
mine practicality and feasibility, and technical and economic evaluations of
the subject methods. The subtask work statements suggest consideration of a
wide variety of categories of methods for achieving PCB emissions reductions
and elimination. These include manufacturing process and. materials handling
changes, use of substitute products, and novel methods of wastewater treat-
ment that might be adapted frcm other industries or emerging technology. For
consideration, a novel technique must show potential for PCBs waste treatment
capabilities beyond that of current technology, at least in some respect. Im-
provements in current methods were to be fully considered. For selection as
a prime or recommended treatment method, a candidate had to be amenable to
capital and operating cost estimates for full scale, plant-sized treatment fa-
cilities at this time. Costs were to be applicable to a practical range of
wastewater flows and levels of PCBs and other pollutants. MDnitoring costs
were to be included.
Treatment methods were to be examined and compared for effective-
ness and reliability of achieving specified concentrations and absolute
amounts of PCBs in ultimate wastewater disposal. Any transfer of PCBs to air,
soil, or groundwater was to be determined for the treatment methods. The
-------
practicality of putting any treatment method into operation within one year
and within three years of the promulgation of Section 307(a) effluent
limitations was to be determined.
2.3 Capital and Operating Costs for Selected Treatment Ted)Bglogieg_
This portion of the study was to assess the costs of each of the
selected control technologies. Capital costs for both equipment and in-
stallation, annual operating costs for labor, materials, utilities., interest
on invested capital, and other recurring costs were to be included. Costs
were to be tied to preliminary design criteria which were to shew capacities.
Estimates of incremental costs per unit of increased production, or unit of
PCBs handling, were to be made.
3.0 JEPORT CONTENT
The report describes the three phase investigation by Versar Inc. into
the practice and potential for control of PCBs emissions from PCB production
and the manufacturing of capacitors and transformers containing PCBs, The
first phase covers on-site plant inspections, industry and waste characteri-
zations and details on practices currently in use. The second phase details
an exhaustive search of present and potential treatment technologies applica-
ble to PCB wastes. Included are selections and recommendations of the most
practical approaches to control measures for installation within one year,
c*nd within three years, to achieve effluent limitations necessary to meet, the
purpose of Section 307(a), including zero discharge of wastewaters. Also in-
cluded are selected technologies expected to contribute valuable approaches to
the zero discharge potential for PCBs, but which are still in the resec-rroh
stages requiring more than five years for full development. The final phase
presents the cost analyses of the recommended PCBs control technologj.es for
wastewaters, scrap oils and solid wastes.
-------
SECTION II
SUMMARY AND CTNCLUSICNS
The table on the succeeding page gives a quantitative summary of daily
PCB waste loads achievable by PCB manufacturer and the two major PCBs users
(capacitor and transformer industries). This information was derived from
data supplied by these industries and presents estimates of amounts of PCB
wastes currently discharged by these industries to the environment via water,
land and scrap oil generated and destroyed by these facilities via inciner-
ation.
The results of this study are grouped into three broad categories. First,
those dealing with industrial characterization, production, use and emissions.
The second group contains those giving an overview of available technologies
for treatment of various PCB wastes and the third summarizes the costs associ-
ated with currently recommended waste treatment methods.
Industrial Characterization
1. The domestic annual production of PCBs in 1974 was 40,466,000 Ibs of
which approximately 22,000,000 Ibs was used by capacitor manufactur'-
ers and 12,000,000 Ibs was used by transformer industries. The
balance reflects inventory changes and about 5,395,000 Ibs of export
sales.
2. There are a total of 37 PCB user (capacitor and transformer) plants
in the U.S. and one PCBs manufacturer. Of the 38 plants, 10 dis-
charge tneir effluents into the water ways while the remainder dis-
charge PCBs into the sewage treatment plants. All plants in these
categories have discharges under heavy rain fall, conditions.
3. In addition to air emissions, there are three types of waste materi-
als generated at these plants that require treatment and proper
handling in order to minimize the PCB entry into the environment.
These are:
-------
TABLE 1
SUMMARY OF WASTE LOADS
Daily Average
KB Discharge Land-Destined Scrap Oils
in Waterways PCB to
or Sewers Wastes Incineration
PCB Manufacturer 3,06 Ibs 301 Ibs 1425 Ibs
Capacitor Industries 5.86 Ibs 4440 Ibs 3968 Ibs
Transformer Industries 0.17 Ibs Unknown 1750 Ibs
-------
(a) waste waters containing trace quantities of PCBs
(10 to 500 ppb PCBs)
(b) waste PCBs, scrap oils and small quantities of
process water highly contaminated with PCBs
(c) burnable and non-burnable solid materials contaminated
with PCBs.
Quantitative estimates of these wastes are given in Table 1.
4. The quantities of land-destined wastes originating at the PCBs and
capacitor manufacturing plants were estimated from sparse data sup-
plied by a few plants. Similar information is not available from
the transformer industry. These plants normally drum their solid
wastes and have incomplete information on the number of drums ac-
cumulated annually. Some plants compact their wastes prior to drum-
ming; others drum their wastes without the benefit of compacting.
There are no adequate records to quantify these wastes. At some
plants, the majority of this waste category, however, exists in an
enormous total waste volume and probably at relatively low concen-
trations of PCBs. Further efforts are needed to adequately define
these wastes.
5. As yet, very little is being done to control air emissions. The
general assumption is that the vapor pressure of PCBs is so low that
there will be essentially no air contamination. A few facilities,
however, were reported to be filtering and chilling exhaust air from
PCBs impregnation areas.
6. While the emissions of PCBs to water are expected to decrease due to
improved pollution abatement of waterborne wastes, the release of
PCBs to air and land is expected to increase, The primary source of
increasing air emissions is the increase in the load of incineration
materials due to proper handling of wastes which were previously
discharged into the waterways or sewers. The quantities of land-
destined wastes will also increase due to improved housekeeping
measures.
-------
7. Since most water used at these facilities is for non-contact cooling
purposes, at most plants it is possible to significantly reduce the
effluent volume by segregation of wastewaters, recycling and proper
housekeeping measures.
8. Most plants have already undertaken PCBs containment programs in
order to minimize the entry of PCBs into the environment.
9. There is no commercial scale wastewater treatment for PCBs removal
being practiced bfeyond those of gravity settling of the heavy PCBs
layers as a sludge from the bottom of sumps or tanks, and skimming
of a contaminated oil-layer from the water surfaces.
10. Wastes in PCBs production and PCB wastes in capacitor and transformer
manufacturing differ primarily in that the producer generates some
higher boiling, highly chlorinated and polycyclic materials referred
as "Montar" that are separated from the Aroclor products and in-
cinerated. Thus the users of PCBs do not receive the higher boiling
materials. Aroclor 1016 is the principal PCB used by capacitor
industry; at some plants Aroclor 1242 and 1221 are also being used
in limited quantities. Mineral oil is the principal coolant oil used
by the transformer industry. PCB transformer oils (blends of 60 to
70 percent Aroclor 1254 or 1242 and 40 to 30 percent trichloro-
benzene) are used in only 5-10 percent of these plants' manufacturing
volume. However, it has been concluded that the characteristics of
the waste materials and effluents generated in all these facilities
are, in general, similar enough so that the same kinds of treatment
technologies can be utilized for purposes of controlling PCB entry
into the environment.
11. Most industrial analysis for PCBs in wastewaters are based on grab
samples and probably have
the one and ten ppb level.
samples and probably have not been more accurate than about - 50% at
-------
12. Rivers receiving PCBs discharges for a number of years vary greatly
in PCBs content with time, apparently depending upon PCB content in
storm water runoff and the degree to which bottom sediment is agi-
tated and suspended.
13. Wherever there have been PCB operations in the past, there are proba-
bly high concentrations in local waterways bottom sediments.
14. All water streams have greatly lowered PCBs contents:
(a) when solids are removed
(b) when dispersed or dissolved oils are removed
(c) When the stream comes in contact with a wide variety of solid
surfaces
(d) when no surface active agents are present.
15. Over the past 45 years, waste PCBs from transformer and capacitor
operations have been used as local road oiling compounds. Sometimes
they were discarded in dumps adjacent to manufacturing facilities.
These are sources of long term leaching of PCBs into waterways, par-
ticularly with storm water runoff.
Overview of Available Technologies for the Treatment of Various PCB Wastes
1. The most advanced treatment technology in use is incineration. The
PCBs manufacturer and one user have plant scale incineration capable
of destroying PCBs with very high efficiency. There are at least
two commercial services (Rollins Environmental Services and Chemtrol
Pollution Services) available, with four incinerator locations in
the Eastern and Southern U.S., that destroy PCBs.
2. Incineration is primarily applicable to waste PZBs and scrap oils
contaminated with PCBs. All installations have the capacity of
"burning" some contaminated wastewater but, of course, the pro-
portion of that water to the exothermic oil burning must be kept
low.
-------
3. Only one commercial incineration service (Rollins) can routinely
handle all kinds of PCBs contaminated transformer and capacitor com-
ponents, sludges, fuller's earth and other solids, as long as they
can be contained in a 47 gallon fiber drum. One PCB user company
incinerates transformer internals for purposes of metal recovery.
4. Adequate methodology is available for those plants wishing to con-
trol the release of PCBs to the environment. Currently available
technologies can be applied to the efficient removal of PCBs from
wastes, or their destruction with the other wastes.
5. The PCBs content of wastewaters can be lowered to the 1 ppb level or
below by carbon adsorption, macroreticular polymer resins and proba-
bly other adsorbents.
6. Carbon adsorption is currently the best available techno3.ogy for
plant scale treatment of PCBs wastewaters. This conclusion is based
on laboratory tests with PCBs in water, and on the long background
of plant scale use of carbon adsorption for removal of organics from
water.
7. UV-ozonation technology is the best demonstrated method, on the labo-
ratory scale, for destruction of PCBs in wastewaters when the streams
occur in large volume, on a relatively continuous flow basis and with
PCBs at the ppb concentration levels. This technology has the po-
tential for conversion of PCBs to CO-, H_0 and HC1.
8. Waste liquid PCBs and scrap oils contaminated PCBs are best handled,
as a guideline, by high temperature (2000-2400 F) and long residence
time (2-3 seconds) incineration. However, because of incinerator
design variables, the conditions should be chosen in each case to
lead to 99.999% destruction.
9. Incinerators should be equipped with low temperature alarms, and low
temperature shut down of PCBs feed. They should have high efficiency
water scrubbers to prevent HCl dissemination.
-------
10. The best incinerator cxxrfoination for handling wastes from these in-
dustries is a rotary burner fired by a liquid burner, and followed
by an afterburner and scrubber system. The rotary burner can be
designed to handle a variety of solid materials, and the liquid
burner can handle both the oily and water type wastes.
11. Polymeric resins (AMBEKLITES) were found in laboratory tests to be
approximately as effective as carbon in removing PCBs from water.
Further pilot scale testing is needed with this newer (than carbon)
technology to accurately assess its potential.
12. Although still in the laboratory stage, catalytic reduction of PCBs
offers the possibility of reduction to biphenyl and HC1; and cata-
lytic oxidation offers a potential for destruction of PCBs to 00.,
*• i
H20 and HC1.
13. Reverse osmosis and ultrafiltration appear to offer a long term po-
tential for a more maintenance-free, and lower operational cost
method of separation of PCBs from water. Success is dependent upon
specialized membrane development now being carried out at several
facilities.
14. It is believed that treatment systems employing activated carbon and
possibly UV-ozonation could produce effluents which would be at or
below the limits of detectability for PCBs with currently available
analytical techniques. However, since no full scale systems for the
treatment of PCBs are in operation at this time, this possibility
cannot be confirmed.
15. Unfortunately, no methodology is presently available which can
guarantee "zero discharge" to the environment. "Zero discharge" ob-
jectives can be best met now by eliminating discharge streams and
developing recycle systems. All streams that are high in pollutants
and can't be treated for reuse and the rainwater runoffs should be
collected and incinerated.
10
-------
Costs Associated with Currently R^gonmended Waste Treatment Methods
I. The Table on the succeeding page summarizes the estimated maximum capital
and annual costs for the treatment of wastes generated in each industry
category- As shown
treatment of all contaminated and potentially contaminated wastewaters,
the incineration of scrap oils and the incineration of contaminated solid
waste generated in the domestic production and utilization of PCBs if
$35.50 million and $9.61 million, respectively. The total annual cost
given above represents $0.28 per pound of PCB utilized within the U.S.,
based on 1974 data.
2. Rainwater contribution to the capital and annual wastewater treatment
costs was estimated at 15 and 25 percent, respectively. If significant
isolation of the contaminated areas can be achieved these costs will be
somewhat reduced.
3. The estimation of total maximum costs anticipated for the manufacturer
and users of PCBs was facilitated by preparing cost estimates for each
plant site visited during the course of this study and then aggregating
to the total industry category on the basis of the percentage of the
total industry represented by the visited plants.
4. The treatment system includes settling for solids removal, oil skimming,
equalization, fine media filtration, terminal PCB treatment, flow measure-
ment and sampling, and discharge to surface receiving waters. All sludges
and free floating oils are disposed of through incineration. All backwash
waters used in the system are recycled back to the gravity settling basin
for treatment. All spent adsorbent is destroyed by incineration (for
wastewater flows up to 300 gpm) or regenerated on-site for reuse 'for
waste flows above 300 gpm).
5. An activated carbon system was selected as the terminal PCB treatment
method for purposes of estimating the maximum total costs. A potentially
viable alternative to carbon adsorption, UV-ozonation, has less well da-
fined capabilities. Additionally, it is believed that unless tiie UV-
ozonation system can do an equal or better job of destroying PCBs at a
lower cost, it will not be selected by these industries.
31
-------
TABLE 2
MAXIMUM PCB INDUSTRY TREATMENT COSTS
Industry
PCB Producer
Capacitor Manufacturing
Transfonrar Manufacturing
Total
Total Cost Per Pound of
PCB utilized $ 0.28
Total Cost Per Pound of
PCB discharged to waterways $ 2896
Capital
Cost
$ 1,600,000
ng 16,600,000
ring 17,300,000
$35,500,000
Total
Annual
Cost
$ 595,000
4,200,000
4,815,000
$9,610,000
Cost Per Pound
Produced or Used
$0.015
0.190
0.401
12
-------
6. A carbon adsorption system will remove PCBs from industrial effluents to
a level of one ppb or less. It is impractical to design an activated
carbon system to obtain a choice of effluent concentrations. The operat-
ing cost savings for operations at various effluent PCB concentrations
are expected to be too small to justify designing carbon adsorbers for
PCBs removal at higher than one ppb level.
7. The work performed by Houston Research and Westgate Research on UV-
ozonation indicates that the final concentration is a function of resi-
dence time, and that UV-ozonation systems can be designed to achieve any
desired effluent PCB concentration.
8. Preliminary comparisons of the capital costs of the two systems show at
least a 100 percent greater cost for the ozone system over the carbon.
Mien pretreatment costs are combined with treatment, the UV-ozonation
process is about 10 percent higher than the carbon process.
9. The design and fabrication of incineration equipment is a highly proprie-
tary industry, it is believed that any attempt to estimate the cost of
such systems on a rational basis would be fruitless. In addition, it
seems quite improbable that the smaller PCB user plants would find it
economically justifiable to install and operate such a system for de-
struction of scrap oils and solid wastes contaminated with PCBs. The al-
ternative to on-site incineration is contract incineration. For purposes
of cost estimation and based on the range of prices quoted by service
companies, a flat rate of $0.10 per pound has been assumed for all scrap
oil and solid wastes even if they are generated by facilities which have
an on-site incinerator.
10. Contract incineration of large volume, dilute aqueous streams does not
seem to be a feasible alternative due to the logistics and transportation
cost. For purposes of achieving "zero discharge", capital costs for on-
site incineration systems were developed. In this case, the accuracy of
capital cost estimates are not deemed as critical in view of tho high,
operating cost, primarily fuel cost, associated with this system,,
13
-------
The U-TiL!. TJSL "zero discharge" system comprises segregation of waters
used _n I'he plant aucl cooling and recycling all non-contact cooling
water.. Al"; other wastewaters will be routed to a central system for
equal!"/ut tor, settling and fine media filtration. The portion of the
water to be. utilized in the plant will be subjected to terminal treat-
ment with carbons, the excess water will be subject to incineration in
a specially designed system which would have the potential of energy
reccivery.
,:.api-laL and total annual costs to achieve "zero discharge", based on the
Tystem described above, was developed for selected plants. It was con-
cluded tnat the total annual costs for "zero discharge" was about 2.0 to
12,0 times the cost of the carbon system (close loop cooling water case)
•when rainwater was included.
Rainwater runoff is a direct function of the locations' rainfall and run-
olf ?rea. The rain runoff area may or may not be dictated by the size of
the production facility. This single flow could provide the largest flow
variation for all production facilities.
For die "zero discharge" system the rainwater contribution to the annual
wastevater incineration cost is very high, Ocsttparison of the two "zero
discharge" cases i (1) inclusion of rainwater runoff as incinerating
rrarerial and i.2) exclusion of rainwater runoff i indicates the annual
cos'; of the ca^e 1 to be 2 to 42 times the cost of foe case 2 for the
vci.:A'jus plants under study.
As a rough approximation, enerjy requirement for waste incineration is
directly proportional r,c the hydraulic load on the unit. Water inciner-
ation is far more energy intensive than scrap oil incineration.
Mnei jy roquirero&jrt of each, plant is proportional to the total hydraulic
.'.x.- of "he pretjreatrrent and teimijial treatment systems.
14
-------
SECTION III
There are indications within the PCBs manufacturer and PCBs user indus-
tries that progress is being made in air and water pollution abatement; how-
ever, little is being done to assure that the land-destined wastes are dis-
posed of in an environmentally sound manner. This shortcoming can be
attributed to higher costs associated with handling large volumes of wastes
further information is needed regarding methods and costs of handling land-
destined wastes.
Based on the summary and conclusions cited in Section II, the following
recommendations are made:
1. Verification sampling should be conducted under controlled conditions
in order to define the levels of PCBs discharged into the environ-
ment with higher degree of accuracy.
2. Promulgation and enforcement of regulations for water pollution
abatement fron stationary sources should be continued. Itegulations
for non-point sources should be initiated to minimize the entry of
PCBs into the environment.
3. The use of more biodegradable PCB (Aroclor 1016) in transformer ap-
plications should be encouraged.
4. New regulations should be developed and enforced to ensure that the
land-destined wastes from these facilities are disposed of in an en-
vironmentally sound manner.
5. .Additional data should be developed to adequately characterize solid
wastes generated by these three industries and then define methods
and costs of handling these wastes.
6. The capability of various incineration facilities should be deter-
mined for PCBs destruction on a given doctored feedstock using a pre-
developed and specific test plan. Analysis of available data indi-
cated that there are differences of opinions as to optimum tempera-
ture level and residence time for PCB incineration.
15
-------
7. Waste PCBs liquids and scrap oils contaminated with PCBs should be
incinerated in approved facilities. New incineration facilities can
be installed in one year.
8. Every effort should be made at each plant to minimize flows. The
source of rain runoff contamination should be traced and efforts
should be directed to isolate the contaminated areas.
9. Disposable foot coverings and outer garments should be provided at
each plant; work clothing should not be allowed to leave the plant
other than via solid waste.
10. All spills should be cleaned via rags or floor dry and disposed of in
waste drums for incineration. All solid waste should be disposed of
in impervious sealed containers.
11. As a housekeeping measure, all sinks in the PCB impregnation and
PCBs handling areas should be contained. The wash water from these
sinks should be properly treated prior to disposal.
12. Wastewaters containing PCBs should be treated Icy the standard sani-
tary procedures of settling the sludges and sediments, removal of
suspended solids, and skinning to remove floating oils; all prior to
specialized terminal treatment for PCBs.
13. Wastewater pretreated as recommended above should then be treated in
carbon adsorption columns for removal of PCBs to the ppb levels or
less.
14. Spent carbon should be regenerated in an environmentally sound man-
ner , or sent to approved incineration along with other PCBs wastes.
15. Waste capacitor and transformer components, absorbent solids, adsor-
bents, soil, cloth, paper, wood and similar solids contaminated by
PCBs should be incinerated in a rotary kiln. The exhaust should be
fed to an incineration unit for liquid PCBs, equipped with after-
burner and scrubber to assure destruction of all PCB vapors. Such
facilities can be installed in one year.
16
-------
16. Exhaust air from PCB impregnation areas should be filtered through
absorbent carbon, with chilling, if necessary, to meet clean air
quality standards. This equipment can be installed in one yaar.
17. Research on the catalytic oxidation of dilute PCB vapors in air
streams should be conducted to determine the feasibility of low
temperature destruction of PCBs by an energy saving process. This
method will take more than three years to develop.
18. Research on UV-ozonation should be continued to achieve the follow-
ing goals:
a) determination of optimum conditions for a continuous flow
reactor system using actual wastewaters
b) conduct on-site demonstration tests using portable equip-
ment at a PCB using plant, operating on a wastewater slip-
stream
c) design, construct and operate a pilot unit, on actual
wastewater, to develop reliable capital and operating cost
UV-ozonization treatment technology can be fully developed and in-
stalled in three years.
19. Pilot scale R&D on reductive dechlorination should be continued to
obtain reliable capital and operating costs. This method will proba-
bly be available for plant scale application in three years time.
20. Laboratory research on catalytic oxidation as a method of PCBs des-
truction should be initiated. The feasibility of wet catalytic
oxidation using suspended commercial catalysts at elevated tempera-
tures (200°-500°C) and elevated pressures (500 to 1000 psi) should
be determined. This method will take more than five years to develop.
17
-------
SECTION IV
WATER USE AND WASTE CHAPACTERIZATION
1.0 INTRODUCTION
This section discusses specific water uses in the manufacture of Poly-
chlorinated Biphenyls (PCBs) and by selected users of PCBs. The process wastes
are characterized as raw waste loads emanating from specific operations in the
manufacturing of these products, and, where known and reported, they are given
in pounds per ton of PCB used. The specific water uses and amounts are given in
terms, of gallons per day for each of the facilities contacted or visited in con-
nection with this study. Where appropriate, the raw wastes in the effluents are
given both in terms of concentration, mg/liter, or in terms of pounds per day
discharged.
2.0 SPECIFIC WATER USES
Water used in these industries fall into four major characterization headings.
The principal water uses are:
(1) Non-Contact cooling water
(2) Process Water - (a) water used in detergent washing of
capacitor components or PCB filled
capacitor exteriors.
(b) water used in scrubbers for air pol-
lution control
(c) water used as vacuum pump seal
(d) condensate from steam jet ejectors
used at some plants to achieve
vacuum
(e) phosphatizing and fluoride baths
used to prepare the metal surfaces
for painting
(f) water used in paint booths for
scrubbing purposes
18
-------
(3) Auxiliary processes water
(4) Miscellaneous water
Non-contact cooling water is defined as that cooling water which does not
cone into direct contact with any raw material, intermediate product, by-product
or product used in or resulting from the process. The largest use of non-contact
cooling water in these industries is for the cooling of equipment such as vacuum
pumps, tanks and air compressors. At many plants the non-contact cooling water
is also contaminated with PCBs as the general practice is to discharge these
streams combined with process waters or to discharge them through troughs or
trenches that were previously contaminated with PCBs.
Process water is defined as that water which, during the manufacturing
process, comes into direct contact with any raw material, intermediate product,
or by-product used in or resulting from the process.
Auxiliary process water is defined as that used for processes necessary
for the manufacturing of a product but not contacting the process materials; for
example, blowdowns from cooling towers and boiler blowdowns. The volume of
water used for such purposes in these industries is minimal. However, when they
are present, they usually contain PCBs.
Miscellaneous water uses vary among facilities with general usage for showers,
hygiene, eye wash stations, laundry water and sanitary uses. This water also
contains PCBs.
3.0 INDUSTRY AND PROCESS WASTE CHARACTERIZATION
For the manufacture of PCBs and each PCB user manufacturing operation the
following information is given:
. Name and the location of companies in each category;
. A description of the processes at the facilities studied
and pertinent flew diagrams, where appropriate
. Raw waste load data per ton of PCB used and sources of
these wastes;
. Water usage data in terms of gallons per day
19
-------
. Treatment and housekeeping measures practiced at the
facilities and on-going PCB containment programs
. Plant waste effluents found and their composition
3.1 Manufacturing Process - Polychlorinated Biphenyls (PCBs)
3.1.1 Process Description
ttonsanto, the sole domestic manufacturer of PCBs, manufactures
this chemical in their Sauget, Illinois plant. The basic raw material is biphenyl
which is manufactured from pure benzene in another Msnsanto plant. The PCB manu-
facturing operation is conducted in two steps. First, biphenyl is chlorinated
with anhydrous chlorine in presence of ferric chloride to produce crude PCBs and
then the crude PCBs are distilled to obtain the finished product. A schematic
flow diagram of this process is given in Figure 3.1.1-1.
The reaction section consists of 6 reactors (3 batch and 3
cascade). Currently Monsanto manufactures four different types of Aroclors (1242,
1016, 1254 and 1221). For the manufacture of any given product, the chlorinator
is charged with proper quantities of biphenyl and catalyst and heated beyond the
melting point of biphenyl. The flow of vaporized chlorine is then started and
the charge is circulated with a pump. Throughout the chlorination, the temperature
is kept above the melting point of the mixture, but below 150°C to avoid excessive
sublimation and plugging of the line discharging the hydrogen chloride produced
by the chlorination. The reaction pressure is maintained near atmospheric. The
degree of chlorination is principally determined by the time of contact with anhy-
drous chlorine. The contact time varies from 12 to 36 hours for the manufacture
of different Aroclor types, ^e degree of chlorination is measured by the specific
gravity of mixture or the ball and ring softening point when the product is
viscous.
The vapors from the chlorinator (HC1 containing PCBs) are
scrubbed with liquid Aroclor and the gaseous HC1 is sent to another plant at the
Sauget complex for purification. The crude product is held at an elevated tem-
perature and blown with dry air for several hours, after which it is sent to the
raw Aroclor storage tank where a•few tenths of 1 percent of alkali is stirred
20
-------
Sm
or
UJ
8E
10
I
CB
iZ
ce
UJ
i~
CC
0
ce
cc
cc
UJ
3
S3
?y
0
IJ
c
i
c
0
|
X
o
m
f
5
UJ
0
cc
(O
z
c
J
3
3
C
J
1
is
cc
0
t-
•z.
K.
3
X
O
\
— «
1— K
< i
°- °M
I
o: I— i-
S ^
5
ID 5
2 <
^ r- H
/^ ^
F <
5 fc?
< «
CO
2
O
1 w
g
O
(E
0
U
O
CONOENS
Jj
tn
j
j
u
ii
m
1-
(0
OT
O
X
g
1
o
>
n
Ul
iZ
i
£
c
;
(
<.
a
u
2
U
C
2
C
I
7
C
P
<
—
tf
C
cc
c
cc
ft-
c
m-
u
g
<
)
c
1
}
•3
r
i
•
!
i
>
>
>
c
i
i
f »
Lts
r
°5
2 0
$
— O4CC
h- X UJ _
5 m
Jj O CD
s: *-2
ISiL.
Q.
o i- S
>- o
K-
^ h- ^
E o ij
-J en ?<
J
f
21
-------
with the material to react with any remaining hydrogen chloride or ferric chloride.
The air from the blower tank is scrubbed with water and vented to the atmosphere
through a demister.
The raw Aroclor is subsequently batch distilled under reduced
pressure to remove the color, and the traces of hydrogen chloride and ferric
chloride. The methods of purification are different for the different types of
end products. Raw Aroclor 1254, 1242 and 1221, each are distilled in stills under
reduced pressure, achieved via steam jet ejectors; the condensate from the still
is the finished product while the bottoms are the Montars which are drummed and
sent to incineration.
The distillation section for the 1016 product consists of a
gas fired retort and a vacuum distillation tower. The latter is used to allow
the separation of the higher chlorinated, less biodegradable compounds from the
relatively lower chlorinated and more biodegradable ones. The raw Aroclor (42%
chlorinated material) is fed into the reboiler. The vacuum in the tower is main-
tained at about 100 mg Hg via steam jet ejectors. The steam is partially con-
densed and the condensate is discharged into the plant discharge sump. The first
cut from this tower is recycled back to the retort. At a preset overhead temp-
erature the 1016 product is collected and sent to the product storage. The high
boiling residue from the tower is sent to a subsequent chlorination cycle and the
resulting raw Aroclor is distilled in a .still. The overhead from this still is
the finished product. The bottoms from this tower are the Montars, which are
sent to incineration.
For special orders, in order to increase electrical resistivity,
the Aroclors are stirred, at an elevated temperature, with a few tenths of 1% of
well-dried fuller's earth and then filtered through paper.
All Aroclors are stored at 150 °F. Steam coils are used on the
storage tanks for heating these tanks.
3.1.1.1 PCS Production and Usage
The William G. Krummrich plant at the Sauget complex
has an annual design capacity of 48 million pounds chlorobiphenyls. Table 3.1.1.1-1
22
-------
TABLE 3.1.1.1-1 PCB MANUFACTURE & SALES
U.S. production
Domestic sales
U.S. export sales
(Thousands of pounds)
1974
40,466
34,406
5,395
First Quarter
1975
8,532
7,986
1,538
Domestic Sales by PCB Grade
Aroclor 1221
Aroclor 1242
Aroclor 1254
Aroclor 1016
57
6,207
6,185
21,955
10
2,201
2,115
3,660
Predominant Utilization of Aroclors
Aroclor 1221
Aroclor 1242
& 1016
& 1254
Capacitor applications
Transformer applications
23
-------
present data from Monsanto related to production and sales on PCBs for the year
1974 and the First Quarter of 1975. Table 3.1.1.1-2 presents approximate molec-
ular composition of these Aroclors as published by Hutzinger et al (0, Hutzinger,
S. Safe and V. Zitko, "Chemistry of PCBs", CRC Press, 1974).
The majority of the PCBs produced in this plant is
marketed domestically. At present, almost all Monsanto's PCB production is being
used in "closed electric systems" (transformer and capacitor applications).
3.1.2 Raw Wastes
The raw wastes from the manufacturing area consist of the
liquor from the scrubber, the condensate from the steam jet ejectors, water used
for showers and eye baths, miscellaneous floor wash downs, waste oil collected
in drip pans and drums, and montars which are the bottom cut from their stills.
The composition and the quantities of the individual waste stream are not monitored.
All effluent streams generated in the manufacturing area are directed into the
sumps in this area. The waste oil collected in the drip pans and the montars are
emptied into 55 gallon drums and sent to incineration.
The raw wastes generated in the incinerator consist of the
venturi scrubber liquor and the water phase from the separator sump in the incin-
erator area. The composition of the combined stream is monitored. However, the
composition of the individual streams is not known. Non-product PCB discharges
are shown in Figure 3.1.2-1. It has been estimated that this plant generates
about 25 Ibs. of scrap oil and Montar per ton of PCB produced. Additionally
the quantities of material sent to landfill approximates to 5.4 Ibs. per ton of
PCB produced. Additionally, reports that the plant's PCB contribution to air
is under 1 Ibs/day.
3.1.3 Plant Water Usage
On the average, the PCB plant uses a maximum of 388,800 gallons
of water and a maximum of 360,000 Ibs. of steam, daily. Water is used for non-
contact cooling purposes in shell and tube condensers, in a water scrubber, for
floor washings, for showers and in eye baths. Steam is used in the steam jet
24
-------
TABLE 3.1.1.1-2
APPROXIMATE MDLECUIAR COMPOSITION OF AROCLORS
Chlorobiphenyl
C12H10
C12H9C1
C12H8C12
C12H7C13
C12H6C14
C12H5C15
C12H4a6
C12H3C17
C12H2C18
C12HlCl9
C12aiO
Aroclor Type or Grade
(percent composition)
1221 1242 1254 1016
11 <0.1 <0.1 <0.1
51 1 <0.1 1
32 16 0.5 20
4 49 1 57
2 25 21 21
0.5 8 48 1
ND 1 23 <0.1
ND <0.1 6 ND
ND ND ND ND
ND ND ND ND
ND ND ND ND
*ND - Denotes non-detectable
25
-------
cc
JJ
CO
cc
o uj
f_ ^
UJ
CO
K (
* I
0
c
•V
UJ £
UJ Z ^ CC
OR FROM JtT tOtUIUKS —
AUSTS FROM THE SCRUBS
FACE AREA EVAPORATION (1
DENSATE FROM STEAM E
-CONTACT COOLING WATE
c
D
3
2
J
5 «
Q c
§ J
z
o
1- _J
S _j
uj iL
z o
0 Z
- -1
to
z
Q. fL
2
(O £L
IIMMO
TOMS FROM SEPARATOR
LECTION FROM ALL DRl
S FOR CLEANING
«r» r\r~L\f
urv ui\ i - -
MT FILTER PAPERS a
1.3:0: zzcc z i- -J o o uj
< x r> ooo ooo < _i a.
> iu co u z co 2 m o cc u. to
3
H
Q
<
U
0
0
u
u
26
-------
ejectors and for steam tracing purposes. The plant uses municipal water and
purchased steam.
The process water from this facility consists of the liquor
from their scrubber and the steam condensates which are discharged in one of the
two sumps in the manufacturing area.
Additionally, 273,600 gallons of water are used in the incin-
erator daily for quenching the hot gases from the fire box. The resulting weak
muriatic acid in the quench pot is used in the venturi scrubber and in the packed
tower.
The type and quantities of water used and discharged at this
plant are summarized below.
Water Balance
Manufacturing Plant
Process Water
in water scrubber
misc. floor wash downs
condensate from steam jet ejectors
Non contact cooling water
showers, eye bath
condensate from steam tracers
Quantities, GPD
Used
Discharged
14,400
7,200
—
360,000
7,200
—
14,400
7,200
14,400
360,000
7,200
28,800
Incinerator
water used for hot gas quenching
water phase from the sump
Total 388,800
273,600
Total 273,600
432,000
273,600
14,400
288,000
3.1.4 Wastewater Treatment and Housekeeping
Monsanto reports significant environmental controls at their
Krummrich, Sauget plant. Since 1969, they have invested more than 22 man-years of
work and millions of dollars in this program. The in-house goals have reduced the
PCB discharges into water to about three pounds per day.
27
-------
A John Zink designed incinerator was erected at Sauget in 197.
to safely dispose of PCBs. A schematic flow diagram of this operation is given
in Figure 3.1.4-1. Aroclor is steam atomized and fed into the fire box. Natural
gas is used for combustion and the feed is incinerated at temperature above 2200 °F
at 5 percent excess oxygen with a retention time of 2-3 seconds. The gases are
quenched with water and the exhausts from the quench pot are passed through a
high-energy venturi scrubber, then through a packed column which is irrigated by
the weak muriatic acid originating from the quench pot. Exhausts are then vented
to atitosphere through a demister. These exhausts as well as the effluent from the
incinerator section are monitored.
In the incineration area, drainage is directed to trenches
and piping which flow into a 10,000 gallon underground concrete basin. The water
layer from this basin is pumped continuously, combined with the scrubber liquor,
metered, monitored and discharged into the sanitary sewers of the Sauget complex
and from there it is sent to the F,ast St. Louis municipal sewers. The organic
phase from the sump is periodically pumped into waste storage t;anks for inciner-
ation.
The incineration unit has a rated design capacity of 10 million
pounds per year. However, since the start of its operation this unit has achieved
a service factor of about 0.60. MDnsanto reports that this incinerator can achieve
a maximum of 6 million pounds of capacity annually; the unit is plagued with various
itechanical problems.
Monsanto uses their incinerator to process both their own wastes
and as a service to other industries. The service charge for incineration is an
average of 5<= per Ib. of material, but the cost appears to be increasing.
The incinerator feed is brought into the plant either by truck
in 55 gallon sealed drums, by tank trucks or by rail. The drums are opened, picked
up by a fork lift and emptied into a concrete pit. The tank truck carrying the
waste liquids enters the incinerator area and the liquid waste is then pumped from
the truck into the pit. The material in this pit is periodically pumped via a
vertical certrifugal pump in one of the four 20,000 gallon, each, incinerator
waste feed tanks.
28
-------
K.
1
or
LJ
H
O
o.
o
U.
«
O UJ
I- W
o a:
DC
UJ
00
to
Ul
00
<
111
p
(O
UJ
V)
I
o
r>
O
_l
o
(M
X
z
o
z
UJ
-UJ
CO
29
-------
The rail car is brought into a designated area close to the
incinerator site. The material from the rail car is normally pumped into a
long term, 500,000 gallon storage tank. The material from this tank is pumped
into the incinerator feed tanks located on the incinerator pad, when required.
Drainage is provided along the rail tracks. These drains
empty into the 10,000 gallon sump located under the incinerator pad.
In the manufacturing area, MDnsanto has taken a number of
significant steps to prevent loss of PCBs to the environment. Drainage is directed
to trenches and piping, and then to one of two concrete 3,000-gallon underground
settling basins. This insures PCB containment in case of accidental spill or equip-
ment failure. Relief valve lines and atmospheric vents are routed through catch
tanks, or are redirected to settling basins.
When small quantities of PCBs are collected in the settling
basins of the manufacturing area they are later pumped into 55 gallon drums, and
eventually incinerated. The overflow from these sumps is combined with the non-
contact cooling water used at the plant, monitored and then discharged into the
Sauget complex's sanitary sewer and from there to the East St. Louis municipal
system.
PCBs are packed and shipped in galvanized-steel 55 gallon drums,
or in railroad tank cars. All tank cars are top loaded. In the drum filling
area spills are cleaned via rags or floor dry and these materials are drummed and
sent to landfill located in the town of Sauget.
In the PCB truck or rail car loading area drainage is directed
into a small concrete pit. The material accumulated in this pit is periodically
pumped into the basins located in the manufacturing area.
Nitrogen blanketing is provided on storage tanks to eliminate
any "breathing" of the tanks and resultant PCB escape.
Mist eliminators have been installed in vapor lines to eliminate
the possibility of PCBs leaving the manufacturing area through these lines.
30
-------
Finally, underground sewers have been replaced with above-
ground sewers, and repaired or combined with others, so that the effluent from
the department can be monitored. In addition, this step will prevent any unknown
buildup of PCBs in the sewer systems or any contamination of PCBs into other
sewers.
A high housekeeping level is maintained in the plant itself.
Housekeeping responsibilities which the operators have assumed are as follows:
. All pumps are checked for leakage on every shift.
Drip pans that collect leaks are emptied into
scrap PCB drums.
. All leaks are reported and documented so that
corrections can be made and settling basins
observed.
. "Floor Dry" is used to absorb any PCBs that
have spilled or leaked. If it becomes necessary
to flush PCBs to the settling basin, a minimum
amount of water is used.
. Sampling drums and scrap PCB drums are quickly
palletized, labelled and transferred to the in-
cineration area.
3.1.4.1 Treatment Facility for the Effluent fromJ3auget
Complex
The processing and incineration departments' aqueous
effluent enters the plant sewer system, and this system discharges into the Sauget
Village waste sewer system. The combined streams then flow to the village primary
treatment plant. The village treatment plant is under expansion to a secondary
chemical treatment plant, scheduled for 1976 completion.
Additionally, evaluations are being conducted to
include the village plant discharge in a projected regional biological treatment
plant.
3.1.5 Plant Effluents
This plant has no point source discharge from their operation.
There is a single discharge from the manufacturing operation (the combined stream
of process and non-contact cooling water) to the main sewer system of the Sauget
31
-------
complex and there is a second discharge from the incinerator area to the same
sewer system. The composition of these streams as reported by Monsanto are as
follows:
Effluent from the Effluent from the
Manufacturing Incineration
Operation area
flow rate, gpd 432,000 288,000
PCBs, ppm 0.75 0.15
PCBs, Ibs/day 2.70 0.36
It has been reported that these effluents are clear liquids
with essentially no suspended solids. The incinerator effluent may contain some
amounts of chloride. However, no information is available on the chloride content.
3.2 Askarel Capacitor Manufacturing Industry
Presently 90-95 percent of all capacitors manufactured in the U.S.
are of the PCB impregnated type. Two important types of capacitors are phase
correctors on power lines and ballast capacitors for fluorescent lighting. Aroclor
1016 is the principal PCB used in this application; at some plants Aroclor 1242
and 1221 are also being used in limited quantities. The principal types of Aroclor
impregnated capacitors and their applications are given below.
A. High Voltage Power
Generally AC capacitors are used to improve the power factor of a circuit.
Power factor is the ratio of true power in watts to the apparent power as obtained
by multiplying the current flowing to the load by the circuit voltage. The power
factor correction can be made directly at the load or at utility substations. In
the latter case high voltage units will be designed for 4,800 to 13,800 volt service.
To the utility engineer the use of capacitors is purely a matter of economics. The
main benefits that result from the use of capacitors are:
1. Reduction of losses associated with the delivery of electrical
power to the point of use.
32
-------
2. Reduction of the investment required in equipment for delivering
electrical power to the point of use, which may be broken down
into:
a. Reduction of line current for the same kilowatt load.
b. Reduction of the kva rating of equipment required to
handle the same kilowatt load.
c. Reduction of the transmission line voltage drop for
a given kilowatt load.
d. Control of delivered voltage if the capacitor kva is varied.
Electric utilities also use capacitor banks in series with distribution
circuits to irrprove voltage regulation.
B. Lew Voltage Power
Capacitors installed in industrial plants at the demand site (typically
large motors and welders) are designed for 230 to 575-volt service. Capacitors
installed near the loads are the most efficient way to supply the magnetizing
current to produce the flux necessary for the operation of inductive devices.
Rates for the sale of power are generally structured to encourage power factor
correction at the site, eliminating the need for the electric utility to trans-
mit both power-producing current and magnetizing current all the way from the
generator to the plant site.
The same considerations apply to induction heating applications, the
principal difference being that capacitors for this application are designed for
operation at 960 to 9600 Hz.
C. Lighting
Capacitors improve the efficiency of lighting systems. A fluorescent or
mercury vapor lamp can be ballasted without the use of a capacitor, but the power
factor of the lighting system would then be in the range of 50 to 60%. For commer-
cial or industrial lighting with either fluorescent or high intensity discharge
lamps, the use of a capacitor in the circuit provides part of the lamp ballasting
and brings system power factor into the range of 90 to 95%.
33
-------
D. Air Conditioning
As in the lighting applications, the capacitor improves system efficiency.
Air conditioners could be made to operate without capacitors, as do hone re-
frigerators, but because of the higher capacity required for air conditioners,
the resultant line current would virtually eliminate home "plug-ins" and would
still further overburden a seriously threatened national power network. Almost
all air conditioner pump motors are of the split-winding type on which the
capacitor provides phase differential for the so-called start winding, thus
delivering good starting torque. The proper size capacitor permits high (90%)
power factor after start-up.
E. Industrial Electronics
This market category is a catchall covering many varied applications, two
important ones being motor run and power supply applications. Motor run appli-
cations are for pumps, fans, and farm feed equipment, and do not differ signifi-
cantly from air conditioning applications. The power supply market uses capacitors
principally to provide high power factor, but through careful design the capacitor
can also provide wave shaping where desired.
3.2.1 Askarel Capacitor Manufacturing Plants
There are seventeen companies in the U.S. which manufacture
askarel capacitors at nineteen plants. The name and the location of these plants
are given in Table 3.2.1-1. Some plants manufacture industrial capacitors only
and others manufacture power capacitors.
Ten major companies were contacted. Detailed information was
obtained on six plants. Four plants were visited. The data herein represents
approximately 50 percent of 1974 PCB usage in capacitor application. The PCB
usage in this category was approximately 22,000,000 Ibs. in 1974. The PCB usage
in individual plants is considered by these companies to be confidential infor-
mation. The range of PCB usage at these plants was 64,000 fold in 1974. Plant
ages range from five to thirty-seven years.
Capacitors used in lighting and air conditioning applications
contain 0.05 to 1.0 Ibs. of Aroclor. The largest power capacitors contain about
34
-------
Table 3.2.1-1
U.S. Capacitor Manufacturing Industry Using PCBs
Company Name
(In order of PCBs Usage)
General Electric Company
Westinghouse Electric Corp.
Aerovox
Universal Manufacturing Corp.
Cornell Dubilier
P.R. Mallory & Co., Inc.
Sangamo Electric Co.
Sprague Electric Co.
Electric Utility Co.
Capacitor Specialists Inc.
JARD Corp.
York Electronics
McGraw-Edison
RF Interonics
Axel Electronic, Inc.
Tobe Deutschmann Labs.
Cine-Chrone Lab, Inc.
Location of the Plant
Hvdson Falls, N.Y.
Ft. Edward, N.Y.
Bloomington, Ind.
New Bedford, Mass.
Bridgeport, Conn.
, N.J.
New Bedford, Mass.
Waynesboro, Term.
Pickens, S.Carolina
North Adams, Mass.
LaSalle, 111.
Escondido, Calif.
Bennington, Vt.
Brooklyn, N.Y.
Greenwood, S. Carolina
Bayshore, L.I., N.Y.
Jamaica, N.Y.
Canton, Mass.
Palo Alto, Calif.
35
-------
77 Ibs. of Aroclor. The most popular size contains 36 Ibs. Capacitors are not
rebuilt and returned to service after failure. They are disposed of and replaced
by new units. G.E. reports that high voltage utility capacitors, low voltage
power capacitors, and induction heating capacitors are manufactured at a rate of
200,000 per year, about 2 to 3% of which are replacements; the balance are for
new installations. The current market for capacitors used in lighting appli-
cations is about 44,000,000 units annually of which 10% are estimated to be
replacement ballasts. The current market for capacitors in air conditioning
application is above 12,000,000 units annually, with 5% of these estimated to
be for replacement usage. The market for capacitors in industrial electronics
applications is estimated at 23,000,000 units per year with no estimate as to
the relative size of the replacement market.
3.2.1.1 Askarel Handling
In most plants PCBs are shipped via tank car to a
rail siding several miles from these plants. Individual plants provide a de-
signated tank truck for the transfer of the PCBs from the rail yard to the manu-
facturing plant. To la'-ge plants, PCBs are brought via rail cars into the plant
site.
In most plants PCBs are unloaded and transferred to
the PCB storage tank without the benefits of any curbs or dikes. At very few
plants, the unloading operation of the PCBs from the tank car to the storage
facility is confined and the receiving area is roofed and diked.
PCBs from the raw storage PCB tanks are filtered
through fuller's earth and stored in finished product storage tanks. PCBs are
next pumped from the storage tanks to the impregantion areas for use. Excess
PCBs from these areas are recycled back to a designated tank and from there either
filtered and reused or, if defective/ it is pumped into the scrap storage tank.
Spent fuller's earth employed at these plants is
either drummed and stored at the plant site or is sent to a landfill.-
36
-------
3.2.1.2 Process Description
Jfost plants in this category manufacture either large
power capacitors or small (including less than 2 Ibs, of PCBs) industrial type
units. The large capacitors are either flood filled or manifold filled. All
small capacitors are flood filled either in a vacuum tank or in an automatic
"carousel" arrangement where loading and unloading occurs at one station, and
the capacitors in each cell are progressively dried, evacuated and filled.
A generalized schematic flow diagram for the manu-
facture of large capacitors and small industrial capacitors are given in Figures
3.2.1.2-1 and 3.2.1.2-2, respectively. The basic manufacturing operation at these
plants can be divided into two major operations.
a) Non PCS related operations consist of the following steps:
. Fabrication of capacitor cans, covers and brackets from
sheet aluminum or steel. Smaller plants, however, purchase
capacitor casings, capacitor terminals, connectors and foil.
. Vapor degreasing or detergent washing of metal components.
At some plants, ultrasonic cleaning is used to clean smaller
components.
. Roll winding the capacitor paper or polypropylene film with
aluminum foil.
. Complete assembly of capacitor components and sealing wi
covers. In some units fill holes are provided for PCB intro-
duction; in others PCBs are introduced through bushings or
a valve.
. Leak testing via Jtiass spectrometer and pressure testing prior to
vacuum drying.
b) PCB related operations
Three completely different impregnating techniques are employed
by this industry.
37
-------
a 5
*£o£
3l*-«
Jf=!o
fel°«
III
2«
3-
g
(0
r
cc —
occ
oo
g»:
H E
o
5°
<
o
~~ 0 ^0
x- gt
o — • • 52
5 8S;
^ <0
Ul
1-
cO
1
K o
° C3 i4
Z —v.
li
XUI
z
) e>
[UJ
• it
Q.
s
—
w~
3
U.
Q
3
j_
OQ
O
0.
UJ
or
UJ
_i
£
C
h
I
1
0
C
F
«
0
i-
.
e
L
C
<
C
C
1
I
t
X
o
K
U,
a
tf
K
o
•
5
j
i
i>
5
r
c
j
e
u
3
t
3
O
J
i.
5
?
00 =
aoo-
3§
oz
"-"• ^
Z-"
1"-
1-
(T
<
UJ
"x
t- UJ
Z -J
UJ-J
0.3
to u.
38
-------
u.
r!
1 TRICHLORO.
I RECOVERING
lu
£c
1 CAUSTIC 1 R
1 IH.
. C~
CAPACITOR
ASSEMBLY
a
SOI DFRINfi
WASTE
»•
^^
•^
H VAPOR DECREASING \
1
1
1
IP
1^
ASSEMBLED U
Z
ij
S
L~l 1-
o^
'1 —
1
o -
?£
H3
a. _
e sasAOOj jo NOIiVOiaaVj f""*" " HflNiwrnv
_ __.J
•ft
El
Si-
$\
»J
n
5
9
£
0 gi1
."d| DETERGENT WASHI
H20
L_t
ric. 'WASTE
SCHARGED
-^v— "5
55
<
0
CO
s
J
'
*.. S
°li
^2
-> x
j
j
Lj
5
j
>
9
j
:
t-
<
a
u
c
u
c
<
a
V
a
k
tr
<
u
_i
<
a:
X
o
s.
to
t
f ^Uss-
:_l-l IS
•S=:|
—1 1
o^ 1 ~"
S§' q «
Z. »vx
~~ X o
>-
on
g
o
CVJ
U-
o
or
O
s
ga
UJ
UJ I
a: H
Is
u, a.
a:
3
u.
a
UJ
M
^
cc
UJ
O
39
-------
(1) Conventional flood filling of capacitors
This type of filling operation is used for impregnating large
power and small industrial capacitors. Here the capacitors are
arranged in baskets or in large tanks which are subsequently
transferred into vacuum chambers. Vacuum is then drawn and the
capacitors are kept under vacuum at an elevated temperature for
a specific period of time in order to evaporate the moisture from
the capacitor interiors. The temperature of the tank is next
lowered and PCBs are introduced under vacuum. The capacitors are
allowed to soak for some time after which they are transferred to
the sealing or crimping station via mobile car, sealed, and excess
PCBs are then drained. The sealed capacitors are next sent to
vapor degreasing or detergent washing for cleaning the exterior
of these units. The clean units are heat soaked and heat tested
in an oven and then sent to 100% electrical testing.
Subsequent to electric testing, the capacitors with steel casings
are pretreated prior to painting. The pretreatment is either vapor
degreasing for painting in an electro-static field or it is
several stages of phosphatizing and rinsing operations, prior to
automatic spray painting, where water is used to scrub the vapors
generated during painting. Most capacitors with aluminum casings
are marketed unpainted. However, at some plaints a portion of capac-
itors with aluminum casings are fluoride treeited and then spray
painted. The capacitors are next either air dried or dried in an
infrared oven prior to packaging and shipping.
(2) Automatic Flood Filling Operation (Carousel)
In this method of capacitor impregnation, the assembled units are
placed in ovens for the removal of moisture from the capacitor
interiors. The baskets containing these capacitors are next trans-
ferre:1 to the carousel chamber which is at the loading position.
The carousel passes through 13 subsequent cyc3.es consisting of
40
-------
various degrees of vacuum applications and then subsequent PCB
filling and soaking positions. The PCBs are next drained out
of the chamber, and the capacitor baskets are removed from the
vacuum chamber and tilted to drain excess PCBs in a sump located
adjacent to the carousel loading and unloading station. The capac-
itors are next transferred to the crimping or sealing stations.
From this point on the remainder of the manufacturing steps are
identical to those employed in conventional flood filling operation.
In general, vacuum in conventional flood filling operations is
achieved via mechanical pumps, whereas vacuum in each automatic
flood filling station is achieved via a vacuum pump and several
stages of steam jet ejectors.
Rejected industrial capacitors are disposed of in 55 gallon drums
either directly or after disecting. Rejected power capacitors
which are flood filled are returned to the salvaging operation.
Here the capacitor covers are cut on an end mill, the PCBs drained
and the interiors examined to find the cause of the electrical
failure. Where possible the rejected units are then repaired and
reprocessed. Otherwise, the capacitor interior is discarded in a
designated solid waste drum. The capacitor casing is sent to vapor
degreasing and it is subsequently sold to a scrap dealer.
(3) Manifold Filling Operation
At some plants, large capacitors are filled by this method. Here,
the capacitors are completely assembled, sealed and vacuum tested;
then each unit is filled individually under vacuum through a mani-
fold. The capacitor valve is next sealed, the exteriors spot cleaned
with trichloroethylene and then these units are transferred into an
oven where they are heat soaked for a given period of time prior
to electrical testing. Those capacitors which pass the electrical
tests successfully are vapor degreased, painted, dried, packaged and
shipped* The rejects are returned to the filling station where
41
-------
the majority of the PCBs are retrieved through, the fill valve
under vacuum. The rejects are next transferred to the salvage
operation.
Mechanical pumps are used throughout the system to achieve
vacuum.
3.2.1.3 Raw Wastes
The raw wastes originating from these plants consist
of scrap PCBs collected in sumps, drums and drip pans, contaminated vacuum pump
oils, the fractionator bottoms from the trichloroethylene recovery, the caustic
bath used, at some plants, for purposes of paint stripping, spent detergent wash
and rinse water from capacitor or component cleaning operations, rinse water used
in the welding and plating operations, steam condensate from jet ejectors, the
seal water used in vacuum pumps, water used in phosphatizing and fluoride baths,
water spray used in the paint booths, boiler blow downs and cooling tower blow downs.
Furthermore, at least in one plant: in this category
water infiltrating at a subgrade elevator shaft creates an additional waste stream.
Solid wastes from these plants consits of scrap capac-
itors or capacitor interiors, floor dry and rags and newspapers used for cleaning
spills, spent fuller's earth and other filtering media used for PCBs filtration
and waterless cleaners used at some plants for hand cleaning purposes.
Non-product PCB discharges, their source and disposition
are given in Table 3.2.1.3-1. Adequate data to quantify the various wastes generated
at these plants is not available. Information on waste loads reported by some
plants are given in Table 3.2.1.3-2.
3.2.1.4 Water Use
The water used at these plants is in the range of
12,500 to 1,260,000 gallons per day. In most plants, water is used primarily for
once-through non-contact cooling in vapor degreasers, in pumps and in vacuum tanks.
Plant 101 recycles 4;heir non-contact cooling water through a cooling tower.
42
-------
« o o S
01 K; is "i
C "i
(q -H
4J 'H
O £
o °i! nj'
.5 q
i< S
Vacu'-n
Ev-Efcra
Vapors
O O
•H H
4J +J
« cfi 2 w
tn o
>i -0
**" v
"*
• _
a fr u- a
0 "J M 4J
ll :' 4a! c
o) ri iJ > 4J >
c -p t; -ri c -H
2 v< d •, g M
S c 6 e B c
jMaJNa
e
w
p
-P
lO
H ^1 H
1
4J Vi
C OJ J-J
(a £> CJ
* t d
gCP nffio-J-J
c: w
3&
n'3
O rH
3d| a
§2|p
Q O
2 fyt) iJ
*d '» G
(U -r-J id -P
•a-Sr3 R
s &3 -a
&88J g
c ,d a
d9yod
ti ,£ « H
& S -:J 8 -1
* k cu u' ^
.<"!>, s-as-
-O 01 H U
u in p H
!„' r
43
-------
TABLE 3.2.1.3-2
QUANTITY OF WASTE LOADS
(Ibs/ton of KB used)
Plants
Land Destined Waste
Wastes to Incineration
Spent Fuller's Earth to
Incineration
Notes:
(1)
(2)
(3)
(4)
(5)
100
(1)
73
100
4.1
101
129
56
None'2'
102
65
98
104
J4>
109
105
127
106
.(5)
300
Unknown
Reported as PCBs. We estimate an additional 175 Ibs/ton of PCB con-
taminated solid material which is being landfilled locally.
PCBs are filtered through packed Porocell cylindrical filters. The
filter media has an estimated life cycle of 20 year.
This plant uses clay for PCB filtration. The quantity of spent clay
sent to incineration is not known.
These plants store their solid wastes consisting of spent fuller's
earth, rejected capacitors, rags and floor dry, on site. Plant 104
generates ten 55-gallon drums of waste and plant 105 forty 55-gallon
drums of waste each week.
The weight of materials sent to landfill is not known. However, this
company reports that during 1974 approximately 650 capacitors were
rejected. The rejected units were drained and then sent to a landfill.
44
-------
Process water used at various plants consists of water
used in detergent washing, in ultrasonic washing, condensate from the steam jet
ejectors, seal water from the vacuum purrps, water used in phosphatizing and in
fluoride baths and rinses, and water used in caustic treatment and painting opera-
tions. Additionally, water is used for sanitary purposes and in some plants it
is also used as boiler feed.
quantities and types of water used at these plants
are given below:
Plants
Intake gallons/day
Water Usage in gallons/day
Non-contact cooling
Process water
Rinse water from hot
solder dip
Detergent washing &
rinsing
Phosphatizing baths
Water used in tin plating
Fluoride bath & rinse
Water used in painting
Pvirp seal water & steam
condensate
Sanitary
Boiler feed make-uo
100
740,'000
101
12,500
102
104
105
655,000 3,000
200
8,000
44,940
40,000
1,500
52,000
25,000*
-
-
-
10,000
30,000
200
400
12,000
50
50
44,000
65,000
15,000
Notes: (1) 560,000 gallons/day of this water is used as process rinse water
in non PCS associated, electrolytic capacitor operation
(2) Make-up water
(3) It has been estimated that only 150,000 gallons of this water is
associated with the plants PCS operation. The balance is used in
the manufacture of Mica capacitors
(4) It includes water used in paint stripping and welding operations.
106
1,000,000 1,260,000 480,000 336,500
388,000 1,195,000 387,000 326,500
5,000
5,'000
45
-------
3.2.1.5 Wastewater Treatment
The najor in-plant PCB wastes which reach water streams
originate in the impregnation areas. Due to the nature of water requirements in
this industry, at some plants it is possible to significantly reduce the quantities
of wastes entering the water systems by adoption of proper housekeeping measures.
Since there is no demonstrated technology available for the terminal treatment
of PCBs in wastewaters, there are no PCB wastewater treatment techniques currently
in use at any facility. Details on wastewater treatment techniques used
at various facilities for purposes of oil recovery or neutralization and on-going
containment programs undertaken by some plants to prevent the entry of PCBs into
the enviroruTient are covered below.
Plant 100
In 1971-75, in-plant modifications were initiated to control PCB discharges
from this facility. These modifications consist of repiping for scrap recovery,
consolidation of processing and PCB handling area, and the installation of an oil/
water separator for the recovery of contaminated oil used in their vacuum pumps.
The water from this separator is discharged into the municipal sewers. The re-
covered oil which contains PCBs is used as supplemental fuel in their boiler.
The plant's water effluent, consisting of primarily once-through non-contact
cooling water, is discharged into a single trough on their property and from there
it enters into a river. Their air conditioning condensate is discharged into a
second trough and from there into the river.
Boiler blowdowns and water from the oil/water separator is combined with
their sanitary water and discharged into the municipal sewers.
The only containment measure underway at this plant for prevention of PCB
spills is at the Aroclor unloading area where provision is being made for a welded
metal pan (31 long x 18" wide x 1" high) to catch pump oil and Aroclor spills.
This plan will have a weather tight cover and will be periodically emptied and
the contonts Irc.eierated.
46
-------
Plant 101
This plant segregates non-oontact cooling water from process water.
The non-contact coolina water is recycled back to the system through a
cooling tower. The process water consists of the water used in the detergent
washing operation. The spent detergent wash is drummed, sealed and sent to a
landfill. The rinse water from this operation is sent to the municipal sewers.
This plant practices "control sink" measures. Their employees are advised
to use waterless cleaners for cleaning their hands prior to washing.
Plant 102
In addition to askarel and non-PCB impregnated capacitors, plant 102 manu-
factures electrolytic capacitors. Therefore, this plant has a three pond system
for the treatment of the plant effluent. The entire plant water discharge, with
exception of their sanitary water, is sent to a large concrete basin for pH
adjustment. The overflow from this basin is routed to a 200 ft x 160 ft x 20 ft
deep lake which has a retention time of 2 to 3 days. The overflow from the lake
is sent to a 380 ft x 230 ft x 20 ft deep aeration pond which is equipped with
mschanical aerators for BCD control. The overflow from this last pond enters the
town creek and eventually ends in a river. Additionally, an oil skinner is in-
stalled upstream to the treatment plant to remove oil from the effluent.
The plant is currently undergoing several modifications to minimize PCB
entry into the river. The containment pits, located within the impregnation areas,
are being sealed and access to the drainage system removed. Cooling water from
pumps and tanks is being rerouted and dumped elsewhere. Vapor condensates, oil
drips from vacuum equipment, and spills and scrub water are collected and held within
these containment pits.
Plant 104
The only effluent treatment currently employed at this plant is a decanting
operation used for the treatment of the paint stripping rinse water and the ground
water infiltration.
47
-------
Recent in-house modifications related to PCS control are as follows:
. A discharge line which passed under the tank farm
area has been plugged and replaced with a new line
bypassing this area to avoid further contamincition
of the effluent by infiltration.
. The discharge from the welding wash operations is
segregated from storm drainage and diverted into
the main plant outfall.
. The caustic bath, which was previously discharged into
the river, is drummed and sent to incineration.
. Curbing and diking have been constructed around the
tank farm area.
. An automatic monitoring station has been installed on
the major outfall from this plant.
Additionally, as a housekeeping measure, two sinks in the impregnation room
and one in the salvage recovery room used by personnel handling PCBs are con-
tained. The wash water from these sinks is drummed and sent to incineration. Other:
sinks in the adjacement areas are designed as controlled sinks and personnel using
these sinks have been instructed to use a waterless hand cleaning material prior
to, washing their hands.
Other containment programs to prevent the entry of PCBs into the environment
are underway at this plant. The projected completion date of the entire program
is by the end of 1976. Major modifications planned are as follows:
. Segregation of cooling water from process water and the
installation of cooling towers for the recycling of the
non-contact cooling water.
. Consolidation of all discharges to achieve a single plant
outfp.ll.
48
-------
. Elimination of underground roof drain systems.
. Provisions will be made for the treatment of plant
exhausts from all PCB related operations using a dry
filtration system.
„ Provisions will be made for a covered sheltered area
for PCB loading and unloading area to prevent the rain
runoffs from this area.
Depending on the success of the above mentioned modifications, a terminal
PCB treatment facility may or may not be installed at this facility.
Plant 105
Recently this plant has made several modifications at their various manu-
facturing steps in order to minimize the PCB entry into the environment. The
detergent wash and rinse water, which were originally sent to the river, are
now contained and incinerated. The trenches near the detergent wash machines
nave been plugged to prevent discharge into the river. These trenches are
periodically pumped out and the material sent to the scrap oil tank. Containment
sink and control sink operations have been instituted in and around the impregna-
tion areas. The condensate from the jet ejectors and the seal water from the
vacuum pumps are directed to a 55 gallon drum which contains kerosene for the
extraction of PCBs from the contaminated water. The decant from this drum is
discharged to the river. At the tank farm area all gravel surfaces have been re-
placed with concrete.
Other containment programs to prevent the entry of PCBs into the environment
are underway at this plant. The projected conpletion date of the entire program
is by the end of 1976. Major modifications planned are as follows:
. Segregation of non-contact cooling and process water and
provision for recycling of cooling water through a cooling
tower.
. Segregation of uncontaminated storm water runoffs, with
direct discharge to river.
49
-------
. Provision for exhaust treatment using a dry filtration
system.
. Collection of all sanitary water and the treatment of
this effluent with activated sludge, aeration, clarifi-
cation and chlorination prior to the terminal treatment
for PCB removal.
. Filtration and carbon-adsorption of the plant's process
water, sanitary effluent, and contaminated storm water
runoff to achieve best practical levels of PCB treatment
prior to discharge of the plant effluent into the river.
Installation of a recycling system on their detergent
washers. This system consists of gravity settling and
sand filtration and is anticipated to minimize the fre-
quency of detergent wash incineration.
. A garage shelter area will be erected at PCB unloading
area. Additionally, the underground scrap PCB tank will
be replaced by an above grade and larger tank.
. A recycling system is also planned for the vacuum pump
seal water to minimize the quantity of process water
discharged.
A three-^week pilot operation of activated carbon adsorption, preceded by
filtration, has been just completed at this plant. The mobile pilot unit was
provided by a vendor. The samples from this operation are currently being
analyzed to evaluate the performance of the system as a terminal treatment system
for PCB removal.
Plant 106
This plant practices no end-of-pipe wastewater treatment techniques. The
in-house PCB control operation is, however, an efficient means of eliminating con-
taminant losses to process or to sewers.
50
-------
At this plant the raw KB storage tank and its appurtenant distribution
pump/filtration system is located within the confines of an approximately 52,000
gallon containment basin. Approximately 35,000 gallons of capacity in this basin
is available for collection of scrap PCBs. The PCB contaminated material from
this basin is ultimately shipped to a contract incineration facility.
The excess PCB from the impregnation system is drained by vacuum through
an overhead pipe to a return sump and is pumped back through a Hilco filter to
the storage tank. The interiors of the impregnation tanks have collection pans
below the capacitors and spilled PCBs are drained from these pans into five gallon
buckets for return to the contaminated storage sump. The PCB spills to the floor
of the impregnation tanks are drained through the piping systen back to the con-
taminated PCB sump. Oil dry is used outside of the impregnation tanks to minimize
tracking.
Any byproduct or spillage resulting from the filtering operation is contained
in the collection sump. The filter media is currently disposed of by landfill.
In addition, no water use is allowed in PCB handling areas. However, the
impact of personal contact (that is, carrying it to wash basins, home, etc. on
shoes, clothing and skins) may be a significant source.
The plant's tank truck unloading area is a curbed bay approximately 15 by 68
feet. Storm water that collects in the unloading bay can either be discharged onto
the ground adjacent to the bay or into the 35,000 gallon contaminated PCB recovery
sump. The valves from the storage bay area are padlocked and chained; therefore,
only authorized personnel operate the valves. The overall goal that plant personnel
try to achieve is to minimize the storm discharge from the bay by allowing evapora-
tion. If a large quantity of water is believed contaminated, the drain valve to
the contaminated PCB collection sump is opened.
Personnel representing plant 106 are of the opinion that there is no way to
achieve zero discharge of PCB type contaminates to the environment. Therefore,
there efforts have been directed toward development of a replacement type dielectric
as opposed to exploring treatment systems for potential discharge problems at their
facility. The plant personnel believe that with cooperation of a dielectric fluid
51
-------
supplier they have developed a suitable replacement product for PCBs. Attached
as Appendix A is the company's positive statement on this si±:>stitute.
3.2.1.6 Effluent Conposition
Effluent from plants in this category range between
2500 to 1,260,000 gallons per day. As indicated in the waste treatment section,
these effluents are sent either into receiving waterways or into municipal sewers
without any PCB treatment. Available information on effluent flow rates and their
PCB contents, for most plants in this category, are given in Table 3.2.1.6-1. More
detailed effluent information were obtained on four plants and Table 3.2.1.6-2
compares the discharges from these plants. Table 3.2.1.6-3 lists PCB concentra-
tion of the plant intake water from various sources.
It is of interest to note that one plant (plant 100)
reported large variations in the PCB content of the receiving waters even at the
time when the plant was not in operation. They attributed 1±ie variations to
storm conditions, stirring-up the sediments from the bottom of the river and
greatly increasing the PCB concentration in the water phase.,
3.3 Askarel Transformer Manufacturing Industry
All plants in this category manufacture transformers using mineral oil
as the dielectric fluid. This dielectric fluid is the principal irtpregnant used
at these facilities. PCB transformer oils (blends of 60 to 70 percent Aroclor
1254 or 1242 and 40 to 30 percent trichlorobenzene) are used in only 5-10 percent
of these plants' manufacturing volume.
In general, a transformer consists of a core and coil immersed in a
dielectric fluid. There are two broad classifications of transformers: distri-
bution transformers, which are used to step down voltages, and power transformers,
which are used primarily to step up voltages. The precipitation power supply units
can be actually classified as a third class of askarel transformer. However, the
larger plants in this category have only two manufacturing departments — distri-
bution and power; and they manufacture the precipitator transformers in their
distribution transformer department. The following types of askarel transformers
are manufactured at these plants:
52
-------
TABLE 3.2.1.6-1 RANGE OP FLOW RATES & PCB
CONCENTRATION IN EFFLUENTS
FROM CAPACITOR MANUFACTURING
PLANTS
Plant H
100
101
102
104
105
106
107
108'
109
110
1113
112*
113
126
Notes:
Effluent
Flow Rate
fo. Outfalls gal /day
001
None
001
002
003
Major outfall 1
Cooling water
6 roof drains
Lab & roof
drains
Compressor cooling
water
Paint strip rinse
Major outfall
From spray piint
booths
None
None
Cooling water
& boiler blowdown
Cooling water
Roof runoff
Storm runoff
001
None
None
None
001
None
800,000
9,000
807,000
165,000
36,000
,060,000
100,000
32,000
72,000
50
650,000
1,500'
336,500
150,000
48,000
129, OOO2
2,900
unknown
100,000
2,570
18,000
10,224
12,000
580
Effluent
PPb
avg/max
17/55
161/161
9
3.2
3.4
21
7.8
4.4
3.7
4800
370
13,300
26.6
1900/4000
45/108
338/705
1360/2370
3/11
57/83
821
8.7
130
14.3
6200 3
PCB Content
Ibs/day
avg/max
0.1134/0.3670
0.0121/0.0121
0.0605
0.0044
0.0010
0.1856
0.0065
0.0011
0.0022
0.0002
2.0058
0.1664
0.075
2.3769/5.7546
0.0180/0.0432
0.3636/0.7585
0.3290/0.0573
0.0475/0.0692
0.0176
0.0013
0.0110
0.0014
0.03
Disposition
of effluent
River
Sewer
River
River
River
River
River
River
River
River
Rive-r
River
Sewer
Sewers
River
Sewer
River
River
River
Sarer
Sewer
Sewer
River
Sewer
(1) Discharged once every 4-5 weeks.
(2) Contains 10,000 gpd sanitary sewage and
8000 gpd of surface water infiltration
6) These facilities recycle most of the non-contact cooling
water used at the plant. These effluents consist of a small
fraction of the cooling water plus their sanitary discharge.
53
-------
vo
8
p1
(Q
C^t .£?
vo ft
• -H
rH O
•
• 0
n
0) o
rH v)
(Q **H
I
I
IT)
S
O
rH
O
rH
O
°
1
t
*3* CO
o in
in
vo
rH
VO H
O CO
vo
P3 CM
ro I
0 0
o •
*r
CO
s ?
o in
o •
in
0
C>J
(M rH
o in
VO
ro x
5
^
o
o
V
!
en
O
CM
Oj
in
vo
vo
H
ro
•
o
o
ro
O
CM
§
0)
rH
*
I
1
54
-------
CM] -rl
rHl O
a
rH| -H
rH O
rHl -H
\0\ -H
0| O
m
CM
CO
J>
•
rH
•
CNJ
*
m
I
5
(0 I
B
o fl)
o
d
r-
in
1
rHI -H
o u
r-
l
I
Ol -H
o
m
•
CM
C^
O
•
o
1
55
-------
A. Askarel Distribution Transformers
1. Network (Receive up to 14,400 volts and deliver 120,
240 and 480 volts)
2. Pad - mounted (Receive up to 14,400 volts and deliver
120, 240 and 480 volts)
3. Pole - mounted (Receive up to 14,400 volts and deliver
120, 240 and 480 volts) The application of these trans-
formers in power distribution systems places a great
premium upon their reliability and high overload
capability.
4. Precipitator power supply (Receive 480 AC volts and
deliver 50-60 kilovolts low amperage DC) These units
are generally installed close to hot gas stacks in
an atmosphere that would be a fire hazard to oil-insulated
transformers and a corrosive hazard to open dry-type
transformers. Sealed dry-type transformers are impractical
for high-voltage DC.
Voltages listed cover the majority of askarel distribution trans-
formers. Product scope includes up to 34,500 primary volts and virtually any
secondary voltage, 34,500 and below.
Quantities of askarel used in this class of transformers are in
the range of 500 to 5,000 Ibs. in each unit depending on -the rating and the size
of the transformer.
A great majority of distribution type transformers have pro-
vision for venting. Many of these units are equipped with spring loaded venting
devices which vent upon a pressure excursion, and diaphragm rupture discs are
offered as a customer option.
B. Askarel Power Transformers
1. Secondary substation
(a) Load center units
(b) Secondary substation generation units
56
-------
(c) Switchboard units
(d) Internal units
(e) Motor control units
These 5 comprise the largest group of askarel-
insulated transformers and they find widespread
application in the automobile, paper, chemical,
textile, steel, nonferrous metal, cement, mining
and petroleum industries. They are used in com-
mercial and public buildings, such as schools
and hospitals; in defense and nuclear energy
installations; and by private and public utilities.
2. Master unit substation
3. Primary unit substation
4. Limited ampere substation
5. Industrial furnace
Furnace transformers used in the hot, dirty atmosphere
in proximity to glass melting and induction furnaces, which require high current,
low voltage power supplies (receive up to 15,000 volts and contain 2000 to 4000
Ibs. of askarel each). Existing technology does not permit construction of
sealed dry-type transformers for these power ratings.
6. Rectifier
Rectifier transformers used for large rolling mills and DC
industrial power supplies (receive up to 15,000 volts AC and deliver low voltage
high amperage DC. Each unit contains about 19,000 Ibs. of askarel). These units
are covered by the same conments given for industrial furnace transformers.
7. Transportation
Railroad transformers used on-board in electric locomotives
or multiple unit electric railroad cars (receive up to 25,000 volts and contain
700 to 2400 Ibs. of askarel in each unit depending on the rating and size of the
transformers).
57
-------
(a) Third rail
These transformers are used for rapid transit
systems, and are basically serving a rectifier
function
(b) Locomotive
Prior to 1932, all on-board transformers were open
dry type. Because of problems with them., railroads
went to askarel-insulated transformers. The changes
in locomotive design since the 1930's would not now
accomodate open dry-type transformers as replacements
for askarel units. A recent trend has been to replace
askarel by oil units, and this will continue unless
new Department of Transportation regulations require
nonflammability.
However, since a tunnel fire in 1940 caused by an oil
filled locomotive transformer, Perm Central
will not allow any oil containing transformer
equipped locomotive into New York City.
(c) Multiple-unit car (MU)
These transformers are mounted under the flat-bed
of passenger cars. They ride along in this location,
about 8 inches above the rail, at speeds up to 150
nph. The transformer must be ruggedly built to with-
stand the impact of flying debris and constant vibra-
tion. Power to the cars is brought in through an
overhead catenary and is fed to the underside of the
car where the transformer, controls, and propulsion
equipment are located.
Space and weight are critical in this application.
There are only about 33 inches above the rail. The
width of the transformer is limited by the width of
the car.
58
-------
Only oil- or askarel-insulated units would provide
the required performance levels in the space avail-
able. As with locomotive applications, present
Department of Transportation regulations do not
restrict the use of flammable liquids, and the use
of askarel units has been dictated largely by the
economic considerations of fire insurance rates,
and by Penn-Central safety regulations.
8. Reactors
These units provide reactance (receive up to 15,000 volts
under normal operating conditions. During normal operating conditions, they
deliver volts and current as received. During power surges they choke the voltage
and deliver the normal output).
9. Grounding,transformers (receive up to 15,000 volts).
3.3.1 Transformer Manufacturing Plants
There are thirteen companies in the U.S. which manufacture askarel
transformers at eighteen plants. The name and location of these plants are given in
Table 3.3.1-1. Some plants manufacture various types of transformers described
above and others manufacture only one specific line of transformers.
All plants in this category were contacted to obtain informa-
tion on their waste water. Detailed information, however, was obtained on two plants,
which were visited. The largest U.S. transformer plant (based on quantity of PCB
use) was not visited because of lack of cooperation. The data base represents
approximately 20 percent of the 1974 PCB usage for transformer application. The
PCB usage in this category was 12,000,000 Ibs. in 1974 (equivalent to 15.0 to 17.0
million Ibs. PCB transformer oil). The PCB usage in individual plants is considered
by these companies to be confidential information. The range of plant PCB usage was
975 fold in 1974. Plant ages range from three to eighty-five years.
The amount of askarel used in individual transformers ranges
from 40 to 1500 gals. (516 to 19,350 Ibs.) with an average of about 232 gals.
(3,000 Ibs.). G.E. estimates that the total askarel-insultated units that have
59
-------
Table 3.3.1-1
^ Transformer Manufacturing Industry Using PCBs
Company Name
Westinghouse Electric Corp.
General Electric Company
Research-Cottrel1
Niagara Transformer Corp.
Standard Transformer Co.
Helena Corp.
Hevi-Duty Electric
Kuhlman Electric Co.
Electro Engineering Works
Envirotech Buell
H.E. Uptegraff Mfg. Co.
H.K. Porter
Van Tran Electric Co.
location of the Plant
South JBoston, Va.
Sharon, Pa.
Rome, Ga.
Pittsfield, Mass.
Finderne, N.J.
Buffalo, N.Y.
Warren, Ohio
Medford, Oregon
Helena, Alabama
Goldsboro, N.C.
Crystal Springs, Miss.
San Leandro, Calif.
Lebanon, Pa.
Scottsdale, Pa.
Belmont, Calif.
Lynchburg, Va.
Vandalia, 111.
Waco, Texas
60
-------
been put into service in the United States since 1932 is 135,000, and virtually
all of these units are still in service. The lifetime-before-failure is often
longer than 30 years, and almost all units that do fail are rebuilt and returned
to service. The current production rate of askarel transformers is about 5,000
units per year.
3.3.1.1 Askarel Handling
Few plants in this category purchase PCBs and tri-
chlorobenzene and do their own compounding. Most plants purchase pre-compounded
askarel from Monsanto. Smaller plants purchase their askarel in 55 gallon drums,
filter the askarel and pump directly into the transformer tanks. To large plants
askarel is shipped via rail cars. Rail cars enter in a designated area at the
plant site. Askarel is then pumped from the rail cars into the raw askarel storage
tank. Any spillage in this area is cleaned by rags or floor dry. The raw askarel
is next filtered through attapulgite clay or fuller's earth and is often passed
through a plate and frame type filter for final cleaning and then stored in the
finished product storage tanks. Spent clay or fuller's earth from this operation
is drummed and stored on site or sent to a landfill.
In some plants, the entire storage area is located within
dikes and curbs designed to contain at least the contents of the largest single
tank plus sufficient free board to allow for precipitation.
At most plants the askarel is distributed from the tank
farm area to the filling station, but at one plant (Plant 103) the finished askarel
is next trucked from a covered/curbed storage area to an uncovered/bermed tank
farm area. Here the truck enters a shelter area, and the askarel is then pumped
from the truck into distribution storage tank.
Recycled askarel from the manufacturing operations are
generally returned through pumps into a storage tank or into 55 gallon drums and
from there it is either filtered for reuse or is sent to incineration if it is
proved to be defective.
61
-------
3.3.1.2 Process Description
Most plants manufacture all the hardware and components
necessary for the transformer assembly. The transformer interiors and the con-
tainers are brought to the askarel filling stations where transformers are assem-
bled, filled and sealed.
Ihe filling operation is done in a designated station.
At plants where small quantities of askarel are handled spills and drips are
cleaned via floor dry or rags. At plants where large quantities of askarel are
handled, filling operation is conducted on gratings located on sumps. All drain-
age is directed into these sumps. The sumps are inspected and cleaned periodically.
All scrap askarel from surops is pumped into drums and sent to incineration.
Various transformer assembling and filling procedures
are being practiced throughout this industry. In general, the transformer assem-
bling and filling operations consist of a predrying step for removing moisture
from the transformer interiors, several stages of askarel filling and askarel
topping, addition of electrical connections and bushings, electrical testing and
sealing.
At plants where relatively small quantities of askarel
are used, the assembly and filxing procedure for distribution and power trans-
formers are alike. Some plants manufacture only a given size: of transformers
(single tank size operations). At larger plants different assembling and askarel
filling procedures are employed for various transformer subcategories. Some of
the askarel transformer assembling and filling procedures practiced within this
industry are described below.
3.3.1.2.1 Assembly and Askarel Filling Procedure for
the Distribution and Power Transformers
A. Vapor Phase Drying Prior to Filling
This procedure is primarily used in
the filling of network and pad-mounted transformers. The drying of the trans-
former internals is a vapor phase treatment with a kerosene-like petroleum distill-
ate. A schematic flow diagram depicting this system is given in Figure 3.3.1.2.1-1.
The assembling steps for these transformers include the following:
62
-------
UJ
cc
OT
1
cc
Ul
u
K
(£
UJ
(CO
syu-o
2S!2z
£C
<
cc
63
-------
. First the transformer interiors are placed in large vacuum chambers
heated with steam coils. The petroleum liquid is admitted, vacuum
and heat are applied. The petroleum vapors rise to be condensed
on the interiors, exchanging latent heat for sensible heat and
raising the temperature of the interiors. After a specific time
the liquid is pumped out; vacuum and heat are continued to dry the
internals-.
. The internals are next removed and placed in their own containers
and flushed with askarel at atmospheric pressure. The askarel
is kept in the tank until the liquid temperature is lowered to
a preset level at which time it is pumped out of the tank. The
purpose of askarel flushing is to chill the transformer interiors
prior to final filling.
. The tank is next placed in a designated vacuum chamber. Vacuum
is drawn and then fresh askarel is metered in slowly to avoid
foaming or flooding.
. The vacuum chamber is next opened, the electrical connections
are made and the liquid level in the tank is then brought to
the operating level at atmospheric pressure. The transformer
tank is then crane lifted and lowered on conveyors and sent to
specified test stations for a series of electrical testing before
the transformer is completely sealed.
. If a transformer fails the electrical tests, it is returned; the
askarel is drained and reprocessed. Otherwise, the transformer is
sent to cover welding operations.
. The tank is next lifted and sent back via conveyor belts to the
filling station where it is topped with askarel through an access
overhead pipe. The transformer is next pressure tested for leaks.
The bottom drain valve is then sealed and the top opening is pipe
fitted and sealed. Sealing is accomplished primarily by teflon
64
-------
tape. However, for sane transformers, the sealing is done by an
air cured sealant at customer's request.
B. Oven Drying of Transfomrer Internals
Prior to Filling
This procedure is used primarily in
the assembly of pole type or precipitator transformers. The drying of internals
for this category of transformers is accomplished in a convection oven. The
internals are then placed in their own casings and the bushings and electrical
connections are next made. Vacuum is then applied via a mechanical pump and the
tank is filled with askarel while the transformer internals are still at a pre-
determined elevated temperature. Exhausts from this operation are pulled through
a venting system.
MDSt transformers in this subcategory
have gasketted covers with barrel band type connections. Subsequent to the filling
operation, the transformers are sealed and pressure tested for leaks. The large
units are liquid pressure tested while the smaller ones are air pressure tested.
A schematic flow diagram depicting this system is shown in Figure 3.3.1.2.1-2.
The contaminated oil from the vacuum
pump is treated as waste askarel and is sent to incineration.
C. CVen Drying of Assembled Units Prior
to the Filling
This procedure is used primarily in
the assembly of the precipitator transformers. Here the transformer internals are
placed in their own casings, the covers are welded on the casings and the assembled
tank is placed in the convection oven all sealed. At an elevated temperature,
vacuum is applied at predetermined intervals. The tank is then transferred to
the filling station where askarel is added through an upper pipe opening under
vacuum. Both the top piping and the drain valve are next sealed and the unit is
then pressure tested. A schematic flow diagram depicting this system is shown in
Figure 3.3.1.2.1-3.
65
-------
a:
O
U.
(f)
z
<
a:
u.
o
o
z
0=
o
Ul
CM
r
cc
o
2
UJ
o
1
u.
K.
UJ
2
o:
£
CO
2
<
oc
I-
66
-------
ac
tu
K- A
el
«0 I
• o: o
UJ
UJCO
ecu
O.H
00
CO
ec
Id
U
a:
1
O
UJ
_J
m
S
UJ
LL
O
CD
o:
o
UJ
10 o:
(Nl
to
to
UJ
Ul
s
X
H-
a. uj
< o:
a: <
o sc
(O CO
a.
o
o
2
CC
UJ
s
o:
e
-------
D. Assembling Procedure for the Askarel
Filled Power Transformers
assembling procedure for all
types of askarel filled power transformers is alike. Here, the transformer
internals are dried in sealed vacuum tanks, then they are placed in their own
containers and the cover is welded and sealed. The unit is then pressurized and
checked for leaks via a helium detector. The tank is next transferred to the
filling station where low vacuum is drawn for a period of 24 hours. The askarel
line is then connected and the tank is filled slowly while under vacuum up to the
operating level. The unit is then sealed and sent to electrical testing. After
the tests the transformer is returned back to the filling station where it is
nitrogen pressure tested just prior to shipping. The schematic flow diagram
depicting this system is shown in Figure 3.3.1.2.1-2.
All power transformers have air space
on the top and are equipped with an air cooled circulation system. Additionally,
each unit has a spring loaded relief valve as standard equipment.
These transformers have a greater
potential for environmental contamination through the vent system, via leaks on
the pump and the valves.
3.3.1.3 Raw Wastes
In general, the process raw wastes from these plants
consist of waste askarel collected in sumps or pans at the filling stations, con-
taminated vacuum pump seal oil, contaminated kerosene like petroleum distillate
(used at least in one plant in vapor phase drying operation) , contaminated askarel
used for transormer interior flushing and contaminants in the plant water effluent.
Additionally, plant 103 has a unique waste stream con-
sisting of contaminated ground water which is being pumped from three caissons at
the plant site. Plant 103 has also a bleedwater discharge from their incineration
system.
Solid wastes from these plants consist of rags and
floor dry used for miscellaneous cleaning purposes and spent clay used for Aroclor
68
-------
filtration. Additionally, at plant 103 it includes the sludge from oil/water
separators.
Non-product PCB discharges are shown in Table 3.3.1.3-1,
Estimates on the quantities of raw waste generated in two plants which were
visited are given below.
Plants
Quantities of Waste Loads 103 114
Wastes to incineration, Ibs/ton of PCB used 115 98.0
Solid waste material stored on site in 55
gallon drums 135 units Unknown
3.3.1.4 Water Use
Water is not an essential component of transformer
manufacturing process. PCB wastes which reach water streams at these plants are
due to inadvertent occasional loss during handling and residuals accumulated
around the drainage systems from past operations when no precautions were taken
in handling and disposal of PCBs.
This category uses water primarily for non-contact
cooling purposes in vacuum pumps, compressors and at some plants as contact cooling
in their welding and plating operations and in phosphatizing of the steel surfaces
prior to painting. All contact cooling operations listed above are primarily non
PCB related operations. Only plant 103 has a PCB related process water which con-
sists of the water used in their waste incineration system for quenching the hot
gases from the reaction section. Plant 103 is also the major water user in this
category.
In most smaller plants all cooling is accomplished by
air. At these plants water is primarily used for hygiene and sanitary purposes and
for occasional compressor or pump cooling purposes. The only water that comes into
contact with PCBs at these latter plants is that used for personal hygiene. The
types and quantities of water used at two plants visited are given below.
69
-------
Table 3.3.1.3-1. Non-Product PCS Discharges
Source 103 114
Plant exhaust Air Air
Incinerator exhaust Air None
Personal hygiene and sanitary water Sewer Sewer
Ground water pumped through caissons TO oil separators None
then river
Water used at plant To oil separators Sewer/river
then river
Contaminated oils and waste PCBs incineration Incineration
Floor sweepings and rags used for cleaning decapsulator Incineration
Rejected transformer interiors for copper
recovery decapsulator Stored on site
Contaminated and defective empty drums To steel furnace Shipped to the
in steel mill company which
handles their
contaminated oils
Sludge frcm the oil separator Stored on site None
Clay used for askarel filtration Stored on site None
70
-------
Plant No.
Water Usage in Gallons/day 103 114*-
Non-contact cooling 1,646,000 38,000
Process water
. Contact cooling 183,000 None
. Detergent washing 2,000 None
. Incineration quench 122,000 None
Hygiene Unknown 300
Boiler make-up Unknown 200
Sanitary Unknown 9,000
Non PCB related
. Water used in a lab Unknown 500
. Water used for external cleanups None 4,000
*Breakdown of the water usage was estimated.
3.3.1.5 Waste Water Treatment
There are no effluent treatment techniques in practice at
any plant in this category for the removal of PCBs. Details on wastewater treat-
ment techniques utilized at one facility (plant 103) for purposes of oil recovery
and on-going containment programs undertaken by two plants (plants 103 and 114) in
order to prevent the entry of PCBs into the environment are covered below.
Plant 103
Significant environmental controls have been reported at Plant 103. Since
1966, 3 men per year (two exempt and one non-exempt employees) are allocated to
this effort.
In order to control oil losses to the environment, water/oil separator facil-
ities were constructed on all major outfalls. Interceptor sewers were also con-
structed to reduce the number of discharges from approximately 35 to 10. All
storm and industrial wastewaters from askarel manufacturing areas are served by
71
-------
two water/oil separators (separators 005 and 006). Each separator has a capacity
of 2 million gallons. Each water/oil separator is divided .into two longitudinal
compartments. The wastewater enters the separator over a grating and passes
through a drum basket type separator for the removal of large particulates. This
drum is equipped with overhead water sprays to flush the particulates from the
basket walls. The wastewater flows through the drum and crosses the length of the
separator basin at which time the sludge settles and the oil is skinmed via two
belt type oil skimmers, one on each separator compartment. This oil is recovered
and used as supplemental fuel in the incinerator. The water phase passes a weir
and enters the river.
The separators are effective during the plant's normal daily operation.
However, during storms, in addition to the increased runoffs from the city areas
the river water level rises and floods the system.
Other containment structures undertaken by this plant since 1966 include re-
location of all bulk storage tanks within dikes designed to contain the entire
contents of the largest single tank plus sufficient free -beard to allow for pre-
cipitation. Tank truck unloading stations include curbs arid sumps adequate to
contain the entire loss of tank truck contents. All new diked storage areas are
drained by manually operated valves which are normally in the closed position.
All askarel storage tanks and distribution lines have; been relocated above
ground for ready leak detection and maintenance. In the few instances where pipe-
lines are below grade, they are enclosed in sleeves to contain leakage.
In areas where the ground water is contaminated with oil, three caissons
have been erected in the ground to pump the ground water into the water/oil
separator for the separation of oil. The ground water from two of the caissons
are treated in oil separator 006 and the water from a third caisson is diverted
into oil separator 005.
Additionally, a John Zink designed thermal oxidation incinerator was erected
at this plant in 1972 to safely dispose of PCBs. A schematic flow diagram of this
unit is given in Figure 3.3.1.5-1. This unit consists of tojo steam atomized
burners and a long cylindrical chamber to provide residence time for thermal de-
gradation. Following the chamber is a water spray quench pot and a counter current
packed scrubber column located at the base of the stack.
72
-------
<
1
(
<
<
<
1
(
1
1
1
1
c
E
<
(!
C
«
u
c
u
1-
V
«
e
jj
i.
X.
J
a
u
f.
ii
>
:>
>
U
c
c
t
3
>
L
C
J
J
J
1
1-
<
?£_
<
u
i
H
<
a
(.
ij
c
a
c
a
T
3
t
I
C
9
J
3
^
>
1
tlJ
CD
*c
<
O
^
I.
e
1 '
or
Ul
Y-
?
a
Ul
Ul
u.
z
Ul
2
<
Ul
^
0
2C
INDUCED
o
Ul
^
o
1
Ul
°K
o
CO
Is
cu.
3
u.
1—
H
u.
u.
Ul
cc
Ul
1
a:
UJ
N
j 5S
< 5
K H
D
UJ
Z
3
o
(O
o o
Ul Ul
DC
1
Ul
2
O
o:
UJ
g
cc
Ul
H
a:
cc
C5>
O
1
U.
aj
I
i
to
to
ff
73
-------
The waste oils are brought into the incinerator site in 55 gallon drums or
by truck trailers. These waste oils are next pumped from the drums or from the
trailers into the incinerator feed tanks. The steam atomizing burners inject
the combustible liquid wastes into the combustion section with air, in such a
manner to create a vortex type turbulence. This produces high heat release and
effective combustion promoting the thermal degradation process. After combus-
tion, the waste gases proceed through the oxidation chamber which provides 3 to
12 seconds of residence time at temperatures 1600 to 1800 °F for the degradation
reactions to go to completion. The flue gases from the chamber pass through a
quench pot which contains a series of water sprays to cool the gases. An induced
draft fan then forces the cooled gases through the packed bed scrubber column.
Kere, acidic ions produced during the combustion process are absorbed in the
scrubber liquor. The scrubber liquor is then neutralized and disposed into the
sewer.
A high temperature decapsulator has been incorporated .in the thermal oxidizer
unit for solid incineration and copper recovery. The exhaust of this unit is routed
to the upper end of the oxidizer chamber.
Additionally, operating personnel are instructed on the operation and main-
tenance of equipment to prevent oil discharges; notices containing instructions
for proper handling of PCBs are posted throughout the plant.
Water/oil separators are inspected and minor maintenance performed on a
daily basis. All discharges are monitored on a regular basis as required by NPDES
permits and data is compiled on a weekly basis.
Plant 114
This plant employs no wastewater treatment technique. However, their tank
farm area is located in a 32 inch deep concrete basin which is completely imprevious.
This basin is epoxy painted and all joints are sealed with Dow Corning "Silastic".
Additionally, there is a second, inner, 12 inch deep basin around the askarel tank.
The storm run-off from the tank farm area is accumulated in these basins. This
wastewater is periodically pumped out and sent to incineration.
74
-------
All oil receiving piping is made of welded steel. In order to prevent spills
and leaks all sight glasses on their tanks have been removed and the fittings
sealed. The oil volume in these tanks is determined by differential pressure.
3.3.1.6 Effluent Oomposition
Effluent flow rates and PCB contents are available
from the three plants which discharge into the rivers and from two plants which
discharge primarily into sanitary sewers. Detailed effluent composition is avail-
able only on Plant 103. This information is given in Tables 3.3.1.6-1 and
3.3.1.6-2, respectively.
The remaining plants in this category which discharge
to sanitary sewers report that their water usage, for reasons other than for
sanitary purposes, is minimal. However, they have no data available on the quanti-
ties and composition of water discharged.
75
-------
Table 3.3.1.6-1
PCS Concentration in Effluents from Transformer
Manufacturing Plants
Plant
No.
103<»
114
115
116(3)
Discharge
Designation
005
006
None
None
002
003
004
SI
*2
13
Effluent
Flow Rate
Gal/Day
1,310,000
550,000
50,000.7.
2,000^'
36,000
24,000
150,000
252,000
378,720
504,000
Effluent PCB Content
117
Batch
discharge
13,500
PPb
Avg/Max
4.9/120
7/75
Unknown
Unknown
-------
Table 3.3.1.6-2. Influent and Effluent
pH yearly average
pH min.-max.
Alkalinity, mg/1
BOD, 5-day, mg/1
Chemical oxygen, 103/1
TS, mg/1
TDS, mg/1
TSS, mg/1
TVS, mj/1
Armenia, mg/1
(AN)
Kjeldahl Nitrogen, mg/1
Nitrate AsN, mg/1
Phosphorus Total, mg/1
Color (As P)
Turbidity, mg/1
Total hardness, mg/1
Organic Nitrogen, mg/1
Sulfate, mg/1
Sulfide, mg/1
Chloride, mg/1
Cyanide, yg/1
Fluoride, mg/1
Aluminum - total, yg/1
Boron - total, yg/1
Calcium - total, mg/1
Chromium - total, yg/1
Compositions of
Incoming
Municipal
Water
lion gpd 1.86
6.4
14
2
*
69
68
1
34
0.05
/I 0.45
0.15
1 0.01
30
<25
44
1 0.5
9.0
<0.1
6
<0.00
0.16
1 100
70
6.5
1 <10
Plant 103
Outfall
005
1.31
7.7
6.4-8.0
Avg/Max Avg
Cone. Ibs/day
25/25 273
3/3 28
23/34 251
70/70 765
67/67 735
3/9 31
19/22 208
<0. 2/1.0 <2
0.8/1.0 9
0.55/0.9 6
0.04/0.05 <1
25/30
<25/<25
31/44 340
0.6/0.7 6.6
14/18 153
<0.1/<0.1 <1.0
6/8 61
O.OOA0.01 0.04
_
500/550 5
-
7.0/7.5 77
30/30 0.3
Outfall
006
0.55
5.5
3.1-7.1
Avg/faax Avg
Cone. Ibs/day
15/20 69
27/30 124
75/92 344
364/497 1671
341/483 1565
23/30 106
89/96 408
2.8/3.0 13
4.6/5.1 21
0.98/1.02 5
54/124 248
30/30
<25/<25
33/44 152
1.8/2.0 8
9.0/9.0 41
<0.1/<0.1 <0.5
11/13 50
<0.00/<0.00 <0.02
9.5/12.0 44
200/200 0.9
9/70 0.04
7.7/8.0 35
310/350 1.4
77
-------
Table 3.3.1.6-2 (Con't)
Cobalt, total, yg/1
Copper - total, yg/1
Iron - total, yg/1
Lead - total, yg/1
Magnesium - total, yg/1
Manganese - total, yg/1
Mercury •- total, yg/1
Molybdenum - total, yg/1
Nic'cel - total, yg/1
Potassium - total, mg/1
Silver - total, y
Sodium - total, mg/1
Tin - total, yg/1
Titanium - total, y
Zinc - total, yg/1
Oil & Grease, mg/1
Phenols, yg/1
Surfactants, mg/1
Chlorinated hydrocarbons, mg/1
(except pesticides)
KBs, yg/liter
Incoming
Municipal
Water
<7
500
330
<30
7
200
<1
<7
100
1.1
<10.0
5
<7
2
30
1.0
3
<0.02
<1.000
Outfall
005
4/4
20/500
150/330
<10/30
3/7
50/200
-------
SBCTICN V
SELBCTICN OF POLLUTANT PARAMETERS
i.o nmoxJcriON
This section discusses the pollutant parameters which are of significance
to the polychlorinated biphenyls CPCB) production industry and those capacitor
and transformer manufacturing operations which utilize PCBs as either dielectric
fluids or cooling fluids. Historically speaking, both the production of PCBs
and the utilization of these materials in electrical components have been con-
sidered industries with relatively inconsequential pollutional effects on the
environment as a whole. However, in the recent past (perhaps the last five years)
the significance of the buildup of polychlorinated biphenyls in the environment
has become apparent. Ironically, the very properties which make polychlorinated
biphenyls desirable for utilization in electrical components, i.e. relative
inertness to physical, chemical and biological attack, contribute significantly
to the undesirability of continuous discharge of this material to the environment.
Following the recognition of the relative stability of this material and the
buildup of PCBs in the environment, a broad spectrum of studies have been con-
ducted by governmental, industrial and private institutions to determine the ef-
fects of PCBs on various levels of the eco-system. It is not the intent of this
section to reiterate and confirm all of the claimed effects. Suffice it to say
that a rather convincing body of information has been developed which indicates
the potential mutogenic, teratogenic and carcinogenic properties of PCBs, partic-
ularly in higher life forms including human beings. In deference to the producers
and users of PCBs, it should be noted that it is only with the benefit of rather
recently acquired scientific data and a good measure of 20-20 hindsight, that the
scientific community today is able to point to the rather poor control measures
of these industries in the past as one of the causes for the relatively high level
of PCBs found in the environment today. In other words, it is believed that the
producers and users of PCBs should not be unduly chastised for their past handling
of this material when little or no information was available to indicate that
stringent control measures should be applied.
79
-------
It can be categorically stated that the principal contaminant in waste-
waters, solid wastes and scrap oils, generated by the manufacture and use of PCBs
are the PCBs themselves. Other constituents are of significance only to the ex-
tent that they CD affect the properties: of the PCBs contained in the wastes,
(2) to the extent that other constituents interfere with, the destruction or re-
moval of PCBs from wastewaters, solid waste or scrap oils or, (3) to the extent
that other constituents are contaminated with PCBs. In other words, there may
be constituents added to wastewaters through manufacturing and production opera-
tions which themselves might need control in order to permit compliance with all
applicable discharge and water quality standards; however, for the purposes of
this evaluation and report, only those constituents which in some way affect the
control of PCBs will be considered.
2.0 SIGNIFICANCE AND RATIONALE FOR SELECTION OF POLLOTANT PARAMETERS
As previously stated, the basic rationale employed in selecting those
pollution parameters which are of significance in the production and utilization
of PCBs was one of identifying those parameters which could in some way affect
the control and treatment of PCBs. Each of the parameters which are considered
of significance are listed in this subsection and the rationale for their inclu-
sion are presented herewith. Again, it is not the intent of this section to list
other pollution parameters which might require control in themselves, separate
from that required for the control of PCB discharges.
2.1 Polychlorinated Biphenyls
As previously stated, PCBs are the principal contaminant of signifi-
cance in all wastes generated in the production and utilization of PCBs. The
principal sources of PCB contaminated waters, solid wastes and scrap oils from
the PCB manufacturing operation and from the capacitor and transformer industry
have been previously described in Section IV. The basic sources can be categorized
as follows: (1) spills, leaks and drips associated with the handling, transport,
storage of PCBs, (2) concentrates originating from the refining of PCBs, (3) scrub
waters employed in systems associated with the incineration of scrap oils and PCJs,
(4) vapors and condensates derived from the vacuum dewatering of PCBs in capacitor
80
-------
and transformer filling operation, 15} vapor degreasirtg arid washing operations
associated with the exterior cleaning of capacitors, which have been flood filled
with, askarel, (6) handwashing and general clean-up by personnel who become con-
taminated with PCBs in the production operations, (71 other production operations
such as paint stripping, painting, etc. which become incidentally contaminated
with PCBs. There is, of course, also the PCB contamination of ground water and
storm runoff from historical operations in which PCBs were employed. In other
words, for existing facilities where PCBs have been employed for many years, con-
tinuous contamination of ground water and storm water runoff is possible through
past losses of PCBs. It as recognized that general losses of PCBs to the atmos-
phere through local and general ventilation with subsequent redeposition of PCBs
on the roofs and grounds of PCB utilizing facilities represent current and con-
tinuous potential contamination of ground and storm water runoff. Finally, de-
pending on the ultimate philosophy regarding the control of PCB discharges from
user facilities, the raw water supply erttployed by each user facility may in fact
represent a source of PCB contamination.
2.2 Chemical Oxygen Demand (COD)
Hie non-specific parameter, chemical oxygen demand (COD), has been
selected to identify other organic contaminants which might arise from the pro-
duction or utilization of PCBs. COD is considered to be a significant parameter
since it can affect the terminal treatment systems which have been identified as
potential means of removing or destroying PCBs from contaminated wastewater. The
two systems which have been identifed for terminal treatment include activated
carbon adsorption and UV light assisted ozonation. In both of these cases, other
organics as measured by the COD test will affect the terminal treatment system.
In the case of activated carbon, such organics will occupy adsorption sites on
the carbon and affect its rate of exhaustion. In the case of the UV-ozonation
system, other organics as measured by the COD test will increase the required
dosage of ozone to destroy the PCB since the ozone will be nonspecific with respect
to the organic that it oxidizes. Some examples of the types of organic contami-
nants and their sources in the production and utilization of PCBs include:
(1) other dielectric fluids which might be used in the same plant as PCBs including
81
-------
dioctylphthalate, castor oil, mineral oil, etc., (2) other crganics which might
be admixed with the PCB such as trichlorobenzene, (3) detergents which may be
employed in cleaning operations particularly in electronic component production
facilities, (4) solvents and organic pigments associated with the surface prepara-
tion and protection of capacitor and transformer cans, (5) vacuum pump oils,
(6) and in sane cases kerosene where aluminum rolling operations are conducted at
the capacitor or trans former manufacturing site. Again, COE) is considered a
significant parameter since it does measure the increase in either carbon exhaus-
tion rate or ozone dose required to effect removal or destruction of PCBs.
2.3 Fats, Oils and Greases
Free floating fats, oils and greases are considered to be a significant
pollution parameter since PCBs are soluble in such materials, In removing these
materials through the application of quiescent separation basins, the materials
collected on the surface of such basins will require separate controlled destruc-
tion or disposal since they will be contaminated with PCBs. The sources of such
materials have been cited under Subsection 2.2 (COD) and that portion of the oily
materials which exceed their solubility in water are being referred to in this
case. The separate handling and disposal of the materials which can be removed
through gravity separation represent an added treatment and disposal cost and
therefore fats, oils and greases are considered a significant pollution parameter.
2.4 Suspended Solids
It is recognized that PCBs have an affinity for adsorption on sus-
pended solids contained in wastewater. There are basically two types of suspended
solids which are of interest in this case. First, there are those which are of
sufficient size and weight to permit their removal by gravity separation. Again,
the sludges generated in the gravity separating operation would be anticipated to
be contaminated with PCBs and would require separate destruction or disposal
thereby adding to the total treatment cost. The second type of suspended solid
which would be of interest j_n this case would be those which are filterable. As
currently envisioned, terminal treatment schemes would include a filtration opera-
tion employing either dual or multi-media filters. The backwash waters from such
82
-------
filtering operations would be high in filtered suspended solids which would be
anticipated to be contaminated with- PCBs and would require, separate treatment
and disposal. The sources of suspended solids in the PCB manufacturing and
utilization industries include, filter media such as diatomaceous earth, clays,
etc. and general floor dust generated in housekeeping operations.
2.5 Dissolved Solids
Dissolved solids or salts are relatively inconsequential with regard
to nominal end-of-pipe treatment systems. However, in the case of "zero" dis-
charge as defined in Section VII, the total dissolved solids level of the raw
wastewater can become of significance when it is recognized that complete recycle
is one means which is being considered for achieving zero discharge of PCBs. In
the case of complete recycle, it is recognized that periodic blowdown of recycled
waters will be required to prevent excessive buildup of the dissolved solids. Since
the frequency of blowdown will be determined by the initial TDS level and the rate
of evaporation of water from the system, the initial IDS level can become signifi-
cant.
2.6 Other Constituents
Other constituents such as acids or alkalies which would affect the
pH, possibly heavy metals, possibly other chlorinated hydrocarbons, etc., may in
themselves represent constituents which would require control to place PCB pro-
ducers and users in compliance with all applicable regulations. However, it is
not believed at this time that any of these parameters affect directly the princi-
pal goal of the treatment systems to be described herein for the removal of PCBs.
It should be recognized, however, that other treatment systems for controlling
such parameters and the costs associated with such systems should not be over-
looked in the total pollution control and wastewater management program for PCB-
using facilities.
83
-------
SECTION VI
WASTEWATER TREATMENT TECHNOLOGIES
1.0 INTRCDUCTICN
An analysis is made of the similarities and contrasts between the oper-
ations and waste control problems of the PCBs producer and capacitor and trans-
former manufacturing users. The major waste management or waste stream cate-
gories are summarized and a description is provided of the current control
practices. Additionally, the candidate treatment technologies that were con-
sidered are presented as well as the rationale used in selecting the recom-
mendations. This section concludes with the recommended control technologies
for use immediately, within three years, and those that will not be available
for five or more years.
1.1 Similarities and Contrasts Between PCBs Wastes and Control Practices
in Their Production, and in Their Use as Dielectrics
Based upon detailed plant inspections and examinations of the process
steps in the production of PCBs, and then their use in capacitor and trans-
former manufacturing, it has been determined that there are four major cate7
gories of wastes that must be considered. The categories will be described in
the next section. It has also been determined that the characteristics of
these wastes in PCBs production and dielectric manufacturing are similar
enough so that the same kinds of control and treatment technologies can be
used.
The manufacturer, MDnsanto, does generate Msntcirs and more highly
chlorinated biphenyls as byproducts of the production process. These are
separated from the product stream as part of the final preparation of the indi-
vidual member products of the Aroclor series. Thus the capacitor or trans-
former manufacturer never encounters these highest boiling and most refractory
PCB hcmologs and related compounds. However, these differences are not gener-
ally critical to the currently recomiended treatment processes. There might
be implications to incineration temperatures. They would have a bearing on
the final development of several of the candidate more novel treatment methods
that are still in the laboratory or pilot plant stages.
84
-------
A minor contrast is that Monsanto produces and handles the pure
Aroclors (mixtures of PCBs) and also blends them with triohlorobenzene to pre-
pare askarels (non-flammable transformer fluids). Capacitor manufacturers
handle only pure Aroclors, and transformer manufacturers handle primarily the
blended askarels.
In production there are a number of ways that water streams come in
contact with PCBs as part of necessary process steps.
However, in some of the dielectric users' manufacturina, there is
rigorous exclusion of water from all steps to preserve the high quality of the
dielectric device. Thus, the chances for those users to pollute water streams
are very minimal and occur mostly through inadvertent drips or spills during
askarel transfer, rather than during manufacturing steps which are all in a
dry environment.
1.2 Sunmary of Waste Management Problem Areas
1.2.1 Waste Liquid PCBs and Contaminated Scrap Oil
The users of dielectric materials have strict requirements
for purity. Typical requirements are:
Inorganic chlorides 100 ppb max
Acidity, mg KOH/g 0.01 max
Water content 30 to 35 ppm max
Resistivity, 100°C, 500 V, 100 to 500 x 109 ohm-
0.1 inch gap cm, min.
Dielectric strength, 25°C 35 KV min
When these or other properties cannot be met at the pro-
ducer's facility, or after being used in the filling operation by the manu-
facturer and it is found that they cannot be restored to the askarel specifi-
cations by filtering and drying disposition is required.
As described in earlier sections, there are several catego-
ries of contaminated liquids:
85
-------
(1) PCBs contaminated with mineral oil;
(2) Mineral oil contaminated with PCBs; .and
(3) Nonreclaimable contaminated transformer askarels -
arced askarels, askarels from manufacturing spills
and sump accumulations, and askarels from holding
basins, drip and drain pans, washings, sample jars
and containers.
1.2.2 PCBs in Wastewaters
Possibly the most universal pathways by which PCBs enter
wastewater streams, are: 1) from operator wash-up after handling PCBs and
2) from outside groundspills that are blended into the neixt rainfall storm
water. The importance of these pathways can be seen when we realize that if
sixteen operators (or 4 operators 4 times a day) washed one ounce per day of
PCB from their hands into the sanitary facilities, that would be one pound per
day discharge. We did see situations occurring where operators get their
hands immersed in PCBs; presumably most of this was wiped off in rags. How-
ever, if each of 100 employees in the area used 50 gallons of sanitary water
in a day, that would be (50 x 100 x 8.34 Ibs/day) or 41,500 Ibs/day. The one
pound of PCBs in 41,500 pounds of water would then give an averaged concen-
tration over that day of 24.1 ppm in the sanitary stream (presumably solu-
bilized by detergents).
Other categories previously discussed are:
1) Incinerator scrubber and quench water
2) Steam jet ejectors in vacuum distillation
3) Capacitor detergent wash solutions.
1.2.3 PCBs ContaiTdnated Solid Wastes
These can be divided into two categories, burnable and non-
burnable:
Burnable Solid Waste Materials Containing PCBs. These ma-
terials can be disposed of by high-temperature incineration and consist of
cellulosic materials, rags, pressboard, wood, sawdust, fuller's earth in bulk
or in cloth bags, blotter papers, nitrile or cork gaskets, and similar ma-
terials.
86
-------
Nonburnable Solid Waste Materials Containing, _or Contaminated
with PCBs. These materials may consist of capacitors and transformer internal
components; steel, copper and aturninum components; filter units of the steel
mesh construction type; and askarel drums and cans.
Materials of this nature should be allowed to drain, with the
liquid collected in. drip pans, before disposal.
1.2.4 Air Emissions ofPCBs
PCBs although of very low vapor pressure can be emitted to
the atmosphere from the following operations and practices:
1) ftroclor scrubbing of air in PCBs manufacture
2) Vapor exhaust from steam jet ejectors
3) Evaporation from accidental spills
4) Evaporation from hot surfaces as part of flood-
filling, or inspection or holding operations
5) Vacuum pump exhausts
6) Evaporation from plant wastewater.
1.3 Summary of Current PCBs Waste Control Practices
In Section IV of this report, the individual practices used by
plants to control dissemination of PCBs to the environment have been described.
This section sumnHrizes the current methods of treatment beina emploved for the
four1 major categories of wastes we have established.
1.3.1 Control of Waste Liquid PCBs and Contaminated Scrap Oils
Plant visits, contacts and follow up investigations have
shewn the following to be the major control methods:
a) incineration
b) disposal in sealed drums to sanitary landfill.
Both incineration and landfill may be carried out by the
facility generating the waste, or the facility may engage a contractor for the
service.
87
-------
{fonsanto, the producer, lias its own John /'link designed in-
cinerator that vaporizes the PCBs liquids and maintains a turbulent burning
gas at above 2200 °F for about 2 seconds, One trans former manufacturer uses
a John Zink designed incinerator, which vaporizes the PCBs and then burns at
1600-1800°F for 3 seconds or longer. This latter facility can also destroy PCBs
soaked transformer intervals, but cannot routinely handle spent fuller's earth.
A number of PCBs users send all kinds of solid wastes to
Rollins Environmental Services facility at Logan Township, N.J. Rollins uses
a specially designed complex with a rotary kiln that exhausts to an after-
burner plus a liquid turbulent burning chairber also exhausting to the after-
burner. Liquids can be burned in both the liquid chairber and the kiln. In
the kiln they are used to incinerate solids. The afterburner is 40 feet long,
providing good residence time, but it is followed by a hot duct of about e-
quivalent length that allows further combustion. Rollins claims a residence
time of 3 to 4 seconds, at a minimum temperature of about 2400°F at -the aft
end of the hot duct. The gases then go to a venturi scrubber and a tower
scrubber for cooling and neutralization. Foe non-PCBs incineration, Rollins
scmetiires lowers combustion temperature', to 2200°F and residence time to 2 to
3 seconds. The facility operates with EPA approval. These residence times
and temperatures were chosen experimentally to get 99.999% PCBs destruction.
This facility also handles all kinds of solids as will be explained later.
Another facility that incinerates liquid PCBs, with New York
State EPA approval, is the Chemtrol Corp.'s facility at Model City, New York.
Liquids going to landfill are sealed in drums and ailed to
landfills that have deep clay bases, and impervious bulkheads to prevent
leaching and seepage. However, incineration seems to be the preferred route
for liquids.
13. 2 Control of PCBs in Wastewaters
Our plant investigations have revealed that there are no
methods being practiced whereby dissolved or otherwise bound PCBs in water
streams are being either extracted from those streams, or destroyed in them.
88
-------
As described previously, separations are made such that PCBs
and sludges are removed from the bottom, and oily phases are removed from the
surfaces of water bodies. However, nothing as yet is being done to the water
layer itself.
This is the area of PCBs wastes control and treatment most
needing investigation, development and reduction to practice of feasible,
practical and economical methods for eliminating PCBs frcm all varieties of
wastewater streams and collected bodies that are encountered.
In general, these streams will run from an undetectable
level of PCBs (less than 1 ppb) up to about 500 ppb; with many instances be-
ing in the 10 to 50 ppb range.
Several plants are engaging in segregation measures to keep
the quantity of PCBs contaminated wastewaters as low as possible. A few
plants planned adsorption tests with carbons and other adsorbents, but there
are no full scale operations.
1.3.3 Control of Solid Wastes Contaminated with PCBs
The two current practices observed for the disposition of
solid or semi-solid PCBs wastes are incineration and landfill. Since we
found only one dielectric device manufacturer with partial incineration capa-
bility, and Monsanto's incinerator cannot handle solids, all facilities ex-
cept one must ship their solid wastes away from the plant site for either
treatment. The one plant with partial capability can incinerate transformer
internals to recover copper, and also can incinerate paper, rags, cardboard
and the like. However, they do not as yet, incinerate fuller's earth, con-
taminated dirt, and similar materials. These are druntned and stored for later
disposition.
Although there are a number of experimental facilities
throughout the U.S. that could undoubtedly incinerate all types of solid ma-
terials, one conmercial venture has received the bulk of the work. Rollins
Environmental Services has been described above for liquids incineration. For
solids incineration of almost all types, the tumble burner or rotary kiln is
89
-------
used. When PCBs contaminated materials are destroyed the kiln temperature is
brought up to 2200°F. All kinds of solids, packed in 47 gallon lined fiber
drums and not posing unusual safety hazards, are accepted. The current (late
1975) fee is about 7^/pound. Additional charges of $3/fiber drum handling,
tranefX.»rtation charges, and possibly other charges for unusual problems might
be made.
Rollins will not accept impact sensitive, radioactive ma-
terials, or heavy metals concentrations in the PCBs wastes of generally great-
er than about 25 ppn. For anything packed in steel drums, Rollins charges
$10/drum handling charge.
As a general rule, anything packed according to the latest
ICC tariff for hazardous materials will be accepted.
As with the liquids incineration, the gases from the kiln at
2200 F pass to the afterburner at a temperature of about 2500°?. They exit
the afterburner to a long hot duct that completes combustion and maintains the
temperature to 2400 F until the gases enter the venturi scrubber, and thence
to the laver scrubber and then the stack. Rollins has experimented with
lower temperatures of 2000 - 2200°F and has not found destruction of PCBs to
the 99.999% level desired. The residence time for gases at the 2400° temper-
ature is 3 to 4 seconds in a turbulent regime.
The other alternative for disposition of solid wastes is a
controlled scientific landfill. Here again, one commercial venture has
handled a large quantity of the PCBs solid wastes under supervised conditions.
The Chemtrol Pollution Services Co. of Model City, N.Y. oper-
ates a landfill located on the shores of Lake Ontario, under New York State
EPA supervision, in a geologic setting that is claimed to be ideal for com-
plete attainment,,
This "scientific landfill" is all located above ground on a
bed of. 40 feet thick clay. On this foundation, they construct cells for re-
ceipt of drummed solid wastes. The cells are lined with 30 mil thick chlori-
nated polyethylene film, and when loaded, they are sealed, or covered with
f:ivc3 ftel: of clay. At the bottom of each cell there is a sump, so that all
90
-------
leachate is collected and removed from the cell. That leachate is pH controlled,
settled, filtered, and treated by flow through a carbon bed. The ground and
surface waters are checked monthly for chemical content by an outside analytical
laboratory. The facility has been approved for usage as a PCBs disposal site by
the New York State EPA, and thousands of drums have been landfilled over the
last five years. An inventory is kept of the contents of each cell.
Chemtrol has the capability of converting semisolids and
sludges to solids by using silicate cement powders and proprietary gelling agents.
Our review and analysis of the industrial situation for PCBs
solid wastes control shews that there are adequate options, at a price, for safe
disposal and destruction. There is no longer justification for open dumping to
the ground or in lagoons.
1.3.4 Control of Air Emissions of_JPCBs_
Most emissions from ambient temperature PCBs are not con-
trolled or collected for treatment in any way. They generally are collected as
part of the overall plant air exhaust, which frequently is ducted to roof ex-
haiosts without treatment. In a few isolated cases there was exhaust chilling
and seme fibrous or granular filtration of air with the potential for PCBs con-
tamination.
In the hot flocd~fAiling of capacitors there is a cool down
at the end of filling,, before the tank is opened to the plant atmosphere. How-
ever, the temperature is still probably 100°F to 150°F for a fraction of an hour,
with a surface area of several square meters, with the filling tank open to the
plant atmosphere. There are individual exhaust ducts at the tanks, but they
lead directly to a roof exhaust system. Again the PCBs covered capacitors are
held in an oven at 100 - 150 F to prevent any moisture condensation before seal-
ing. Vapors from this storage are ducted to the roof and vented.
In seme cases, waste scrap oils containing PCBs are burned
together with fuel oil in standard boiler systems. In these instances much of
the PCB content; is probably vaporized and exhausted to the air, rather than in-
cinerated.
91
-------
2.0 CANDIDATE PCBs WASTE TREATMENT TECHNOLOGIES CONSIDERED
As part of the examination phase of this study of alternative methods, we
have contacted a spectrum of equipment suppliers, developers, and even re-
searchers, for both U.S. and foreign technology.
Key information was obtained about potential methods of PCBs removal
from, and destruction in, wastewaters, by cooperative testing performed between
Versar Inc. and several materials suppliers and process developers. Their assis-
tance is gratefully acknowledged.
We visited the U.S. producer, JYbnsanto, and several •transformer and ca-
pacitor manufacturers to ascertain the nature, characteristics and quantities
of plant effluents that now contain measurable amounts of PCBs, or might con-
tain PCBs following a spill or other incident. This has provided the back-
ground for weighing the advantages and disadvantages of all available tech-
nology for potential application to PCBs wastewater treatment. We discussed
and evaluated the control methods they now use, their possible shortcomings,
and ideas for better optimized systems.
In order to find the new technology and approaches that might be forth-
coning; we conducted a computer search of Chemical Abstracts from 1972 to cur-
rent 1975, which eliminated PCB synthesis, manufacture, cr toxicology data
and focused on bringing out degradation, decomposition and waste treatment data.
Searches were also made of the Lockheed Engineering Index, National
Technical Information Service; and Predicasts for information on new business
developments.
The text reference used as a source of summarized data prior to 1972 was:
The Chemistry of PCBs, by Hutzinger, Safe and Zitko; published by CRC Press in
1974.
The candidate treatment technologies have been divided into the same four
categories used under 1.2 and 1.4, above. Also, as we evaluated technologies,
they were placed into categories indicating their readiness for usage as fol-
lows:
92
-------
1. Dexranstrated Full Scale Treatments
2. Pilot Plant Scale Methods
3. Research Approaches to PCBs Removal or Destruction.
Some of what is reported is work based on treatments of conpounds similar
to the PCBs, i.e., chlorinated hydrocarbons and other refractory organics.
The full scale plant treatments are ready for application to PCBs waste-
problems now. The pilot scale methods are expected to take one to five years
to be developed to the point where they will be determined suitable, or un-
suitable, for plant scale PCBs applications.
The research methods are expected to take three years or more before
being ready for plant scale applications decisions.
In the latter part of this section (2.5) the applicability of various
treatments to the zero discharge situation is discussed.
2.1 Treatment of Waste Liquid PCBs and Contaminated Scrap Oils_
Only two kinds of treatment or disposal methods were considered suit-
able for this waste category; and they are the ones now being used. However,
there are improvements and design features that are highly desirable to pre-
vent PCBs transfers from this liquid state to water, air or land.
2.1.1 Incineration
A standard configuration of a long cylindrical steel com-
bustion chamber, with a length to diameter ratio of 4/1 to about 8/1, well in-
sulated with high alunina refractory brick is suitable for the primary chamber.
Two burners should be provided; one for a high BTCJ exothermic combustion ma-
terial such as gas, or a fuel oil, and one for the PCB waste liquid. Primary
air or steam is provided to vaporize the oils, and secondary air is added to
complete combustion and provide a measured excess of air. These feeds should
be made tangentially, or with adequate baffling to assure turbulent flew
throughout, and to prevent hot or cold spots in the chamber. It is obvious
that aJjnost any residence time may be caI:-< lated for a gas flowing through a
cylindrical chamber. However-, there wil .a a minimum flew rate and a maximum
93
-------
residence tine for which turbulence is assured throughout the chamber. Turbu-
lence throughout can be measured by unifoim wall temperatures in all parts of
the chamber. Most fuel flames will provide adiahatic combustion temperatures
of about 3000 F. When the more endothetmic PCBs are heing destroyed, the
flame can be rapidly cooled to 2200-2500"F. However, it has been found in a
number of instances that for the very high destruction efficiencies desired,
99.999% or higher, this is the minimum temperature range required for a resi-
dence time of 3 to 4 seconds. Following the combustion chamber there should
be allowance for afterburning, which nay be accommodated in insulated ducts.
These can be of sufficient length to assure the total residence time of 3 to
4 seconds. Following the combustion phase there must be rapid cooling of the
gases with sufficient cold water, usually as a peripheral ring of sprays or
jets. Also, at this point neutralization should be accomplished so that hot
hydrochloric acid contact with further components is negated. At 'this point,
some systems introduce a high energy venturi scrubber. This is particularly
helpful if particulates are expected in the gas stream. In all systems the
gas stream is then washed counter currently with water', or water plus neutral-
izing agent, in some form of a packed tower. Usually there is a demister as
the last element contacted by the gas, before it is exhausted through the
stack.
The stack gas emissions and the scrubber effluent liquid
should, be monitored periodically for PCBu levels.
Hie incinerator should ha\r3 adequate automatic controls to
prevent PCBs from being emitted undestroyed. The chief failsafe control is to
have a burner shutdown if the pyrometer at the far end of the combustion
/->
chamber or duct drops below about. 2150 r. There are a number of other controls
that can be provided to warn that such an occurrence might take place. These
would be flow rate controls on the primary and secondary feeds. Low fuel flow,
or high PCBs-to-fuel flew ratio could lead to a temperature drop. Inadequate
air flow could lead to the same difficulty. All such parameters should be
nonitored, and adequate operator warxJ.ngs provided. Cf course, to prevent de-
struction of the venturi and scrubber factions, shutdcv.p should be automatic
if cold water flow is ever insuff ici,~
94
-------
2.1.2 Sanitary or Scientific Landfill
This method is an alternative to incineratior. for a PCBs user
that cannot justify an incineration facility just for I'CBs , ar..d has no waste
pickup service organization within reasonable distance. Ther^ is little rorni
for variation in the kinds of landfill that will adequately contain PCBs. It
should have all the desired features described in Section 1.3,3 above , but in
addition, there must be high assurance that the liquid containers will be
sealed and leak proof, and have long life in the landfill.
2 . 2 Treatment of Wastewaters Containing PCBs
This category of treatment has received the most attention in our
study, because it has received the least attention in practice thus far. Our
survey has found a variety of methods applicable to reduction or "zero dis-
charge" of the quantity of PCBs disseminated to the environment by wastewaters,
2.2.1 Carbon Adsorption
This is a water purification and industrial product puri; i-
cation technique that has been widely used over a good part of the 20 tb century,
The plant in Marshall, Texas that makes "DAROO" brand carbon for IC1--US was
originally built in about 1912.
The early consideration of carbon adsorption ao a FFHFIS of
removing -PCBs from wastewater is logical and reasonable. Carbon adsorption .Is
most effective in removing high molecular weight, non-polar, and relatively
insoluble compounds from water. All these requirements art-1 met by the PCBs.
As will be discussed later, we have cjondiiat-ed recent, tests,
in cooperation with carbon suppliers, showing the effectiveness of ^rerovs.l of
PCBs from wastewaters such that reasonable effluent requirements ''unld K-» met;
say of the order of 1 ppb.
The apparent disadvantage in using carton, {-'ne'. .
investigation, is that once PCBs are adsorbed on the carbon, ,u ",v " !>>,vr ;o
undergo destruction by high temperature incineration, rather. t;.-r '
such as is normally practiced.
95
-------
A great advantage of carbon treatment is that recent studies
by Calgqn and others have shown that a wide spectrum of other toxic organics
can be effectively removed from water by activated carbon.
Fran practical, operating and economic points of view, carbon
adsorption is favored since it is currently being used in large scale municipal
water purification systems. This gives good confidence in predicting capital
and operating costs, and reliability factors.
At the 68th Annual AIChE Ifeeting in November 1975, G.
Strudgeon of Zurn Industries described design features for a 30 million gallon
per day carbon treatment system for municipal wastewater at Garland, Texas.
This entire system will cost $5 million, including carbon regeneration. Oper-
ating costs are expected to be $6,000/day.
Although carbon adsorption is thus far the only mechanism that
has been proven effective in removing PCBs, that same meeting provided data on
three other mechanisms by which carbon has removed other jxjllutants from waste-
water: 1) by biodogradation, when bacteria coated carbon particles, and then
attacked pollutants in the wastewater; 2) by catalytic action, i.e., by pro-
viding a very active surface, and holding pollutant molecules while other
degradation or oxidation reactions took place; 3) by chemical reaction, in
which the carbon in the bed is actually depleted as part of a chemical reaction
that removes pollutants.
The utility of any of these modes of activity for PCBs removal
remains only a potential at this time.
As ail adjunct to this evaluation study, we have conducted
cooperative laboratory adsorption isotherm tests with Carborundum Company, and
ICI-US. This gave new data on the removal of higher concentrations, and lower
concent rah ions, of PCBs from wastewater for combination with the published data
by Ca'Lgcn,
96
-------
Laboratory adsorption isotherm testing is a reliable technique
for determining the feasibility of adsorption treatment for PCBs removal from
wastewaters. This testing procedure specifically indicates:
1) The effluent levels of PCBs concentration
obtainable by adsorption treatment.
2) The weight of PCBs that will be adsorbed by
the adsorbent at the concentrations being
studied.
On the other hand, the laboratory test is not used to deter-
mine the necessary contact time for granular carbon beds to effect the desired
reduction. This determination is usually performed in small pilot carbon beds
under dynamic hydraulic flow conditions—also, a well-established procedure is
reported in the literature.
Laboratory and pilot plant tests of PCBs removal must be care-
fully carried out to prevent inadvertent losses of PCBs. The root of the problem
is the very low, parts per billion, solubility of PCBs in water. Examples of
pathways of experimental loss are:
a) Evaporation to the air; as described in 1.2.4; and
b) Adsorption on a variety of solid surfaces, sediments
and the like.
2.2.1.1 PCBs .Adsorption Testing by Carborundum Company
Versar Inc. together with Carborundum Co. conducted
a preliminary study to determine the ability of an experimental, coal-based,
activated carbon in removing PCBs from water. This study also provided an ana-
lytical check in that both companies ran electron captive gas chrcmatographic
analyses on the same samples; and there was close agreement on results.
The activated carbon had a surface area of 950 to
1050 square meters per gram. The iodine number was approximately the same as
the surface area, indicating that almost all of the pores had a diameter greater
than 10 to 15 angstroms.
97
-------
The control, and test samples were filtered before
extraction and gas chrcmatographic analysis. The removal of PQBs by solids,
surfaces, filter media, and the like, was appreciated before these tests, thus
a high PCBs concentration in the filtered control was the target.
The PCBs used were the Aroclor 1254 mixture. It was
solubilized by methanol to give a 1000 ppm stock solution. This was diluted
with distilled water to prepare the 1000 ppb test solutions.
Table 2.2.1.1-1 presents the data. The filtered
control level is considered the actual ppb level that the carbon was adsorbing,
rather than the prepared level. The filter removed approximately 85 percent
of PCBs in these test runs.
As shown in Table 2.2.1.1-1, the percent PCBs re-
moval from a rather concentrated water solution (160 ppb), at low carbon
dosages was quite good since during our plant Surveys we found PCBs concen-
trations in wastewater destined for river discharge to rjange from 50 to 500 ppb,
the level treated in this study is an intermediate value. It is, however, the
highest value treated in the cooperative studies under this program.
Carborundum Co. plans to continue this effort to
determine the effects of other carbon dosage levels on various feed PCBs con-
centrations. The above reported data is only considered a preliminary effort.
2.2.1.2 PCBs Adsorption Testing by ICI-US
In a cooperative program with Versar, ICI-US per-
formed preliminary adsorption tests to determine the ability of powdered carbon
in removing Aroclor 1254 from water. Versar conducted all analytical tests for
this program.
The Aroclor was solubilized with methanol so that a
1000 ppm concentration of Aroclor 1254 in water was achieved. Two types of
granular activated carbon were tested: a lignite base, and a coal base. Both
types were ground to a fine powder (90% through 325 mesh) before adsorption
testing. Prior to grinding, the lignite base and coal base carbons had a sur-
face area of about 650 square meters per gram and 1000 square meters per gram,
respectively.
98
-------
vo
00
•<* IT)
• •
CO 00
0>
cn o
IT) (N
* ro"
vo o
• •
(N
IN "n"
vo •»»
CN
-------
The test solutions were made up with distilled, de-
ionized water to a volume of one liter. Four levels of Aroclor concentrations
were treated: 10 ppb, 100 ppb, 500 ppb and 1000 ppb. Both the control and
treated solutions were filtered before analysis. The great affinity of PCBs
for all solid surfaces was not fully appreciated, during these tests. The
filtration was uniform for all samples through 2.4 on Reeve Angel fiber glass
discs (Grade 934AH) . It can be seen in Table 2.2.1.2-1 that most of the PCBs
were removed in filtration. However, by using the filtered "control sample"
as an approximation of the amount the filter removes fram all feed samples,
we can get a good preliminary estiirate of the PCBs removal ability of the two
kinds of commercial carbons at varous wastewater pollutant levels, and at
various carbon dosages.
The results of these laboratory tests indicated that
powdered carbon was highly effective in removing PCBs from, water at four
levels from 1 to 40 ppb. There was not a great deal of difference between the
effectiveness of the two carbons. The lignite carbon has relatively large
pores, and a smaller surface area than the coal-based carbon. The lignite-
based carbon simulates the activity of the coal-based carbon after a number of
thermal regenerations. Thermal regeneration tends to increase the pore size,
and lower the surface area of any given carbon. These kinds of activated
carbons, however, tend to stabilize at about 550 square meters per gram, even
after many regenerations.
The conclusion from these results is that carbon-
adsorption can be an effective method of removing PCBs from wastewaters, even
after many thermal regenerations.
It is quite significant that in all these tests, the
treated effluent PCBs levels ranged from 90 to 320 parts per trillion, which
are well below the target maximum PCB level of 1 ppb.
Further testing is needed to get confirmatory iso-
therm data and column test data. In column tests, the granular form of carbon
100
-------
m CN
oo oo
rH O
oo en
en en
oo r-
en a\
(S
en en
en en
m
14H
W
^*ft
nr
r- o
r- en
en
o
-------
is vised, and, therefore, sate of the interior portions of the carbon are not
as accessible as they are in the ground form. Thus, the tests with powdered
carbon given on a weight fraction basis, i.e., pounds PCBs removed per pound
of carbon is the maximum weight fraction of PCBs that can be removed in a
scaled-up, commercial system.
2.2.1.3 PCBs Adsorption Testing by Calgon Corp.
Adsorption isotherms were run on Aroclor 1242 and
1254. Stock solutions of each compound were prepared in acetone at 100 mg/1.
•o
Filtrasorb 300 (FS-300) activated carbon was used throughout. Carbon was
added from a stock suspension of 1 g or 2 g of pulverized ES-300 per liter of
distilled water.
Isotherms of Aroclor 1242 and 1254 were prepared by
the following method. Exactly 1 ml of PCS solution was added to seven flasks,
each containing slightly less than 1 liter of distilled water. Measured
volumes of the 2 g/1 carbon stock solution was added to each flask to give
carbon concentrations of 0,2,5,10,25,50 and 100 mg/1. The total volume of
each flask was 1000 ml. After four hours agitation on a wrist shaker, each
solution was filtered through 0.45 micron millipore pads and stored in a re-
frigerator in quart amber glass bottles with Teflon-lined caps until analyzed.
A nickel-63 electron capture gas chromatograph was
used for analysis. All samples were extracted and concentrated approximately
100 times before analysis. The method is described in the 1971 EPA report,
"Methods for Organic Pesticide Analysis in Water and Wastewater."
Table 2.2.1.3-1 shows the PCBs removal data, and
gives comparative data for Aldrin. It appears that the Aroclors are removed
similarly to the pesticides. Removal froti the 50 ppb level to the 1 ppb level
seems possible.
Figure 2.2.1.3-1 shows the Calgon data plotted to
give the weight percent of PCBs that Filtrasorb-300 carbon can pick up, at
levels down to 1 ppb. It can be seen that the curves take a downward break
at about 2 to 3 parts per billion, indicating that the; weight of activated
102
-------
TABLE 2.2.1.3-1
RESULTS OF CALGON CORP. LABORATORY ISOTHERM
TESTS FOR CARBON REMOVAL OF PCBs
Carbon
Dosage (mg/1)
Residual (ppb)
Aroclor 1242
Control 45
1.0
2.0 7.3
2.5
5.0 1.6
10.0 1.1
12.5
25.0
50.0
Aroclor 1254
49
—
37
—
17
4.2
—
1.6
1.2
Aldrin
48
—
26
—
15
12
—
6.3
4.4
103
-------
VI
Q>
41
INN i i i » mn IMI
uiii
O
104
-------
carton required to achieve removal of a unit weight of PCBs is rising rapidly.
If this kind of data is confirmed with larger scale column testing, it would
mean that removal of PCBs from water in the parts per trillion range is far
more difficult than removal in the parts per billion range.
Also from the figure it can be seen that the initial
concentrations (Co) of the Aroclors to be tested were about 100 ppb. However,
filtration before testing removed PCBs so that the starting concentrations for
the tests were 45 and 49 ppb.
Hager and Rizzo of Calgon Corp. have described the
essential elements of full scale adsorption systems in a paper presented to
the EPA Technology Transfer Session on Treatment of Toxic Chemicals; Atlanta,
Ga., 4/19/74.
Each system is comprised of three basic functional
components:
1. The adsorption treatment of the wastewater
2. The carbon reactivation equipment
3. The carbon/water transport arrangement.
For PCBs, the practicality of reactivation must be
determined.
2.2.1.3.1 Adsorption Treatment of the Wastewater
The adsorbers hold the granular activated
carbon beds through which the wastewater flows. They can be designed for
pressure or gravity flow to achieve the desired contact time of the water with
the carbon. Suspended solids and space limitations also must be considered in
the adsorber configuration. Flow rates usually fall under 10 gpm per square
foot of carbon bed surface area. Contact times for industrial wastewater
mixtures usually are in excess of 60 minutes, which is about twice the time
employed for purification of domestic sewage. When suspended solids are
present, they will be filtered out of the carbon bed. This dual purpose of
carbon beds can be usefully employed as long as the adsorbers are designed to
accommodate periodic backwashing and bed-cleaning procedures, such as air scour
105
-------
and surface wash. For PCBs service, minimum backwashing would be desirable,
as it creates quantities of concentrated wastewater for incineration or treat-
ment. Some settling and prefiltering will be required even when subjected to
incineration.
A well-designed water distribution or
underdrain will insure good backwashing performances as well as an even distri-
bution of the water. Well established filtration design practices can be ef-
fectively employed in carbon bed design systems. Carbon beds must be periodi-
cally removed via water slurry and obstructions to the flow of carbon from the
adsorber should be avoided in adsorber design.
Carbon beds are normally in excess of 10
feet deep and usually fall into one of four basic configurations:
1. moving beds
2. beds in series
3. beds in parallel
4. expanded beds.
The type of configuration selected de-
pends on a number of variables; principal among them are total water flow,
suspended solids, and degree of contaminant reduction desired.
2.2.1.3.2 Reactivation of the Granular Carbon
Thermal oxidation is generally employed
to reactivate the exhausted carbon for reuse using either multiple hearth
furnaces or rotary kilns. The size of the thermal reactivation equipment is
based on the carbon exhaustion rate, i.e., pounds of carbon exhausted per
thousand gallons of wastewater treated, and the weight of contaminant on the
carbon. Excess capacity is designed into the thermal reactivation unit to al-
low for variances in carbon use rate due to changes in the wastewater flow and
organic loading.
The exhausted granular activated carbon
is heated to 1600-1800°F to effect volatilization and oxidation of the dis-
solved organic contaminants. Oxygen in the furnace is normally controlled at
less than 1% to effect selective oxidation of contaminants over activated
106
-------
carbon. A 5% loss of activated carbon per reactivation cycle is an acceptable
bench mark upon which granular carbon system economics may be based. Particu-
larly for the PCBs which will require an incineration temperature in excess of
about 2200°F, and a few seconds residence time, the reactivation equipment must
include an afterburner. An air scrubber with HC1 neutralization will be the
last element of the reactivation train. Thus far, the regeneration of carbon
that has seen PCBs adsorption service has not been proved.
2.2.1.3.3 Carbon Transport
Granular spent carbon is usually trans-
ported between the adsorbers and the reactivation equipment by water slurry.
Various pumping designs can be employed including centrifugal and diaphragm
pumps as well as movement using either hydraulic or pneumatic pressure. The •
transport piping should include flush ports and wide radius bends. About one
to three pounds of carbon per gallon of water are used, depending upon dis-
tance and elevation considerations.
2.2.1.3.4 Materials of Construction
Special consideration should be given to
the selection of materials of construction used in the adsorption system from
two standpoints:
1. Galvanic Corrosion - any tendency
toward galvanic corrosion due to water characteristics, such as conductivity
and pH, will be enhanced in granular carbon beds. Mild steel tanks or ad-
sorbers holding granular carbon under water should be lined; tanks can also
be constructed of cement or reinforced synthetic resins or plastics.
Carbon-in-water slurry piping, which ex-
periences only periodic exposure, is usually exempted from and special cor-
rosion considerations.
2. Erosion - periodic replacement of the
carbon beds can cause lining failures at exit ports in carbon tanks and ad-
sorbers. Special grades of stainless steel are usually employed for such wear
points.
107
-------
2.2.1.4 Carbon Regeneration Alternatives - Wet Catalytic
Oxidation
For smaller installations, it might prove feasible
to dispose of the spent carbon by incineration. The carbon supply companies
also offer services of removal of spent carbon, and off-site regeneration,
when delivering fresh carbon.
However, there is a new wastewater treatment tech-
nique that might be applicable to carbon regeneration, and should be evaluated
soon. "This involves wet catalytic oxidation. Some of the Laboratory studies
in wet catalytic oxidation are described in Appendix D. DD<:kheed Missiles and
Space Co. (Waste Treatment Systems Section) of Sunnyvale, California, announced
in January, 1976, that they have a process ready for pilot plant evaluation
and conmercial scaleup.
In a very preliminary look at the PCBs destruction
potential of this method, a sample of wastewater from an unnamed user of PCBs
was subjected to treatment. The PCB mixture approximated Aroclor 1221 most
closely. The wastewater sample was very concentrated, 5 pprn. Since this is
much higher than that Arcolor's solubility in pure water it is believed that
solubilizing agents or particulates were carrying most of the PCBs into the
mixture. At any rate, within 20 minutes of treatment time, about 90% of the PCBs
had been destroyed. No large quantities of reaction products were detected.
However, this is just one data point and much further confirmation is needed.
The test reactor was a continuously stirred batch system, with air continuously
sparged in. The test conditions used were; reactor pressure, 1500 psig;
temperature, 550T1; catalyst concentration, 0.5 gram; and reactor volume, 1.5
liters.
A proposed use of this catalytic system is for
carbon regeneration when it is in the spent water slurry form.
Since the capital and operating costs of the catalytic
system are tied to the hydraulic load, it might be practical to trap PCBs with
carbon in wastewater, and then treat the carbon slurry by catalytic oxidation
108
-------
to destroy the PCBs. By this method, the PCBs would be removed from the waste-
water by the highly efficient Carbon adsorption method, and the volume handled
by the catalytic system would be greatly reduced.
At this point, feasibility testing for regeneration
of activated carbon by catalytic oxidation is needed. The activated; carbon
industry has long sought alternatives to thermal regeneration of carbon, be-
cause of its exposure and the loss of carbon by oxidation in that regeneration.
Whenever they have tried solvent regeneration, or acid or base treatment, the
first regeneration only produces about 40% of the active surface. This then
drops to 30% usually after a few more treatments. Such a degradation is un-
acceptable. Thus, the points to be determined are: a) can this process de-
stroy the adsorbed PCBs on carbon, converting them to GCL, Ho^ anc^ HCl? and .
will the regenerated carbon have, say 90-95% of its original activity? and
c) can the process be operated economically? The latter involves the neces-
sary pressure (500 to 2000 psi), operating temperature (300 to 650°F), source
of oxygen (air, CL or ozone). type and lifetime of catalyst, residence time,
materials of construction, and similar variables.
2.2.1.5 Further Applications Data
ICI-United States provides an excellent booklet en-
titled "A Symposium on Activated Carbon", providing considerable detail on ap-
plications. They also provide information on special graphical procedures
helpful in the scale-up from isotherm to column testing.
2.2.2 Ultraviolet-Assisted Ozonation
Both UV radiation, and ozone, separately, have been used in
water purification for some time. It is only over the last five years or so
that the synergism of the conbination toward the destruction of organics in
water has been appreciated.
Two conmercial organizations: Houston Research, Inc., of
Houston, Texas; and Westgate Research .Corp. of Marina Del Rey, California
are engaged in development of UV-assisted (or catalyzed) ozone oxidation of
109
-------
refractory organics. Both organizations have cooperated in preliminary tests
of PCBs destruction, and the method has shown great premise as a viable, large-
scale, economic water treatment method. They are working mostly with the 2537
Angstroms "near" UV radiation; but plan to investigate the "far" UV mercury
radiation at 1845 Angstroms.
Of particular interest is the fact that these methods hold
premise of destruction of hydrocarbon organics completely to C0? and water.
Thus they are likely candidates for zero discharge and total recycle systems.
Related to this work is that of Dr. Lawrence of Environment
Canada at Burlington, Ontario. He is concentrating on the near UV radiation
of about 3650 Angstroms, available from sunlight, but using titania or alumina
as photocatalysts, rather than the combination with ozone. The ultraviolet
region extends to 3800 Angstroms, and then blends into visible violet. Other
related work using visible light is described in later subsections.
Generally the UV radiation is provided by conmercially avail-
able tubes, which when operated at low power (and low pressure in the emitting
tube) are quite efficient in transferring UV radiation to water. The path
length must be kept quite short, probably of the order of a few inches maximum,
as the absorption in the water, and by even small amounts of particulate, will
be nearly complete in that length.
Ozone is sparged into the reactor, and vigorous stirring is
provided, to the point that increasing turbulence or mixing power input does
not give an increase in reaction rate.
Various kinds of efficiency values are given to rate different
systems, flows, arrangements, etc. Total organic carbon removed per watt-
second of UV input, or per gram of ozone introduced, are examples of efficiency
measures.
Since ozone generation is expensive, work is directed toward
maximum use of the ozone introduced. The expense is almost completely the
electric power expense of the silent discharge tube method of making ozone
frcm air. Of course the UV expense is also an electric power expense.
110
-------
The following subsections cover discussions on the general
attributes of UV radiation and its molecular interactions, the photodegrada-
tion of PCBs, and finally UV-assisted ozonation experiments with PCBs.
2.2.2.1 Ifolecular Responses to Ultraviolet Region Energy
Energy absorbed in the ultraviolet (UV) region (100-
3800°A; or 10 to 380 nm) produces electronic transitions within the molecule.
The principal characteristics of an absorption band are its position and in-
tensity. The position of maximum absorption (X max) corresponds to the wave-
length of radiation having energy equal, to that required for an electronic
transition. A molar absorptivity at maximum absorption (E max) of 10,000 or
greater is regarded as high intensity absorption. Low intensity absorption is
considered an E max of less than 1,000. Biphenyl in alcohol at 250 nm has an
E max of 18,000; and in hexane at 246 nm, an E max of 20,000.
This would seem to indicate that the basic structure
of the PCBs would be readily activated by a mercury lamp generating 253.7 nm
radiation. However, this absorption will be modified by the number of
chlorine atoms attached to biphenyl in forming the PCBs. If at any point in
the destruction of PCBs saturated hydrocarbons are formed, they will be un-
responsive to 253.7 nm radiation. Saturated hydrocarbons contain sigma
electrons exclusively. Since the energy required to bring about ionization of
the sigma bonds is of the order of 185 k cal per mole it is only available in
the far ultraviolet, below 200 nm. Single carbon-carbon and carbon-chlorine
bonds have about the same high bond energy, and are similarly stable to rupture.
However, there is a mercury emission at about 185 nm, which should be capable
of activating those bonds.
When irradiated by the proper UV wavelength, excita-
tion of bonding electrons takes place, and free radicals are produced which
are highly reactive in the presence of ozone.
Thus it appears theoretically possible to produce
the excited states in PCB molecules necessary to make them highly receptive to
oxidation by the emissions from mercury vapor tubes. Since ozone is a power-
ful oxidizing agent, and it too is excited in the same regions making it even
111
-------
more effective, the possibilities for a wide range of oxidations are present.
This would seem to make complete destruction of PCBs to CD-, water and HC1
feasible.
2.2.2.2 Photodegradation of PCBs
Because of the strength of the C-C1 bond, it
had been assumed for some time that little photochemical cleavage would occur.
Further, in the photochemical breakdown of DDT and related compounds, cleavage
of the aromatic C-C1 bond is usually not involved. However, Safe and Hutz-
inger reported in Nature in 1971, that a hexachlorobiphenyl photolyzed rather
readily in organic solvents when irradiated at 310 run. It gave products which
are formed by stepwise loss of chlorine, rearrangement, and condensation. A
number of additional researchers have found that some' pure chlorobiphenyls
and some PCB mixtures can be decomposed by laboratory UV sources and by sun-
light over long time periods.
The wavelength of the UV source appears to be quite
critical to successful photochemical deccrnposition. Low pressure mercury
lamps emitting 2537 Angstroms (254 nm) seem to be several times as effective
as the UV from sunlight which is greater than 2950 Angstroms (295 nm).
Studies have been carried out in organic solvents and water, cind in both
liquid and vapor states.
Much of the past work has been on organic solvent
solutions of PCBs to get higher concentrations to work with. Vapor phase
studies have been run to simulate treatment of atmospheric emissions in en-
vironmental studies.
Reductive dechlorination, the main photoreaction of
chlorobiphenyls, is faster in hydroxylic solvents such as methanol and iso-
propanol. Even chlorobenzene loses chlorine rapidly on irradiation in iso-
propanol. Nishiwaki, et al, observed complete dechlorination of a PCB mixture
in 15 minutes in an alkaline isopropanol solution, using a mercury lamp.
Biphenyl and sodium chloride were identified in the reaction mixture.
112
-------
Photosensitizers can increase the decomposition rates.
Tryptophan, diethylaniline , benzophenone and triphenylene sensitized the photo-
reaction of dichlorobiphenyl . The reaction was quenched by n-hexyl mercaptan
and di-n-butyl sulfide. Experiments in water solution have been hampered by
adsorption of the PCBs on the walls of reaction containers, particularly glass.
Attempts have been made to use surfactants to keep
the PCBs in suspension. Adsorption on solids like calcium carbonate, silica,
soils, etc. in water suspension has been tried to aid in photolysis studies.
However, UV radiation will not penetrate deeply into solids, so that adsorbed
PCB must be kept close to the surface.
2.2.2.3 Experiimental Factors in UV-Assisted Ozone Oxidation
o£
For several decades the advantages of ultraviolet
radiation in the sterilization of aqueous and dry media have been known. Like-
wise, the powerful oxidizing effects of ozone, which produce sterilization,
have been known. It is only with the relatively recent advent of requirements
for removing refractory organics from water, and the ability to detect these
organics in parts per trillion and even lower concentrations , that the pressing
need for more powerful oxidizing capability has been felt. It was quite natural
to combine the two effects into a single or staged treatments. It has been
found that there is a strong synergism in many cases.
Enough data is now accumulating to show that this
category of wastewater treatment is a powerful method for removal of refractory
organics from water. It has the potential of removing organics to an effective
zero discharge level. However, as the removal of very small concentrations be-
comes exponential with time, the residence time required in UV-ozone treatment
equipment becomes a critical factor. The following variables have been identi-
fied as affecting residence time:
- MDlecular structure of the organic
- Concentration
- UV transmissivity of the wastewater
- UV intensity
113
-------
- UV wavelength
- Ozone concentration
- Liquid turbulence and gas-liquid contact
(transfer coefficients)
-EH
- Temperature
The following sub-sections will present experimental
findings about the importance of these variables.
2.2.2.4 Destruction of PCBs and Refractory Qrganics at
Itesearch, Inc.
2.2.2.4.1 PCBs Destruction Date
In preliminary experimentation to shew
feasibility of PCBs destruction, a high pressure 650 watt mercury tube gener-
ating 2537 Angstrom radiation was used in a 21 liter reactor. Figure
2.2.2.4.1-1 shows the arrangement for a smaller reactor. Oxygen was sparged
in. at 3 liters per minute, with an ozone concentration of 2%. This was felt
to be excess ozone usage, but a good starting point for tests. A 1000 ppm
solution of Aroclor 1254 in methanol was used to get a 514 ppb solution of
Aroclor 1254 in water. Otherwise the solubility of Aroclor 1254, which is
about 50% pentachlorobiphenyl, is only 12 ppb in water. It was theorized that
significant destruction of this compound would forecast even greater de-
struction of the other Aroclors now produced, which are all less chlorinated
and more soluble in water.
Figure 2.2.2.4.1-2 is a plot of the
normalized residual concentration of Aroclor versus time of UV- assisted
ozonation. it can be seen that in one hour about 2/3 of the PCBs had
been decomposed; and in three hours, only about 7 percent of the original
PCBs remained.
2.2.2.4.2 Operating Data Obtained from Refractory
Qrganics l-3
During the past four years work has
progressed at Houston Research on ozonation for water purification . Comparison
tests have shown that the addition of UV radiation enhances the reaction rate
114
-------
Variable
Speed
Mixer
BCS
0V Light
or
liter
ity)
~~-
L
1 1
Ex
^_ — , —
r
i
i i
iaus t ^^
Gas
Te
pll
an
Temperature
Control
Moni toring
Sampling
Ozone
Gene rator
Air
Oxygen
Figure 2.2.2.4.1-1.
Lab Scale Apparatus for Reaction and Mass
Transfer Studies at Houston Research, Inc.
115
-------
for the destruction of organics by 10 to 100 fold. Further it was found that,
for the most refractory compound they had tested prior to the PCBs, acetic
acid, there was essentially no reaction without UV assistance. However, with
UV radiation, the oxidation proceeds rapidly at room temperature .
Figure 2.2.2.4.2-1 shows the effects of
UV and tenperature on acetic acid destruction. The ordinate is the fraction
of total organic carbon remaining, showing that for the 30°C or 50°C tests,
with UV, there is nearly couple te destruction in four to five hours to CD- and
water. The fact that the curves are displaced to the right of the mass trans-
fer limiting line shews that some chemical reaction rate improvement can be
sought.
Figure 2.2.2.4.2-2 shows the oxidation
improvement achieved by doubling the UV power input, with near constant
temperature,
Further description may be found in:
"Ozone/UV Process Effective Wastewater Treatment", by Prengle, Ma.uk, Legan
and Hewes; Hydrocarbon Processing, October 1975, 82-87.
These are a sampling of the kinds of opti-
mization experiments that must be run with the PCBs oxidation system to obtain
good economy of design and efficiency of operation.
2.2.2.5 Destruction of PCBs and Refractory Organics at
- "
2.2.2.5.1 PCBs Destruction Data
Cooperative testing and research between
Versar Inc. and Westgate Jtesearch, Inc. was performed to determine the ef-
fectiveness of Westgate ' s UV-assisted ozonation process in destroying PCBs .
The following experimental conditions were used in
the treatment of synthetic wastewaters containing Aroclor 1254 and Aroclor
1016:
116
-------
iW CC+)
o
*H
in
E!
!' I I I'l I-.I I i I'I I''! Mll| :--Li. J._!.|..:..l_!_l_
Figure 2.2.2.4.1-2.Aroclor 1254 Destruction by UV-Assisted ;
I i ! I I I '
IiiitialJItoncentrataorL-qf-Aroclor - 514, ppl
-i—I
117
-------
TO
o
•rt
0
o
6
41
M
3
U
rt
o
H
r:
Qj
60
s
o
-------
01
at
a
c
at
e
o
o
o
119
-------
Reactor volume = 3 liters
Reaction time = 4 hours (excess time used to
give best chance of destruction)
UV Source = one 43 watt, low pressure Hg lamp
operating at 2537 Angstroms
Reactor type = vertical, cylindrical, 18" long,
3" diameter (1% UV path length)
Pressure = atmospheric
Temperature = 23°C
£H = 6.2
Ozone feed = 70 milligrams ozone/minute in
3.4 liters per minute of oxygen;
or about 1.4% by weight ozone in
oxygen
Reactant preparation = The pure PCBs (Aroclor
1254/ were mixed with an
equal proportion of
methanol, and then with
distilled water to get an
apparent true solution. A
Hamilton syringe, with
vernier calibration to de-
liver down to a microliter,
was used to prepare an
estimated 200 ppb concentra-
tion in a beaker. This
solution was added to the
reactor (3 liters). Then
a 200 ml sample was with-
drawn to get the "before"
sample,.
When preparing the Aroclor 1016 solution,
a more concentrated PCBs solution of about 800 ppb was achieved.
The before and after treatment concentra-
tions of PCBs are shown in Table 2.2.2.5.1-1.
120
-------
Table 2.2.2.5.1-1
UV Ozonolysis Destruction of Typical Capacitor and
Transformer PCBs at Westgate Research
Aroclor 1016 Aroclor 1254
Initial (influent) Concentration (ppb) 790 200
Final (effluent) Concentration (ppb) 0.5 1.5
Table 2.2.2.5.1-2
Destruction of Aroclor 1016 by UV-ozonation at Westgate Research
Chlorinated
UV-Ozonation Time Aroclor 1016 Cone. Products Cone.
(minutes) (ppb) (ppb)
0 237.0 0.00
15 14.7 33.8
30 7.76 25.86
45 1.73 21.80
60 1.73 18.47
75 0.52 17.63
90 5.2 21.43
120 <0.1 12.64
150 <0.1 16.91
180 <0.1 9.12
210 <0.1 23.07
240 <0.1 12.58
121
-------
The data show that the destruction was
highly effective in these pioneering tests; over 99% of the original PCBs were
destroyed. In addition, the final concentration was at the desired level of
about 1 ppb. Such an effluent would be expected to be a reasonable stream for
recycle operations and intermediate term zero discharge potential.
Of course the contact time and ozone ex-
penditure were overly large, but the goal of this preliminary testing was to
demonstrate destruction of two key Aroclors in use today.
A more comprehensive study of the UV-
ozonolysis of Aroclor 1016 was run, with samples removed from the reactor
every 15 or 30 minutes for up to 4 hours. As shown in Table 2.2.2.5.1-2, the
initial PCBs concentration was 237 ppb. Within 45 minutes the PCBs level had
been degraded to 1 to 2 ppb, a 99 percent decomposition. By two hours the
PCBs concentration was less than 100 ppt (parts per trillion). The last
column in the Table labeled "Chlorinated Products Concentration" is of parti-
cular interest. This is the first time quantitative data on the residual com-
pounds in the reaction mixture have been compiled. These residual compounds
occur at 33.8 ppb level within 15 minutes from the start of the test and then
erratically and slowly drop in level to about 10 to 20 ppb over the four hour
period.
The only statements that can be made
about these residual compounds is that they are non-PCBs but chlorinated
materials. Possibly an altered UV wavelength, or ozone concentration, or use
of catalyst or other agent could degrade these other products to the non-
detectable level. However, further testing is required.
2.2.2.5.2 Pilot Scale Tests of Refractory Organics
Decomposition
In a paper presented at the 68th annual
meeting of the AIChE, November 20, 1975; "UV-OX (TM) Process for the Effective
Bemoval of Organics in Waste Waters", Zeff has described pilot scale tests of
UV-assisted ozonation using 2537 Angstrom radiation. This work has grown from
a patented invention for a household appliance to purify tap water. The
122
-------
effectiveness of the method is demonstrated by the reduction in TOC of 500 ml,
of tap water from 5 mg/1 to 1 mg/1 in one minute, using 0.07 mg of ozone and
0.1 watts of 253.7 nm UV. The bacteria plate count was also reduced by a
significant amount.
In these studies, batch reaction con-
ditions were optimized to get closest to complete TOC destruction with least
UV energy input and best 0, use. Then continuous and two stage operations
were investigated. As shown in Figure 2.2.2.5.2-3, a six inch path length of
UV was compared with a three inch path. In a batch test, the longer path con-
dition took 3 to 4 times as long as the snorter path to achieve the same TOC
reduction, in approximate accord with the inverse square law of intensity re-
duction.
The laboratory scale equipment arrange-
ment is shown in Figure 2.2.2.5.2-2. Table 2.2.2.5.2-1 presents TOC re-
ductions for a 5 part organic mixture under two experimental conditions, show-
ing ozone usage efficiencies, and energy usage efficiencies, with a simulated
two stage continuous operation. Zeff has calculated the prime energy costs
for UV assisted ozonation based on a scaleup of a pilot reactor. The pilot
reactor handled 20 gph, so that 210 of these units in parallel would handle
100,000 gpd. The plant area needed for this equipment would be only 18 feet
square, with a height of 3 feet.
Based on pilot tests of this reactor, it
might be assumed that it could convert wastewater of 40 mg/liter of TOC to
less than 5 mg/liter, using six-43 watt UV lamps. It is assumed that a UV
energy requirement of 5 watt-^ninutes per milligram of carbon oxidized will
suffice. Also, the ozone requirement will be 2 moles per gram-atom of carbon
oxidized, but at a 75% efficiency, or 2/.7S = 2.67 moles ozone per gram-atom
of carbon. Assuming a power cost of 1.5C/KWH; the ozone generating power cost
would be $53.40 and the UV power would be $18.93; or a total of $72.33 per
100,000 gal. of water treated.
123
-------
CM
"4-1 C -^
o o E
•H
4J iJ W M >-, CO
C O
o o,
M£ I T3S-5
•H 4J n QJ CO
HO, N •
O.D
B -H
o> x
O
o
i-l
CM
O
• CO
.CM
r-l
C
'e
I
-------
Ozone
Generator
Oxygen
or
Air
Supply
Flowmeter
Flowmeter
Purified
Water
Figure 2.2.2.5.2-2
SCHEMATIC OF BENCH REACTION SYSTEM AT WESTGATE RESEARCH CORP.
125
-------
o
pi
o
H
O
<:
w
0/001
O
03ui/ui-« xopuj AH
UTUl/ViUI
001 3°
. °L
0/001
UTIU/2U1
001 3°
T/3iu
001
utm/3ui
0
ouoo
uxui/3w
SSBH e
• 0/001 pa-tlddns
03tu/ui-M xapui
UTUi/3ui
001 jo
1/Sui
001 ^
ufui
tl'tUI/'JlU
SSL'H ^o
ouoo 0
u"tw
001
01103
ro
vo
O
o
VA
in
CNI
0
a-
o
o
m
vO
in
cs
o
CO
0
cr>
vt
m
in
_co_
in
r--
ro
0
vo
O
CN
1—1
O
r-l
C-J
OO
r-l
LO
O
o
J=
o
i-
o>
-M
«s
en
>o
X
11
c
o
Q.
ro o
I CJ
CJ
• in
in
• (O
CM
CM C
o
o
M
o
i_ t-
o o
o. ex
c c
ro CO
-------
Zeff points out that the 210 reactor
nodules would require 1260 of the 43 watt lamps, at 6 per reactor. This ap-
pears to be a large number of lamps, but it is actually very practical. Such
arrays of lamps are regularly used for room illumination in factories and
large offices, and the low-pressure mercury lamps used here are nothing but
fluorescent lamps without the phosphor coating. An extended array such as
this is much more energy-efficient than fewer high-power, high-pressure, "in-
dustrial" UV irradiators; and maintenance and replacement of lamps is much
simpler. The life of the low-power lamps is 7500 hours as compared to 1000
hours for the high-power lamps.
2.2.2.6 Laboratory Test Results fron AJResearch Corp.
In these tests, high pressure 450-watt mercury lamps
were used. The test solutions were very dilute mixtures of methanol, ethanol,
isopropanol, and acetone in water. The ozone was introduced into the 4 liter
organic mixture in an oxygen stream, that was bubbled in at 2 liters per
minute, at an ozone concentration of 33 rag/liter. With the expenditure of 7.9
grams of ozone, the mixture of alcohols and a ketone was reduced from 115 rag/1
ODD to 10 mg/1 in 2 hours.
2.2.2.7 Comments on UV-ozone tests
These highly encouraging test results by three dif-
ferent companies in the destruction of some of the most refractory organics
encountered in wastewater treatment, give confidence that scaled-up systems
could adequately destroy PCBs in industrial wastewaters. The next step re-
quired is detailed cost effectiveness testing in pilot scale equipment.
2.2.3 Non-Carbon Adsorbents for PCBs
A variety of non-carbon materials, some well known for treat-
ing water, and some that seem to be quite specific for PCBs removal have been
found. Limited cooperative laboratory testing of PCBs removal has been con-
ducted in order to gain an insight into potential effectiveness.
127
-------
Materials considered or examined were:
1) Rohm and Haas Amberlite AD series resins. These
were laboratory-tested for PCBs removal.
2) Polyvinyl chloride. This was tested by Canadian
investigators for PCBs removal.
3) Clays and Humus. These were tested by Monsanto
for PCS removal.
4) Polyurethane. This was tested by Canadian,
Swedish and other investigators for PCBs removal.
5) Sphagnum Peat. This is used in corimercial water
purification, but has not been tried with PCBs.
6) Pplyelectrolytes as floccing agents. These are
not really adsorbents, but could be" aids to re-
moving finely divided adsorbents firm treated
wastewaters. They were not tested,
7) Coal. This was not tested with PCBs, but is
being experimentally used for water treatment.
8) Molecular Sieves. These were not tried with PCBs,
since they were judged to be of improper character.
They would be expected to preferentially remove
water from PCBs.
9) Miscellaneous Sorbents. There are a number of
proprietary Oil Sorbents, such as the "3M Brand"
series, that were not checked with PCBs, but might
have seme application.
2.2.3.1 The Amber lite- XAD Series of Macroreticular Resins
2.2.3.1.1 PCBs Adsorption Testing
Cooperative preliminary experimental work
was carried out between Versar and Robm and Haas to test the PCBs adsorption
capacity of XAD-4 resins. New data was provided which ODnfirmed the effective-
ness of this resin, which had been previously shown by Lawrence (as described
in Appendix D of this report).
Since carbon adsorption is the more es-
tablished technology for removal of organics, it was felt that a side-by-side
comparison of a carbon and an Amberlite(R) resin would c*3d valuable information.
128
-------
These tests showed that the resin and the
carbon have comparable effectiveness in PCBs removal. The resin method en-
visions on-site regeneration, with the concentrated waste PCBs treated by in-
cineration.
Details on the apparatus and materials
used by Rohm and Haas and the results of their experiments are given in Ap-
pendix B.
2.2.3.1.2 Prpoe_ss_Conoept for Besin Adsorption of
PCBs
Based upon the experiments described in
Appendix B and the general Rohm and Haas experience with removal of organics
from wastewater, they have envisioned the following as a workable plant scale
process, subject to further experimentation. The wastewater to be treated is
passed through one or more columns, each containing polymeric adsorbent.
Once the resin is loaded to capacity with PCB, it is taken off line for re-
generation.
A water miscible solvent is usually used
for regeneration, and is in turn displaced from the adsorbent by water. The
stream resulting from this operation is carefully fractionated to optimize
solvent recovery. The final rinse usually contains a very low level of
solvent in water and this must be collected as a PCBs wastewater.
The distillation column permits solvent
recovery at high purity, leaving water and PCB in the bottoms from the column.
To minimize distillation costs, a patented variation of the process, called
"superloading", is used to maintain a high PCB concentration in the final
toxic material to be disposed. Part of this process is illustrated as a sepa-
rator from which the organic phase is PCB while the aqueous phase is recycled
to "superload" the AMBEKLITE adsorbent. A completely enclosed system can
readily be designed to insure minimum operator exposure to PCB.
129
-------
A process concept flow sheet of FCB re-
moval system is shown in Figure 2.2.3.1.2-1.
The experimental work recommended so that
this concept may be moved to plant design stages is as follows:
1. Mare extensive leakage data should be gathered for XAD-4
and other AMBEKLITE polymeric adsorbents, encompassing several influent con-
centrations of PCBs. Other Aroclors should also be tested.
2. The ability of the AMBEKLITE polymeric adsorbents to be
solvent regenerated should be demonstrated and the optimum solvent determined.
3. The capacity of the AMBEKLITE polymeric adsorbents should
be determined over a number of loading/regeneration cycles to see the effect,
if any, of long term operation on capacity.
The description of XAD-4 resin and its
comparison with other Rohm and Haas resins, which also might have application
to PCBs removal, are given in Appendix C.
Further descriptions of non-carbon ad-
sorbents are presented in the Appendix D. They contain relevant experimental
work on PCBs and refractory organics, and are included to give a more complete
picture of the options assessed under this program. Also included in Appendix D
is summaries of catalytic reduction, catalytic oxidation, microorganisms
studies, and ultrafiltraticn and reverse osmosis. For the removal of PCBs
from wastewater, these are all considered to be in the research stages.
Several of them may have potential for contributing to zero discharge tech-
nology.
2.3 Treatment of PCBs - Contaminated Solid Wastes
2.3.1 Incineration
This technology has already been best described under sub-
section 1.3.3. For the variety of solids ranging from granular particulate
like fuller's earth, to large chunks of solids, like tranformer internals, all
130
-------
ADSORPTION
COLUMNS
(2 OR 5)
Rgure 2.2.3.1.2-\ '
PROCESS CONCEPT FLOW SHEET
BY ROHM AND HAAS COMPANY
131
-------
impregnated with PCBs, the best destruction method is ci rotary kiln. This
should be followed with adequate afterburning to prevent: PCBs vaporization.
2,3.2 Sanitary Landfill
The best technology for segregation of the PCBs solid wastes
so that spillage, leakage to waterways, or emissions to the atmosphere will
not occur, is described under section 1.3.3.
2.4 Treatment of Air Emissions
2.4.1 Condensation Methods
Our plant survey showed at least one location that was
practicing chilling of exhaust gases from PCBs processing areas. With the low
vapor pressure of PCBs even at room temperature, it might be expected that
they could be effectively swept out of chilled air with the condensing water.
Uhdoubtably this does occur to some extent; however, the high activity co-
efficients of PCBs would tend to keep the PCBs vaporized at a level somewhere
near their pure liquid vapor pressure at that low temperature. Previous work
by Versar has shown that PCBs in stack emissions from sludge incineration are
not removed by water scrubbing (EPA Contract 68-01-1587).
2.4.2 Granular Adsorption Methods
Although no data was found for collection of PCBs from an air
stream by any form of granular filter, it would be expected that 'such treat-
ment would be effective. In fact, it would be expected that the same adsorb-
ents discussed under 2.2.1 and 2.2.3 above would be the most effective. Acti-
vated carbon removal of organic vapors has been practiced in such widely di-
vergent circumstances and devices as gas masks, kitchen range hood systems,
and submarine air recycle systems.
2.4.3 Catalytic Oxidation of Organics in Evaporated Effluents
The following studies of vapcr phase oxidation seen to offer
a potential for PCBs destruction in air exhausts at much lower temperatures
than incineration levels. Catalysts would have to be resistant to HCl vapors,
132
-------
but fuel savings and insulation savings would be large. It appears that this
procedure would be amenable to all proportions of water and crganics in such
air streams.
Borkowski ("The Catalytic Oxidation of Phenols and other Im-
purities in Evaporated Effluents," Water Besearch !_ 367 (1967) passed the
vapors over a catalyst at elevated temperatures. Copper oxide was the most
active of a large number of catalysts tested and oxidation appeared to be com-
plete to carbon dioxide and water at temperatures over 300 C and a residence
time of about 0.08 seconds. Without the catalyst 1000-1200°C was required to
achieve the same degree of removal.
Walsh and Katzer ("Catalytic Oxidation of Phenol in Dilute
Concentration in Air", Ind. Eng. Chem. Process Design Develop. 12: 477 (1973))
studied contaminated air-water vapor streams over supported copper oxide and
showed that the rate was first order in phenol and relatively rapid between
150 and 270°C. At 150°C and a space velocity of 4100 hr the phenol con-
version was 99.6%, and there was little evidence of any intermediate organics
in the condensate.
Again, these methods of air purification are in the research
stages, and actual testing of PCBs in air is required.
2.5 The Potential for Zero Discharge
Hie best method of achieving zero discharge, in the face of the
practical problem of defining what is a zero concentration, is to establish
total recycle. This appears feasible for wastewater, but not for solids or
air emissions. Fortunately, for solids, we have incineration technology which
promises very high efficiency of destruction simply by setting the temperature
and residence times high enough.
For air emissions, although vapor pressures are low for PCBs, there
is a potential for large surfaces giving off small quantities of PCBs. The
research technology shows approaches to powerful adsorption and destruction
methods for organics in air. If the latter ca_~. achieve destruction at lower
temperatures through catalytic action, ec-ruivalent to that achieved by in-
cineration, near zero emissions are predictable.
133
-------
For wastewater recycle, the powerful methods of adsorption and cata-
lytic destruction premise PCBs reduction to levels low enough such that reuse
is practical. Our survey has shown some municipal and fresh waters have PCBs
concentrations of one or more ppb; and river water can be many times that.
Thus the recycled water at one or more ppb level of PCB would be very suitable
for use.
3.0 RATIONALE MD SELECTIONS OF CURRENTLY RECOMMENDED WASTE TREATMENT METHODS
Based upon our plant surveys of the PCBs-using capacitor and transformer
manufacturers, the PCBs manufacturer and the waste treatment equipment suppliers,
and upon the analysis and evaluation of all available technologies whether in
commercial use, pilot plant, or research states, we have daveloped recom-
mendations for the most practical treatment methods to use now, and pre-
dictions of what could be applicable over the short and long term future. Our
current reccmmendations are based heavily on technology that is currently in
cormercial scale use, and at the same time is doing an excellent job of PCBs
destruction or rsnoval, or else it holds great premise of doing that job
based upon success in similar applications.
3.1 Incineration Recommended for Liquid PCBs and Scrap Oils
For this categoryf we determined only two candidate methods: in-
cineration and sanitary landfill. It is possible that scarne of chemical de-
gradation methods discussed under wastewater treatment micjht later become ap-
plicable to the concentrated PCBs liquid, but the prospect was not clear at
this time.
Sanitary landfill is not recommended for liquids when incineration
is available. The potential for liquids escaping in large quantities from
ruptured containers, caused by any of a number of circumstances, and then
causing massive leaching and liquid control problems at the landfill, is felt
to be too great,
Incineration, on the other hand, offers a straightforward and
physically simple method of final destruction. Incineration facilities that
134
-------
have successfully handled PCBs liquids are available in Massachusetts, New
York, Delaware, Illinois, Texas and Louisiana. Other pilot or experimental
facilities are available in other parts of the country. With the increase in
requirements for disposing of many other liquid organics, it is expected that
new facilities suitable for PCBs destruction will be added.
We therefore reconmend incineration and particularly if there is a
choice of the kind of disposal facility to erect at a new location.
3.2 Carbon Adsorption and UV-AssistedOzonation.Recommended for PCBs in
Wastewater
Our survey of wastewater treatment technology was most extensive,
and excellent potential for current, near and long term methods was found.
The longer term pilot or research scale methods hold great promise of allowing
zero discharge.
In this category, our prime recommendation is carbon adsorption.
This technology has been proved over most of this century in a wide variety of
industrial adsorption problems. It is being applied successfully to the re-
moval of new organics from water on a continual basis. Our cooperative labo-
ratory work with several suppliers has confirmed preliminary published re-
ports of success in removing PCBs. All of the aspects of commercial carbon
adsorption from favorable capital and operating economies to reasonable oper-
ating methods, materials of construction, and lack of transport of pollution
to air or land, have been proven for PCBs-like materials. There is every
reason to believe that there will be commercial success with PCBs removal from
wastewater.
Potential disadvantages are the collection of backwash water and the
spent carbon for incineration or other treatment. With these limitations in
mind, a search was made for an alternate not having those limitations.
As an alternative, we recommend consideration of UV-ozonation. It
is appreciated that this technology is still somewhere in the pilot plant to
research stages. However, our cooperative testing with two equipment suppliers
shows the method to be effective in destroying PCBs. It offers the potential
of also degrading breakdown products all the way to CO-, water and .'-:.!.. This
135
-------
kind of a process, that generates no solids or liquids for later disposal, must
be considered for application where no wastewater treatment of any kind now
exists, and where facilities for incineration of carbon system wastes are not
convenient.
The major factors yet to be determined for the UV-ozone systems are
costs and operating practicality. Separately, UV and ozone systems are being
used for commercial systems. It is therefore anticipated that the combination
will be practical. Choices of the proper UV radiation wavelength and power
levels still needs to be made; as well as methods of improving ozone use ef-
ficiency. It appears that commercial UV and ozone generators are suitable for
this use.
3.3 Incineration and Landfill Recommended for Contaminated Solids
For PCBs contaminated solids, although incineration is recommendea
because of its final destruction capability and prevention of any long term
problems, sanitary or scientifically controlled landfill must be considered a
close second choice. At present, the only known facilities for handling a
wide variety of solids are those of Rollins Environmental Services in Delaware,
Texas and Louisiana. This limitation on locations for treatment requires that
the alternate, landfill, be included.
Landfill, as practiced by Chemtrol Corp., appears perfectly suitable
for containment of PCBs, at least over a medium term. Our reservation with
this method is that it might be relegating a problem to the future. Our con-
cern is that some decades in the future, when a landfill is closed for any
reason, no agency will be prepared to handle the maintenance, sump emptying
and the like to prevent leaching. We anticipate that some time in the future,
many landfills will have to be mined, and final destruction or recovery made
of the materials, for land use or hazard reasons.
3.4 Dry Carbon Filter Adsorption Reoonmended for Control of Air Emissions
For much the. same reasons as in 3.2, above, we recommend that current
emissions of PCBs in plant air be trapped in carbon-containing filters. It is
recogn.17?^ i.h<;c- other and better adsorbents may emerge from research such as
136
-------
described in section 2.2.3, but it is felt that such advances will be readily
applicable to any kind of filter pack, screen, cartridge or the like.
Carbon's long record for proven capability in this area is an over-
whelming factor in its choice. There is the generation of contaminated carbon,
and it must be incinerated or treated. For this reason, over the longer term,
we see the use of low temperature catalytic oxidation methods as described in
section 2.4.3. Those research methods hold the promise of near zero discharge,
with the generation of no solid wastes.
137
-------
SECTION VII
COST OF TREATMENT AND CONTROL TECHNOLOGIES
1.0 SUMMARY
The estimated maximum total annual cost for the treatment of all contamin-
ated and potentially contaminated wastewater to a PCB concentration of 1 ppb or
less, the incineration of scrap oils and the incineration of contaminated solid
waste generated in the domestic production and utilization of polychlorinated
biphenyls (POBs) is $9,610,000. Expressed in other terms, this maximum total
annual cost represents $0.28 per pound of PCB produced arid utilized within the
United States. The basic assumptions and rationale employed in developing this
estimated maximum annual cost can be summarized as follows:
1. The raw water supply of the only domestic producer and all trans-
former and capacitor manufacturers is assumed to be contaminated
with PCBs. In addition, the assumption is made that ultimate
PCB discharge regulations will be based on gross PCB discharge
and not on net additions. The basic assumption regarding the
contamination of raw water supplies is believed justified on the
basis of plant visits conducted during this study since raw
water supplies were reported to contain from 1 to 57 parts per
billion PCBs. The assumption regarding the ultiinate development
of PCB discharge regulations on a gross discharge basis is open
to question, but is believed to be a reasonable assumption for
the purposes of estimating maximum annual cost.
2. All runoff resulting from precipitation on the buildings and
grounds of domestic PCB manufacturers and users will be con-
taminated with PCBs and will therefore require treatment prior
to discharge. This assumption is based on two additional
assumptions including (first) that historical use of PCBs has been
relatively uncontrolled and that buildings and grounds can be
assumed to be contaminated with PCBs, and (second) that all plants
138
-------
regardless of age, employ local and general exhaust ventilation
systems in PCB use areas and that PCBs are thereby redeposited
continuously on the buildings and grounds from these ventilation
systems. It is recognized that the potential exists for control
of ventilation systems at the source through the application of
adsorbant materials such as activated carbon. However, it is
again believed that this assumption results in the worst case
and thereby permits estimation of the maximum annual cost for
PCB discharge control.
3. All non-contact cooling waters are contaminated based on
assumption 1 and are employed on a once-through basis thereby
requiring treatment prior to discharge.
4. All sanitary wastes, with the exception of waters employed for
personal hygiene, are segregated and discharged to separate
treatment facilities (the costs of which are not included in
this estimate) or to municipal facilities.
5. All other process waters, condensates, boiler blowdowns, etc.
are collected and subjected to treatment.
6. All waters derived from the sources described in items 1 through
5 are subjected to a treatment system comprising quiescent
settling for settleable solids removal and free oil skimming,
equalization, fine media filtration, carbon adsorption, flow
measurement and sampling, and discharge to surface receiving
waters. All sludges and free floating oils are disposed of
through incineration. All backwash waters from the carbon ad-
sorption system and fine media filtration are routed back to
the gravity separation basin and thence through equalization, etc.
All exhausted carbon is destroyed through incineration for waste
flow up to 300 gpm (this assumption was made since it is not
believed that existing commercial regeneration facilities are
adequately equipped to permit complete destruction of PCBs employing
139
-------
standard carbon regeneration practices), Above 300 gpm,
specially designed on-site carbon regeneration is employed.
7. A flat fee of $.10 per pound has been assumed for the in-
cineration of scrap oils and solid wastes contaminated with
PCBs. This assumption is made on the basis of the maximum
reported rate for incineration of such wastes at commercial
waste disposal facilities.
1.1 Plant to Plant Cost Variations
The estimate of the total maximum costs anticipated for the manufacturer
and users of PCBs was facilitated by preparing cost estimates for each plant site
visited during the course of these studies and then aggregating to the total in-
dustry category on the basis of the percentage of the total .Industry represented
by the visited plants. The total estimated cost was based on the assumptions
listed under 1.0. This method of aggregating to the total industry is believed to
provide a reasonable estimate of the total cost to be anticipated for the industry
as a whole. However, other than for the plants actually visited, it is extremely
difficult to make reasonable projections of the cost variations from plant to
plant. The key variables affecting variation in cost from plant to plant are
summarized and discussed below:
1. Most end-of-pipe treatment systems are significantly in-
fluenced with regard to size (and therefore cost) by the
volumetric flow rate of the wastewater to be treated. In
addition, in a case such as PCB contaminated wastewater,
where the level of contamination is expected to be reason-
ably constant, the total treatment costs are essentially
directly proportional to the volumetric flow rate. In
reviewing the water usage rates at the plant sites visited
during this study, it is apparent that the volumetric flow
rate varies significantly from one plant to another with
little apparent relationship between the production rate
and volume of wastewaters. This situation results in a
140
-------
considerable variation in cost from plant to plant
for terminal wastewater treatment which virtually
defies rational evaluation.
2. The extent of plant site contamination with FOB will
impact on the quality and quantity of runoff water which
will require treatment. It is anticipated, although
difficult to quantify, that the extent of plant site
contamination with PCB would be related to the plant age.
However, it is conceivable that the extent of building
and ground contamination with PCBs even in relatively new
facilities could be rather high depending on the house-
keeping control practices for handling spills, leaks,
and otherwise contaminated materials.
3. Again, with regard to the volume of runoff water to be
treated, the geographic location and terrain can influence
this volume due to differences in level of precipitation
and the local runoff coefficients.
4. It is believed that land availability will only be a signifi-
cant factor when storm water runoff treatment is required.
The principal land area requirement for the treatment
facilities envisioned is in equalization and holding facil-
ities which will be controlled by the quantity of runoff
which must bo captured and contained prior to treatment.
In general, land availability and the cost of such land
is too variable as to estimate. Of the plant sites actually
visited, it is believed that most would have sufficient land
available for all required treatment facilities. However,
at least one plant was relatively congested and was in an
area which was so highly developed that land availability
could become a problem.
5. In the event that the raw water supply is not contaminated
with PCBs, significant variations from plant to plant could
141
-------
occur in the cost to effect segreg
cooling waters which could be dire*
as previously indicated, of the pic
none reported that their raw water
PCBs, However, it is possible that
developed on a net addition basis n
charge basis and therefore all plant
charge non-contact cooling water wit
1.2 Activated Carbon Terminal Treatment Systt
The actual PCB level attainable through a
not well defined at this time. Theory would indicate
not be attained. However, it is quite conceivable th<
perly designed and operated carbon treatment system wi
analytically detectable limit (1 ppb) of PCBs. The ra
activated carbon system for estimating the maximum tot;
included the following considerations:
1. Although this system falls short of "ze
sense of having absolutely no PCB discharge, it is believed
quite probable that it would approach zero discharge from
the point of view of being below the detectable limits.
2. Activated carbon adsorption is a field proven technology
and the associated costs for construction and operation are
well enough defined that industries or agencies reviewing
the cost estimate could understand and have confidence in
the cost data generated.
3. The only other potentially viable alternative, UV light
assisted ozonation, has less well defined capabilities and
is much more difficult to evaluate accurately for construction
and operating costs.
4. Until such time as the exact capabilities of potential altern-
atives and the costs for these alternatives are accurately
142
-------
known, i.t is believed that the activated carbon system
would be installed in most cases. In any event, unless
the UV/ozone system can do an equal or better job of
destroying PCBs at a lower cost, it would not be selected
by industry. Therefore, the costs estimated on the
basis of an activated carbon system should represent
the maximum anticipated costs,
1 . 3 Costs Based on Volume Flow
As previously indicated, the costs for the terminal treatment facil-
ities considered herein are essentially proportional to the volumetric flow rate.
The estimated total maximum annual cost for PCB wastewater treatment presented in
this summary was based on "worst case" conditions. By developing all costs on
the basis of volumetric flow, it is possible to examine the effect of removing
certain wastewater categories from the total treatment requirement and evaluating
the effect of this exclusion on the total treatment costs. In other words, if
one desires to evaluate the reduction in costs which would be anticipated by assuming
that storm water was not contaminated and that it could be directly discharged, one
could remove this flow category from the total volume of wastewater to be treated
and quickly evaluate the reduction in total treatment cost resulting from this
change in assumptions. Several cases of this type have been examined in Section 3.6.
2.0 COST REFERENCES AND RATIONALE
This subsection summarizes and discusses the cost references and rationale
employed in developing capital and operating cost estimates for the various control
and treatment schemes presented in Section 3.0.
^•l Interest Costs and E^uj^ JTijiancing_Charges
For purposes of amortizing capital costs estimated in this section, an
interest rate of 8% per annum has been assumed. The actual method of financing
the construction of pollution control facilities are extremely varied and will un-
doubtedly differ from corporation to or.irporat.ion- See Sections 2.3 and 2.6 for
additional discussion of the arrortixa •-''.< >n of ca^xi t--)] costs.
-------
2.2 Time Basis for Costs
All cost estimates contained in this report are based on first quarter
1976 dollars for both capital and operating costs.
2.3 Useful Service Life
The useful service life of major equipment items and treatment units
projected for use in the treatment schemes presented herein may vary from 10 to 50
years. This assumes that the equipment is properly maintained and that it is
employed for the service intended. However, it is believed that service life is
of limited utility in this case since the long term use of PCBs is questionable.
For purposes of amortizing capital costs, a payout period of ten years has been
assumed in all cases. It is, of course, conceivable that certain equipment items
such as stainless steel vessels could be salvaged at the conclusion of service for
PCB contaminated wastewater treatment. However, it is not believed that t^e
salvage value will be high enough to have any significant effect on the overall
costs for construction and operation of the treatment systems.
2.4 Depreciation
Straight line depreciation was used over a ten year period.
2.5 Capital Costs
Capital costs as employed herein include the purchase price for major
equipment and materials, delivery costs, installation costs, site piping, yard
work and engineering fees associated with design and construction supervision.
Capital costs do not include royalties or licensing fees which might be associated
with any of the treatment systems which are under patent. Capital cost estimates
do include reasonable profit and overhead for any contractors performing install-
ation or construction services. Where possible, all capital costs for major equip^
ment items have been based on vendor quotes for specific items of equipment. For
items such as excavation and concrete construction, actual quantity estimates have
been made and current unit costs have been employed.
2.6 Annual Capital Costs
All capital costs have been amortized over a period of ten years at
8% interest to define annual capital costs.
144
-------
2.7 Land Costs
Land costs have not been included in any of the capital cost estimates.
Except in rare instances where a production facility is actually "land locked"
it is not believed that the cost of land would have a significant impact on the
total treatment facility cost. A land area requirement curve has been developed
to relate land area required to the volumetric flow of wastewater to be treated,
see Figure 2.7-1.
2.8 Operating Expenses
Where available, curves relating labor requirements to various unit
operations have been employed to estimate labor requirements. When such curves
were not available for a specific unit operation, rational estimates of the
operating labor requirements were made based on experience. All operating costs
for labor have been estimated at $15 per manhour. Electric power costs have been
estimated at $.02 per kilowatt hour. Commercial disposal of scrap oils, sludges
and solid wastes have been estimated at $.10 per pound. Fuel costs have been
estimated on a range from $1.00 to $2.00 per million BTU's depending on the use
rate. (The exact value used is identified in each specific instance.) Costs for
virgin activated carbon have been estimated at $.50 per pound.
2.9 Rationale for Incineration^Costs
Incineration at a temperature of 2200°F for a residence time of 2-3
seconds has been identified as a means of completely destroying polychlorinated
biphenyls. At the present time there are at least three distinct types of PCB
contaminated wastes which are candidates for incineration. These include:
1) Pure organic streams, which are commonly referred to as
"scrap oils". These may range from essentially pure PCB
liquid to solvents with only low levels of PCB contamina-
tion.
2) Solid wastes which are contaminated with PCBs including
reject electrical components, paper, rags, diatomaceous
earth and clays used to filter PCBs, spent activated
carbon, etc.
145
-------
FIGURE 2.7-1
REQUIRED TREATMENT PLANT AREA FOR
REMOVAL OF PCB FROM WASTEWATER
REQUIRED
PLANT AREA
(l,000sq.ft.)
-200n
100-
90-
80-
70-
604
40.
30.
9-
6*
7.
6-
5^
10
K.
/
40 60 80 100
aoo
400 6OO BOO IOOO
20OO
FLOW (gpm)
146
-------
3) Aqueous waste streams containing from a few ug/1 PCBs up
to tens of mg/1 PCBs (only small quantities of this type
material are currently incinerated, but this type of
stream becomes very important in one of the proposed
schemes for achieving "zero discharge". See Sections
2.11 and 3.3.3).
Of the producers and users of PCBs, only two, Monsanto and one trans-
former manufacturer, currently have on-site incineration facilities. The Monsanto
facility for incineration is described in Section IV - 3.1.4. This system handles
only organic wastes in the liquid form. The system capacity is approximately six
million Ib. per year.
The transformer manufacturer's incinerator also has a rated capacity of
six million Ibs. per year and is capable of accepting solid wastes such as trans-
former internals, paper, rags, and cardboard, but not diatomaceous earth, soil,
etc. see Section TV, 3.3.1.5
In spite of the fact that these systems have only minor differences
in configuration and essentially the same operating capacities, their present day
estimated installed costs range from $420,000 to $1,100,000. The more costly system
does include significantly greater storage capacity which may explain some of the
cost difference. It should be further noted that both of these incineration systems
were supplied by the same vendor. The point being made here is that there is no
single representative system with a typical cost per unit of capacity.
In attempting to develop capital costs for incineration equipment,
several leading suppliers were contacted. All suppliers were hesitant and virtually
refused to provide guidelines for capital costs estimating unless very specific
details of the system requirements were provided.
Since the design and fabrication of such equipment is a highly pro-
prietary industry, it is believed that any attempt to estimate the cost of such
systems on a rational basis would be fruitless. In addition/ it seems quite impro-
bable that the smaller capacitor and transformer manufacturing facilities would
find it economically attractive or technically practical to install and operate
incineration facilities for scrap oils and solid wastes contaminated with PCBs.
147
-------
The alternative to on-site incineration is shipping scrap oils and
solid wastes to commercial disposal firms. The handling practices and tolls
associated with commercial disposal operations have been described in Section VT
for both scrap oils and solid wastes. For purposes of estimation in this report,
a flat rate of $0.10 per pound has been assumed for all scrap oils and solid
wastes even if they are generated by facilities which have an on-site incinerator.
The off-site incineration of dilute aqueous streams in most cases
may not be a feasible alternative basically due to the logistics and cost of trans-
porting rather large volumes of water. In this case very rough estimates of capital
costs for such incineration systems have been prepared and are presented in
Section 3.3.3. In this case the accuracy of capital cost estimates are not deemed
as critical since the operating costs, principally in terms of fuel consumed,
are relatively high. It should be noted that large scale incineration of dilute
aqueous streams is only considered as a portion cf a proposed system to achieve
"zero discharge" (see discussion in Section 2.10.2).
2.10 Definition of Levels of Treatment
2.10.1 Effect of Varying Effluent PCB Concentrations
At the inception of this study it was decided that treatment costs
would be evaluated for effluent PCB levels of 1, 5 and 10 ug/1. Adsorbers using
activated carbon may remove PCBs to concentrations less than one part per billion.
An activated carbon system could be designed to obtain a choice of effluent con-
centration. However, it is very unlikely that such a design would produce a given
level of PCBs on a continuous basis. The breakthrough of PCBs is anticipated to
be relatively fast as compared to the actual run time of the adsorber; therefore'
the operating cost savings for the various effluent PCB concentration are expected
to be too small to justify designing carbon adsorbers for PCB removal at higher
than one ppb levels. The costs for this system were evaluated at two influent COD
levels, i.e., 30 and 50 mg/1.
For the second potential alternative, UV-ozonation, the work per-
formed by Houston Research and Westgate Research indicates that the final concentra-
tion is a function of residence time and ultraviolet light catalyzed ozonation
systems apparently can be designed to achieve any desired residual concentration
of PCB.
148
-------
For example, assuming the data can be extrapolated to very
low PCB concentrations, the following information extrapolated from Houston Re-
search data indicates the different residence times that could be involved.
Original Concentration, Co Final Concentration, C Residence Time, t
(ppb) (ppb) (min)
500 10 260
500 5 310
500 1 420
200 10 200
200 5 250
200 1 360
50 10 110
50 5 160
50 1 270
These extrapolated data indicate that both residual concen-
tration and final concentration dictate the residence time required. The data can
be approximately represented by the equation
In C/0.9CO = -0.01468 t
where:
C = final concentration, ppb
Co = initial concentration, ppb
t = residence time, minutes
The actual test data may be seen in Section VT - 2.2.2.4. For
this system, effluent PCB levels of 1 and 10 ug/1 and influent COD levels of 30
and 50 mg/1 were evaluated.
2.10.2 Zero Discharge Definition
Since the passage of Public Law 92-500, which establishes as a
long term goal the achievement of "zero discharge" of wastes to the waters of the
United States, the practical interpretation of the term "zero discharge" has been
149
-------
discussed and debated. The roost stringent interpretation would be that of
absolutely no discharge of either treated or untreated wastewaters. A second
interpretation would allow discharge of treated wastewaters provided that con-
tandnants in the wastewater were reduced to acceptable levels.
In the case of hazardous and toxic materials, the definition
of the acceptable level of discharge is normally tied to the limits of detectability
of analytical techniques available at that time. Unfortunately, this type of
standard presents a moving target to dischargers since analytical techniques are
continuously being improved.
In the case of PCB discharges being evaluated herein it is
believed that treatment systems employing activated carbon and possibly UV cata-
lyzed ozonation and polymeric adsorption could produce effluents which would be
at or below the practicle limits of detectability for PCBs with currently avail-
able analytical techniques. However, since no full scale treatment systems are
in operation at this time, this possibility cannot be confirmed,
For purposes of this study, zero discharge is evaluated for
a system approaching the most stringent definition of zero discharge.
3.0 PCB WASTEWATER TREATMENT PLANT DESIGN PARAMETERS AND COSTS
3.1 Introduction
Any treatment system designed to remove PCBs from wastewater should be
capable of handling all possible contaminated flows. These contaminated flows
can include:
1. Rainwater runoff which washes PCBs from roofs- and grounds.
2. Washdown water used to clean production areas; where PCBs
are used and manufactured items.
3. Washwater used for personal hygiene, but not toilet
usage.
4. Water that has come into contact with air contaminated
with PCBs, including cooling tower blowdown, steam jet
ejectors, vacuxsn pump water seals and scrubber water.
159
-------
With the exception of rainwater runoff, the volume of water is a function of the
size of the production facility and attitude of labor and management toward
water usage and/or conservation. These flows may or may not be relatively con-
stant day to day. It is anticipated they will vary widely throughout the day.
Rainwater runoff is a direct function of the location's rainfall and runoff area.
The runoff area may or may not be dictated by the size of the production facility.
This single flow would provide the largest flow variation for all production
facilities. Since this variation exists, a flow equalization basin is necessary
to provide a near constant flow to any downstream treatment unit operation,
Possible contaminants which have to be dealt with in any treatment
system are:
1. PCS
2. Solvents - water miscible and immiscible
3. Detergents
4. Oils and other hexane soluble substances other than
PCB
5. Suspended solids
All of these substances will be affected by their residence in an equal-
ization basin. PCB present in quantities above about 200 ppb will separate from
the wastewater and, unless acted upon by some other contaminant, settle to the
bottom. Suspended oils and immiscible solvents will separate and rise to the
surface. Suspended solids will either rise, fall or remain suspended depending on
their densities and/or particle size. It has been found that suspended solids
will absorb PCBs on their surface and these must be removed either in the equaliza-
tion basin, a subsequent clarifier or by filtration. For the purpose of the design
described in this section, it has been assumed that their concentration is low
enough that they can be removed by separation in the first section of the separa-
tion basin and any remaining suspended material would be removed by multimedia
filtration. If the solids are greater than 100 rng/1, a clarification system should
be installed to relieve possible solids handling problems in the basin. A bar
151
-------
sreen or trash rake should be used to prevent large objects from entering the
basin and creating problems with the downstream process feed pumps.
The wastewater is assumed to flow by gravity from its various sources
to a point where it would be pumped into the equalization basin. Any heavy material
collected from the bottom of the basin would be drummed and sent to an incinera-
tor for decomposition of PCBs. The floating liquids would be removed by a belt
skimmer and drummed for subsequent incineration, since these oils will contain
high concentrations of PCBs. A schematic flow diagram for this treatment system
is given in Figure 3.1-1.
The presence of water miscible solvents and detergents will most likely
alter the level of PCB solubility in water depending on their concentration. There
has been no comprehensive study to determine the affect these substances have on
solubility due to the impossible number of contaminant systems that can exist.
For the purpose of this report it will be assumed that PCBs will enter the equal-
ization basin at an unknown concentration and leave it at 200 ppb and filtration
will reduce the PCBs level to 50 ppb. Subsequent terminal treatment systems, i.e.,
carbon adsorption, UV-ozonation, etc. will reduce the 50 ppb PCB concentration to
one ppb or less.
Any treatment design should incorporate unit operations that would com-
pletely contain PCBs and not allow them to enter the surface water. This requires
a guarantee of the integrity of all drains, lines, pump stations and treatment
operations.
The soluble substances which are not removed by filtration can be
removed by a number of methods. Activated carbon adsorption and light catalyzed
ozonation will be discussed in depth in this section. Due to the low effluent
concentration required for PCBs, only those systems which can be purified by
oxidation have been considered.
Activated carbon can be either incinerated or regenerated and the
evolved PCB oxidized in a burn chamber at 2000° to 2200°F. This will insure that
the treatment medium will not become contaminated to the point, that it will be
a significant source of PCBs to the environment. It should be noted that activated
carbon adsorption systems are numerous and full scale operation has been documented.
152
-------
153
-------
Ozonation using ultraviolet light radiation has been shown to be an
alternative treatment method. Although, as a waste treatment method it has not
found application, it has been used in chemical reaction technology.
The main difference between the operations of the two systems is the
quality of the effluent and the need for regeneration for the carbon. The acti-
vated carbon adsorption process will adsorb all substances which are capable of
entering the particles' pores. This means that any PCB present in the waste-
water has the potential of being completely removed until breakthrough of the
bed occurs. The ultraviolet light catalyzed reaction appears to approximate a
first order reaction and the effluent concentration is therefore a function of
the residence time in the reaction zone.
3.2 Pretreatment
3.2.1 Equalization and Separation Basin
The equalization basin's primary function is to equalize flow
so the feed to the downstream treatment units is nearly constant. Since the flow
variation could not be computed except on a case by case basis, an arbitrary sizing
calculation was used. The volume was calculated using three times the design flow
and a 24 hour residence time. This basis reflects consideration of discharge data
for two of the manufacturing plants visited. The maximum liquid depth was assumed
to be ten feet with an allowance of three feet of freeboard to prevent slop-over
due to wave action.
Several methods can be used to ensure that collected wastewater
does not enter the subsurface ground water. These are:
1. Utilize existing soil conditions where the
soil strata is relatively impermeable
2. Line the basin with bentonite
3. Line the basin with an impermeable organic
membrane
4. Construct the basin with reinforced concrete.
154
-------
The wastewater entering the basin will probably contain sus-
pended PCBs, oils and solvents. It is expected that PCBs will attack and deteri-
orate a synthetic organic liner. At this time the alternative that has the best
chance of succeeding in containing the wastewater appears to be reinforced con-
crete construction. A sloped bottom at the head end of the basin will allow
settleable solids and oils to accumulate for subsequent removal by pumping. It
is anticipated that the settleable solids collected will be small compared with
collected PCBs and this material will be pumped to a decant tank. Here water
can be more easily separated and the heavy material can be drummed for later
incineration.
A belt skimmer should be provided to remove floating oils.
These can be accumulated in a drum; and when filled, it can be removed for incin-
eration and replaced with an empty drum.
The basin influent would be pumped utilizing a lift station de-
signed for the design flow plus runoff from a one hour, 25 year storm
vary between two and four inches of rain per hour depending on geographical location.
This figure may be obtained from the Rainfall Frequency Atlas of the United States,
Technical Paper No. 40, U.S. Dept. of Commerce. The total flow, of course, will
depend on the area drained. It is assumed that this area is not a function of pro-
duction facility size and any attempt to size an influent pump station would intro-
duce an unknown error. For this reason, the cost of this operation is not included
in the cost estimates. It is believed that this cost will not affect overall
capital costs significantly. The effluent pump station was designed on the design
flow rate at 270 feet of water discharge head.
The decant tank can only be sized knowing the volume of oily
material to be separated. The sizes were arbitrarily picked as 1000, 5000 and
10,000 gallon vessels. The flows through 80 gpm would utilize a 1000 gallon tank,
for 160 through 640 gpn a 5000 gallon vessel and for 1700 qpm the 10,000 gallon tank
would be utilized. The water layer would be recycled to the equalization basin
while the underflow would be drummed for subsequent incineration.
155
-------
3.2.1.1 Equalization Basin Capital and Operating Cost
The fact that reinforced concrete has been chosen to
provide a barrier between the water environment and the basin makes this one
nodule the single most expensive item in the total capital cost of a treatment
system for removal of PCBs from wastewater. If rainwater runoff is inconsequential,
then the size of the basin can be reduced significantly. If ventilation of the
PCB handling areas is controlled and all such areas are enclosed, provided there
is no prior ground contamination, rainwater runoff will be nonexistent in the
sizing of an equalization basin. If this is the case, the capital cost commitment
with this operation can be reduced by about a third. However, at present this
is not the case.
The capital cost is shown in Figure 3.2.1-1 and the
backup data for these costs is exhibited in Table 3.2.1-1.
3.2.1.2 Equalization Basin Costs
The capital cost of the equalization basin was cal-
culated using $5.00 per cubic yard of earth removed and $285 per cubic yard of
reinforced concrete for construction.
The annual operating and maintenance costs for the
basin were calculated using 1.0% of the capital cost. The capital cost was
amortized using an interest rate of 8% and a project life of ten years.
Pump operating and maintenance costs were figured at
2% of the total station capital cost.
3.2.2 Multimedia Filtration
Multimedia filtration using beds of sand and anthracite was
chosen as a final suspended solids polish system. It would be advantageous to
remove these solids to as low a concentration as possible to prevent them from
passing through any subsequent treatment system designed for removal of soluble
contaminants. Depending on the settleability of the solids in the equalization
basin, chemical and/or polyelectrolyte addition to the filter feed may be desirable.
It should be recognized that chemical addition such as alum, lime or iron salts
156
-------
FIGURI: 3.2.1-]
FLOW EQUALIZATION BASIN INSTALLED CAPITAL COST
5000
3000
TOTAL
BASIN
COST
$ 1,000)
1000.
900.
eoo.
700.
600-
500.
400-
300-
200. -
100,
9O.
80.
TO.
6C.
50.
10,
30.
20,
10
20 30 40 60 80 100
200 300 400 600 800 90O
2000 3000
FLOW (gpm)
157
-------
TABLE 3.2.1-1
FLOW EQUALIZATION BASIN
SIZING AND COSTS
BASIS:
FLOW (GPM)
LIQUID VOI
WIDTH (ft)
1. 24 hour retention
2. 3 times normal flow
3. 18" concrete
4. 12 ft. depth
5. 10 ft. water depth
6. L/W = 2.0
M)
OLUME (1000 gal)
t)
ft)
ON COST ($1000)
COST ($1000)
SIN COST ($1000)
D SUMP ($1000)
20
86
24
48
2.6
45.6
48
20
40
176
34
68
5.1
75.6
81
20
80
345
48
96
10.2
128
138
21
160
690
68
136
20.5
224
245
22
320
1380
96
192
41
402
443
27
640
2760
136
272
82
742
824
32
1700
7340
222
444
218
1820
2038
42
TOTAL BASIN & PUMP
COST ($1000) 68 101 159 267 470 856 2080
158
-------
will increase solids handling problems and all solids must be sent to incineration.
It is anticipated that the source of solids will be soil washed into the treat-
ment system by rainfall and carried into production areas by workers. Care should
be used so that other processes which might supply organic solids and are free of
PCB contamination are not included in the PCB pretreatraent system.
There is a possibility that chloride concentration in the filter
feed would be such that mild steel construction would nor be advisable. Possible
absorption of organic contaminants in an organic liner used to preclude corrosion
is expected and therefore all stainless steel (Typt.- 304) construction is suggested.
Although pilot testing of a filter system is suggested to ob-
tain optimum flow rates and bed depths, this is impossible for the systems des-
2
cribed in this report. Superficial velocities of 0.5 to 6 gpm/ft. have been
used with success. For the system design described in this section, superficial
2
velocities of 1.5 to 3 gpm/ft. were used.
It was assumed that a 12 foot diameter vessel would be the
largest size used due to the difference in shop and field fabrication costs. Only
in the case of the 1700 gpm flow would this pose a problem .in making the decision
to provide shop or field fabricated vessels. Since the choice of stainless steel
construction was made, it was found that for this case, shop fabrication would be
less expensive even when the cost of added valuing and piping was considered.
Filters would be rnanged for parallel flow, and in all cases
through 640 gpm, the number of filters would be two. Each filter would be capable
of treating the entire flow when the other was off Line for backwash. This time
period would be a maximum of 30 minutes per day,
2
Backwash flow would be at 20 gpm/ft. and increased from 10 to
2 2
20 gpm/ft. in 5 gpm/ft. increments lasting about two to three minutes. When
backwash has been completed, a rinse cycle would be used for about ten minutes to
settle the bed using clean water from the clean water reservoir.
The flow rate through each filter would be Iialf that through
each carbon adsorber. The filter requirement would be the same as the adsorber
159
-------
requirement. For example, if two, seven foot diameter carbon adsorbers are re-
qiired in series, then two, seven foot filters would be required arranged for
parallel flow.
3.2.2.1 Filtration Costs
The capital cost of the filtration system was assumed
to be equal to that of the carbon adsorber system since like vessels would be used.
The operating and maintenance costs were calculated from work performed by Burns
(4)
Roe. The operating and mair
be represented by the equation
(4)
aid Roe. The operating and maintenance cost curve which was for March 1969
0 & M Cost - 50 (Q)~°'364
wi ere :
Cost (=K/1000 gal
Q (=)gpm
Ccsts were increased by 70% changing the equation to:
Cost = 85(Q)-°'364
The combined costs of equalization, decant and filtra-
tion constitute the total modular cost. The total installed pretreatment system
ccsts and annual operating costs are exhibited in Tables 3.2.2.1-1 and 2, respectively.
Graphical representation of these costs can. be seen in Figures 3.2.2-1 and 2.
Capital cost was amortized using an interest rat-, of 8% and project life of ten
3.3 Activated Carbon Treatment System for Removal of PCBs from
Wastewater
Activated carbon has the property of adsorbing substances, organic in
nature, from water. This adsorption may be caused by three mechanisms: (1) Van der
Walls forces, (2) electrostatic factors, (3) hydrogen bonding. Van der Waals forces
(tie attraction between molecules that produce a liquid such as carbon tetrachloride
instead of a gas at standard conditions) are the predominant mechanisms of ad-
so -ption by activated carbon .
160
-------
TABLE 3.2.2-1
TOTAL INSTALLED PRETREATMENT SYSTEM COST
($1000)
FLOW (gptn)
EQUALIZATION BASIN
FILTER SYSTEM
DECANT TANK
TOTAL MDDULAR COST
PIPING AND INSTRUMENTATION
(25%) 26 36 56 88 149 267 673
TOTAL INSTALLED EQUIP-
I'ENT COST 129 180 280 440 764 1330 3365
ENGINEERING (10%) 13 18 28 44 75 133 334
TOTAL PRETREATMENT
PLANT COST 142 198 308 484 821 1463 3699
20
68
32
3
103
40
101
40
3
144
80
159
62
3
224
160
267
78
7
352
320
470
120
7
597
640
856
200
7
1063
1700
2080
600
12
2692
161
-------
TABLE 3.2.2-2
ANNUAL OPERATING COST FOR PRETREATMENT
FOR REMOVAL OF PCB FROM W&STEWATER
($1000)
FLOW
AMORTIZATION (8%-10 yr)
OPERATION & MAINTENANCE
POWER (2C/KWH)
ANNUAL PRETREATMENT COST 25
COST/1000 GAL ($)
20
21
4
-
25
2.38
40
30
6
1
37
1.76
80
46
8
2
56
1.33
160
72
13
4
89
1.06
320
122
20
8
150
0.89
640
218
32
17
267
0.79
1700
551
63
45
659
0
162
-------
FIGURE
3.2.2-1
PRETREATMENT SYSTEM CAPITAL COST
5000
4000
3000
2000
COST
(flOOO)
IOOO.
900.
800.
700.
600-
500.
400.
300.
200'
100'
10
20
40 60 80 100
200
400 600 800 IOOO 2000
FLOW (gpm)
163
-------
FIGURJ; 3.2.2-2
ANNUAL COST FOR PREIREATMENT
80O
ANNUAL
COST q£
($1000) TO
in-
(
—
Y*
—
xf
— -
x1
/
-
X
-
s
s
_
k
-
--
>*
r<
'/&>
./*
'
^
I/
- -
/
--
X
—
p
/
-
/
o
f
10
20
GO 80 100
200
400 GOO 800 1JOO 20OO
FLOW (gpm)
164
-------
The pore size distribution is a crucial factor in adsorption. For
treatment of wastewater, a large range is required due to the complexity of the
wastewater where everything from methanol to long chain hydrocarbons are possible.
If the pore size is too small, larger molecules will be screened since they will
not be able to move into the carbon particle. A typical value of the surface area
2
of carbon is 1000 m /g and this .indicates that most of this area is on the inside
of the carbon granule. Because of this large internal surface area, organic
molecules may be removed from water solution by the adsorption process. Desirable
pore diameters range from 20 to 80 Angstroms, since most wastewaters do not con-
tain an individual organic component dissolved in water, a number of molecules
must compete for the available carbon surface area. As may be seen in Table 3.3-1,
a chemical oxygen demand (COD) of 30 mg/1 is anticipated, with an effluent con-
centration from the carbon system expected to be 10 mg/1.
The PCB concentration is expected to be 50 parts per billion and this
is anticipated to be removed to a level of 1 ppb or less by the activated carbon
beds. At present, there are no known tests that have been perfomad that provide
a reliable figure on the exhaustion of activated carbon in column adsorption using
PCB contaminated wastewater. Data using PCBs was developed by ICI-US, Carborundum
and Calgon Corporation (a subsidiary of Merck & Co., Inc.) by isotherm testing.
PCBs were the major organic contaminant in the water during these tests. Methanol
was present as a solvent in ppm quantities. These tests can be said to be static
in nature, whereas a column test would be dynamic, and would approximate the con-
ditions expected in a full scale system. However, since this is the only usable
data available, it will be used in this study. The carbon's capacity for COD is
anticipated to be 0.25 pounds of COD per pound of carbon. Whereas, the PCB
capacity is anticipated to be .006 pounds of PCB per pound of carbon. Based on
the anticipated concentrations from Table 3.3-1, the usage rate of carbon based on
COD alone would be 0.667 pounds per 1000 gallons, while that for PCB would be
0.0695 pounds per 1000 gallons. From this it can be seen that COD and not PCBs
would be the controlling factor on the operational cost of any activated carbon
system since it is believed that the activated carbon would become saturated with
COD and at that point PCBs would not be rvreterentially adsorbed. This is a
-------
TABLE 3.3-1
CARBON SYSTEM FEED CHARACTERISTICS
Feed Effluent
PCB (PPB) 50 < 1
HEXANE SOLUBLE SUBSTANCES (mg/1) 5 < 1
COD (mg/1) 30 10
SUSPENDED SOLIDS (mg/1) 10 < 1
-------
gross assumption and may be disproved at a later time when lengthy column studies
have been completed. It should be noted that since COD is assumed to be the
limiting factor in this case, it would be to the user's advantage to limit
the quantity of organics that enter any system designed to protect the environ-
ment from PCBs.
3.3.1 Activated Carbon System Design
The parameters used for this study to design an activated carbon
adsorption system may be seen in Table 3.3.1-1. These parameters were arbitrarily
chosen and were not developed by testing using PCB waste. They are, however,
based on past experience from a number of studies in actual treatment facilities.
The material of construction was chosen as Type 304-L Stainless Steel, since the
activated carbon will cause electrolytic corrosion of any mild steel vessel in
which it is used. Mild steel vessels can be internally, organically coated;
however, it is anticipated that as the activated carbon becomes saturated with
PCBs, that this may tend to dissolve in the organic coating, loosen the bonding
between the coating and the steel and cause removal of this coating from the
vessel wall. At this point severe corrosion may begin with ultimate failure of
the vessel. It is also foreseen that periodic internal cleaning of the vessels
will be necessary and this organic coating would be a hindrance to the use of a
number of solvents that may best remove PCBs from the surface.
Since PCBs are not considered acidic or basic in nature, pH
should not have a strong affect on the adsorption process, and therefore does not
necessarily have to be stringently controlled. However, for protection of external
piping and valving, it is suggested that the pH should be in the neighborhood of
6.5 to 9. At pH's above 9, there is a possibility of producing a non-breakable
emulsion of oils which may tend to pass through all systems untreated and of
damaging the equalization basin.
Due to the expected low usage of activated carbon in this system,
only static bed contacting systems will be investigated. These will include two
adsorbers arranged in series operation. They will be operated until PCBs are
detected in the effluent of the lead column. At this point the lead column would
16V
-------
TABLE 3.3.1-1
CARBON SYSTEM DESIGN CRITERIA
SUPERFICIAL VELOCITY (gpn/ft2) 3-6
CARBON VOID FRACTION 0.4
CARBON DENSITY-BftCKWASHED .,
AND DRAINED (#/ft ) 25-28
SUPERFICIAL RESIDENCE TIME (min) 15
/CTUAL RESIDENCE TIME IN BED (min) 6
DESIGN MAX PRESSURE (Psig) 50
ACTUAL MAX PRESSURE (Psig) 30
PCB/CARBON CAPACITY (Ih/lb) .006
COD/CARBON CAPACITY (Ib/lb) .25
L68
-------
be removed from the line, the carbon removed and virgin or regenerated carbon
installed. At this point the polish column would be valved into the lead posi-
tion and the previous lead column would be in the polish position. When the
carbon becomes spent, it can be discarded or regenerated thermally.
The virgin carbon replacement should be added to an adsorber
water level approximately one-third the depth of the vessel to prevent massive
air pockets from forming. This carbon should be allowed to wet for approximately
24 hours either before addition or after addition to the adsorber. When the
carbon charge has been completed the charge opening should be closed and the
2
adsorber placed on backwash. The recommended backwash procedure would be 5 gpm/ft
2 2
for five minutes, 10 gpm/ft for five minutes and 12 to 15 gpm/ft for ten minutes.
This will allow a controlled expansion of the bed and prevent massive disruptions
2
at the surface as would be experienced if the full backwash of 12 to 15 gpm/ft
were used initially. A number of possible flow situations have been investigated
and the data relating to these may be seen in Tables 3.3.1-2 and 3.3.1-3 and
Figure 3.3.1-1. It has been found that field erection of stainless steel vessels
is considerably higher than shop fabrication. For this reason no adsorber larger
than 12 feet in diameter has been investigated.
Although no suspended solids buildup in the bed is foreseen,
a surface wash is recommended for any possible upset of the upstream facilities.
This will be a rotating pipe arrangement that will expand the top 6 to 12 inches
of the bed without full backwash. The carbon will act as a filter and remove
suspended solids from the wastewater to a point where breakthrough occurs. At
this point it is anticipated that the solids, which possibly contain absorbed
PCBs, will cause a premature breakthrough of the carbon bed. It is therefore re-
commended that backwashing be attempted at least on a weekly basis. If COD con-
centration is high and sanitary facilities are involved in the contribution of
this COD, then septic conditions could occur. If this happens, backwash should be
performed daily.
A number of underdrain systems are commercially available. How-
ever, it is strongly suggested that any underdrain containing plastic materials
169
-------
TABLE 3.3.1-2
CARBON ADSORPTION SYSTEM
SIZING AND COSTS
FLOW Cgpm) 20 40 80 160 320 640 1700
SUPERFICIAL VELOCITY
(gpnVft2)
VESSEL DIAMETER (ft)
RESIDENCE TIME
(min/adsorber)
BED DEPTH (ft/adsorber)
VESSEL HEIGHT (ft)
BED VOLUME
(ft3/adsorber)
CARBON WEIGHT
(Ib/adsorher)
CARBON COST
($/adsorber)
UNIT ADSORBER COST
($1000)
UNIT INSTALLATION
($1000)
TOTAL UNIT COST ($1000)
NUMBER OF ADSORBERS
TOTAL ADSORPTION SYSTEM
COST ($1000) 32 42 66 86 138 236 708
3
3
6.1
5.7
9
40
90
3
4
6.0
6.4
10
81
2170
3
6
6.3
6.0
10
170
4600
4
7
6.0
8.3
13
320
8650
5
9
6.1
10.2
16
650
17550
6
12
6.3
12.0
19
1360
36700
6
12
6.3
12.0
19
1360
36700
55
13
3
16
2
1085
16
4
21
2
2300
25
6
33
2
4325
31
8
43
2
8770
48
L2
69
2
18350
80
20
118
2
18350
80
20
118
6
170
-------
TABLE 3.3.1-3
ACTIVATED CARBON ADSORPTION SYSTEM
ANNUAL OPERATING COSTS
($1000)
FLOW (gpm) 20 40 80 160 320 640 1700
YEARLY CARBON COST 3.51 7.Q1 14.Q1 28.Q1 4?2 702 13?2
CARBON SYSTEM AMORTIZATION 4.8 6.3 9.8 12.8 21 35 106
(8% - 10 yrs)
LABOR AND MAINTENANCE 10.3 10.9 12.5 14.8 17 27 76
PUMPING-COST (2C KWH) 0.2 0.3 0.7 1.3 3 5 14
ANNUAL OPERATING COST 18.8 24.4 37.0 56.9 88 137 333
COST/1000 GAL ($) 1.79 1.17 0.88 0.67 0.52 0.41 0.37
Spent carbon replaced by virgin carbon @ 50£/lb.
2
On-site regeneration.
171
-------
FIGURE 3.3.1-1
ACTIVATED CARBON SYSTEM INSTALLED COST
1000'
800
600
400
200
SYSTEM
COST I0°
$1000) so
60+
20
10
20 40 60 80 100 200 400 60O 80O 1000 20OO 3OOO
172
-------
should not be used for this particular application, due to the possible impregna-
tion of the plastic with PCBs. Stainless steel should be used as the underdrain
system and this can either be installed by bolting or welding to the vessel. This
underdrain performs three important tasks.
1. Support of the activated carbon bed.
2. Allow passage of the processed wastewater from the
bed to a penum located at the bottom of the absorber.
3. Uniform distribution of the backwash water.
Three basic types of underdrains are presently in use. They include:
1. Well screen
2. Neva-Clog (a product of Multi-Metal Wire Cloth, Inc.)
3. Individual nozzles which provide even distribution of
the backwash water; installed at predetermined spots
in a support plate.
For the purpose of this study, removal of activated carbon will not be performed
through the support plates or any underdrains. The activated carbon will be re-
moved from an access port at the side of the vessel, either hydraulically or
physically. Hydraulic removal is recomnended and this can be done with an eductor
2
while backwashing at approximately 2 to 3 gpm/ft . Slurry concentrations of 1 to
4 Ibs/gal have been experienced but 1 Ib/gal is average. Suggested pipeline velocity
is 3 to 6 feet/second.
Typical granular activated carbon specifications may be seen in
Table 3.3.1-4. Molasses decolorizing index and iodine number have been included
to provide the user with a relationship that would give information as to the
macropore volume and micropore volume of the activated carbon. The iodine number
may be related to the total pore volume of the activated carbon, including micro
and macropores whereas the molasses decolorizing index provides information as to
the larger pores (greater than 30 Angstroms in diameter). Activated carbons with
large molasses decolorizing indices tend to provide a higher rate of absorption
173
-------
TABLE 3.3.1-4
TYPICAL ACTIVATED CARBON SPECIFICATIONS
MOLASSES DECOLORIZING INDEX (min) 6.3
IODINE NUMBER (min) 900
ABRASION NUMBER (min) 70
MOISTURE WHEN PACKED (70, max) 2
ASH (%, max) 8.0
PARTICLE SIZE (U.S. Sieve Series) 8 x 30
OVERSIZE (%, max) 8
UNDERSIZE (%, max) 5
174
-------
over those with lower numbers. However, the total working capacity of the
activated carbons that have the higher numbers might tend to be less than those
with the lower numbers if the iodine numbers are essentially the same. For
virgin carbons, the iodine number tends to be the major source of absorptive
information available to the knowledgeable consumer and the use of the molasses
decolorizing index may be dropped without any appreciable decrease in quality of
the activated carbon.
Granular activated carbon has been chosen for this application
since the requirement is for very low to zero discharge of PCBs. Powdered acti-
vated carbon is also available; however, it is questioned whether it may be used
successfully for this application.
3.3.1.1 Regeneration of Spent Carbon
It has been mentioned that due to the low usage rate
of the activated carbon, it should be used on a throw away basis. There is
a possibility of thermal regeneration for this activated carbon. Other possible
regeneration techniques are solvent extraction, chemical regeneration and steaming.
Due to the high efficiency required of the carbon beds, it is strongly suggested
that solvent extraction not be utilized. Since the nature of the PCB molecule is
not acidic or basic, chemical regeneration will not apply here. Due to the high
boiling point, and therefore extremely low vapor pressure of PCBs, steaming at
any available steam pressure is not recommended for removal of these PCBs. This
leaves only thermal regeneration.
Carbon regeneration costs were developed using a 30
inch, six hearth furnace at an installed capital cost of $180,000, including carbon
handling system. These costs are shown in Figure 3.3.1.1-1.
A number of physical and chemical operations are in-
volved in thermal regeneration. Carbon used in liquid service is dewatered by
mechanical means to 30 to 40% moisture and fed to the thermal regeneration furnace.
The first section dewaters the carbon completely by vaporization at temperatures
up to 300°F. As the carbon passes through the furnace, the temperature increases
175
-------
UJ
o
<
CO
D
CO
>
f-
co
o
o
o
t-
<
o:
UJ
2
^ UJ
^S
-} o:
m
" Z
63 O
§ CD
d cc.
E<
u
_)
<
H
i Og
i3 z o
s < P.
I O
•^ LU
u 5
ID < UJ
x Z (- .
* Q: in >• O
ro u.
(rf
i.o CD
O
O
•s
U
I 2
CD tr v> v)
o: ui o (-
< z o 2
UJ
(T
176
-------
due to countercurrent flew of carbon and gases. The next step is removal of
volatile organic compounds which is performed at temperatures of 300° to 600°F.
Those organic compounds which are not removed by volatilization remain in the
carbon pores and are thermally decomposed, leaving an amorphous carbon residue
at temperatures of 600° to 1200°F. This can be called a charring step. The
last and most important step of the thermal regeneration process is removal of
the char from the pores and can be called reactivation.
Since this char cannot be physically removed from the
carbon, it must be chemically removed. This can be done by selective oxidation.
Oxygen reacts with carbon to produce gaseous carbon monoxide and carbon dioxide
which can be removed with the combustion gases. This reaction occurs at temperatures
of 600°F and is exothermic. It is difficult to control since the oxygen attacks
the carbon granule as well as the char in the pore. In fact, the action is so fast
that at temperatures above 1000°F the oxygen reacts with the first carbon with
which it comes in contact. This means that it will preferentially oxidize the
carbon of the granule and deteriorate it.
Another method of removing the charred deposit is to
use steam. Steam reacts with the carbon to produce gaseous carbon monoxide and
hydrogen which can be removed with products of combustion in a direct fired system.
This reaction is endothermic and tends to be quite controllable. Between 1200 and
1500°F, the rate is quite low and not very effective for removal of char. At
temperatures above 1500°F it is effective and between 1500° and 1700°F the steam
reacts preferentially with the amorphous char deposit in the pores. The reaction
with the crystalline, graphitic structure of the original granule is much slower.
As the temperature increases beyond 1700°F, the reaction becomes less selective.
Thus, the effective temperature range for the char removal step with steam is 1550°
to 1700°F. A steam to carbon ratio can be as high as 1:1 but normally it is 0.5:1.
At temperatures above 1700°F some of the fine micro-
pores are sacrificed for large micropores. In most cases, the small micropores
(less than 30 Angstroms diameter) are the pores which contribute to the absorption
of low molecular weight organic compounds. It is anticipated that PCBs will fall
in this range. When these are widened, the compounds are not held as tightly, and
177
-------
therefore the capacity of the carbon for the simple organic compounds decreases.
The optimum regeneration keeps the micropore volume constant. In most cases it
is advisable to sacrifice some of the micropore volume (tip of the pore) in order
that the granule is not subjected to over regeneration. Usually this amounts to
about 5% of the virgin micropore volume.
It has been shown that destruction of PCBs at temper-
atures above 1600°F is effective in about 99.997% efficiency/3^ Since most
comnercial PCB users utilize the lower boiling (about 400°F) PCBs, then most of
these compounds would be vaporized from the activated carbon before it reached
the 1600° to 1700°F portion of the reactivation furnace. These will then be ex-
pelled with the combustion gases from the furnace before they could be decomposed.
For this reason an afterburner would be required that would have a minimum temper-
ature of 1600° and maximum temperature of 2000°F. Retention times in the after-
burner have been shown to be between 0.10 and 0.11 seconds. Commercial installations
utilizing retention times of 2 to 3 seconds for decomposition of PCBs are in use.
The combustion gases from a reactivation furnace should pass through such an after-
burner system. This should be scrubbed to remove any particulate contaminating PCBs
and HC1 formed in the decomposition steps.
Any thermal regeneration process should be strictly
monitored for residual PCBs in the gaseous effluent and also any remaining in
the granular activated carbon that has been regenerated. It is strongly suggested
that any regeneration facilities designed for reactivating PCB carbon should isolate
that carbon from any other carbon to be regenerated. In other words, the reactiva-
tion facility should be dedicated to PCB use only. This will ensure that carbon
that has not been contaminated with PCBs previously will not be cross contaminated
with that which has been used for PCB cleanup.
3.3.1.2 Economic Analysis of Carbon Treatment of PCB
Contaminated Wastewater
Figure 3.3.1.2-1 graphically depicts the different
operating costs between a waste stream containing only PCBs and one with other
organic contaminants that provide COD adsorption differentials of 20 and 40 mg/1
178
-------
500
400
30O
200
100
90'
ANNUAL sc
($1000) so,
4O
FIGURE 3.3.1.2-1
ACTIVATED CARBON
WASTEWATER TREATMENT SYSTEM
TOTAL ANNUAL COST
BASED ON DIFFERENT
CONTAMINANT PARAMETERS
400 600 800 1000
2000 3000
FLOW (gpm)
179
-------
(feed concentration of 30 and 50 mg/1). The only cost differential contemplated
is the difference between the carbon usage for the three streams. This cost
differential indicates that it would be to the treaters' benefit to isolate and
treat only those streams contaminated with PCBs. It could also be to their bene-
fit to remove any stream which is free of PCBs, but containing carbonaceous COD
from the treatment scheme. It is evident from Figure 3.3.1.1-1 that carbon
usage rates less than 200 Ibs. per day do not justify on-site regeneration when
a virgin carbon cost of $0.50 per pound is used. Using the PCB isotherm data,
it was found that no flow in the range studied would justify on-site regeneration.
When using the assumed COD adsorption rate of 20 and 40 mg/1, it was found that
regeneration on-site is economically feasible at flows above 210 gpm.
Toll regeneration could be utilized for systems using
less than 200 Ibs/day of carbon if the regeneration facility is equipped with
sufficient safeguards to prevent contamination of other carbon by PCBs and ensure
that air and water pollution does not occur. If this is possible, the operational
costs for flows less than 210 gpm can be reduced accordingly since these were
calculated using virgin carbon replacement of the spent carbon. This holds for
all figures used to develop the PCB curve in Figure 3.3.1.2-1, also.
At flows above 640 gpm, parallel adsorber operation
was required because of on-site fabrication cost of stainless steel vessels. Shop
fabrication is less expensive than on-site fabrication of stainless steel pressure
vessels. Twelve (12) feet is the maximum transportable diameter.
For a total treatment cost, the exist of equalization,
filtration and incineration must be added to the capital cost for the adsorption
system and $180,000 for on-site regeneration capital cost for flows above 210 gpm.
Figure 3.3.1.2-2 includes all amortized capital costs
associated with the activated carbon system except disposal of spent activated
carbon used in the incremental flow cases between 20 and 160 gpm.
3.3.2 UV-Ozonation - Potential Alternative
3.3.2.1 Description of System Components and Rationale
for Cost Evaluation
180
-------
COST
(fl/IOOO
GALLON)
300'
20.O
01
10
FIGURE 3.3.1.2-2
COST PER 1000 GALLON FOR
PCB REMOVAL FROM WASTEWATER
WITH CARBON
TOTAL TREATMENT COST
PRETREATMENT COST
CARBON TREATMENT
40 60 80 IOO
200
400 600 800 1000
aooo
FLOW (opm)
181
-------
The technology employed in a proposed PCB treatment
system utilizing UV catalyzed ozonation has been described previously. Some
description was also provided on scaling up Westgate Corporation's batch reactors
for a continuous treatment system. Further discussion with Jack Zeff of Westgate
and Dr. Robert Legen of Houston Research led to the conclusion that scaling up
batch reactors for purposes of developing a PCB treatment system cost evaluation
was not the best approach. Unfortunately, however, with the process still in its
infancy (not much beyond laboratory development) no straight forward guidelines
exist to generate cost data.
With this in mind, the approach taken in developing
approximate cost data was to:
1) Design a reactor similar to a chlorine
contact chamber;
2) Size this reactor based on estimated required
residence times extrapolated from Houston
Research Corporation's batch studies with PCBs;
3) Estimate the required 0., production rate based
upon recommendations by Houston Research for
wastes containing only PCB plus the additional
ozone demand exerted-by the non-PCB COD in
the waste;
4) Estimate the required number of UV lamps
for a given reactor volume based on re-
commendations by Westgate for continuous
systems.
The reactor shown in Figure 3.3.2.1-1 is a three stage
covered concrete tank. The length to width ratio was chosen as 2 to 1 to minimize
short circuiting. The depth was chosen as 12 to 14 feet to maximize 0~ transfer
efficiency. Spargers, located near the tank bottom, provide the required mixing
as well as CL dissolution.
182
-------
0?
d
*"
0
u ur°
-J
$
rf 1
a: X
*— f
i
3
h-
3
^^^
IT
-o:
2UJ
BIT,
39
L<
J 3NOZO
g
if
N
N
CXI
cxa
-nxi
J
Q
UJ
ffi
|
^0-3*
F3
Gjj
y
2
S9IJDA
SPARGER
\
\
1
1
|
H
^f
n
Lj
.81
w
TJ
is .-
z
o
UJ
UJ
z
o
cc
O
o
o:
e
<
UJ
o:
Q
LJ
o
a.
o
or
a.
183
-------
UJ
2
JBI
If
J L
<\j">
UJ
.1 , ,1
-I I
*
^"
5
CVJ
M
(d
U
5|
4J S
2 C
ro O
„ O
•H -.
Cn O
I , , I
I I
*
<
UJ
o:
o
UJ
C/)
o
CL
•o
cc
Q_
184
-------
Low pressure, non-phosphur coated fluorescent tubes
(not shown in the sketch) would be stacked vertically throughout the tank to
provide the required UV light. For a full scale system, Westgate suggested use
of six 43-watt tubes for a reactor volume of 54 cubic feet. For conservation in
the cost analysis, the number of lamps was determined based on use of one lamp
per five cubic feet of reactor volume.
The reactor residence time was determined from a
logarithmic plot of Houston Research batch study data. See Figure 3.3.2.1-2.
The data show that the reaction rate approximates first order kinetics. This
implies that the rate of PCB oxidation by this process is proportional to the
PCS concentration and that very long residence times are required for final PCB
concentrations on the order of 1 to 10 ppb. Extrapolating the logarithmic plot
is necessary to determine the residence time necessary to get to these very low
final concentrations. Houston Research batch studies did not go to PCB levels
below about 30 ppb. It must be cautioned that extrapolating the plot might be
producing very erroneous data. In other words, residence times far in excess of
those calculated might be necessary to actually achieve the desired final PCB
concentration. On the other hand, both Houston Research and Westgate claim that
the process could be optimized such that shorter residence times would be re-
quired. Jack Zeff, though, stated that an optimization study of at least 4 to
6 months duration would be necessary before meaningful design residence times
could be generated. For purposes of this cost analysis, the residence times
generated from the extrapolated data will be used. Four cases were looked at:
1) Filtration not included in pretreatment.
Influent 200 ppb. Effluent PCB concentra-
tion 10 ppb. Residence Time 210 min.
2) Filtration not included in pretreatment.
Influent PCB concentration 200 ppb.
Effluent PCB concentration 1 ppb. Residence
Time — 365 min.
185
-------
Figure 3.3.2.1-2
PLOT OF HOUSTON RESEARCH TEST DATA FOR DESTRUCTION OF PCS
USING ULTRAVIOLET LIGHT CATALYZED OZONATION
C/Co
01.
one
0.06
OO4
0.02.
o.oi.
0.008-
OJDO&
0.004
OJ002
Co='>00ppb
V
\
\
\
I ITT
\
\
\
100 200
TIME (min)
joo
400
186
-------
3) Filtration included. Influent PCB concentra-
tion 50 ppb. Effluent PCB concentration 10
ppb. Residence Time — 110 min.
4) Filtration included. Influent PCB concentra-
tion 50 ppb. Effluent PCB concentration 1 ppb.
Residence Time — 265 min.
The Houston Research recommendation for the ozone
requirement for a waste containing only PCBs is:
1) 4 rog/1 O_ concentration in effluent from reactor.
2) 0.1 mg/1 per min. of residence time to account
for auto decomposition of CL.
For waste containing PCB in combination with non-PCB COD, an additional 0., demand
will exist. For the sake of conservatism, it will be assumed that the non-PCB
COD must be completely removed before the PCB can be oxidized. It will also be
assumed that this can be done within the residence times required for the PCB
removal. From "Ozone in Water and Wastewater Treatment" by Evans, an estimate of
two to three Ibs. of O_ per Ib. of COD was determined. This estimate is thought
to be valid at COD concentrations below 50 mg/1. The ratio 3 will be used.
1 Ib. COD
The above parameters were used to develop the following
equation:
I A mrr/i 4- M nKT/1) (Residence Time (min)) + (3 Ibs) (COD (mg/1)]
1 4 mg/j. t- ^^ ^ j
X IFIOW -^11 fl.44 X Hf3 ^ (8.434)/1 ^ °3 Produced \]
gpm/ 1.7 Ib03 dissolved/J
The estimate of dissolution efficiency of .7 was also arrived at from "Ozone in
Water and Wastewater Treatment".
187
-------
3.3.2.2 Capital Costs
The breakdown of installed capital costs for a UV
ozone system are shown for various residence times and COD values in Table 3.3.2.2-1.
The system installed costs vs. flow rates are also plotted in Figure 3.3.2.2-1
through Figure 3.3.2.2-4.
Table 3.3.2.2-2 contains total capital costs for the
0., generator and dissolution equipment. The numbers were furnished by the Union
Carbide Corporation for various 0., production rates. The installed ozonator costs
include the generator, air preparation equipment (recycle equipment and 0« pro-
duction equipment when 0 is used as feed gas) and all related systems. The in-
stalled dissolution equipment includes spargers, piping blowers, etc. and was
estimated at 5% of the installed ozonator costs. Electrical and instrumentation
was estimated at 15% of the ozonator and disolution equipment. Recorrroended
building sizes were furnished by Union Carbide and the cost was estimated using
$60 per square foot. The total capital was plotted in Figure 3.3.2.2-5 to generate
capital costs for all required O., production rates.
For production rates in excess of 2000 Ibs/day, O_ was
•~J
assumed to be the feed gas.
The reactor costs were generated by determining reactor
dimensions for the various cases studied. Once the geometry was determined, the
cubic yards of excavation was calculated..and $5.00 per cubic yard was used. The
installed reinforced concrete yardage was determined and a price of $285 per cubic
yard used.
The number of UV lamps was determined based on reactor
volume and a price of $20 per lamp installed was used. Jack Zeff indicated that
the lamps usually cost about $15.
3.3.2.3 Operating Costs
The operating costs for the UV-ozone system are broken
down into two parts. The 0_ generation end dissolution power costs shown in
Table 3.3.2.3-1 were developed from estimates of installed power in KW supplied by
188
-------
OP fri
3.3
M
-f flj
H C
fO -rn
-p tp
£ a
to co
-P 0
o o
rH 4J
-P H
C -P
^
5 J
> J
D *
rH
Jj
^ Vi
QJ fi
rH H
cfl -p
4J W
in q
PI
H UH
ti .
sg
-P rH
d -u
•ss
-H -tJ
Q
•J) O
o
R
O T3
C \
o in
N £1
O rH
o.y
6^
fi<
5
I
^ 2
Q 5)
•P 13
O -rl
Id M
o o
o o
o o
in vo
vo
CO VO
in co
in en
o o
o o
CO O
co r~
H m
o o
*3* CO
en co
rH
o o
o o
O 0
o en
co ^r
VO i — (
O rH
r- ^r
" ^
o o
0 O
0 0
o o
rH 0
in en
CO VO
co r~-
Ln i — i
rH
CD CD
CO VD
O O
o o
o o
rH *y
en r^
in en
01 CN
0 0
o o
CN CD
in ^p
in o
CO 1—
CN CN
O O
O 0
CN O
in o
[^ O
rH
O O
VO O
r- o
co in
o o
0 O
o o
o ^r
CO O
rH
CN CO
O! cn
co cn
CO <•
rH CN
0 O
0 O
o o
o o
0 O
CN in
CN CN
CN O
m o
CN .0
o o
CO O
CN r~
rH rH
O O
0 O
o o
VO CO
CN -31
OJ ^J*
(N •*
rH CN
rH 01
O O
O 0
O O
O fO
rH rH
rH CO
VO CN
in rH
H
O O
o o
o o
o m
in r^
O rH
•*}• CO
0 0
CN T
o
ro
2
CN
O
O
O
CO
o
CN
O
0
O
a*
CO
rH
O
O
O
-------
-a
8
H
en
>
D
I
g
sj;
Ol-P
01 Ul
ui C
*-3 LO
1SJ
0 T3
O r-l
y
O rH
rH Hi
DM O
§>
H§
O O O
0 O O
o o o
rO CO >O
CO CT* m
n vo CN
rH
O O O
o o o
o cr> co
CO ^f rH
•*r ro ^r
ro ^D i — I
rH
O O O
o o o
O CJ1 CO
CO IT) rH
rH n r-
OO VD CN
co r^ in
o o o
O 0 0
O 0 O
O CTi O
^ CT^ CTl
^ co r^
•-1
000
0 O O
o o o
O O 0
rH in CTl
m in en
(N •* CO
m ^r co
ro VD CN
rH
o o o
rH ro vo
f >
o
o
CO
ro
CO
(N
O
O
r^-
C7v
r^
m
04
0
0
r~
m
rH
«;
m
rH
O
0
0
vo
fO
rH
ro
ro
O
CNrHCOOCNOOr-
cN^rcor^oin rHCNjro^DiHt^cri
rH rO rO rH CN CN
OOOOOO OOOOOOOO
OOOOOO OOOOOOOO
o o c°> oo r~~ o~°> 0^1 "*o ro ^ ^j* p^ *d* ON
cr\cou-)rHroo CNinrHCNiroorHO
H rH rH CN
in en 0*1 cr\ oo ^y ^j co vo ro vo ro ^o ^j*
^y co r^- m i — i m rH CN in i — I oj in o o
rH
OOOOOO OOOOOOOO
OOOOOO OOOOOOOO
O O O O O O C"-l (O CO i — 1 O <'") <^ i — 1
o o o ^o (N co r~~ ^i* vis r~~ i — 1 **o u>
"^ rH rH rH rH
siisSS ssgsssss
oj^cor-mr-- rHo4ini^ri^-vu'Jji^ rHi'jui^^-i'in
rH 04 ro rH OJ 01
oJ"=fcncncnrH tHrovijror-'m
rH ro tn rH oj ro
OOOOOO OOOOOOOO
rH rH rH rH
O
ro
CM
190
-------
u
rH
04 O
ro EH
s|
"*" cij
•d .3
-U tji
si
H
Ij -P
-P U5
O O
EH O
rH CO
rn3 H
CO -P
c w
^8
M
ro w
.
a =*
rH
no 4-)
O en
•H c
^
JJ 4->
4J '-H
tS •
~£
4J
iH •
W -U
CO CO
C M
<-3 -P
CO
M
a 5T
C ^\.
O M
¥
01 'i
rH
fa O
rH
Q \
8£
8
q §
43 -a
O -rH
rd M
_JJL
O O O
o o o
o o o
•^ cr> u">
co r- M
rH n
o o o
O 0 O
rH C~l rH
VO Osl ^O
f"-* ^i co
rH (N
O O O
O 0 O
(N in rH
rH
rH (N VD
•^ CO VD
rH CN IT)
O 0 0
O O O
CN 01 CO
ro r^ "=T
rH .-H CN
O -H ro
I*- "^ CO
rH CN
O O O
O 0 O
O O O
O O 0
VD ^J* L")
iH f^
p^ r*~ ro
VO CN LO
rH CN
O O 0
CN ^ CO
S
CM
O
O
0
en
in
o
0
CO
CO
CTi
O
o
04
O4
^
ro
rH
rH
O
O
rH
^D
ro
CO
vo
in
o
o
0
o
*y
r-^
o
in
0
rH
O
o
0
m
o
rH
O
0
O
m
en
o
o
ro
CO
OJ
01
o
o
o
r-^
in
ro
ro
rH
rH
O
o
o
o
in
CO
*y
rH
O
rH
O
0-4
ro
o
0
o
o
rH
m
CN
o
o
o
CTl
^Jf
ro
in
*r
o
o
CO
rH
^
CN
r-
OJ
O4
0
o
o
o
o
rH
CN
r^
04
o
CN
o
vi>
o
o
o
CN
in
o
o
ro
CO
04
ro
o
o
rH
CO
rH
CO
O
o>
o
o
CTi
VD
in
rH
^
ro
in
^
o
0
o
0
o
<2\
CN
•^
in
o
-a-
0
CO
CN
O
o
0
rH
O
o
o
o
o
p^
ro
10
ro
o
o
01
o
04
^0
O
oq
1-1
o
o
rH
UD
en
rH
CN
ro
OJ
o
vo
o
o
o
o
o
OJ
ro
ro
ro
in
o
o
r^
rH
191
-------
oooo
oooo
OOOO
ooooooooooo
ooooooooooo
OOOOOOOOOOO
oooo
oooo
'
o
o
"
o
o
oo
oo
•s .
oooooo
oooooo
o
o
oo
oo
r—i
a
\j
ts
oo
oo
ooooooo
ooooooo
oooo
oooo
-
ooo
ooo
OOO
OOO
oo
oo
OO
O
o
LO
O-j
oooc^ooooo
M c:
192
-------
Figure 3.3.2.2-1
UV-OZONE SYSTEM INSTALLED COST
10000
eooo
6000 •
4000
2000
1000.
800
SYSTEM goo
COST
($1,000) 400
200
100-
eo:.
60
40-
20.
10 •
10
REACTOR RESIDENCE 110 min
20
40 60 80 100
200
\.
400 6OO BOO 1000 2000 40OO 60OO 0000
FLOW (gpm)
O-C.O.D.-3Omg/l
A -C.QD.-5Omg/l
193
-------
Figure 3.3.2.2-2
UV- OZONE SYSTEM INSTALLED COST
10000
6000
4000
20OO
IOOO'
800'
SYSTEM goo
COST
($1.000) 400
200,
80-.
60.
20.
10
10
/CX7
REACTOR RESIDENCE ZIOmin
X
20
40 60 80 100
200
>Y •
X
X
40O 6OO 800 1000 2000 40OO 6000 80OO
FLOW (gpm)
O-COD.-30mg/l
A -C.O.D.-50mg/l
194
-------
Figure 3.3.2.2-3
UV-OZONE SYSTEM INSTALLED COST
10000
eooo •
6000 .
4000 •
2000
1000
800 -
SYSTEM 600
COST
($1,000) 400
200
100-
80-
•
60-
40-
20.
10'
10
REACTOR RESIDENCE 265mm
20
40 60 80 100
ZOO
X
400 600 800 1000 EOOO
I
4000 60OO 80CO
FLOW (gpm)
G-C.O.D.-30mg/l
A-C.OD.-50mg/l
195
-------
K3OOO
8000 •
6000 .
4000
2000
1000-
800 -
SYSTEM 6oo
COST
($ 1,000) 400
zoo
100-
60'.-
60-
40-
20.
10-
Figure 3.3.2.2-4
uv-OZONE: SYSTEM INSTALLED COST
REACTOR RESIDENCE 365min
10 20 / 40 60 60 100 ZOO 400 6OO 800 JOOO 2000 4000 6000 8000
FLOW (gpm) O-C.O.D.-30mg/l
A-CO.D.-50mg/l
196
-------
TABLE 3.3.2.2-2
OZONE GENERATION AND DISSOLUTION COST DATA - CAPITAL COSTS
Ibs. 0,.
Day
Feed Gas
Mr
130
260
510
1020
2040
4080
8200
Feed Gas
°2
2040
4080
8200
Installed
Ozonator
114,000
204,000
372,000
660,000
1,200,000
2,224,000
4,176,000
1,860,000
2,340,000
3,720,000
Installed
Diss. Equip
6fOOO
10,000
19,000
33,000
60,000
112,000
209,000
93,000
117,000
186,000
Electrical &
Instrumentation
18,000
32,000
59,000
104,000
189,000
350,000
658,000
293,000
369,000
586,000
Bldg.
3,800
7,500
15,000
22,500
30,000
45,000
75,000
30,000
45,000
75,000
Total Cap.
142,000
254,000
465,000
820,000
1,479,000
2,731,000
5,118,000
2,276,000
2,871,000
4,567,000
197
-------
Figure 3.3.2.2-5
OZONE PRODUCTION AND DISSOLUTION CAPITAL COSTS
10'
806-
60-
402
20.
DOLLARS
10
80*1
60-
40-
205-
100
200
r-02 As
\Feed Gas
•Air As
Feed Gas
400 600 80o 1000
2000
Ibs 0.
day
40OO 6000 8000 IO
198
-------
TABLE 3.3.2.3-1
OZCNE GENERATION AND DISSOLUTION COST DATA - POWER COST
Ibs 0-.
Day
Feed Gas Air
130
260
510
1020
2040
4080
8200
Feed Gas
o2
2040
4080
8200
Installed KW
64
128
256
510
1020
2040
4100
445
880
1750
KWH 0., Prod
Day
1536
3072
6144
12240
24480
48960
98400
10680
21120
42000
KWH 0., Dissolution
Day
79
154
307
612
1224
2448
4920
1224
2448
4920
Total KWH
day
1615
3226
6451
12852
25704
51408
103320
11904
23568
46920
Cost @ $.02
day KWH
32
65
129
257
514
1028
2066
238
471
938
199
-------
Union Carbide for various production rates. The estimate for dissolution power
was made by taking 5% of the KW for 03 generation. A price of $.02 per KWH was
used. Costs for all production rates are plotted in Figure 3.3.2.3-1.
The power costs for the UV lamps and the operation and
maintenance costs are shown in Table 3.3.2.3-2 for all the conditions studied. Each
lamp was rated at .043 KW. For flow rates up to 640 gpd, it is assumed that one
man would be required for 24 hours at $15 per hour. For flow rates 640 gpd and
above, two men are assumed necessary.
3.3.2.4 Total Treatment System Cost for Zero Discharge System
Because of the; very high capital and operating costs
associated with the generation of 0_ which is primarily affected by COD and re-
actor residence time, it is assumed that full scale systems would incorporate
filtration before UV-ozone treatment. This would reduce the required residence
time as the influent PCB concentration wovild be reduced to 50 ppb.
Because of the desirability to cipproach "zero PCB
discharge" an effluent concentration of 1 ppb is necessary. Measuring concentra-
tions below 1 ppb is questionable so for t:he sake of this cost analysis, 1 ppb
wi31 be assumed to mean "zero discharge".
The total system cost will be based on the following
parameters:
Influent PCB = 50 ppb
Effluent PCB' 1 ppb
Reactor Residence Time = 265 rain
Non-PCB COD Cone. = 30 rng/1
The total installed system cost including pretreatment costs to accomplish the
above appears in Table 3.3.2.4-1. The total annual operating cost and cost per
1000 gallons are presented in Table 3.3.2.4-2.
3.3.2.5 Total Wastewater Treatment Plant Costs Using Activated
Carbon or Ultraviolet Light Catalyzed Ozonation
The costs developed in Sections 3.2, 3.3.1 and 3.3.2
were used to ca.l.:ulate the total treatment system costs. For the activated carbon
200
-------
FIGURE 3.3.2.3-1
OZONE GENERATION AND DISSOLUTION POWER COSTS
10
200
40O
600 800 1000
Ibs03
day
2OOO
4000 6OOO 8OOO IOOOO
201
-------
TABLE 3.3.2.3-2
U.V. LIGHT POWER AID SYSTEM 1-MSTTENANCE COST
Residence
Time (min)
110
210
265
365
Flow
Gal/day
20
40
SO
160
320
640
1280
1700
20
40
80
160
320
640
1280
1700
?0
40
80
160
320
640
1280
1700
20
40
80
160
320
640
1280
1700
U.V. P^wer Operation & Maint.
KWH
day
62
124
249
495
970
1940
3880
5160
116
232
464
927
1853
3707
7416
9849
145
290
584
1170
2340
4679
9358
12,431
201
402
804
16"5
3221
6444
12,890
17,120
Cost
day
1
2
5
10
19
39
78
103
2
4
S
19
37
74
148
197
3
6
12
23
47
94
187
248
4
8
16
32
64
129
258
342
I Men
1
1
1
1
1
2
2
2
1
1
1
1
1
2
2
2
1
1
1
1
1
2
2
2
1
1
1
1
1
2
2
2
Cost
day
360
360
360
360
360
720
720
720
360
360
360
360
360
720
720
720
360
360
360
360
360
720
720
720
360
360
360
360
360
720
720
720
Total
Cost/day
361
362
365
370
379
759
798
823
362
364
368
379
397
794
868
917
363
366
372
383
407
814
907
968
364
368
376
392
424
849
978
1062
202
-------
TABLE 3.3.2.4-1
FLOW (gpm)
Ozone-UV System
Installed
Inclading Engineering
Total Pretreatmsnt
Plant Cost
Total Plant Cost
TOTAL INSTALLED SYSTEM COST FOR
U.V.-OZONE SYSTEM
($1000)
20
40
73 113
142 198
215 311
80
221
308
529
160
320 640 1700
428 748 1300 3231
484 821 1463 3699
912 1569 2763 6930
203
-------
TABLE 3.3.2.4-2
ANNUAL OPERATING COST FOR U.V. OZONE SYSTEM
($1000)
FLOW (gpm) 20 40 80 160 320 640 1700
Total Plant Cost
Amortization 32 46 79 136 234 412 1033
(8%-10 yr.)
Ozone Generation &
Dissolution ^ 24 58 110 146
Power Costs @ $.02/KWH y 1J- -17 /4 ^b ilU 14b
U.V. Light Power
@ $.02/KWH and System 132 134 136 140 149 297 353
Maintenance Cost
Pretreatment Operation
& Maintenance & Power 4 7 10 17 28 49 108
@ $.02/KWH
Annual Total
Treatment Cost 197 238 322 477 789 1508 3340
Cost/1000 GAL($) 18.7 11.3 7.6 5.7 4.7 4.5 3.7
204
-------
system, the total installed system cost is given in Table 3.3.2.5-1. The total
annual operating cost and cost per 1000 gallons are presented in Table 3.3.2.5-2.
Figures 3.3.2.5-1 and 2 graphically depict these costs. Similar information was
presented for UV-ozonation in the previous subsection (subsection 3.3.2.4). In
each instance, pretreatment was considered necessary so the only variation in
cost other than that caused by flow is the difference between the carbon and ozone
systems.
To compare the two, a feed stream containing 30 mg/1
COD and 50 ppb of PCB was used for the final treatment section.
Comparing the terminal treatment capital costs of the
two systems shows at least a 100% greater cost for the ozone system over the car-
bon system. When treatment costs are combined with the terminal treatment costs,
the UV-ozonation process is about 10 percent higher than the carbon process.
This, leads to the recommendation that all efforts be
made to reduce flows.
3.3.3 Cost of Implementing Carbon Adsorption Treatment for Selected
Plants
Based on the unit operations costs for carbon adsorption and
pretreatment operations, including flow equalization, sedimentation/flotation and
filtration, capital and total annual costs were projected for selected manufacturing
plants. Table 3.3.3-1 presents data for a number of plants based on pretreatment
of total plant wastewaters with terminal activated carbon treatment and subsequent
discharge to surface waters. Table 3.3.3-2 introduces the alternative of segre-
gating and recycling non-contact cooling water in a closed loop air cooled system,
while providing pretreatment and activated carbon treatment for the process and
rainwater runoff waters prior to discharge. A comparison of the costs given in the
above tables indicates that the capital cost of the complete treatment alternative
versus that utilizing segregation of cooling water varies between 1.3 and 4.6,
favoring segregation of closed loop cooling water. The total annual costs likewise
reflect cost ratios of 1.2 to 3.4 favoring the alternative based on segregating
the cooling water in a close loop system.
205
-------
FLOW (gprn)
EQUALIZATION BASIN
FILTER SYSTEM
CARBON ADSORPTION
CARBON REGENERATION
DECANT TANK
CLEAN WATER RESERVE
TOTAL MDDULAR COST
PIPING & INSTRUMENTS
(25%)
TOTAL CAPITAL COST
ENGINEERING (10%)
TOTAL PLANT COST
TABLE 3.3.2.5-1
TOTAL INSTALLED PLANT COST
WITH ACTIVATED CARBON ($1000)
20
68
32
32
-
3
7
142
36
178
18
196
40
101
40
42
-
3
10
196
49
245
25
270
80
159
62
66
-
3
16
306
77
383
38
421
160
267
78
86
-
7
19
457
114
571
57
628
320
470
120
138
180
7
27
942
236
1178
118
1296
640
856
200
236
180
7
39
1518
380
1898
190
1088
1700
2080
600
708
180
12
39
3619
905
4524
452
4976
206
-------
FLOW (gpm)
AMORTIZATION (8%-10 yr)
CARBON
LABOR & MAINTENANCE
POWER
TOTAL
COST/1000 GAL($)
TABLE 3.3.2.5-2
TOTAL ANNUAL OPERATING COST FOR
PCB REMOVAL FROM WASTEWATER
WITH ACTIVATED CARBON ($1000)
20
29
4
14
-
47
4.47
40
40
7
17
1
65
3.09
80
63
14
21
3
101
2.40
160
94
28
28
5
155
1.84
320
166
47
37
11
261
1.55
640
284
70
59
12
425
1.26
1700
715
137
139
59
1050
1.18
207
-------
FIGURE 3.3.2.5-1 - Carbon System
TOTAL INSTALLED TREATMENT PLANT COST
FOR REMOVAL OF PCB FROM WASTEWATER
4000-
1000*
COST *oo
f&ir*nn^ Ann-
inn-
>, ^
)/
S
x
X
jf
^r
•'
s(°)
~s?
s^
J^
^
J
x
^^
/
/
9
/
10
20
40 60 80 100
ZOO
400 600 800 1000 2000
FLOW (gpm)
208
-------
FIGURE 3.3.2.5-2
TOTAL ANNUAL TREATMENT COST FOR REMOVAL OF PCS
FROM WASTEWATER USING CARBON AS FINAL TREATMENT
2000-
coo-
800'
600'
400-
COST 200.
($1,000)
100-
80-
60- -
40'
20-
10'
10
20 40 60 80 100 200 400 600 800 1000 2000 3000
FLOW (gpm)
209
-------
TABLE 3.3.3-1
COSTS FOR TREATMENT OF ALL FLOWS AND
DISCHARGING TO RECEIVING STREAM
WITH CARBON FINAL TREATMENT
($1000)
PLANT NO. 100 101 102 103 104 105 106 114
Capital Costs 1700 220 1400 3500 2500 1400 1100 310
Annual Costs 400 57 310 950 6QO 310 260 73
210
-------
TABLE 3.3.3-2
COSTS FOR ENCLOSED RECIECULATION AND COOLING
OF NON-CONTACT COOLING WATER AND TREATMENT
OF ALL OTHER FLOWS USING CARBON AS
FINAL TREATMENT
($1000)
PLANT NO.
Segregation Costs
100
101
102 103 104
105
106
114
Capital
Annual
Dry Cooling Tower
Capital
Annual
Treatment Costs
Capital
Annual
Total Costs
Capital
Annual
74
11
Costs
70
52
480
110
624
173
43
6
20
11
220 1000
57 250
220 1063
57 270
194
29
190
131
1100
260
1484
420
82
12
82
76
380
91
544
179
69
10
50
34
600
145
719
189
39
5
28
25
470
110
537
140
23
3
9
6
170
47
202
56
211
-------
Although segregation of the cooling wate^r in a closed loop
system appears to offer economic advantages, site restraints for segregation may
prove uneconomical at a particular plant. Therefore, some plants may choose to
treat the entire plant water and undergo the maximum costs for treatment.
3.3.4 Zero Discharge
Two alternative systems considered for zero discharge are de-
scribed below.
1. All plant waters, including rainwater' runoff would be
pretreated. The portion of the water to be utilized
in the plant would be subjected to terminal treatment
with carbon for purposes of recycling, while the re-
maining or excess water would be subjected to incinera-
tion in a specially designed system employing natural
gas as auxiliary fuel. Incineration would be conducted
under conditions which would result in essentially zero
aqueous discharge from PCBs producer and users.
2. All non-contact cooling water which accounts for about
90 percent of the water (in most plants) would be seg-
regated and totally recycled. All other waters in-
cluding rainwater runoffs would be united to a central
system for pretreatment. The portion of pretreated
water to be utilized in the plant would be treated with
carbon and recycled. As in the alternative above, the
excess waters would be subjected to incineration.
In both of the above cases non "wash up" sanitary waters are
to be handled separately and are not considered in the zero discharge costing. In
essence, the alternatives differ only in that non-contact cooling waters are
either carbon treated for recycling or segregated in a closed loop cooling system.
3.3.4.1 Description of Dry Cooling Tower
The segregated non-contact cooling water is to be
cooled in an air-cooled tower consisting of finned tubes and then recycled.
212
-------
Forced draft evaporative cooling towers were not considered to avoid the potential
for either water to air or air to water contamination with PCBs. In addition, the
air cooled system eliminates the need for significant blowdown of potentially con-
taminated cooling waters.
3.3.4.2 Design Criteria
Air cooled towers are proprietary systems and the
design criteria will be left to the vendor. Equations for estimating capital
and operating costs for these systems were obtained from the Marley Company.
The incineration system is envisioned to consist of
a baffled firebox with auxiliary fuel burners using a two second residence tine.
Water to be incinerated will be atomized directly into the firebox. No scrubber
is suggested since effluent gases should be very low in solids and HC1. The
exhaust gases could be used to heat a boiler, but no provision has been made for
this. The installation of a heat recovery boiler would reduce the operating
cost; however, it is believed that the energy produced, either as high pressure
steam or electric power, would be much more than required for the facility. The
operating temperature would be 2200°F.
3.3.4.3 Capital Costs
Capital costs for the incineration system were developed
on the basis of the size of burn chamber required for the desired residence time.
Using a 20% heat loss and 25% excess oxygen, it was estimated that about 300 ft
of gases would be exhausted at 2200°F and atmospheric pressure per pound of water
fed to the incinerator. Capital costs are presented for both the incinerator and the
dry cooling tower in Figures 3.3.4.3-1 and 2 and should be considered only approximate.
3.3.4.4 Operating Costs
The annual costs of incineration are most affected by
the quantity of gas required. Cost of gas was assumed to be $0.1083 per 100,000
BTU and the enthalpy of water at 2200°F at 2200 BTU/lb. Annual costs for the
dry cooling tower and incineration system are presented in Tables 3.3.4.4-1 and 2.
213
-------
FIGURE 3.3.4.3-1
INCINERATION COST AS USED FOR ZERO DISCHARGE
COST
3000
2000
ANNUAL OPERATING
COST (g 1,000) -—
I
-CAPITAL COST
($1,000)
COST/l,OOOgal($)
6 8 10
40 60 80 100
200 300
INCINERATOR FEED (gpm H20)
214
-------
400
Figure 3.3.4.3-2. COST FOR TOTALLY ENCLOSED COOLING SYSTEM
200
100
80
60
CAPITAL COST
($1,000)
40
GO
O
o
10
8
-ANNUAL OPERATING
COST ($1,000)
i i i
10 20 40 60 80 100 200 400 600 800 1000 2000 3000
FLOW (gpm)
215
-------
TABLE 3.3.4.4-1
INCINERATOR ANNUAL OPERATING COST
FOR ZERO DISCHARGE. METHOD
($1000)
FLOW (gpm) 12 4 8 20 40 80 160
AMORTIZATION
FUEL
LABOR AND MAINTENANCE
TOTAL
COST/1000 GAL ($) 103 62 49 35 29 26 24 22
13
10
31
54
13
21
31
65
14
42
47
103
18
84
47
149
37
209
62
308
64
418
62
544
101
836
62
1000
157
1640
94
1890
216
-------
TABLE 3.3.4.4-2
ANNUAL OPERATING COST FOR
TOTALLY ENCLOSED COOLING WATER LOOPS
C$1000)
FLOW (gpm) 20 40 80 160 320 640 1700
AMORTIZATION
POWER
MAINTENANCE (20%)
TOTAL 6 7 12 21 37 69 195
COST/1000 GAL («) 57 33 29 25 22 21 22
2
2
2
2
3
2
3
6
3
5
11
5
8
20
9
15
37
17
43
104
48
217
-------
3.3.4.5 Cost of "Zero Discharge" for Selected Plants
The projected capital and total annual costs for
selected plants to achieve "zero discharge" are shown in Tables 3.3.4.5-1, 2 and 3.
Costs presented in Table 3.3.4.5-1 (carbon treatment of cooling water) and Table
3.3.4.5-2 (closed loop segregation of cooling water) result from the pretreatraent
of all wastewater including rain water runoff.
The system employing carbon treatment of cooling water
for purposes of recycling demonstrates a total annual cost disadvantage ranging
from 1.01 to 1.74 times of the annual cost for the system using closed loop segre-
gation of cooling water. However, unique segregation situations may ultimately
favor non-segregation for a particular plant.
For comparison purposes Table 3.3.4.5-3 was developed.
This table presents the estimated capital and total annual operating cost for the
"zero discharge" system (closed loop cooling water) presented in Table 3.3.4.5-2
with the exception that the rainwater runoff is excluded.
For a "zero discharge" system the rainwater contribution
is extremely high. Comparison of the above two "zero discharge" cases [(1) inclusion
of rainwater runoff as incinerating material and (2) exclusion of rainwater runoff]
indicates the annual cost of the case 1 to be 2 to 42 times the cost of the case 2
for the various plants under study.
3.3.5 Comparison of "Zero Discharge" with Discharge to Surface Waters
As mentioned before the least cost "zero discharge" alternative
for a number of plants is exhibited in Table 3.3.3-2 while the least cost carbon
adsorption system (closed loop cooling water) for the same plants, is exhibited in
Table 3.3.4.5-2. The comparison of these two modes of wastewater treatment demon-
strates capital cost ratios between 1.1 and 2.3 favoring discharoe to waterways.
The differences in the annual operating costs for these two modes of treatment
are even more pronounced, demonstrating ratios ranging between 1.9 to 11.8 also
in favor of discharge to waterways.
218
-------
TABLE 3.3.4.5-1
COSTS FOR ZERO DISCHARGE
(CARBON TREATMENT OF NON-CONTACT COOLING WATER)
($1000)
Plant No. 100 101
Pretreatmant Costs
Capital
Annual
Carbon Adsorption Costs
Capital
Annual
Rainwater Incineration Costs
Capital
Annual
102
103
104
105
106 114
Total Capital Cost
Total Annual Cost
1400
220
210
130
its
140
165
170
30
20
14
250
300
1050
170
80
48
1500
3000
3000
500
500
250
850
1500
2100
340
260
180
280
320
1050
170
150
85
460
610
820
150
120
70
700
1000
210
40
40
22
140
165
1750 440 2630 4350 2640 1660 1640 390
515 344 3218 2250 840 865 1220 227
219
-------
TABLE 3.3.4.5-2
COSTS FOR ZERO DISCHARGE
(SEGREGATION AND RECYCLE OF
NON-CONTACT COOLING WATER)
C$1000)
PLANT NO.
Segregation Costs
Capital
Annual
Dry Cooling Tower Costs
Capital
Annual
Pretreatment Costs
Capital
Annual
Carbon Adsorption Corts
Captial
Annual
Rainwater Incineration Costs
Capital
Annual
Total Capital Cost
Total Annual Cost
100 101
102
103 104
105
106
114
74
11
70
52
330
62
62
36
140
165
-
—
_
—
170
30
20
14
250
300
43
6
20
11
800
145
39
22
1500
3000
194
29
190
131
830
150
110
62
850
1500
82
12
82
76
270
50
44
26
280
320
69
10
50
34
490
88
66
39
460
610
39
5
28
25
330
61
21
15
700
1000
23
3
9
6
150
25
21
15
140
165
676 440 2445 2174 758 1135 1118 343
326 344 3187 1872 484 781 1106 214
220
-------
TABLE 3.3.4.5-3
COSTS FOR ZERO DISCHARGE
(SEGREGATION & RECYCLE OF
NON-CONTACT COOLING WATER)
EXCLUDING RAHMWATER RUNOFF
C$1000)
PLANT NO.
100
Segregation Costs
Capital 74
Annual 11
Dry Cooling Tower Costs
101 102 103 104 105 106 114
43 194 82 69 39 23
6 29 12 10 5 3
Capital
Annual
Pretreatment Costs
Capital
Annual
70
50
310
59
-
-
110
17
20
14
180
34
190
131
600
120
82
76
210
40
50
34
340
63
28
25
110
17
9
6
110
17
Carbon Adsorption Costs
Capital
Annual
Total Capital Cost
Total Annual Cost
62
36
516
158
20
14
130
31
39
22
282
76
110
62
1094
342
44
26
418
154
66
39
525
146
21
15
198
62
21
15
163
41
221
-------
3.4 Scrap Oil Incineration
Based on plant site visits and interviews conducted during this study,
data relative to the quantities of liquid PCBs routed to incineration were collected.
These data are reported for capacitor and transformer manufacturers in Table
3.2.1.3-2 and in subsection 3.3.1.3 of the Waste Characterization Section (Section
IV) of this report. They can be surtmarized as follows:
Plant Wastes to Incineration
Number Ib/ton PCS Used
100 100
101 56
102 98
104 109
105 127
106 300
103 115
114 98
Avg. 125
Based on the data presented in Section IV, Table 3.1.1.1-1, 17,200
tons of PCBs were sold in the domestic market in 1974. Employing the average value
of 125 UD. PCBs incinerated per ton PCBs used and the domestic sales of 17,200 tons,
it is estimated that 2,150,000 Ibs. of PCBs are incinerated annually by the
capacitor and transformer manufacturers.
The rationale for estimating costs associated with incineration of
PCB contaminated solids and liquid was discussed in Section 2.9. It was determined
that a flat rate of $0.10/lb. would be assumed. This cost is at the high end of
reported toll incineration charges by commercial disposal firms (reported range
$0.05 to $0.10 per lb.). Employing the highest toll charge should reasonable account
for transportation and handling charges which are otherwise difficult to evaluate.
At the $0.10/lb. rate, the projected cost to the transformer and
capacitor manufacturing industry for scrap oil incineration would be $215,000 per
year.
222
-------
As previously indicated, -the only domestic producer of PCBs, Monsanto,
has an on-site incinerator with a capacity of six million pounds of PCBs per year.
It is reported that a toll of $0.05 per Ib. is charged for incineration of PCBs
returned to the producer by its customers. It can be reasonably assumed that this
is the "breakeven" cost for actual incineration at their facility. Thus the
maximum cost per year would be estimated at 6,000,000 Ibs. times $0.05/lb. or
$300,000. Ihis, of course, would have to be reduced by the amount of toll charges
received by Monsanto. Assuming 80% of the total liquid PCBs incinerated by the
capacitor and transformer manufacturers are actually returned to Monsanto, the
tolls received would be 0.8 x 2,150,000 Ib/yr x 0.05 = $86,000. Thus the net
annual operating costs for Monsanto's incinerator operation should approximate
$214,000 per year.
The total annual cost for scrap oil incineration for the producer and
users of PCBs should approximate $429,000 per year.
3.5 Solid Waste Incineration
The data relative to solid wastes (PCB contaminated) generated by the
capacitor and transformer industry is very limited. Only three plants reported
actual quantities of solids now sent to landfill or incinerated in relation to the
amount of PCBs used. These can be summarized as follows:
Solid. Pastes
Plant Number Ibs/ton PCB Used
100 248
101 129
102 65
Avg 147
Employing the average solid waste generation rate of 147 Ibs. per ton
PCB used and the domestic PCB sales volume of 17,200 tons per year, the projected
pounds of solid wastes to be incinerated per year would be 2,528,000 Ib.
Again, assuming an average cost for incineration of $0.10 per pound,
the estimated annual cost for solids incineration would be $253,000.
223
-------
3.6 Total Industry Treatment Costs
The maximum treatment costs for the PCBs users (capacitor and trans-
former) industries and for PCBs producer (Monsanto) are given in Tables 3.6-1,
3.6-2 and 3.6-3. The costs for the capacitor and transformer industries were
developed using the following assumptions.
1. Raw water supply is contaminated with PCB.
2. All groundwater runoff is contaminated with PCB.
3. All non-contact cooling water is contaminated and
will be used on a once through basis.
4. Sanitary wastes with the exception of wash water
would be discharged to municipality or treated
separately as required by State and local ordinances.
5. All process waters will be treated.
6. Incineration charge of $0.10 per pound will be
charged for PCB oils and contaminated solids.
7. Pretreatment followed by carbon final treatment and
discharge will be used.
Capacitor plants for which data are available as to flows (see Table 3.2.1.3-2
in Section IV) were used to develop costs for this segment of the PCB user industry.
These are plants 100, 101, 102, 104, 105 and 106. These are said to be 50% of
the capacitor industry. Transformer plants 103 and 114 were used to develop costs
for this industry segment and are said to represent 20% of this category. The
capacitor industry used 22,000,000 Ibs./yr PCBs while the transformer industry
used 12,000,000 Ibs/yr in 1974. These rates are equivalent to 65 and 35% of the
domestic sales , respectively.
224
-------
TABLE 3.6-1
MAXIMUM CAPACITOR INDUSTRY COST
Capital $16,600,000
Liquid Incineration 140,000
Solid Incineration 160,000
Annual Wastewater Treatment 3,900,000
Total Annual Cost $ 4,200,000
TABLE 3.6-2
MAXIMUM TRANSFORMER INDUSTRY COST
Capital $17,300,000
Liquid Incineration 75,000
Solid Incineration 90,000
Annual Wastewater Treatment 4,650,000
Total Annual Cost $ 4,815,000
TABLE 3.6-3
COST FOR MCNSANTO TREATMENT
Capital $ 1,600,000
Annual Incineration 214,000
Solid Incineration 11,000
Annual Wastewater Treatment 370,000
Total Annual Cost $ 595,000
225
-------
TABLE 3.6-4
MAXIMUM PCB INDUSTRY TREATMENT COST
Capital $35,500,000
Annual $ 9,610,000
Cost per Pound PCB Sold $0.28
Rainwater contribution to the capital and annual wastewater treatment
costs was estimated at 15 and 25% respectively. If significant isolation of
the contaminated areas can be achieved this will reduce the costs somewhat.
TABLE 3.6-5
MAXIMUM PCB INDUSTRY TREATMENT COST
NEGLECTING RAINWATER RUNOFF
Capital $31,000,000
Annual Incineration 700,000
Annual Wastewater 7,000,000
Total Annual $ 7,700,000
Cost per Pound PCB Sold $0.22
226
-------
BIBLIOGRAPHY
1. Hager, D. G., Rizzo, J.L., "Removal of Toxic Organics from Wastewater by
Adsorption with Granular Activated Carbon", Paper presented to EPA
Technology Transfer Session on Treatment of Toxic Chemicals, Atlanta, Ga.
(April 19, 1974).
2. Hutzinger, 0., Safe, S. Zitko, V., The Chemistry of PCBs, CRC Press.
3. Lombana, L.A., Napoleon, J.M., "An Experiment on PCB Destruction by Inciner-
ation", (Sept. 27, 1972), BSP Division, Envirotech Systems, Inc., Brisbane,
California.
4. Burns and Roe, Inc., Process Design Manual for SuspendedSolid Removal,
Environmental Protection Agency Technology Transfer, p ll-lO,(Oct. 1971).
5. Hershfield, D.M., Rainfall Frequency Atlas of the United States, Technical
Paper No. 40, U.S. Gov. Printing Office (May, 1961).
6. Daniels, F., Alberty, R.A., Physical Chemistry, John Wiley, and Sons, Inc.,
(April 1963), p 611.
7. Process Design Manual for Carbon Adsorption, U.S.E.P.A., Technology Transfer
(Cot. 1973), Chapter 5.
8. Evans, L., Ozone in Water and Wastewater Treatment, Ann Arbor Science
Publishers, Inc. (1972).
9. Zeff, J.D., UV-OX Process for the Effective Removal of Organics in Waste-
Waters, Presented at 68th Annual Meeting of the AIChE (Nov. 16-20, 1975).
10. Prengle et al, Ozone/UV Process Effective Wastewater Treatment, Hydrocarbon
Processing, pp 82-87 (Oct. 1975).
227
-------
APPENDIX A
Plant 106's Position Statement on
Dow's XFS-4169 Capacitor Dielectric Liquid
We believe that the Dow XFS-4169 liquid will be a viable replacement for
Polychlorinated Biphenyls in capacitors based on limited tests involving all-
film capacitor dielectric systems. These data indicate discharge inception
voltage about equivalent to the PCB all-film system and discharge extinction
voltages again equivalent to PCB systems. Typically, for a given design, these
figures would be:
Discharge Inception Voltage Discharge Extinction Voltage
PCB 2.5kv PCB 2.0kv
XFS-4169 2.75kv XFS-4169 2.0kv
Another critical electrical characteristic of capacitor dielectric sys-
tems is dissipation factor which our tests show typically would be .02% for PCB
film systems and .02% for the XFS-4169 liquid film system.
The dielectric constant of the systems also are just about equal as in-
dicated by our test results.
We have completed a number of small sized test units of the XFS-4169
liquid all-film system which are under test. We are continuing our investiga-
tion by producing additional small units to investigate the statistical perfor-
mance of this dielectric system. The initial results indicate satisfactory
life performance.
We have also begun construction of full sized prototype units and have a
few under test.
At least one year will be required to produce and put into field trial a
statistically significant number of full sized units. At that time, we will
have reasonable verification of the performance of XFS-4169 liquid all-film
capacitor dielectric system.
A-l
-------
More extensive testing was done on Dow XFS-4162. The electrical
characteristics were very similar to XFS-4169.
1. 100-KVAR prototype units have operated at elevated stress for
over two years.
2. Laboratory snail sized units have operated at 200% of rated
design stress for thousands of hours.
3. Long-term tests on small sized units have been conducted at
low and high temperatures and with switching surges.
A-2
-------
APPENDIX B
PCB Adsorption Testing by XADM Resin
Experiments and Results
The apparatus and materials used by Rohm and Haas were:
IWo glass columns - 1/2" in diameter
Adsorbent volume in each column - 50 mis
Adsorbent bed height in columns:
XAD-4 AMBERLITE polymeric adsorbent - 16.5"
Activated carbon (Piltrasorb 300) - 15"
PCB material used - Aroclor 1254 (manufactured by Monsanto)
In the procedure, a feed solution representing a PCB-contaminated waste
stream was made up containing around 160 ppb of PCB. Aroclor 1254 is a very
viscous material, so it was solubilized in methanol before being dispersed in
the water. In order to maintain a constant flow throughout the experimental
run, two batches of feed solution had to be prepared. The composition of each
feed solution is presented in Table 1. Methanol was used to increase the
solubility of PCBs in water.
Table 1
Composition of Feed Solutions
Feed Solution A - 13.25 liter
0.0022 gm of PCB or 166 ppb
2 ml of.methanol or 119 ppm
Feed Solution B - 13.25 liter
0.0029 gm of PCB or 218 ppb
2 ml of methanol or 119 ppm
The influent solution was simultaneously passed through the column of
AMBERLITE polymeric adsorbent and activated carbon at a flow rate of 2 bed
volumes/hour (0.25 gpm/ft resin). Samples were collected from each column
daily; so that each sample nominally represented 48 bed volumes of effluent.
B-l
-------
Liquid passed through both columns for five days, or until 240 bed volumes of
effluent (12 liter) had been collected from each column.
For analysis, the FCBs in each effluent sample were extracted into a
volume of hexane equal to one tenth the volume of the original sample. The
influent material was sampled four times during the run, and likewise extrac-
ted into the same ratioed amount of hexane. This single stage extraction
removed upwards of ninety five percent of the PQBs present in the aqueous
sample. The effluent• samples were then evaporated to 5 mis of hexane, further
concentrating the PCBs present. These PCBs in hexane samples were then sent
to Versar, Inc. for analysis, with the results given in Table 2.
Table 2
Reductions of PCBs (Aroclor 1254) Concentrations Through
Day
1
2
3
4
5
A significant amount of PCBs 'was adsorbed onto the walls of the influent
container. Sample A, for example was prepared with 166 ppb of PCBs. This loss
by adsorption on equipment surfaces has been also detected in other tests, and
must be provided for.
Both AMBEKLITE polymeric adsorbent and activated carbon reduced the con-
centration of PCBs in water to less than 0.05 ppb. The higher leakage of
material for the first day's passage of effluent could be due to an incomplete
conditioning of the resin beads, and some material leaching out of the resin.
The high value of leakage for the last sample from the polymeric adsorbent must
be viewed with some suspicion, especially since it indicates a leakage of PCB
higher than that in the influent.
AMBERLITE
Nominal
Throughput (BV)
1-48
49-96
97-144
145-192
193-240
Polymeric Adsorbents
Influent
Concentration (ppb)
25 (Feed A)
none detected
21.1 (Feed B)
0.69
0.69
and Activated Carbon
Effluent Concentration (ppb)
AMBERLITE Adsorbent
0.246
0.031
0.023
none detected
3.478
Carbon
0.050
0.055
0.025
none detected
0.045
B-2
-------
AMBERLITE polymeric absorbents are usually solvent regenerated. This
is because the energy of adsorption is much higher for activated carbon than
it is for these resins. Hence adsorbed solute can be removed simply by
passing a solvent through the resin whose solvent-adsorbed solute extraction
can overcome the Van der Waals forces. Carbon requires a more energy inten-
sive process of regeneration, such as thermal rejuvenation. Work performed
by Musty and Nickles (J. Chromat., 89:185 (1974)) with PCBs and a solution of
10% diethyl ether in hexane as a regenerant indicates that 76 percent of
adsorbed PCBs on an AMERLITE polymeric adsorbent can be recovered in the mixed
solvent. Undoubtedly, a more efficient solvent could be found that could
quantitatively remove PCBs from polymeric adsorbents. Generally Rohm and Haas
has found that simple alcohols or ketones are effective solvents for these
resins.
The ability to solvent regenerate AMBERLITE polymeric adsorbents in situ
would provide the advantage of generating only the more-readily-handled-by
incineration liquid phase PCBs wastes.
B-3
-------
APPENDIX C
Description of Macroreticular Resins from Rohm and Haas Co.
Ion exchange resins have the capability to selectively recover ionic
constituents, both inorganic and organic, from water by an ion exchange phenon-
enon. Organic conpounds are often exchanged or adsorbed irreversibly onto an
ion exchange resin. This may cause a decrease in capacity so that the operating
life is diminished. The more recently developed macroreticular type of ion
exchange resins are polymeric adsorbents that are used specifically for adsorb-
ing aromatic and aliphatic compounds from water.
Structures
The macroreticular structures are characterized by having unusually
large surface areas as compared with those of conventional gel structures.
Using the newer macroreticular polymerization technique, it is possible
to vary widely the particle pore size, pore size distribution, and surface area.
Polymers with very small pores (50 Angstroms or less) and high surface areas (in
the range of 800 square meters per gram) can be prepared. At the other end of
the spectrum, pore sizes in the range of 300,000 Angstroms, which are visible
under modest magnification, are possible.
The macroreticular polymerization technique is applicable to a wide
variety of monomers. It is possible to introduce functional groups onto the
surface of the pre-formed macroreticular polymers. Thus, a great range of
surface types is possible limited only by the availability of monomers or the
applicability of reactions to introduce functionality. The full line of
macroreticular adsorbents includes a spectrum of surfaces from the least polar
to the most polar. For PCBs removal, the non-polar and intermediate polarity
adsorbents should be best. The chemical structure of AMBERLITE XAD-2 and
AMBERLITE XAD-4 in Figure 1 is representative of the non-polar adsorbent series
Figure 2 shows the acrylic-ester composition of AMBERLITE XAD-7 and AMBERLITE
XAD-8, the intermediate polarity adsorbents.
C-l
-------
...
- CI1 - CH2 - Cli - CHj - C1I
i t
CH2 - CH - CH2 - CH - CH2 - CH
i i
Stvucli'rc of Aml.m-llin XAD-? tm* />™Lnr1:to
Figure 1
C-2
-------
C113 C1I3
t f
- CI!2 - C - Cl!2 - C -
i t
Cf^ f* f\
-0 C~0
I I
0 0
I I
R it
i i
0 0
I I
c=o c=o
I I
.. CH2 - C - CI)2 - C -
i t
CH3 CH3
Structure of Anberlitc XAD-7
CH3
•
- CH2 - C -
i
C = 0
1
0
1
Ri
CH3
i
- CH2 - C
i
C = 0
1
OR'
CH3
i
CH2 - C
i
Cf\
- 0
1
0
1
R
t
0
t
C = 0
1
CH2 - C -
|
CH3
Ci!3
i
CH2 - C -
i
C -
i
0
1
R'
CH3
i
CH2 - G
|
C =
i
OR'
0
0
Structure- of /r.ibcrl i to XAP-.^
Figure 2
03
-------
The physical properties of these AMBEKLITE XAD adsorbents are summarized in
Table 1.
TABLE 1
TYPICAL PROPERTIES OF AMBEKLITE POLYMERIC ADSORBENTS
XAD 1
XAO?
XAD 4
XAD 7
XAD 8
Helium P
Chcmicjl Nature J Volume %
orosi'y
ccAir am
Surf.TT Arcn
nr /'ji.nii
1 SUcI-;..! ! No,,,ma|
Avci.i-jn Porn Din I Ocnv!,- ' f^^,;,
Aivj-itoms Q-am-,/rc 1 Sizes
Nonp:i!'ir
Pcilyvyrcrte
Polysivrne
Polystyrene
37
42
51
Aaylic Eslcr
Acrylic Ester
15
«
OG9
OG9
099
Inter mod
1.08
082
100
330
7bO
,:tn PpLuity
.0
140
200
90
'..0
GO
250
1 OG | 20 in &0
1.00 j 20 lo 00
109 ; 20 to 50
1.25 | 20loLO
1 26 j 75 to SO
Macroreticoalar Adsorption Phenomena
An important aspect of the AMBERLITE adsorbents is the nature of the
different surfaces. The phenomenon of adsorption on solids involves van der
Waals1 forces which bind the adsorbate to a solid surface. Many types of
interaction such as hydrophobic bonding, dipole-dipole interaction and hydrogen
bonding are important. It is not possible to. predict accurately just which
materials will be adsorbed well by a given adsorbent; however, from a practical
point of view, the general concept that hydrophobic or nonpolar molecules or
portions of molecules are attracted to hydrophobic surfcices and hydrophilic or
polar materials to hydrophilic or polar surfaces is a useful concept. Pictorial
views of these interactions are presented in Figure 3. If each organic molecule
is thought of as having both a hydrophobic and a hydrophilic end, then the
hydrophobic end will be attracted to hydrophobic adsorbents such as AMBERLITE
XAD-2 and AMBERLITE XAD-4. This is particularly true when the adsorption takes
place from aqueovs solution.
C-4
-------
SORPTION ON AfiOMATIC SORBLK'IS f ROM POLAR SOLUTIONS
MICUOSPHERC
tn TIVT.NT
f> H A 5 C
SORPTION ON ALIPHATIC SORBENTS FROM POLAR SOLUTIONS
MiCROSPHtrtE
• POLAR SOLVENT
PHASE
SORPTION ON ALIPHATIC SOROilNTS FROM NON - POLAR SOLVENTS
NON I'Ol Alt
SOLVCUT
Figure 3
0-5
-------
For PCBs, it would be expected that the biphenyl portion would be typi-
cally aromatic and hydrophobic, and attracted to the arornatic resin. Increasing
chlorination reduces the water solubility of both benzenes and biphenyls, and
thus would not be contributing anything to what little polar character the PCBs
might have. Thus, the PCBs would not have any "polar end" as shown in the
figure, like a sulfonated aromatic, but would be strongly repelled by the water
phase, and strongly attracted by the resin.
A recent study by Dr. James Fritz, et al, of Iowa State University,
reported a method for extracting trace organic contaminants from water using
macroreticular resins. He also demonstrated the feasibility of selective
desorption using appropriate eluants so that the organics could be identified.
He also established the performance or retention efficiency in isolating these
compounds. A summary of these laboratory results is tabulated below:
Adsorbent: AMBERLITE XAD-2 unless otherwise indicated
Particle Size: 100 - 150 mesh
Flow Rate: 10 BV/hr. (1.25 gpm/cu. ft.)
Retention
Test Compounds Influent Effluent Efficiency, %
Benzene 100 0 100
Benzene sulforic acid 3.0 2.1 31
Phenol 0.4 0.22 45
Phenol (AMBERLITE XAD-7) 0.4 0,06 86
Aniline (AMBERLITE XAD-7) 4.C 0 100
Naphthalene 0.05 0 100
It can be seen that AMBERLITE XAD-2 and AMBERLITE XAD-7 were most effic-
ient for recovering the nonionic organic compounds with 100% efficiency. These
results predict good success with PCBs adsorption. On the other hand, ionic
solutes as well as strongly ionized compounds such as benzene sulfonic acids,
p-toluene sulfcnic acids, etc. were not retained with the same efficiency. It
was noted that retention efficiency increases with increasing molecular weight
in a homologous series. This would indicate that the higher chlorinated PCBs
would be the best adsorbed.
C-6
-------
APPENDIX D
Non-Carbon Adsorbtion and Other Research Stage
_ PCB Treatment Technologies _
1.0 POLYVINYL CHLORIDE AND POLYUKETHANE FOAMS
Dr. John Lawrence and co-workers of Environment Canada have reported on
preliminary tests with these polymers, and with carbon and the XAD resins for
removal of PCBs from synthetic waste water solutions and from actual raw sewage.
H. D. Gesser in Analytical Letters (12:883 (1971)), reported that a poly-
ure thane foam column quantitatively adsorbs PCBs from water. Lawrence found the
carbons, polyurethane foams and the XAD-2 strongly adsorbed PCBs from aqueous
solutions, but were much less effective in raw sewage. He found that thw PVC,
however, was very effective in removing PCBs from raw sewage.
The following were his test procedures and results. Dr. Lawrence is the
only investigator known to have worked with PVC, and the comparative data on the
other materials may be used to relate these studies to those in other sections
of this report.
1.1 Experimental Method for PCBs Adsortion Tests by Environment Canada
The tajo stock solutions of Aroclor 1242 and 1254 were prepared by
vigorously mixing an excess of each Aroclor with water for eight hours, allowing
the solutions to stand overnight and carefully decanting off the true aqueous
phase. The water used was double distilled, the second distillation being from
an all-glass system. The concentration of these solutions determined by gas
chromatography was 45 ± 10 ppb which is consistent with the published solubility
for Aroclor 1254 of 56 ppb.
All solvents used were glass distilled pesticide grade (Caledon
Laboratories Inc.). The activated carbons employed were lignite based hydrodarco
400 (ICI-United States) and anthracite based Filtrasorb 400 (Calgon Corporation) .
These were pretreated by heating to 300°C for 12 hours, cooling and twice extract-
ing each 500 gm with 2 liters of hexane. The extracted carbon was then filtered
D-l
-------
and air dried. The polyurethane foams used were DiSPo plugs (Canlab Supplies
Ltd.) and Foams 1115 and 2328 (B.F. Goodrich Ltd.). The first two digits relate
to the density, i.e., 1.1 and 2.3 Ib/ft and the second two to the hardness.
The foams were shredded and successively washed with n-hexane (several times),
acetone and distilled water. They were then air-dried. The pretreatment was
developed to remove trace organic contaminants from the surface of the foams.
The macroreticular polystyrene resins AMBERLITE XAD-2 and XAD-4 (Rohm and Haas
Company) were pretreated by successively washing each 500 gm of resin with one
liter batches of water, methanol, and water. The cleaned resins were stored in
sealed glass containers under methanol to prevent them from drying out. Poly-
vinyl chloride chips (lybnsanto Company) were washed severed times with n-hexane
and air dried.
To determine the adsorption characteristics, 100-ml aliquots of stock
Aroclor solutions were stirred rapidly with weighed amounts of adsorbent for
30 minutes and the adsorbent removed by filtration through a Millipore prefilter
pad. Five milliliters of n-hexane were then vigorously stirred with the filtrate
for 45 minutes and the organic extract withdrawn. These extracts were analyzed
with a gas chromatograph (Varian series) equipped with an electron-capture
detector (Ni ). The gas column (1.8m x 1.5 mm i.d.) was packed with 4% OV-101
80
and 6% OV-210 on Chromosorb W HP /-\QQ mesh- Nitrogen was used as a carrier
gas at 50 ml/min. The injection port, column and detector temperatures were
250°C, 200°C, and 300°C respectively. Extraction and analysis of water samples
spiked with known amounts of Aroclor indicated that greater than 98% of the PCBs
were detected by this method.
Wastewater was collected from the Hamilton Sewage Treatment plant at
the raw sewage inlet pipe. Sampling was carried out using all glass containers
to insure against adsorption onto container walls. The samples were stored at a
constant temperature of 3°C and in all cases were treated and/or extracted within
24 hours of collection. In the evaluation of PCB adsorption from sewage, the
procedure described previously for pure Aroclor solutions was followed except
that 100 ml samples of raw sewage were stirred vigorously with the adsorbent for
one hour and the adsorbent separated by filtration through a 60 mesh, stainless-
D-2
-------
steel screen. After washing, the screen did not retain any raw sewage and,
with the exception of activated carbon, 100% separation of adsorbent was
achieved. The samples were then twice extracted with 50 ml of n-hexane in 500
ml separatory funnels. The aqueous portion was discarded and the organic phase,
after being dried through 15 g of Na-SO, was reduced to approximately 3 ml using
a rotary evaporator. The sample was purified by liquid-solid chromatography on
a florisil support column using petroleum ether to elute the PCB fraction and
then the eluate was evaporated to 3 ml. Prior to injecting the sample into the
gas chromatograph, it was shaken with 0.2 ml mercury to remove residual sulfur
compounds.
The PCBs in the samples were identified by comparison with chromato-
grams of standard Aroclors. The total concentrations of PCB in raw sewage
usually averaged 9.8 ± 4 ppb.
1.2 Experimental Results of PCBs Adsorption from Sewage and Synthetic
Wastewater
.Adsorption data for Aroclor 1254 and 1242 on PVC, activated lignite
carbon, anthracite carbon, two polyurethane foams, AMBEKLITE XAD-2 and XAD-4 are
shown in log-log form in Figure 1.2-1. The weight of PCB adsorbed per unit
weight of adsorbent is expressed as a function of the equilibrium concentration
of PCB remaining in solution. The sets of data do not follow any of the common
isotherm expressions (e.g., Langmuir, Freundlich, BET), and consequently a
theoretical interpretation of the results has not been attempted. It is evident
that the two carbons and XAD-2 have the greatest adsorption capacities but a
residual concentration of less than 3 ppb could not be obtained with lignite
carbon. Both polyurethane foams appear to be good adsorbers with relatively
high adsorption capacities and low residual levels. DiSPo polyurethane foam
plugs were also evaluated but these had identical adsorption properties to the
Goodrich foam 1115. The lower efficiency of XAD-4 is surprising in view of the
similarity between XAD-2 and XAD-4 (they differ only in pore diameters; 90 A for
XAD-2 and 50 A for XAD-4). The lower efficiency of PVC can be explained by the
lower surface area of this adsorbent. The surface area per unit weight is report-
2 —1 2 —1 2 —1
ed as 500-2000 m gm for carbon; 750 m gm for XAD-4 and 330 m gm for
D-3
-------
UJ
o
to .
Q -1-
I
LLI
CO
o:
o
§ .01
CO
li-
en
.001-
.0001
1
10
A PVC
o LIGNITE CARBON
A GOODRICH 1110 I-OAM
o GOODP.IC!-I 2328 I-OAM
n XAD-4
re XAD-2 •::-
o ANTI-:RACITI-: CARBON
100
1000
PCB REMAINING (ppb)
Figure 1.2-1.Adsorption of Aroclor 1254 on P.V.C., Lignite Carbon,
Anthracite Carbon, Polyurethane Foams and AMBERLITE XAD-2 and XAD-4.
* Indicates adsorption of Aroclor 1242 rather than 1254
D-4
-------
-3 2 -1
XAD-2 but only 2 x 10 m gm , for PVC chips - there being no macroreticular
structure in PVC. This gives an area ratio XAD-2/PVC of approximately 10 .
There are two complicating conditions associated with adsorbing PCBs
from raw sewage rather than from synthetic aqueous solution: (a) sewage con-
tains other hydrophobic organic matter which competes for the active sites on
the adsorbent and (b) much of the PCB has already adsorbed onto the suspended
solids by the time the sewage reaches the treatment plant. Condition (b) can
easily be demonstrated by filtering raw sewage and monitoring the change in PCB
concentration. With typical raw sewage.containing 10 ppb PCB, vacuum filtration
through a Millipore prefilter pad resulted in the removal of about 75%. It is
therefore necessary to find an adsorbent which is not only relatively specific
to PCBs, but which also has sufficient affinity for the PCBs such that the PCB-
suspended-solid equilibrium is reversed.
Table 1.2-1 shows the percentage of PCB (including both Aroclor 1254
and 1260) adsorbed from raw sewage by five different media. With the exception
of the PVC, approximately 1 gm of each media was stirred with 200 ml of raw
sewage for 45 minutes; approximately 10 gm PVC were used because of the lower
surface area. To minimize the inconsistency of raw sewage, the data have been
averaged over several determinations on different days and with different
samples of sewage. The data indicate that PVC and XAD-4 are more effective than
carbon or polyurethane foams in terms of percentage of PCB removed from raw
sewage. This is somewhat surprising since the graphs for PCB adsorption for
pure Aroclor solutions (Figure 1.2-1) would predict the opposite to be true.
Lawrence believes the likely reasons for this apparent anomaly are: (a) the
active sites on carbon are preferentially occupied by hydrophobic species in
sewage other than PCBs and (b) suspended solids adhere to the surface of carbon
and foam acting as a barrier to further adsorption.
The above results indicate that PVC is superior to the other media for
removing PCBs from sewage.
The work at Dr. Lawrence's laboratory is continuing, with studies of
methods of scale-up and continuous or multistage operations with PVC. Optimum
D-5
-------
Table 1.2-1. Adsorption of PCBs from Raw Sewage
ADSORBENT % PCB
ADSORBED*
Lignite Carbon 46
Polyurethane Foam 35
Amberlite XAD-2 23
Arriberlite XAD-4 60
PVC 73
* Includes both Aroclors 1254 and 1260. Data are averaged over
several determinations to minimize the variations in raw sewage.
D-6
-------
retention time, and methods of continuous fresh addition and removal of spent
PVC are under study.
The strong competition between organic solids and non-organic media
is further emphasized in the following section on clays and humus from soils.
1.3 PCBs Msorption on Clays and Soil Organics; Adsorbent Competition
E. S. Tucker and coworkers at Monsanto reported on studies of migra-
tion of PCBs (Aroclor 1016) through various soils as induced by percolating
water. (Bulletin of Environmental Contamination and Toxicology 13 (1):86 (1975)).
This was for application to landfill leaching estimates.
Their tests first led them to believe that the higher the clay content
of a soil, the better its retention of PCBs.
The experimental procedure employed consisted of percolating water
through a column packed with soil coated with Aroclor 1016 and then monitoring
the effluent water for PCBs. The soil columns employed were approximately 3" in
diameter by 12" in height and were dry packed in layers. Each soil layer being
3" thick - first uncoated soil, followed by soil coated with 2.5% (w/w) of
Aroclor 1016, and finally another layer of uncoated soil. An acetone solution
of Aroclor 1016 was used to coat the air dried soil, followed by removal of the
acetone in a rotary evaporator.
Three different types of soils were used in this study. The character-
istics of each are shown in Table 1.3-1.
The intent was to simulate the various soil types which could be en-
countered at different landfill sites. The soils and the procedure employed have
been used to evaluate the soil mobility of agricultural chemicals.
Distilled water was fed from the reservoir at a constant pressure to
each soil column. The flow rates were observed to increase, the first few days,
and then decrease and level out. Apparently, after the wetting phase some
channeling occurs until the soil becomes compressed in the column. This effect
was most pronounced with the silty soils. The average flow rates in liters/day
D-7
-------
TABLE 1.3-1
COMPOSITION OF SOILS USED IN STUDY
Soil
% Sand
% Silt
% Clay
% Organic Carbon
Norfolk
Sandy Loam
82.5
11.0
5.5
1.0
Ray
Silty Loam
6.2
83.2
9.6
1.0
Druittner
Silty Clay Loam
2.8
55.4
35.8
6.0
D-8
-------
for Norfolk Sandy Loam, Ray Silty Loam and Drummer Silty Clay were 0.26, 0.53
and 0.32, respectively.
The effluents were then quantitatively adsorbed on polyurethane
columns, and then extracted and analyzed by electron capture gas chromatography.
1.3.1 PCBs Adsorption Results
PCBs adsorption results are given.in Table 1.3.1-1. These
results seemed to indicate that clay was the strong adsorbent for PCBs. This
would have been in agreement with the work of R. Hague and co-workers (Environ-
mental Service Technology, 8:139 (1974)), in which clay was found to have a
high affinity for PCBs.
However, in later work with pure clays, the retention was not
found to be as good as expected. Attention was then directed to the organic
portion of the soils. They have tentatively decided that the organic fraction
is more important to the adsorption of PCBs, and it can be seen from the data
that the high-clay-content soil was also the high-organic-content soil.
Further investigation of some clays might be warranted. Dr. S.
Pearson of Hercules, Inc., in a personal communication stated that bentonite clay
was very effective in removing other pesticide wastes from water. The finely
divided clay was then rapidly removed from the water with "Hereofloc".
1.4 Sphagnum Peat (Lignin-Cellulose) or an Adsorbent
The above work by Tucker, and the general finding that sewage sludge
solids contain many times the amount of PCBs as that in the water phase they
contain, leads to the conclusion that natural organic materials from the earth
might make good adsorbents for PCBs.
Although it has not yet been tested on PCBs, there is a commercial
method of continuous wastewater treatment that makes use of sphagnum peat. It
is the Hussong/Couplan Water Treatment System. The peat is formed into a contin-
uous mat on a mesh belt through which the water is sprayed. The system has shown
high effectiveness in removing certain organics and metals from wastewaters.
D-9
-------
TABLE 1.3.1-1
PCBs FOUND IN PERCOLATING WATER
Norfolk
Sandy Loam
Total
Effluent
Volume (£)
1.3-8.1
10.1
13.5
25.5
48.1
ppb
PCBs
ND
ND
23
63
63
Ray
Silty Loam
Total
Effluent
Volume (£)
2.7-16.4
20.7
27.6
51.9
98.1
ppb
PCBs
ND
65
92
153
136
Drummer
Silty Clay Loam
Total
Effluent
Volume (£)
1.6-9.9
12.5
16.6
31.4
59.2
ppb
PCBs
ND
ND
ND
ND
ND
ND = None detected, < 1 ppb
D-10
-------
Capital exists for a system average about 60£ per gallon of daily
capacity. Operating costs when treating a dye house effluent were 7C to 14% per
1000 gallons.
2.0 CATALYTIC REDUCTIOS
In the review of PCB literature two approaches were found to modifications
of chemical structure that could aid in waste control. One approach was the
complete chlorination of PCBs, to give decachlorobiphenyl. The rationale here
was that the completely chlorinated biphenyl would have the least solubility in
water of any PCB, and thus would be much easier to adsorb and remove. However,
this is in the direction of creating the most refractory PCBs also.
Another approach, presenting more favorable aspects, is the dechlorination
of PCBs to give biphenyl or bicyclohexyl. Berg, et al (in the Bulletin of
Environmental Contamination Toxicology 7:338 (1972)) state that PCBs can be de-
chlorinated quantitatively with hydrogen over platinum or palladium catalysts to
give one product: bicyclohexyl.
T. Sawai of Japan, in Genshiryohu Kogyo 18(12):43-7 (1972), reported the
degradation of PCBs using the cobalt 60 isotope. Gamma irradiation at a level
18
of 10 ev/gram produced chain dechlorination in an alkaline propanol solution
saturated with nitrogen. Alkaline concentrations of 0.01 molar gave about 40
times the dechlorination that a neutral solution gave.
A process of reductive dechlorination, more amenable to low cost operation
and commercial scaleup, is being developed by Sweeny, Saltonstall and co-workers
at Envirogenics Systems Co. of El Monte, California.
2.1 Reductive Dechlorination of PCBs at Envirogenics Systems Corp.
Envirogenics, originally working in chlorinated pesticides, has develop-
ed catalyzed reduction process methods for the following compounds:
DDT aldrin Aroclor 1221
ODD CHLORDANE Aroclor 1242
Kelthane dieldrin Aroclor 1254
Perthane endrin Aroclor 1232
Methoxychlor heptachlor Aroclor 1248
Lindane toxaphene Aroclor 1260
D-ll
-------
The reductive dechlorination reaction has been run at ambient tempera-
ture and pressure, by flowing the liquid chlorinated hydrocarbon through a
column containing metallic iron granules with a metallic copper coating, blended
with sand. The copper exerts a catalytic action. pH is maintained at nearly
neutral, since with low pH, say 2.0, the iron conversion rate to ferrous
chloride goes up 10 times, but the reaction is not that much faster. That is,
the effluent FeCl? content goes up to 50 ppm from 5ppm. The chlorinated hydro-
carbon is converted to a hydrocarbon.
PCBs, in this process, seemed to lose chlorines stepwise, leaving un-
identified halogens.
Envirogenics is now under EPA contract #68-03-2364, to develop and
demonstrate at the bench-scale level (1-3 gpm), an effective, low-cost/ practical
process for the treatment of dilute (ppb to 1 ppm) aqueous manufacturing and
processing wastes containing polychlorinated biphenyls.
Specific objectives and guidelines include:
1. Reduction of PCBs to 5 ppb or less, with levels of < 1 ppb
desirable.
2. Sufficiently low projected treatment cost to be economically
attractive.
3. Low expected toxicity of all effluent products, including both
degraded PCBs and any added reagents.
4. Use of readily available materials, both in construction and as
reagents.
The process is simply pumping the liquid to be treated through the
catalytic column. Process characteristics that might be proposed for the scaled-
up PCB process, based on what has been learned about the other pesticides thus
far, are:
catalyst; 100 mesh metallic iron granules coated with 0.1 milli-
equivalents of copper per gram of iron
support; 60 to 100 mesh (approx) sand
number of beds; 4, operable in series or parallel
bed composition: 500 pounds iron catalyst, plus 3500 pounds sand
-------
flow rates: 2 to 10 gpm per square foot
pressure drop: about 10 psi
pH; Kept neutral through caustic
piping; 304 ss, with teflon tape seals
tanks: Steel with epoxy coating inside
pretreatrnent; Wastes are assumed to be free of undissolved solids and
oils; They are fabric filtered before entering process
The expectations for PCB reductions in wastewater with a pilot plant
like above, are based on small one- to two-inch diameter column lab tests with
PCBs that gave reductions from 50 ppb to less than 0.1 ppb.
The capital cost of the equipment to process 100 gpm is estimated at
about $65,000, plus tank storage and erection costs. Operating costs, including
amortization of capital, is estimated at 72£ per 1000 gallons of effluent.
3.0 CATALYTIC OXIDATION AND MISCELLANEOUS REACTIONS
This category of PCBs destruction includes oxidation by: air, oxygen,
ozone, hydrogen peroxide and chlorine dioxide; usually assisted by catalysts or
reaction sensitizers.
The PCBs are very resistant to chemical attack. Monsanto states that they
are not affected by boiling sodium hydroxide, or by long contact, say 10 days,
with concentrated sulfuric acid at ambient temperature. There is no apparent
reaction in a bomb at 250 psi of oxygen and 140°C.
3.1 Strong Acids
Russian workers have reported nitric acid decomposition of PCBs. They
used nitric acid of density = 1.4, and refluxed two PCBs, a pentachloro and a
heptachloro homolog for periods ranging up to 100 hours. They found dichloro and
trichloro benzoic acids from the former; and trichloro and tetrachloro benzoic
acids from the latter (heptachloro). Less concentrated nitric acid would not
oxidize these compounds, nor would potassium permanganate or chromic acid. How-
ever, mono-, di- and tri-chlorobiphenyls can be oxidized to the corresponding
chloro-benzoic acids with chromic anhydride and acetic acid.
D-13
-------
The less rigorous conditions can also produce a large mixture of
nitrated chlorobiphenyls.
3.2 Electrochemical Oxidation
Notwithstanding these examples of oxidative resistance, J. D. Stuart
and co-workers at the Dept. of Chemistry of University of Connecticut have con-
ducted laboratory electrochemical oxidations of PCBs at ambient temperature and
pressure. They oxidized PCBs with one to ten chlorine atoms, using very high
anodic potentials in a dry methyl cyanide solvent. They believed that a series
of reactions occurred, starting with hydrolysis from traces of water present,
followed by oxidation.
A number of other catalytic oxidation investigations of refractory
organics are being conducted throughout the U.S., some being sponsored by
branches of EPA. Although PCBs have not been tested in these cases yet, the
methods have sufficient power and flexibility to predict that PCBs could be
destroyed. In addition, these are methods that have the potential of zero dis-
charge since they have converted refractory organics to C00 and water.
3.3 Catalytic Ozonation and Ultrasound Decomposition of Organics
Prof. Gerard Smith and J. W. Chew of Southern Illinois University pre-
sented results of catalytic ozonations of non-PCB organics in aqueous systems at
the 68th annual AIChE Meeting in November 1975. They tried catalyst with oxygen,
and ozonation without catalyst, and neither method had the effectiveness of
catalytic ozonation. An Fe^Oo catalyst was used, and phenol and ethyl aceto-
acetate were used as model compounds in water. With a liquid retention time in
their flow reactor system of 25 minutes, and a gas flow of 0.1 liter/minute of
30 mg/liter concentration of ozone, % TOC removed was 95% for an initial TOC of
100 mg/1 and about 85% for an initial TOC of 400 mg/1. The packed reactor had
the aqueous solution flowing down, and the ozone flowing up.
In experiments comparing the effectiveness against refractory organics
of ozone plus Raney nickel and ozone plus ultrasound, Prof. Smith found that the
reactions were similar. Also, he found that ultrasound and oxygen gave similar
D-14
-------
reactions. It appeared, however, that all of these effects could not be added
together to get greater oxidative reaction. For example, adding ultrasound to
an ozone catalytic reaction did not materially change the reactivity of just
ozone and the catalyst with the organics. Ultrasound was applied at 800 KHz,
and 4 to 5 watts/sq. on. He also found that at these ambient temperature and
pressure conditions, although phenol disappeared rapidly, other organics formed
were oxygenated aromatics rather than those formed from rupture of the aromatic
ring.
3.4 Wet Catalytic Oxidation of PCBs
Dr. L. W. RDSS at Denver Research Institute has studied the wet
catalytic oxidation of strong wastewaters having GOD values of 3,000 to 15,000.
Tests showed that Fe2 (SO.)- - Cu SO. - H202 at pH 5.0 and Fenton catalyst at
pH 2.5 gave COD reductions of 95.4% and 96.0% respectively after 2 hours at
400 F. Platinum oxide as a solid catalyst also gave good results. Most of the
organics were cellubiosics.
3.5 Wet Catalytic Oxidation and Catalyst Durability
Prof. J. F. Katzer and co-workers at University of Delaware have been
studying the elevated pressure and elevated temperature catalytic air-oxidation
of refractory organics. They are currently being supported by EPA in a study of
the durability of catalysts in wastewater environments. In a paper soon to be
published in the Journal of the Water Pollution Control Federation, Katzer re-
ports on studies of the complete oxidation of phenol to CO- and water in an
aqueous medium using a supported copper oxide catalyst. Rapid degradation rates
were found, and the rate data were used to run preliminary design and costs for
commercial-sized waste treatment plants. His conclusion is that catalytic oxi-
dation of wastewaters is cost competitive with other physical chemical treatment
techniques.
Katzer found that pressures of 10 to 20 atm, and temperatures of 100 to
200 C were required to get adequate reaction rates and give up to 99% destruction
of organics to GO- and water. It was found that ambient temperature and pressure
generally caused more adsorption on a variety of catalysts, than any reaction.
D-15
-------
3.6 Dye-Sensitized Visible Light Photo-Oxidation of Organics
Prof. R. L. Sanks and co-workers at the Civil Engineering Dept. of
Montana State University are working on the dye-sensitized, aerobic, photo-
oxidation of refractory organics. They were able to rapidly break apart the
benzene ring of cresol using dyes like methylene blue or bengal rose. The key
to the process is that in the presence of sunlight and air, methylene blue can
produce singlet oxygen from the 0~ in air. There are two forms of this singlet
oxygen, with a half life of only a millisecond, but with the ability to easily
shatter benzene rings. Prof. Sanks visualizes a process whereby a lagoon con-
taining wastewater with refractory organics, could have the dye molecules
attached to a long chain alkyl or a floe of some kind to keep it at the surface.
The sun would provide the photolysis, and little dye would be consumed as it is
a catalyst in the reaction. Such a process, if it can be developed, suggests a
route to a very low cost method of obtaining zero discharge, and water recycle
systems.
3.7 Chlorine Dioxide Oxidation
International Dioxide, Inc. of New York City is offering a stabilized
chlorine dioxide. It is being used by municipalities for taste and odor control
of water. It is used as an adjunct to chlorination and can decompose chlora-
mines. It oxidizes and destroys phenol and does not chlorinate. Unstabilized
ClO? is a dangerous explosive, thus this stabilization offers a very powerful
oxidizing agent for potential degradation of organics.
4.0 DESTRUCTION OF PCBs BY MICROORGANISMS
"The Chemistry of PCBs", by Hutzinger reports only two pure cultures of
micro organisms that had shown metabolic activity on individual chlorobiphenyls
up until the end of 1973. One culture, Rhizopus japonicus only converted the
mono-or dichlorobiphenyls to chlorohydroxy biphenyls or possibly multi-hydroxy-
biphenyls.
More promising results were obtained with two species of Achromobacter,
isolated from sewage effluent. Under aerobic conditions, 4-Chlorobiphenyl was
D-16
-------
converted to 4-chlorobenzoic acid. It appeared that the organism was able to
attack the non-chlorinated benzene ring. Most of these degradations take quite
a number of hours to occur. More recently, up to pentachloro-biphenyls have
been oxidized by Achroroo bacter. Monsanto was able to demonstrate a signifi-
cant reduction in the one, two and three chlorobiphenyls of Aroclor 1242 after
72 hours of treatment with activated sludge. The higher chlorinated honologs
did not seem to be affected.
The degradation of PCBs by microbial action from sludges and pure cultures
seemed to give results similar to animal metabolism studies. The lower chlorin-
ated species, again up to about the trichloro hcmolog were decomposed to
phenols, catechols and related compounds.
The importance of oxidative conditions was demonstrated in tests with
silage containing PCBs. After several months of storage of Aroclor 1242 with
silage that had undergone fermentation reactions under anerobic conditions, no
change in any of the PCBs had taken place.
The use of DDT degradation as a model for PCB degradation conditions may
not be too helpful. Results of a study by Johnsen (PCB Newsletter, January,
1973) showed that di- to hexachloro biphenyls were incubated for one month with
soil and soil containing cattle manure. No indication of PCB metabolism was
found. Under these conditions; p,p' DDT degrades almost completely.
It has also been found that biphenyl itself is more easily degraded or
hydroxylated, than any of the chlorinated biphenyls.
5.0 REVERSE QSMDSIS AND ULTRAFILTRATION
No tests for PCBs removal by these technologies were found in our survey.
Generally, these methods have found most success in removing dissolved
salts from water. However, DuPont reports in a private communication that they
have achieved 90% removal (rejection) of organics in water at the 1000 to 2000
ppm level. They have developed the hollow fiber permeator system, and the
fibers are aromatic polyamides. The system is reported to work well on organics
D-17
-------
with molecular weights greater than 100. This would predict good success with
the PCBs, all of which have molecular weights about 2 to 4 tiroes that high.
U.O.P.'s Fluid Systems Division reports in a private cxanmunication, that
based on their success with wound experimental polyamide membranes operating on
chlorinated hydrocarbons in water, they would predict a 95% to 97% removal of
PCBs at the 50 ppb level in water. They point out however, that the reject
water stream with concentrated PCBs would contain 10% to 20% of the original
volume of wastewater. A certain amount of recycle could be built into the sys-
tem to reduce the generation of concentrated wastewater. R.,0. units require no
suspended solids in the feed wastewater. They can operate for 6 months to one
year before any maintenance for cleaning is required.
Dr. W. L. Short of the Chemical Engineering Dept. of the University of Mass.
reports in a private communication that ultrafiltration has achieved 50% rejec-
tions of phenols and chlorinated phenols. He believes that rejection of a given
compound can be improved by attaching some of that compound, or a similar
material, to the membrane to act as a sort of electronic barrier.
D-18
-------
APPENDIX E
Energy Requirement for Various Treatment Options
The attached Tables 1 through 6 present the estimated annual fuel oil
requirement in therms (100,000 BTUs) and power requirements in kilowatt-hours
for the various treatment options included in the Task II report. This infor-
mation is provided in order to facilitate the assessment of the energy impact
of the various levels of effluent treatment.
The estimated total PCB user industries', (capacitor and transformer
industries) energy requirements were facilitated by preparing energy require-
ments for each plant site visited during the course of these studies and then
aggregating to the total industry category on the basis of the percentage of
the total PCB industry represented by the visited plants.
The capacitor plants visited are Plants 100, 101, 102, 104, 105 and 106
which represent 50 percent of the PCB user capacitor industry. Transformer
plants visited are Plants 103 and 114 and they represent 20 percent of PCB
user transformer industry.
E-l
-------
TABLE 1
Base Case (No Wastewater Treatment)
Estimated Annual Energy Requirement for
The Incineration of Liquid and Solids Wastes
PLANT NO. 100 101 102 103 104 105 106 114
Fuel Oil
Requirement,
Therms 5300 1800 2900 3100 7300 14,500 1800 2800
Estimated PCB User Industry Fuel Oil Requirement, Therms 100,000
E-2
-------
TABLE 2
Estimated Annual Energy Requirements for
Treatment of All Flows and Discharging
to Receiving Stream with Carbon
Final Treatment1
PLANT NO. 100 101 102 103 104 105 106 114
Power Requirement,
million Kw hr 1.25 0.14 1.13 3.35 1.60 1.13 0.83 0.20
Fuel Requirement,2
Therms 5300 1800 2900 3100 7300 14,500 1800 2800
Estimated PCB User Industry Power Requirement, million Kw-hr 30
Estimated PCB User Industry Fuel Oil Required, Therms 100,000
Notes; ^osts for this treatment option are presented in Section VII,
Table 3.3.3-1, page 210.
2Fuel requirement is same as the base case.
E-3
-------
TABLE 3
Estimated Annual Energy Requirements for Recirculation
and Cooling of Non-Contact Cooling Water and
Treatment of Ml Other Plows Using Carbon as
Final Treatment1
PLANT NO. 100 101 102 103 104 105 106 114
Power Requirement,
million Kw hrs 1.80 0.14 1.30 4.70 2.30 1.60 1.10 0.24
Power Requirement2
neglecting rainwater
million Kw hrs 1.70 0.10 0.72 4.20 2.10 1.40 0.87 0.21
Fuel Requirement,3
Therms 5300 1800 2900 3100 7300 14,500 1800 2800
Estimated PCBs User Industry Power Requirement, million Kw-hr 40
Estimated PCBs User Industry Fuel Oil Requirement, million therms 36
Notes; ^osts for this treatment option are presented in Section VII,
Table 3.3.3-2, page 211.
2This information is included for purposes of determining the
rainwater contribution to the power requirement.
3Fuel requirement same as the base case.
E-4
-------
TABLE 4
Estimated Annual Energy Requirements for Zero Discharge
(Carbon Treatment of Non-Contact Cooling Water)L
PLANT NO. 100 101 102 103 104 105 106 114
Power Requirement,
million Kw hr 1.40 0.11 0.63 3.30 1-70 1.10 0.74 0.18
Fuel for incineration,2
million therms 0.925 1.841 24.903 12.003 2.207 4.415 7.801 0.923
Estimated PCBs User Industry Power Requirement, million Kw-hr 28
Estimated PCBs User Industry Fuel Oil Requirement, million therms 148
Notes; lCosts for this treatment option are presented in Section VII,
Table 3.3.4.5-1, page 219.
2Includes the fuel requirement for solid and liquid PCBs waste
incineration (base case fuel requirement).
E-5
-------
TABLE 5
Estimated Annual Enerqy Recruirements for Zero Discharge
(Segregation and Recycle of Non-Contact Cooling Water)l
PLANT NO. 100 101 102 103 104 105 106 114
Fewer Requirement,
million Kw hr 1.80 0.12 1.10 4.50 2.30 1.50 1.00 0.22
Fuel Oil Requirement,2
million therms 0.925 1.841 24.903 12.003 2.207 4.415 7.801 0.923
Estimated PCBs User Industry Power Requirement, million Kw~hr 40
Estimated PCBs User Industry Fuel Oil Requirement, million therms 148
Notes: 1 Costs for this treatment option are presented in Section VII,
Table 3.3.4.5-2, page 220.
2 Includes the fuel requirement for solid and liquid PCBs waste
incineration (base case fuel requirement).
E-6
-------
TABLE 6
Estimated Annual Energy Requirements for Zero Discharge
(Segregation and Recycle of Non-Contact Cooling Water)
Excluding Rainwater Runoff1
PLANT NO. 100 101 102 103 104 105 106 114
Power Requirement
million Kw hr 1.80 0.12 1.10 4.50 2.30 1.50 1.00 0.22
Fuel Oil Requirement,2
therms 5300 1800 2900 3100 7300 14,500 1800 2800
Estimated PCBs User Industry Power Requirement, million Kw-hr 40
Estimated PCBs User Industry Fuel Oil Requirement, therms 100,000
Notes; 1Costs for this treatment option are presented in Section VII,
Table 3.3.4.5-3, page 211.
2Sauce as the base case.
E-7
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1 REPORT NO
3. RECIPIENT'S ACCESSION-NO.
TITLE AND SUBTITLE Assessment of Wastewater Management,
Treatment Technology, and Associated Costs for Abatement
of PCBs Concentrations in Industrial Effluents
6. PERFORMING ORGANIZATION CODE
REPORT DATE ---,
January 30, 1976
7 AUTHOR(S)
Gayaneh Centos, Robert L. Durfee and E. E. Hackman, III
(Versar Inc.) and Kenneth Price (Clark, Dietz & Assoc.)
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Versar Inc.
6621 Electronic Drive
Springfield, Va. 22151
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
68-01-3259
12. SPONSORING AGENCY NAME AND ADDRESS
U.S. Environmental Protection Agency
Office of Toxic Substances
Washington, D.C. 20460
13. TYPE OF REPORT AND PERIOD COVERED
Final Report, Task III
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT
This document p> rs.ents this findings oi a study of available wastorat?r nioriqr.-
itt'nt and treats-tnt technology for the purpose of determining toxic polluunt effluents
concentrations and daily load achievable ill three industrial categories: polychlorinated
biphenyls (PCBs) rr,.Tnufacairing; capacitor nor.afacturing; and transformer Tar.ufacturing.
All plants in the above categories have PCB discharges to either waterways or
sewagp treatment plants, under normal operating conditions. All plants have dis-
charges to storm Covers or directly to waterways under heavy rainfall conditions.
Extensive survey of wastewater treatment technologies and cooperative laboratory
nork with several suppliers of treatment equipment and research facilities has con-
firmed that carbon adsorption technology is the best current candidate for s^ccess-
ful rerroval of PC3L frcsn the vastewaters. Ap an alternative lF"-o--onatior was con-
sidered. Ohis techrjo]og}f is still in the research stage; /icv^ver, it offers potential
of complete deFtriictioii cf KBs all the way to CX>2, water and HC1.
Arothcr aclsorberit t<5chnology new in the development stage, AHBERLITE polymeric
adsorbents, has derronstrated a PCBs remDval efficiency roughly equivalent to carbon
during laboratory tests. Further testing is needed with this adsorbent to accurately
assess its potentiality.
For scrap oils and burnable solid wastes generated at these plants, high temper-
ature, controlled incineration offers a straightforward rethod of destruction, where-
as scientific landfilling appears to be the best suited mode of disposal for non-
burnable contaminated solids.
Zero discharge objectives can be best achieved by eliminatinq discharge streams
and developing recycle systems. All non-contact cooling water would be seoregatod,
cooled, and recycled. All other Wastewater streams would be pretreatod. "Ine portion
of the pretreated water which would be used in the plnnt wold be treatod w-th carDon,
while th3 excess water would be incinerated in a specially designed systtn which would
allo*- for energy recovery.
STju'^rrinq data, rationale for tho selection of above recommended treatment
technologic": aid asc-ociateJ costs are coi.tamod in UTIS, report.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
x>lychlorinated biphenyls (PCBs) carbon adsorption
PCBs Manufacturer-MDnsanto UV-ozonation
ielected PCBs use-capacitors & transformers levels of treatment
PCBs water pollution costs of treatment
PCBs in land destined wastes
PCBs incineration
wastewater treatment technologies
c COSATI Field/Group
3 DISTRIBUTION STATEMENT
telease unlimited
19 SECURITY CLASS (This Report}
unclassified
21 NO. OF PAGES
281
20,SECU3'TY-CLASa (This page)
D SECURITY.CLASS,
unclassified
EPA Form 2220-1 (9-73)
-------
-------
m
D
C
<
3J
o
-n
O
3J
H
C
00
i
m
S
TJ
o
m
33
N ° 33
^ I. S
-*• c
8 .3
0) ro
Q.
Q. n;
- ^^
Z O
i*
fro
Q- H
&
8 8
D CO
O I!.
E 5L
ID
o
<
3J TJ
O O
Z ^
z
H
CO
CO
Ul
s
H
z
o
------- |