Un,;ed States Industrial Environmental Research EPA-600/2-79-104
Environmental Protection Laboratory May 1979
Agency Research Triangle Park NC 27711
Research and Development
Symposium Proceedings:
Textile Industry
Technology
(December 1978,
Williamsburg, VA)
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the ENVIRONMENTAL PROTECTION TECH-
NOLOGY series. This series describes research performed to develop and dem-
onstrate instrumentation, equipment, and methodology to repair or prevent en-
vironmental degradation from point and non-point sources of pollution. This work
provides the new or improved technology required for the control and treatment
of pollution sources to meet environmental quality standards.
EPA REVIEW NOTICE
This report has been reviewed by the U.S. Environmental Protection Agency, and
approved for publication. Approval does not signify that the contents necessarily
reflect the views and policy of the Agency, nor does mention of trade names or
commercial products constitute endorsement or recommendation for use.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/2-79-104
May 1979
Symposium Proceedings: Textile
Industry Technology (December 1978,
Williamsburg, VA)
Frank A. Ayers (Compiler)
Research Triangle Institute
P. 0. Box 12194
Research Triangle Park, N. C. 27709
Contract No. 68-02-2612
Task No. 37
Program Element Nos. 1AB604and 1BB610
EPA Project Officer: Max Samfield
Industrial Environmental Research Laboratory
Office of Energy, Minerals, and Industry
Research Triangle Park, NC 27711
Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Research and Development
Washington, DC 20460
-------
ABSTRACT
The proceedings for the "Textile Industry Technology" symposium
constitute the final report submitted to the Industrial Environmen-
tal Research Laboratory. The symposium was held at
Williamsburg, Virginia, December 5-8,1978.
The purpose of the symposium was to bring about an exchange of
ideas in three interrelated areas—emission control, energy conser-
vation, and material recovery. Also a session on assessment
methodology was presented. This was the first symposium on tex-
tile industry technology sponsored solely by the United States
Environmental Protection Agency. This meeting was EPA's most
extensive technology transfer activity in the textile area, an
activity dedicated to support the attainment of pollution control and
energy and material conservation.
11
-------
5 December 1978
CONTENTS
Page
Keynote Address 1
Stephen J. Gage
Session I: IMPACT OF LEGISLATION ON POLLUTION
CONTROL IN THE TEXTILE INDUSTRY
Max Samfield, Session Chairman
Air Pollution Control in the Textile Industry
Donald F. Walters1"
Water Pollution Control in the Textile Industry 7
Robert B. Schaffer
RCRA and the Textile Industry 11
Harry W. Trask
The Textile Industry Looks at Pollution Control
Don K. Hill1"
Session II: WASTEWATER POLLUTION CONTROL TECHNOLOGY
W. Wesley Eckenfelder, Session Chairman
Origin and Internal Control Options for Priority Pollutants in the Textile Industry 17
Edwin L. Barnhart
Contact Filtration for Textile Wastewater Pollution Control in Poland 23
Jerzy Kurbiel* and
Wieslaw Zymon
Use of Powdered Activated Carbon for Textile Wastewater Pollution Control 37
David G. Button* and
Francis L. Robertaccio
Activated Carbon Adsorption for Textile Wastewater Pollution Control 51
E. J. Schroeder* and
A. W. Loven
•Indicates speakers
'Indicates papers presented at the
conference but not submitted for publication
iii
-------
CONTENTS
Page
Biological Treatment of Textile Waste by Aeration 61
T. A. Alspaugh,
John Hodges, and
Arthur Toompas*
The Use of the Coagulation-Clarification Process in the Treatment
of Textile Wastewaters
John H. Koon*
6 December 1978
Session II: WASTEWATER POLLUTION CONTROL TECHNOLOGY (continued)
A Fundamental Approach to Application of Reverse Osmosis for Water
Pollution Control 73
S. Sourirajan* and
Takeshi Matsuura
Hyperfiltration of Nonelectrolytes: Dependence of Rejection on Solubility
Parameters 107
H. G. Spencer* and
J. L. Gaddis
Separation of Toxic Substances by Hyperfiltration 115
J. Leo Gaddis* and
H. Garth Spencer
Session III: AIR POLLUTION CONTROL TECHNOLOGY
Donald F. Walters, Session Chairman
Incineration and Heat Recovery in the Textile Industry 123
William H. Hebrank
Fixed Bed Carbon Adsorption for Control of Organic Emissions 129
W- Macon Sheppard
Fluidized Bed Activated Carbon Adsorption 141
Ram Chandrasekhar* and
Carmen M. Yon
Catalytic Oxidation of Hydrocarbon Fumes 153
Richard E. Githens* and
Donald M. Sowards
Radiation Curing for Textile Coating 163
William K. Walsh* and
B. S. Gupta
iv
-------
CONTENTS
Page
Control of Airborne Particulates in Textile Plants 171
Mansour H. Mohamed* and
Arthur C. Bullerwell
Session IV: SOLID WASTE POLLUTION CONTROL TECHNOLOGY
P. Aarne Vesilind, Session Chairman
Textile Sludge Characterization 177
Robert G. Shaver* and
Deborah K. Guinan
Dewatering of Waste Activated Sludge 183
Eugene Donovan. Jir.
7 December 1978
Session IV: SOLID WASTE POLLUTION CONTROL TECHNOLOGY (continued)
Characteristics of Activated Sludge Affecting Solid-Liquid Separation
Joseph A. FitzPatrick1"
Rationale for Sludge Management in a Wastewater Treatment Plant 197
P. Aiarne Vesilind
Textile Sludge Treatment and Disposal 201
David H. Bauer,*
John .P. Woodyard, and
Stephen P. Shelton
Session V: ASSESSMENT METHODOLOGY
Dale A. Denny, Session Chairman
Level 1 Measurement Procedures for Effluent Characterization 213
R. G. Merrill, Jr.,*
L. D. Johnson, and
J. A. Dorsey
Short-Term Biotesting of Environmental Effluents 225
Shahbeg S. Sandhu
The Testing of Environmental Samples for Mutagenicity and Carcinogenicity
Using Microbial Assay Systems 231
Larry D. Claxton,*
Joellen Huisingh, and
Michael Waters
-------
CONTENTS
Page
Cellular Toxicity of Liquid Effluents from Textile Mills 239
James A. Campbell,
Neil E. Garrett,*
Joellen L. Huisingh, and
Michael D. Water
Repeated-Exposure Toxicology Studies 249
Robert J. Weir
(Presented by Ross Hart)
Assessment of Textile Waste Toxicity with Marine Bioassays 251
Gerald E. Walsh* and
Lowell H. Bahner
Freshwater Toxicity Tests Related to the Textile Industry 273
William B. Horning, II* and
Timothy W. Neiheisel
Effluent Guidelines Division Procedures for Measuring Priority Pollutants 281
Roger D. Holm
Organic Sampling and Analysis for Environmental Assessment -/... 297
Philip L. Levins and
Judith C. Harris*
Measurement of Color in Textile Dyeing Wastewaters 307
Linda W. Little
Use of Bioassay Screening Techniques for NPDES Permits 311
W. H. Peltier* and
L. B. Tebo, Jr.
8 December 1978
Session VI: ENERGY AND MATERIALS CONSERVATION
John R. Rossmeissl, Session Chairman
Reclamation of Warp Sizes Using Thermal Precipitation 317
Warren S. Perkins,*
Robert P. Walker, and .
Leo J. Hirth
An Industrial Wastewater Recirculation System for the Fibrous Glass
Textile Industry 323
S. H. Thomas* and
D. R. Walch
VI
-------
CONTENTS
Page
Optimization of Hyperfiltration System for Renovation of Textile Wastewaters 331
S. M. Ko* and
P. G. Grodzka
Solvent Recovery from Textile Dryers 345
Nathan R. Shaw
Solvent Dyeing of Nylon Carpet Using the STX Solvent Dyeing Process 347
Charles C. Wommack* and
M. Pierre Favier
Reconstitution and Reuse of Dyebaths 363
Wayne C. Tincher,*
Fred L. Cook,
L. Howard Olson, and
W. W. Carr
vii
-------
KEYNOTE ADDRESS
Stephen J. Gage*
When I was sent an advance copy of the pro-
gram for this symposium, I was pleased for
many reasons. First of all, I noted that this is a
"multimedia" conference indicating that we in
EPA and, hopefully, the textile industry as
well, recognize that potential pollution prob-
lems exist in the air, in the water, and on land.
The interrelationship among these three media
must be recognized in solving all pollution
problems. For example, one may solve an air
pollution problem only to discover that a more
severe water pollution problem has been cre-
ated in the process. Also, we have found in-
stances where toxic materials improperly dis-
posed of in landfills have leached out and
polluted local ground waters.
Second, I was pleased to see that govern-
ment agencies other than EPA's Office of
Research and Development are represented:
The Department of Energy has a definite role
to play, indicating recognition of the need for
energy conservation along with requirements
for adequate pollution control. Similarly,
EPA's Office of Air and Waste Management
and the Office of Health and Ecological Effects
are represented, recognizing the interde-
pendence of our environmental problems.
Third, I realized in studying the program
that government (both Federal and State), the
textile industry, universities, and equipment
manufacturers are represented in good bal-
ance. No one has a monopoly on knowledge,
and it is necessary that these four segments of
our society share their knowledge and strive
toward a common goal: namely, to eliminate
whatever pollution problems industry may
have and do so in the most efficient and cost-
effective manner.
Fourth, I note an entire session is devoted to
methodology, with respect both to analytical
techniques and to bioassays. We sometimes
tend to fall in love with numbers without fully
•Assistant Administrator for Research and Develop-
ment, U.S. EPA, Washington, DC.
realizing how these numbers were derived and
without regard for their accuracy. We must
continue to be critical of our methods and
refine them to the point where we are confi-
dent of the results obtained.
The purpose of setting industrial emissions
standards is to protect the health of our people.
These standards must, therefore, be the best
reflection of what we know about the short-
term and long-term health effects of the pollu-
tants. When new data become available, it is
essential that this information be transferred
to other government agencies, to private scien-
tific groups, and to industries at the earliest
possible moment.
The setting of industrial pollution discharge
standards is a very complex process involving
the legislative and judicial arms of the govern-
ment, numerous government and private scien-
tific groups, industries, and the public at large.
It is thus difficult to clear away the smog of
emotions surrounding pollution issues and
make decisions on a rational scientific basis.
Conferences such as this serve as a sounding
board for scientific thought, where industry,
universities, and government may exchange
ideas and develop new ones. Court cases drain
the resources of both industry and the tax-
payer. When possible, it is better to resolve
problems at the scientific roundtable than to
testify in court!
I am well aware of the spirit of cooperation
that currently exists between the Environmen-
tal Protection Agency and the American Tex-
tile Manufacturers' Institute (ATMI), the
Northern Textile Association, and the Carpet
and Rug Institute. We hope that such coopera-
tion will continue to exist and that out of the
mass of data being accumulated will come real-
istic, cost-effective controls, which both gov-
ernment and industry will consider to be in the
best interests of the nation.
I know there is much skepticism on the part
of many persons concerning the work of EPA.
There have been charges by industry that we
are moving too fast and too soon with too little
data on which to base decisions. We are further
-------
accused of putting industry "out-of-business"
because of what is termed excessive pollution
control costs.
The record shows otherwise: The cost of pol-
lution control contributes less than one-half of
one percent to the annual rate of inflation (ref.
1). Pollution control creates more jobs as it
displaces. Even after adjusting for inflation,
corporate profits continue to rise and unem-
ployment is the lowest it has been in several
years.
As a result of pressures by EPA to reduce
pollution, industry is finding less harmful,
substitute materials in many instances. In still
other cases, money is actually saved through
simultaneous material recovery and energy
conservation resulting from efforts to elimi-
nate pollutant discharges to rivers, lakes, and
municipal treatment plants. An example of this
is the recovery of polyvinyl alcohol sizing
agents through the use of ultrafiltration. Judg-
ing from your program, many such useful
emerging technologies will be discussed during
the next 3 days.
The textile industry has been a major con-
tributor to our American standard of living.
There are about 7,000 textile plants in the
United States, and most of these are on the
eastern seaboard. These plants employ almost
one million persons and produce products
whose gross sales amount to $40 billion an-
nually, representing 0.6 percent of our gross
national product on a value-added basis, (ref. 2)
The production of all these goods requires
water —lots of water. The textile industry dis-
charges 1 billion gallons of water each day into
lakes, streams, and municipal treatment sys-
tems. The total value of this water, including
the dissolved solids contained in it and the
energy in the form of heat, amounts to over $1
million per day. What a golden opportunity to
combine pollution control with energy and ma-
terial recovery!
Industry, however, is not the only segment
of our society to cast stones at EPA. Envi-
ronmental groups, by contrast, accuse EPA of
not moving fast enough. They say little has
been accomplished. Again, the record shows
otherwise: Fish are returning to our lakes and
streams. As an example, salmon are returning
to the Connecticut River for the first time in
years. Fish and other marine life are returning
to the Willamette River, the Buffalo River, the
Detroit River, and the Houston Ship Channel.
(ref. 1)
Similar, though less dramatic, improve-
ments have been made in reducing air pollu-
tion: Carbon monoxide levels have fallen, par-
ticulate emissions have been reduced, and SOX
levels are measurably smaller. Hopefully,
these trends will continue.
The emphasis on pollution control has been
changing in recent years, with regulations
becoming more and more stringent as addi-
tional environmental hazards are recognized. I
will not dwell on any of these regulations since
I am certain these will be discussed by the
overview speakers.
In 1976, a Consent Decree settling a court
suit filed under the Federal Water Pollution
Control Act was signed by EPA and four envi-
ronmental groups in which EPA agreed to set
effluent standards for 65 classes of compounds
for 21 industries, including the textile in-
dustry. As a result of the Consent Decree, the
Effluent Guidelines Division of EPA published
a list of 129 organic pollutants and metals on
which they proposed to set regulations. These
have become known as the "priority toxic pol-
lutants." Subsequently, amendments to the
water pollution legislation, known as the Clean
Water Act of 1977, converted the terms of the
consent decree into law. Specifically, the 1977
Act requires the application of best available
technology economically achievable (BATEA)
to control the priority pollutants by July 1,
1984.
These regulations, plus the advent of the
Resource Conservation and Recovery Act, the
Toxic Substances Control Act, and the forth-
coming pretreatment standards of the Clean
Water Act, place a great burden on the textile
industry. However, we are also aware of the
progress the textile industry is making toward
solving some of these problems. Most textile
plants now have installed biological aeration
systems, clarifiers, and finishing ponds, which
have enabled the majority of them to meet the
1977 guidelines.
Tertiary treatment systems are being sys-
tematically investigated. One such large-scale
investigation is, of course, the ATMI/EPA joint
project to study the best available technology
economically achievable (BATEA), scheduled
to be completed this spring. Economic as well
as technical studies are being conducted. Some
-------
textile plants have already installed advanced
water pollution control equipment such as
reactor/elarifiers and carbon columns in an-
ticipation of the regulations.
Thus far, my primary emphasis has been on
water pollution problems, but I did not inten-
tionally mean to neglect potential air pollution
problems. I am pleased that your program in-
tends to cover this aspect. As you well know,
the textile industry uses solvents and resins of
many types; in dyeing and finishing operations
these are volatilized. Once again, ATMI and
EPA are cooperating in attempting to deter-
mine the nature and extent of these air emis-
sions. Several textile finishing plants have
volunteered to have an EPA contractor mea-
sure the amount and composition of their emis-
sions. Under other contracts, a state-of-the-art
survey for air pollution control technology is
being made, which may result in a demonstra-
tion grant.
As you are aware, on August 7,1977, Presi-
dent Carter signed PL 95-95, the Clean Air Act
Amendments of 1977. This brings on new res-
ponsibilities for the operators of stationary
source facilities. This will also change the con-
ditions for obtaining permits for expanded
plants and new plants.
Some textile plants have already recognized
that they have a serious air pollution problem
and have installed incinerator-heat recovery
systems or carbon columns. Thus, they also
recognize the opportunity for energy and,
possibly, material recovery along with air
pollution control.
I know that the list of suspected harmful
compounds grows larger and larger and that
the "pollutant of the month" phase is now
uttered jokingly. It is, however, no joking mat-
ter, for the incidence of cancer continues to
rise significantly. It is, nevertheless, a compli-
ment to our society that we are now balancing
our mad rush to produce more and more con-
sumer goods, regardless of long-term conse-
quences, with concern for the long-term health
of our population.
Improved epidemiological studies, better
analytical techniques, and more sophisticated
bioassays are beginning to pinpoint the cul-
prits in our environment, and we are beginning
to do something about it. Rules and regulations
governing such substances as asbestos, PCB's,
PBB's, vinyl chloride, phenols, chromium, and
sulfides are only the beginning.
We are fully cognizant, however, of the eco-
nomic problems facing the textile industry,
such as its low profit margin and the threat of
foreign imports. It is definitely not EPA's
policy to ignore the economic aspects of an in-
dustry.
The Toxic Substances Control Act states, in
one of its three policy statements, that
"authority over chemical substances and mix-
tures should be exercised in such a manner as
not to impede unduly or create unnecessary
economic barriers to technological innovation."
(ref. 3) This meeting, sponsored by EPA, is
testimony to the fact that EPA intends to pro-
mote technological innovation when such tech-
nology is consistent with our planned policy of
pollution control at minimum cost to the indus-
try and to the public.
REFERENCES
1. Remarks by the Honorable Barbara Blum,
Deputy Administrator, U.S. EPA, pre-
pared for delivery before the Water Pollu-
tion Control Federation, Washington, DC,
March 29,1977.
2. Information obtained from Mr. O'Jay Niles,
Director of Technical Services, American
Textile Manufacturers' Institute.
3. Quoted from the Toxic Substances Control
Act by the Honorable Douglas M. Costle,
Administrator, U.S. EPA, at the Center for
Continuing Education Seminar on the Fed-
eral Regulation of Environmental Carcino-
gens, the Capital Hilton Hotel, Washing-
ton, DC, April 12, 1977.
-------
Session I: IMPACT OF LEGISLATION ON POLLUTION CONTROL
IN THE TEXTILE INDUSTRY
Max Samfield, Session Chairman
-------
WATER POLLUTION CONTROL IN THE TEXTILE INDUSTRY
Robert B. Schaffer*
Abstract
The textile industry was one of 27 categories
of point sources for which effluent limitations
and standards were established under the Fed-
eral Water Pollution Control Act Amendments
of 1972. The first regulations promulgated by
the Administrator of EPA affecting the textile
industry appeared in the Federal Register in
July 1974. Pretreatment regulations for exist-
ing sources in the textile point source category
appeared in the Federal Register in May 1977.
These regulations set forth effluent limitations
for existing sources, new source performance
standards (NSPS), andpretreatment standards
for the textile industry. The pollutant param-
eters addressed are biochemical oxygen de-
mand (BOD), chemical oxygen demand (COD),
suspended solids, phenols, sulfides, chromium,
color, and pH. And for the wool scouring seg-
ment of the industry, oil and grease were also
limited.
The emphasis of the water pollution control
program changed on June 7, 1976, when the
EPA signed an agreement that settled four
lawsuits brought against the Agency by the
Natural Resources Defense Council and the
Environmental Defense Fund. As a result of
this settlement agreement, EPA is charged
with the task of developing and promulgating
effluent limitations guidelines and standards
for 65 toxic substances. This is to be done for
21 major industry categories of which textiles
is one.
Table 1 is a list of the 21 major industries set
forth in the settlement agreement. Table 2 is a
list of the 65 chemical substances for which.ef-
fluent limitations and guidelines and N8P8 and
pretreatment standards are to be developed.
Since the list of 65 chemical substances is
somewhat generic, the Agency was faced with
the problem of how to identify specific chem-
*Director, Effluent Guidelines Division, U.S. En-
vironmental Protection Agency, Washington, DC.
ical substances that would meet the require-
ments for addressing the 65 substances listed
in the settlement agreement. Table 3 is a list of
129 specific chemical substances (114 organics,
13 heavy metals, cyanide, and asbestos) repre-
sentative of the 65 generic substances set forth
in the settlement agreement.
The 1977 Amendments to the Clean Water
Act recognized the significance of the settle-
ment agreement and emphasized the impor-
tance of the Agency's toxics control strategy.
The new Act directed EPA to concentrate on
the pollution problems causing the most con-
cern and to reevaluate the areas where prog-
ress had already been made. Of the 129 chem-
icals mentioned earlier, 29 have been deter-
mined to be most significant in textile effluents
(see Table 4). Others have been suggested as
possible constituents of textile mill wastes by
industry sources but have not been detected in
the field sampling program and are not pres-
ently considered as important as the 29 chem-
icals in Table 4. The control of these sub-
stances will be the subject of the best available
technology (BAT), NSPS, and pretreatment
regulations for toxic pollutants.
The progress in controlling nontoxic dis-
charges through the earlier effluent limita-
tions was reflected in Section 304 (b)(4) of the
Act which mandated that conventional pollut-
ants such as BOD, COD, suspended solids, and
pH be controlled under best conventional pol-
lutant control technology (BCT). This replaced
BAT for control of conventional pollutants. If it
is determined that effluents from some textile
subcategories contain only conventional pollut-
ants, EPA would issue BCT regulations for
those subcategories when the BAT guidelines
for toxics are issued. Congress specified that
BCT could be equal to or more stringent than
best practicable treatment (BPT). BCT will be
determined for each industrial subcategory
after balancing the costs with the effluent
reduction benefits. BCT permits would not be
eligible for economic variances or water qual-
ity variances.
Not all pollutants are classified as conven-
tional or toxic pollutants under the Act. Pollut-
-------
TABLE 1. INDUSTRIES COVERED BY CONSENT DECREE
Timber products processing
Leather tanning and finishing
Steam electric power plants
Iron and steel manufacturing
Petroleum refining
Nonferrous metals manufacturing
Paint and ink formulation and printing
Auto and other laundries
Paving and roofing materials
Ore mining and dressing
Coal mining
Textile mills
Soap and detergent manufacturing
Plastic and synthetic materials manufacturing
Pulp and paperboard mills; converted paper products
Rubber processing
Miscellaneous chemicals
Machinery and mechanical products manufacturing
Electroplating
Inorganic chemicals manufacturing
Organic chemicals manufacturing
TABLE 2. SETTLEMENT AGREEMENT POLLUTANTS
Acenaphthene
Acrolein
Acrylonitrile
Aldrin/dieldrin
Antimony and compounds
Arsenic and compounds
Asbestos
Benzene
Benzidine
Beryllium and compounds
Cadmium and compounds
Carbon tetrachloride
Chlordane
Chlorinated benzenes
Chlorinated ethanes
Chloroalkyl ethers
Chlorinated naphthalene
Chlorinated phenols
Chloroform
2-chlorophenol
Chromium and compounds
Copper and compounds
Cyanides
DDT and metabolites
Dichlorobenzene
Dichlorobenzidine
Dichloroethylenes
2,4-dichlorophenol
Dichloropropane and dichloropropene
2,4-dimethylphenol
Dinitrotoluene
Diphenylhydrazine
Endosulfan and metabolites
Endrin
Ethylbenzene
Fluoranthene
Haloethers
Halomethanes
Heptachlor and metabolites
Hexachlorobutadiene
Hexachlo rocyclohexane
Hexachlorocyclopentadiene
Isophorone
Lead and compounds
Mercury and compounds
Naphthalene
Nickel and compounds
Nitrobenzene
Nitrophenols
Nitrosamines
Pentachlorophenol
Phenol
Phthalate esters
Polychlorinated biphenyls
Polynuclear aromatic hydrocarbons
Selenium and compounds
Silver and compounds
2,3,7,8-tetrachlorodibenzo-p-dioxin
Tetrachloroethylene
Thallium and compounds
Toluene
Toxaphene
Trichloroethylene
Vinyl chloride
Zinc and compounds
-------
TABLE 3. SPECIFIC CHEMICAL SUBSTANCES TO BE EXAMINED
1, Acenaphthene*
2. Acrolein»
3. Acrylonitrile*
4. Benzene*
S. Benzid'me*
6. Carbon tetrachloride*
Chlorinated benzenes*
7. Chlorobenzene
8. 1,2,4-trichlorobenzene
9. Hexachlorobenzene
Chlorinated ethanes*
10. 1,2-dichloroethane
11. 1,1,1-trichloroethane
12. Hexachloroethane
13. 1,1-dichloroethane
14. 1,1,2-trichloroethane
15. 1,1,2,2-tetrachloroethane
16. Chloroethane
Chloroaikyl ethers*
17. bis(chloromethyl) ether
18. bis(2-chloroethyl) ether
19. 2-chloroethyl vinyl ether
Chlorinated napthalene*
20. 2-chloronaphthalene
Chlorinated phenols*
21. 2,4,6-trichlorophenol
22. Panachlorometa cresol
23. Chloroform*
24. 2-chlorophenol*
Dichloro benzenes*
25. 1,2-dichlorobenzene
26. 1,3-dichlorobenzene
27. 1,4-dichlorobenzene
Dichlorobenzidine*
28. 3,3-dichlorobenzidine
Dichloroethylenes*
29. 1,1-dichloroethylene
30. 1,2-trans-dichloroethylene
31. 2,4-dichlorophenoJ*
Heptachlor and metabolites*
100. Heptachlor
101. Heptachlor epoxide
Hexachlorocyclohexane*
102. a-BHC-alpha
103. b-BHC-beta
104. r-BHC (lindane)-gamma
105. g-BHC-delta
Polychlorinated biphenyls (PCB'sj*
106. PCB-1242 (Arochlor 1242)
Dichloropropane and dichloropropene*
32, 1,2-dichloropropane
33. 1,3-dichloropropylene
34. 2,4-dimethylphenor
Dinitrotoluene*
35. 2,4-dinitrotoluene
36. 2,6-dinitrotoluene
37. 1,2-diphenylhydrazine*
38. Ethylbenzene*
39. Fluroanthene*
Haloethers*
40. 4-chlorophenyl phenyl ether
41. 4-bromophenyl phenyl ether
42. bis(2-chloroisopropyl) ether
43. bis(2-chloroethoxy) methane
Halomethanes*
44. Methylene chloride
45. Methyl chloride
46. Methyl bromide
47. Bromoform (tribromomethane)
48. Dichlorobromomethane
49. Trichloroflouoromethane
50. Dichlorodifluoromethane
51. Chlorodibromomethane
52. Hexachlorobutadiene*
53. Hexachlorocyclopentadiene*
54. Isophorone*
55. Naphthalene*
56. Nitrobenzene*
Nitrophenols*
57. 2-nitrophenol
58. 4-nitrophenol
59. 2,4-dinitrophenol*
60. 4,6-dinitro-o-cresol
Nitrosamines*
61. N-nitrosodimethylamine
62. N-nitrosodiphenylamine
63. N-nitrosodi-n-propylamine
64. Pentachlorophenol*
65. Phenol*
107. PCB-1254 (Arochlor 1254)
108. PCB-1221 (Arochlor 1221)
109. PCB-1232 (Arochlor 1232)
110. PCB-1248 (Arochlor 1248)
111. PCB-1260 (Arochlor 1260)
112. PCB-1016 (Arochlor 1016)
113. Toxaphene*
114. Antimony (total)*
115. Arsenic (total)*
116. Asbestos (fibrowl*
117. Beryllium (total)*
118. Cadmium (total)*
Pthlate esters
66. bis(2-ethylhexyl) phthalate
67. Butyl benzyl phthalate
68. di-n-butyl phthalate
69. di-n-octyl phthalate
70. Diethyl phthalate
71. Dimethyl phthalate
Polynuclear aromatic hydrocarbons*
72. 1,2-benzanthracene
73. Benzo(a)pyrene (3,4-benzopyrene)
74. 3,4-benzofluoranthene
75. 11,12-benzofluoranthene
76. Chrysene
77. Acenaphthylene
78. Anthracene
79. 1,12-benzoperylene
80. Fluorene
81. Phenanthrene
82. 1,2,5,6-dibenzanthracene
83. lndenol(1,2,3-cd)pyrene
84. Pyrene
85. Tetrachloroethylene*
86. Toluene*
87. Tryichloroethylene*
88. Vinyl chloride (chloroethylene)*
Pesticides and metabolites
89. Aldr'm*
90. Oieldrin*
91. Chlordane*
DDT and metabolites
92. 4,4'-OOT
93. 4,4'-DDE (p,p'-DDX)
94. 4,4'-DDD (p,p'-TDE)
Endosulfan and metabolites*
95. a-ensodulfan-alpha
96. b-endosulfan-beta
97. Endosulfan sulfate
Endrin and metabolites*
98. Endrin
99. Endrin aldehyde
119. Chromium (total)*
120. Copper (total)*
121. Cyanide (total)*
122. Lead (total)*
123. Mercury (total)*
124. Nickel (total)*
125. Selenium (total)*
126. Silver (total)*
127. Thallium (total)*
128. Zinc (total)*
129. 2,2,7,8-tetrachlorodibenzo-p-dioxin (TCDOI*
'Specific compounds and chemical classes as listed in the consent decree
Note: Ambiguous compounds or classes of compounds are underlined.
-------
TABLE 4. MOST SIGNIFICANT TOXIC POLLUTANTS IN TEXTILE WASTEWATERS
Acrylonitrile
Benzene
1,2,4-trichlorobenzene
2,4,6-trichlorophenol
Parachlorometacresol
Chloroform
1,2-dichlorobenzene
Ethylbenzene
Trichlorofluoromethane
Naphthalene
N-nitrosodi-n-propylamine
Pentachlorophenol
Phenol
bis(2-ethylhexyl) phthalate
Tetrachloroethylene
Toluene
Trichloroethylene
Antimony
Arsenic
Cadmium
Chromium
Copper
Cyanide
Lead
Mercury
Nickel
Selenium
Silver
Zinc
ants such as color do not fall into either cate-
gory and are called nonconventional pollut-
ants. If it is determined that effluents from tex-
tile subcategories require color removal, EPA
would issue BAT regulations for this noncon-
ventional pollutant when the BAT guidelines
for toxics are issued. Unlike regulations cover-
ing conventional or toxic pollutants, economic
variances and water quality variances can be
granted.
It is clear that the effluent regulations
presently under consideration for the textile
industry include toxic, conventional, and non-
conventional pollutants. While the effort has
become known as the BAT Revision Program,
the regulations include NSPS and pretreat-
ment standards in addition to effluent guide-
lines for dischargers to receiving waters. The
development and implementation of pretreat-
ment standards for mills discharging into pub-
licly owned treatment works (POTW) is the
subject of a separate paper. However, a few
comments may be helpful.
The general pretreatment regulations were
promulgated on June 26, 1978. The major com-
ponents of the Agency's pretreatment strat-
egy include: (1) technology based pretreatment
standards for toxic pollutants, (2) strong en-
couragement, including financial incentives for
local enforcement of pretreatment standards,
and (3) modifications of the standards for
POTW with documented pollutants removal ef-
ficiencies. The pretreatment standards will be
developed by applying BAT decision criteria to
indirect dischargers in accordance with the
Agency's pretreatment strategy with an aim
toward limiting the impact of textile dis-
chargers on POTW's and receiving waters
while avoiding to the extent possible redun-
dant treatment by industry and municipalities.
The settlement agreement set forth a sched-
ule for the textile industry program to com-
plete the technical and economic studies and to
propose and promulgate regulations. The tech-
nical report that contains the findings of our
contractor, Sverdrup & Parcel, St. Louis, Mis-
souri, has been circulated to other Federal of-
fices, States, industry, and other public in-
terest groups for review and comment. The
draft economic analysis has also been com-
pleted by Development Planning and Research
Associates, Inc., Manhattan, Kansas. Soon the
technologies available for control and treat-
ment of various pollutants will be evaluated
with the cost and economic impacts of the
various options and with the toxicity and ex-
posure levels of the various pollutants. This in-
tegration of technical, economic, and environ-
mental considerations will culminate in a pro-
posed regulation for the textile industry in
May 1979, with final promulgation expected in
December. As we proceed to develop these
standards, we plan to continue to invite active
participation —public, industrial, community,
as well as environmental groups —and we will
continue to solicit industrial input and assist-
ance.
10
-------
RCRA AND THE TEXTILE INDUSTRY
Harry W. Trask*
Abstract
Congress passed the Resource Conservation
and Recovery Act (RCRA) in October 1976. The
stated objectives of RCRA are to promote the
protection of health and the environment and
to conserve valuable material and energy re-
sources. The Act intends that these objectives
be achieved by: providing technical and finan-
cial assistance to State and local governments
for development and implementation of solid
waste management plans; prohibiting future
open dumping on land and requiring upgrading
or closing of existing open dumps; regulating
the treatment, storage, transportation, and
disposal of hazardous waste; promulgating
guidelines for solid waste management prac-
tices and systems; conducting a research and
development program for improved solid waste
management and resource conservation tech-
niques; demonstrating improved solid waste
management and resource conservation and
recovery systems; and establishing a cooper-
ative effort among Federal, State, and local
governments and private enterprises.
The Act statutorily establishes the Office of
Solid Waste within EPA to guide the imple-
mentation of the law and establishes a Fed-
eral/State/local government partnership to
share the implementation. The major thrusts of
the efforts that will be required by this partner-
ship are: land protection through regulation
and control of wastes and waste disposal oper-
ations; regulations to control the hazardous
waste stream from "cradle to grave;"reduction
of the waste stream through increased resource
recovery and waste reduction efforts; public
education programs with rapid dissemination
of all types of solid and hazardous waste
management information materials; and exten-
sive public participation in the development
and improvement of solid waste management
throughout the Nation.
*Program Manager, Hazardous Waste Management
Division, U.S. EPA, Washington, DC.
The regulatory part of the Act, the part deal-
ing with control of hazardous wastes, will be
the subject of the remainder of this paper. But
first, consideration will be given to the overall
problem of hazardous wastes.
THE HAZARDOUS WASTE PROBLEM
According to the results of 17 recent EPA in-
dustry studies, an estimated 34 million metric
tons (wet basis) of potentially hazardous in-
dustrial waste were generated in 1977. The
amount generated is expected to increase with
industry growth and more pollution control.
Approximately 80 percent of this waste is man-
aged onsite by the generators, the remaining
20 percent being hauled offsite for disposal by
contractors.
In the textile industry, certain wastewater
treatment sludges may be hazardous wastes
depending on their constituents. For example,
sludges from wool scouring and certain fabric
dyeing and finishing operations may contain
toxics such as lead, chromium, arsenic, cad-
mium, mercury, etc. It is estimated that about
1.7 million metric tons (wet weight) of hazard-
ous or potentially hazardous wastes were gen-
erated by the textile industry in 1977.
The method most used for disposal of haz-
ardous industrial waste is lagooning in unlined
surface impoundments, accounting for nearly
half of the total. The second most common
practice is dumping in nonsecure landfills.
Together these receive almost 80 percent of all
hazardous wastes. An additional 10 percent is
accounted for by uncontrolled incineration.
Thus, 90 percent of the potentially hazardous
wastes generated by the 17 key industries are
managed by practices inadequate for protec-
tion of human health and the environment.
The consequences of illegal or inadequate
disposal can be quite harmful. Well-docu-
mented cases of groundwater contamination
by leachates, surface water contamination by
runoff, direct contact poisoning, various forms
of air pollution, and damage from fire and ex-
plosions have occurred and will continue to oc-
11
-------
cur as long as unacceptable disposal practices
are used. Most incidents result from open
dumping of hazardous wastes on isolated
tracts of land or from indiscriminate accept-
ance of all wastes regardless of hazard by
municipal landfills.
For example, in 1972 eleven people in Per-
ham, Minnesota were striken with the symp-
toms of arsenic poisoning. Upon investigation
it was discovered that they had been drinking
water from a new well that had been drilled
near a 30-year-old, unmarked deposit of left-
over grasshopper bait. Water samples from
the well were found to contain more than 400
times the amount of arsenic maximally allow-
able in drinking water in the United States.
EPA's Office of Solid Waste has compiled
over 400 case studies of such incidents primar-
ily from data gathered by State environmental
regulatory agencies. It is clear that open en-
vironmental abuses occur more frequently in
States without such programs and without ade-
quate documentation of damages. To this ex-
tent, the cases can be said to understate the
problem.
SUBTITLE C
Subtitle C of RCRA gives EPA authority to
set standards for generators and transporters
of hazardous waste and for hazardous waste
treatment, storage, and disposal facilities.
Subtitle C takes a pathways or "cradle-to-
grave" approach to regulating hazardous
waste. This means regulation of generation
and transportation of hazardous waste as well
as its ultimate disposition. For example, each
hazardous waste load shipped offsite must be
accompanied by a manifest filled out by the
generator of the waste. On this manifest, the
generator designates the facility to which the
waste must be taken. The manifest goes with
the load to the facility and the signed original
is returned to the generator by the facility.
This returned manifest closes the loop of re-
sponsibility for the generator. The generator is
responsible for maintaining these manifests as
evidence that he has properly managed his haz-
ardous waste. National standards have been
proposed for transporters that require deliv-
ery of all hazardous waste to the facility
designated on the manifest. And to adequately
control final disposition and prevent unsound
management, facilities for treatment, storage,
and disposal of hazardous waste must acquire
permits.
The manifest system, the transporter stand-
ards, and the permit system should halt dump-
ing in isolated fields and nonsecure landfills, or
in muncipal sewer systems, because the gener-
ator of the hazardous waste must be able to
show that the waste was sent to an acceptable,
permitted facility. If attempts are made to
falsify returned manifests, a cross-check with
the permitted facility should quickly show that
the load never reached that facility. The facil-
ities receiving hazardous wastes must meet
certain criteria for technical soundness and
financial responsibility if they are to obtain
permits. This prevents designation of simple
pits or fields as disposal sites on manifests
because such sites could never obtain permits.
Currently many sites legally accepting haz-
ardous wastes are not environmentally suit-
able. Such sites include open dumps and munic-
ipal and sanitary landfills as well as "secure"
landfills which were not adequately located or
constructed. Under RCRA, open dumps must
be upgraded to the standards for acceptable
landfills or be shut down. Only chemical land-
fills, other land disposal techniques, and treat-
ment facilities shall be permitted to manage
hazardous wastes.
Taken together, the manifest and permitting
requirements should force most hazardous
wastes to appropriate management facilities at
the same time they close environmentally un-
sound facilities, which will decrease current
capacity. This should substantially increase de-
mand for acceptable hazardous waste manage-
ment.
STATE IMPLEMENTATION
The intent of Congress and the policy of
EPA is to make the hazardous waste manage-
ment program resulting from the Resource
Conservation and Recovery Act as much a
State-run operation as possible. Under the Act,
the Administrator must authorize a State to
operate its program in place of a Federal
operation, if certain conditions are met.
To acquire full authorization, a State pro-
gram must be equivalent to the Federal pro-
gram under Subtitle C, be consistent with
Federal programs under Subtitle C, be consist-
12
-------
ent with Federal or State programs applicable
in other authorized States, and provide for ade-
quate enforcement of compliance.
Consistency with Federal or State programs
applicable in other States means consistency
with other authorized programs. EPA believes
the free movement of hazardous wastes be-
tween States to be an integral part of a consist-
ent, national hazardous waste program and for
this reason is considering not authorizing State
programs which include bans on the importa-
tion of hazardous wastes from other States.
States applying different standards to wastes
originating outside the States will not be
authorized.
IMPLEMENTATION PROBLEMS
There are several potential problems that
must be dealt with in order to successfully im-
plement this national hazardous waste man-
agement program. Some States may be unable
or unwilling to take up the program and thus
place the responsibility for implementation on
the EPA regional office. Public opposition to
hazardous waste management facility siting
may prevent opening a new capacity and
threaten continued operation of some existing
capacity. Also, it will take time for the rela-
tively small hazardous waste management in-
dustry to grow sufficiently to handle the ex-
pected volume demand. Thus, for the interven-
ing period of indefinite length, a capacity short-
fall may occur.
A State may fail to pick up the program for
other reasons. It may simply lack the legisla-
tive authority to control hazardous waste.
Some State legislatures meet only every 2
years; thus, even a State desiring to accept the
program may experience lengthy delay in re-
ceiving the authority to do so.
Some States will lack the resources to
operate a program equivalent to the Federal
one. Although there will be grant money avail-
able, this must be matched at the 25 percent
level by State funds which may not be avail-
able. Resources include personnel and such
things as laboratory and data processing cap-
abilities which may take time to develop even
when sufficient funds are available.
Finally, some States may not agree with the
regulatory program adopted by EPA and thus
may choose not to meet EPA requirements for
authorization. These States may, for example,
ban importation of hazardous wastes from
other States thus forcing each State to need-
lessly duplicate costly facilities to manage all
wastes. Such inconsistency could prevent a
particular hazardous waste from reaching the
facility best able to manage it. It could also pre-
vent the industry from realizing economies of
scale present in large facilities that serve wide
areas.
Public reaction to the siting of hazardous
waste management facilities has been largely
adverse. For example, in Wilsonville, Illinois,
what the State has determined to be a techni-
cally acceptable facility has met with strong
local opposition due to its proximity to the
town. In general, public opposition is based on
an understandable fear for health and safety
and reduced property values. Most people do
not want hazardous waste mangement facili-
ties nearby although they see the need for
them someplace else.
Another implementation problem is the
large number of potential generators who
generate relatively small quantities of hazard-
ous waste, or who generate in an erratic fash-
ion. Chemical containers, for example, al-
though considered to be hazardous unless
properly drained and rinsed or reconditioned,
in total weight are not a significant source.
Still they are ubiquitous and are generated in
many areas that do not have convenient, appro-
priate disposal facilities. This argues, of
course, for the development of methods and
practical procedures that can be established in
these areas to decontaminate the containers
and render them fit for recycle, reuse, and the
like. Retailers who occasionally break a case of
chemicals or have "leakers" are another
special group.
Our current approach to the retailer and
other "small" generator problems has two
facets. First, we identify as generators only
those who produce and dispose of 100 kg of
hazardous waste per month. This quantity,
(about one-half drum of most liquids), excludes
over half of the potential generators, but
"loses" only about one-half percent of hazard-
ous industrial wastes. Since RCRA prohibits
open dumping, this small quantity will be
forced to acceptable landfills that meet the
Subtitle D criteria.
Secondly, we are considering relieving
13
-------
retailers from the manifest reporting and rec-
ordkeeping burdens if they too dispose of
waste in Subtitle D or C facilities.
There will be hazardous waste management
regulations nationwide regardless of which in-
dividual States accept the program contained
in RCRA. However, implementation will be
slower and may be less responsive to local
needs if EPA must institute the program in
most or all of the States.
The ability of the hazardous waste manage-
ment industry to overcome local opposition to
siting and expansion activities is not clear. At
the least, such influences will delay implemen-
tation of the Act. At worst, it could stymie de-
velopment of necessary sound capacity, caus-
ing a serious shortfall and forcing hazardous
wastes into the fields and sewers, to the detri-
ment of both public health and the environ-
ment. In all, public opposition to siting may
become the greatest obstacle to implementa-
tion of a program for sound hazardous waste
management and deserves our most thoughtful
concern.
14
-------
Session II: WASTEWATER POLLUTION CONTROL TECHNOLOGY
W. Wesley Eckenfelder, Session Chairman
15
-------
ORIGIN AND INTERNAL CONTROL OPTIONS
FOR PRIORITY POLLUTANTS IN THE TEXTILE INDUSTRY
Edwin L. Barnhart*
Abstract
To date, textile industry has been able, with
some minor exceptions, to meet the criteria set
down for water pollution control. The industry
has developed within itself the technology and
methodologies necessary for the operation and
control of treatment plants that achieve best
practical treatment of higher levels of removal.
By and large, the movement within the indus-
try has been quite rapid and orderly.
By comparison, the introduction of priority
pollutant control will cause major problems for
the industry. The reasons that disruption will
be caused are important to understand. The
major problem is that identification and con-
trol of the priority pollutants depends on a
level of chemical analysis expertise and skill
which exceeds the level required by manufac-
turing demands of the industry. BOD's, COD's,
and even the more complicated nitrogen analy-
ses required by conventional pollution control
can be performed by a skilled plant chemist,
but the introduction of complex gas chroma-
tography or gas chromatography/mass spec-
trometry (GCMS) exceeds the general level of
background of these individuals. Further, the
type and class of equipment required for these
advanced analyses exceed the requirements for
the normal operation of the manufacturing
facilities. Although this is not true for the
selected few high technology companies within
the industry, the majority of the operating
companies will find themselves in this situa-
tion.
In a similar manner, physical and chemical
phenomena associated with the operation of a
secondary or even a tertiary plant are easily
comprehended by a typical plant engineer. The
complexities of the systems that may be re-
quired to control priority pollutants are by
nature intrinsically involved with the chem-
istry of the reactions and may very well require
•President, Hydrosciences, Inc., Arlington, TX.
special expertise unlikely to be available
within any of the smaller companies.
The textile manufacturing industry's prob-
lem is further complicated by the fact that, by
and large, it does not itself manufacture the
chemicals used in its reactions, nor has it been
common practice for the suppliers to list, in a
meaningful generic fashion, the components of
the various mixtures involved. Even more fun-
damentally, many of the dye stuffs and other
materials commonly used in the industry are
not analyzed, even during their manufacture,
to the point of determining the various in-
termediates that can exist.
The variety and complexity of the chemicals
employed in an average textile manufacturing
operation are somewhat mind boggling. In pre-
paring for this paper, typical operating records
for several mills were examined. During an
average year, the mills examined used from 100
to 300 specified compounds. Of the compounds
used about one half were identified by trade
names rather than by their specific chemical
formula or name. Several randomly examined
trade named compounds contained from 9 to 15
components as part of the manufacturers' spec-
ifications of the particular compound. Many of
the individual components either appear on the
priority pollutant list or are likely to make the
list as it expands. But even more importantly,
the one sample of a compound that was sub-
jected to further actual analysis showed many
components not listed by the supplier. Trace
amounts of a wide variety of chemical inter-
mediates were found. These compounds also
were either on the list or are subject to early in-
clusion. It is against this background, with an
understanding that thousands of compounds,
many of them unsuspected, pass through a
typical textile mill in a given year, that we
must consider the problem faced by the indus-
try.
IDENTIFYING COMPOUNDS OF CONCERN
The first step of any program to control the
discharge of priority pollutants must be the
17
-------
identifications of the compounds present in the
effluent. This is not always as straightforward
as it would seem. For convenience, a list of the
compounds presently designated as priority
pollutants is presented in Table 1. The list in-
cludes 114 organic compounds, 13 heavy met-
als, cyanide, and asbestos. EPA is presently
hard at work developing discharge standards
for each of the materials for a variety of in-
dustrial categories. At present the only reason-
able assumption is that each of the specific
compounds will be controlled at some low level
in the given effluent. Specific compounds, such
as the heavy metals, and the more carcinogenic
compounds may require complete removal.
Perusals of the list will quickly reveal that
many of the compounds are commonplace in
the textile industry. The cause or occurrence of
these compounds may be generally attributed
to three main factors—direct production use,
indirect production use, and nonproduction
use.
Direct Use
A study of chemical purchasing lists will
quickly identify the chemicals of concern that
are in quantity use in the facility. Although
such materials as benzidine and many heavy
metals are passing out of use, carbon tetrachlo-
ride, benzene, ethylbenzene, and the like will
remain common chemicals for at least a fore-
seeable future. Materials identified through
purchasing analysis can be handled in a direct
manner compatible with concern.
Indirect Use
Chemicals introduced into a plant through
indirect use are more difficult to detect. The
first and most significant source of these ma-
terials is the so-called proprietary chemical.
Where a specific functional chemical or dye is
identified only by its trade name, the compo-
nents of that compound must be identified by
working with the manufacturer. But further, it
must be noted that in many cases, even the
compoundor does not really have information
on the materials that are present. Heavy met-
als, traces of reaction intermediates, and the
like are persistent in the widest variety of pro-
prietary chemicals purchased within the indus-
try.
Further introduction of priority pollutants
may occur through nonmanufacturing uses
such as materials used to preserve and stabi-
lize as well as sprays to protect against mold,
etc., and in the packaging area. A variety of
phenolic type materials, for example, are used
as sealants and are present in a variety of
packaging schemes. Isolation of priority pollut-
ants from such activities requires detailed in-
vestigation and is often time-consuming and
frustrating. By way of example, the compo-
nents of a typical carrier are listed in Table 2.
A gas chromatography/mass spectrometry
(GCMS) analysis of this material showed in ex-
cess of 25 discrete compounds.
Nonproduction Use
By far, the most difficult materials to control
come from nonproduction uses; areas that
would never have been of concern before. Ex-
amples of these contaminants include: runoff
from sprays used to control shrubbery and
weeds around waste treatment facilities and
for general ground control, sterilizing agents
used by kitchen and cleaning services, and ma-
terials employed in general construction
around the plant site. At first these uses may
seem trivial, but understand that in a typical
manufacturing plant consuming a million gal-
lons of water a day, 1 pound of a chemical
causes a concentration in the effluent in excess
of .1 mg/liter or 100 ppb. This is significantly
above the level at which many of the priority
pollutants will be controlled.
In a recent investigation a particular ma-
terial of concern was found to be a component
of a commercial cleaning solution used by con-
tract labor over each weekend. The concentra-
tion of this material reaching the treatment
plant was sufficient to contaminate the effluent
for several days of the following week. This
compound that "couldn't possibly be there"
was only found and eliminated after a system-
atic, time-consuming evaluation of every pos-
sible cause.
From the above discussion, it should be clear
that priority pollutants will enter a plant from
the widest conceivable range of activities car-
ried on at the plant site. Working to control
these substances then must envision a much
broader approach than has previously been
conceived within the industry.
18
-------
TABLE 1. COMPOUNDS PRESENTLY DESIGNATED AS PRIORITY POLLUTANTS
Phenols
Phenol
2-ChlorophenoI
4-Chloro phenol
2,4-Dichlorophenol
2,4,5-Trichlorophenol
2,4,6-Trichlorophenol
Chlorocresol
Pentachlorophenol
2,4-Dimethylphenol (xylenol)
2,4-Dinitrophenol
4,6-Dinitrocresol
2-Nitrophenol
4-Nitrophenol
Substituted aromatics
Chlorobenzene
1,2-Dichlorobenzene
1,3-Dichlorobenzene
1,4-Dichlorobenzene
2-Chloronaphthalene
1,2,4-Trichlorobenzene
Nitrobenzene
2,4-Dinitrotoluene
2,6-Dinitrotoluene
Benzidine
3,3-Dichforobenzidine
1,2-Oiphenylhydrazine
PCB's
Hexachlorobenzene
Aromatic hydrocarbons
Acenaphthene
Benzene
Ethylbenzene
Fluoranthene
Naphthalene
Toluene
Tetrachlorodibenzo-p-dioxin
Benzanthracene
Dibenzathracene
Benzopyrenes
Benzofluoranthenes
Chrysene
Aromatic hydrocarbons (con.)
Indenopyrene
Anthracene
Phenanthrene
1,12-Benzoperylene
Fluorene
Organic
Acrolein
Acrylonitrile
bis (chloromethyl) ether
bis (2-chloroethyi) ether
2-Chloroethyl vinyl ether
4-Chlorophenyl phenyl ethers
4-Bromophenyl phenyl ethers
bis (dichloroisopropyl ether)
bis (2-chloroethoxy) methane
Polychlorinated diphenyl ethers
Isophorone
Phthalate esters
Vinyl chloride
N-nitrosodimethylamine
N-nitrosodiphenylamine
N-nitrosodi-n_-propylamine
Chlorinated solvents
Carbon tetrachloride
Chloroethanes
1,2-dichloroethane
1,1-dichloroethane
1,1,1-trichloroethane
1,1,2-dichloroethane
hexachloro ethane
Chloroform
Dichloroethylene
1,1-dichloro-
1,2-dichloro-
Dichloropropane
Dichloropropene
1,1,2,2-tetrachloroethane
Chloroethane (ethyl chloride)
Methylene chloride
Chlorinated solvents (con.)
Methyl chloride
Methyl bromide
Bromoform
Trichlorofluoromethane
Dichlorodifluoromethane
Hexachlorobutadiene
Hexachlorocyclopentadine
Hexachlorocyclohexane
Hexachloroethylene
Tetrachloroethylene
Trichloroethylene
Dichlorobromomethane
Chlorodibromomethane
Pesticides
Aldrin + metabolites
Dieldrin + metabolites
Chloradane + metabolites
Endrin + metabolites
PCB's
DDT+metabolites
Endosulfan + metabolites
Heptachlor* metabolites
Toxaphene + metabolites
Lindane
Metals
Antimony (total)
Arsenic (total)
Asbestos (fibrous)
Beryllium (total)
Cadmium (total)
Chromium (total)
Copper (total)
Cyanides (total)
Lead (total)
Mercury (total)
Nickel (total)
Selenium (total)
Silver (total)
Thallium (total)
Zinc (total)
19
-------
TABLE 2. CARRIER X
Xylene
o-Dichlorobenzene
Benzole ester
1,2,4-Trichlorobenzene
Toluic ester
Methyl salicylate
Butyl benzoate
(3-Methyl naphthalene
a-Methyl naphthalene
Diphenyl
o-Phenylphenol
Unknowns (as diphenyl)
Dimethyl phthalate
CONTROL OPTIONS
There are four options available for priority
pollutant control: chemical control; contractor
control; in-plant control; and treatment.
Chemical Control
Substitution —
The first and most obvious method of elim-
inating priority pollutant problems is the sub-
stitution of acceptable or treatable chemicals
for those presently that are undesirable. As ob-
vious as this action seems, many major compa-
nies have not yet begun a search of their pur-
chased chemicals for this purpose. The devel-
opment of alternate procedures and chemicals
is a long, time-consuming job, and the sooner
begun, the better the probability it will be
finished in time to avoid costly treatment.
But the textile industry will never be able by
itself to achieve the chemical substitutions re-
quired. Rather, the industry as a whole should
request that its chemical suppliers provide
alternates. The economic incentive of the mar-
ket is a sufficient driving force for the large
chemical companies and others to dedicate
research dollars to the solution of the
industry's problems. It is the industry's re-
sponsibility, however, to inform these people
in a clear manner of their absolute intention to
make such changes and to create the driving
force.
Process Change-
Process changes to minimize or prevent
pollution emissions have been an ongoing proc-
ess in the textile industry for the last decade.
A new mill designed under today's specifica-
tions uses 30 to 40 percent of the water that
would have been considered normal usage in
the recent past. Similar and possibly expanded
rates of process modification will be necessary
to completely close certain areas of the plant
that contain priority pollutants. Zero discharge
of undesirable materials should in fact be an at-
Investigations of process change require a
detailed understanding of the economic alter-
natives and the ability to invest money in ap-
propriate systems to eliminate long-range
problems. Such studies, as in the case of chem-
ical substitution, require significant lead time,
and the industry even now should be turning
its attention to these areas of control.
Studies start with a detailed material bal-
ance of the processes and require a high level
of analytical interrogation of the various proc-
ess streams. In virtually all cases, companies
will be forced to employ a higher level of ana-
lytical technology than is presently available
at the mills. This will undoubtedly call for a
much higher level of participation by the com-
panies' research groups.
Contractor Control
From the environmental point of view, em-
ployers are responsible for the activities of
contractors on their site. In the past, little or
no attention has been paid to the miscellaneous
materials used by such people. In the future,
tight specifications governing the use of envi-
ronmentally related chemicals must be exer-
cised. Contractors must be made aware of the
pollution control programs carried on by the
company and their activities must be closely
monitored. Contractors, in this sense, include
those performing everything from pest control
to construction and erection. Many of these
areas are particularly difficult since the con-
tractors one deals with are small, independent
individuals who are unaware of the composi-
tion of many of the products they employ. This
is an area where possible broad-based indus-
trial cooperation could lead to the evolution of
satisfactory standards.
20
-------
In-Plant Control
It is not the purpose of this discussion to deal
with the technical details of the problem; how-
ever, it is appropriate to at least mention them
considering the approaches that need to be
taken.
In-plant control begins with strict regulation
of water use. Virtually any compound is more
easily handled in higher concentration, and in-
creases in concentration often lead to econom-
ical recovery. Closed-loop solvent systems, re-
cycle of appropriate waters, filtration, and
reuse of streams present opportunities in vir-
tually every mill. Particular progress can often
be made if one recognizes that only certain
components of a particular stream are of in-
terest for reclamation. Consider dyeing solu-
tions, for example. Considerable attention has
been focused on the problem of dye removal. In
fact, a detailed analysis of discharge from
these operations shows that the principal
chemicals of concern are carriers, wetting
agents, dispersing agents, etc. Pursuing recov-
ery of these products is far more practical and
opens up the door to realistic reductions of
pollutants.
At-source treatment is also appropriate in
certain circumstances. Essentially this is the
philosophy of intercepting concentrated
streams before they enter the main sewer.
Often a 5,000- or 10,000-gal/d stream treated at-
source can eliminate the necessity of providing
tertiary treatment for a million gallons a day.
Such systems generally rely on adsorption or
related phenomena for the containment of par-
ticular materials. In evaluating such processes,
however, one must be certain to include the
full cost including the disposal of the collected
residues from each of the possible schemes.
Fabric finishing is an excellent example of an
area where at-source control is practical in
many cases.
It is worthwhile to note that the textile in-
dustry has a particular advantage in the recov-
ery of chemicals. In many industries the chem-
icals that are used undergo significant modifi-
cation or reaction during processing. In gen-
eral, the conditions of textile finishing are not
sufficiently severe to modify the basic chemical
structure of the materials involved in the proc-
ess. Therefore, the resulting byproducts are
basically the same as the starting materials in-
volved in the process except they are reduced
in concentration and contaminated with impur-
ities. This lack of change of chemical structure
should provide the basis on which recovery
programs are based.
Treatment
If all else fails, a company will be faced with
tertiary treatment for the removal of specific
organics. To date, activated carbon and other
specific adsorptive resins appear to be the
most promising methods of priority pollutant
control. But even these methods, in many
cases, allow the pass through of specific unde-
sirable products. Over the next several years a
variety of new treatment processes will be in-
troduced into the field to handle such prob-
lems. In virtually all cases, the processes will
be complex systems with high energy require-
ments. Great care must be expended in choos-
ing the appropriate device; looking in consid-
erable detail at its ability to meet the ever-
changing production mix present in this indus-
try.
SUMMARY AND RECOMMENDATIONS
Enforcement of priority pollutant removal
criteria will introduce a new era in the continu-
ing environmental problems of the textile in-
dustry. These pressures will also introduce
new opportunities for advancement among the
companies wise enough to find better solutions
to these problems. To be able to cope with this
situation in an effective manner, the following
course of action should be followed.
1. Know Your Problem
The solution of any problem of this nature
must be based on a fundamental under-
standing of the physical and chemical
properties of the species we are dealing
with. Detailed analysis of particular
materials is a costly, time-consuming pro-
gram and the temptation to bypass it is
always there. However, you must clearly
understand that sound fundamental infor-
mation is the foundation on which any suc-
cessful program must be based. Being
penny-wise and pound foolish is a very
costly business in the environment.
21
-------
2. Choose Your Ultimate Control System
First
Look down the road to ultimate control of
your effluent and determine what must
be done to achieve that position. Then, at
any point in time, do only that which is
necessary to meet the instantaneous re-
quirements of law. Do nothing incompat-
ible with the long-range plan. In this man-
ner, capital will not be wasted. Regula-
tory agencies are open to discussion of
evolving regulation when a company can
clearly demonstrate a long-range satisfac-
tory plan.
3. Use Purchasing and Technical Develop-
ment to Solve Your Problems
As discussed above, the best solutions are
based on elimination of the problem with-
in the manufacturing process, giving this
area the high priority research dollars it
deserves to avoid much greater costs in
capital later.
4. Build Only as a Last Resort
Too often companies assaulted by envi-
ronmental problems take the easy out —
spending money. This satisfies the regula-
tions for the time being and eliminates a
problem; but, in the long run, the treat-
ment plant consumes dollars every day
for the rest of that plant's life. If you can
use ingenuity and development rather
than assuming capital and operating
costs, you will be by far the winner.
22
-------
CONTACT FILTRATION FOR TEXTILE WASTEWATER
POLLUTION CONTROL IN POLAND
Jerzy Kurbiel, Wieslaw Zymon*
Abstract
Textile industry wastewater treated conven-
tionally in a mechanical-biological treatment
plant contains considerable amounts of refrac-
tory pollutants such as dyes, detergents,
and other synthetic compounds. Beyond
mechanical-biological treatment, they require
further treatment by means of physicochemi-
cal processes, especially when the receiving
stream is rather small and high water quality
must be maintained.
INTRODUCTION
In the investigations presented in this paper
the wastewater came from a textile factory
with annual fabric production of 9,000,000 kg,
in which the percentage participation of par-
ticular raw materials was: cotton, about 70 per-
cent and synthetics, about 30 percent. The
plant discharges about 6,000 m3/day of concen-
trated wastewater coming mainly from two
fabric finishing plants. Over 1,500,000 kg of
organic chemical compounds and over
4,600,000 kg of inorganic chemicals are used
per year in the production process.
After pH control by means of stack gases
recarbonization and after primary sedimenta-
tion, textile wastewater was mixed with
municipal wastewater in a 1:1 proportion or
higher and was directed to the pilot plant,
which contained an activated sludge unit and
tertiary treatment devices.
The coagulation in a contact filter was one of
the tertiary physico-chemical processes inves-
tigated within the Project PR-5-532-3 spon-
sored by EPA. In practice the coagulation of
textile wastewaters is often applied before
biological treatment, but in that case the re-
quired doses of the coagulant are very high
reaching, even to 1,000 g/m3.
*Cracow Technical University, Poland.
CHARACTERISTICS OF CONTACT
FILTERS
The contact filters were applied in practice
for the first time by Mine (ref. 1) in 1954 for the
treatment of water with constant and relative-
ly small contents of suspended solids (below
150 g/m3), including solids formed during
coagulation. The basic aspects of the contact
filter, usually constructed as single-medium,
multilayer, upflow filters, lies in the combina-
tion of two processes in one chamber: the sur-
face coagulation and filtration.
The surface coagulation, as opposed to the
conventional process, which can be called
volume coagulation, is characterized by a form-
ing gel on the surface of solids, in this case on
the grains of the filter medium. Surface coagu-
lation has two specific properties that make it
different and more favorable than volume co-
agulation: it is faster and^the arising gel, in the
form of floes, is separated from water
simultaneously with the .coagulation itself.
Also the required dose of coagulant is con-
siderably reduced, at least 30 percent in com-
parison to conventional coagulation.
The contact filters for the treatment of
municipal wastewater which have already
been treated biologically were first applied in
England in the early 1960's (ref. 2). England
and the Soviet Union have the greatest
research and practical experience in this area.
The theory of filtration has been evaluated
most comprehensively during the IWSA Con-
gress of 1966 by Mine (ref. 3) and the Congress
of 1969 by Ives (ref. 4). The objective of
research on the contact filtration done so far
was to find a filtration bed of a perfect struc-
ture needed to create the so-called deep filtra-
tion, in which the fixing of suspended solids
would take place within the whole mass of the
bed along its depth. In practice the construc-
tion of such a perfect bed was only partly suc-
cessful, in spite of many attempts.
In the downflow filter there is a domination
of the upper layer, which fixes most of the sus-
23
-------
pended solids. Even in case of properly se-
lected multimedia filters, the accumulation of
suspended solids at the boundary of layers
could be observed. However, the use of several
bed layers enables the increase of the capacity
of the bed, and as a result of it the filter cycle is
prolonged.
In England the "immedium filter" is used for
tertiary wastewater treatment. It is an upflow
contact filter with the granulation decreasing
as the water flows upwards and containing the
screen bounding from the top. Granulation and
filter bed thickness of an immedium filter, the
Boby Filter, is given in Table 1. Effectiveness
of those filters operated on a full scale (ref. 4),
determined by suspended solids and BOD5 con-
centration in effluent, was in both cases 7 mg/h.
TABLE 1. GRAIN SIZE AND FILTER BED
THICKNESS OF BODY FILTER (UK)(REF. 2)
Filter bed characteristics
Filter media
Gravel
Gravel
Gravel
Sand
Total thickness
Layer thickness
(m)
0.15
0.25
0.25
1.58
2.23
Grain size
(mm)
40-50
8-12
2-3
1-2
The research on the application of the con-
tact filters for tertiary treatment of biological-
ly treated wastewater has been carried out in-
tensively in the Soviet Union since the mid
1960's (refs. 5,6). Table 2 shows bed characteris-
tics and flow rate recommended by the Soviet
Union for wastewater treatment. The applica-
tion of contact filters for tertiary treatment of
municipal wastewater has been studied also in
Poland. Table 3 shows granulation and layer
thickness of contact filters at the Zywiec pilot
plant. This bed structure was proven best and
recommended as design criteria for waste-
water application (ref. 7). Removal of particular
pollution parameters on contact filters oper-
ated without addition of coagulants, according
to the results obtained at the Zywiec pilot
plant (ref. 7) was as follows: suspended solids-
up to 60 percent, BOD-28 percent, and COD-26
percent. When the coagulants were added, the
removal percentage increased considerably.
For the parameters listed above increase of
removal averaged 36 percent, 20 percent, and
24 percent, respectively. In addition to this, 80
percent of phosphorus removal was obtained.
The filtration cycle lasted up to 72 hours, but
because of oxygen depletion it was limited to
24 hours. It was proven that upgrading of con-
ventional municipal wastewater treatment
with the use of upflow contact filters is more
economical than any other system with the
same effectiveness with respect to the refrac-
tory pollutants, i.e., conventional coagulation.
METHODS
The range of investigations carried out on
application of contact filters for tertiary treat-
ment of mixed textile-municipal wastewater in-
cluded determination of the following factors
affecting the process and its effectiveness:
• the kind and dose of the coagulant on the
removal of specific refractory contami-
nants from the wastewaters,
• the influence of variations in the pH reac-
tion of the wastewater during coagula-
tions,
• the possibility of increasing the coagula-
tion efficiency by preliminary and post
TABLE 2. GRANULATION, BED THICKNESS, AND FLOW RATE RECOMMENDED
FOR UPFLOW CONTACT FILTERS APPLIED IN USSR (REFS. 5,6)
Filter bed characteristics
Filter media
(from bottom to top)
Gravel
Gravel
Gravel
Gravel
Sand
Grain size.
dmin
20
10
5
2
1.2
mm
dmax
40
20
10
5
2
Layer thickness
(m)
0.2-0.25
0.2-0.3
0.3-0.4
0.5-0.7
1.3-1.5
Filtration rate (m/h)
Normal
loading
7-8
Increased
loading
9-10
24
-------
TABLE 3. GRANULATION AND LAYER THICKNESS OF CONTACT FILTERS
AT ZYWIEC PILOT PLANT (REF. 7)
Filter media
(from bottom to top)
Gravel 1
Gravel 2
Gravel 3
Gravel 4
Sandl
Sand 2
Total thickness
Grain
dmin
20
10
4
2
1.5
0.8
size, mm
dmax
40
20
8
4
2
1.5
Layer
thickness
(m)
0.2-0.25
0.2-0.30
0.8-1.00
0.1-0.15
0.1-0.15
0.6-0.80
2.0-2.5
oxidation,
• the influence of the wastewater load ex-
pressed by refractories concentration,
and
• the type of installation assuring an op-
timal course of the process.
For the contact filters the following factors
were determined:
• filtration rate and washing flow rate, and
• grain size and depth of filtration bed.
In parallel, investigations were carried out
on a laboratory and pilot scale. A schematic
plan of the treatment plant comprising pri-
mary sedimentation of two separated streams,
textile and municipal wastewaters, activated
sludge unit for combined wastewater, and ter-
tiary treatment facility is shown in Figure 1.
The wastewaters, after biological treatment,
were coagulated by aluminum using two
methods:
1. Surface coagulation on contact filters
(Figure 2). The addition of coagulant salt
to the wastewaters introduced from the
bottom of the gravel-sand contact filter.
The coagulation took place in this case
directly in the filtration bed. A detailed
scheme of the contact filter is shown in
Figure 3.
2. Conventional coagulation in the reactor
with suspended floes is shown in Figure
4. After this process the wastewaters
were filtered through a downflow
anthracite-sand bed.
Supplementary test investigations were also
carried out on the flow conditions of the
wastewatcrs through filtration columns used
for contact coagulation. During the investiga-
tions physico-chemical analyses of the waste-
waters before and after coagulation were car-
ried out on daily averaged samples. The results
plotted on the scale of cumulated probability
gave, as a rule, a straight line that proved their
normal distribution.
FACTORS AFFECTING THE
COAGULATION PROCESS
Influence of the Kind of the Coagulant
Three basic coagulants FeS04 x 7 H20,
FeCl3 x 6 H20, and A12(S04)3 x 18 H20 were
used. Applying FeCl3 x 6 H20 and A12(S04)3 x
18 H20 in the same manner, the optimum dose
for removal of turbidity and color was 100
mg/liter. For FeS04 x H20 the dose exceeded
200 g/m3 in the majority of cases. Moreover,
ferrous salts induced a secondary pink color.
Further investigations on the coagulation proc-
ess were carried out using A12(S04)3 x 18H20.
The decrease in pH within the values 7.5 to
5.8 causes an increase in color removal from
the wastewaters. At the same time the pre-
liminary oxidation by means of NaOCl with a
dose up to 16 mg Cl2/liter increases the efficien-
cy of color removal in relation to the results ob-
tained without preliminary oxidation as well as
in relation to the oxidation itself, especially for
lower pH values. The doses of the oxidant ex-
ceeding 16 mg Cl2/liter do not give any increase
in color removal, especially at coagulant doses
above 200 g A12(S04)3 x 18 H20/m3.
25
-------
s.
1. Screen
2. Grit chamber
3. Primary clarifiers for municipal wastewater
4. Primary clarifiers for textile wastewater
5. Biological pilot plant
a PUot installation for tertiary
treatment processes
7. Overflow for control of industrial
to municipal wastewater ratio
JL J
temporary line
and pump
for study
I
I
Figure 1. Schematic plan of the primary treatment plant and the pilot plant in Andrychow.
26
-------
biological
effluent
o
o
§
I
3
•8
o
s
I
—® sampling
points
Figure 2. Schematic diagram of the pilot upflow filtration installation.
Influence of Wastewater pH,
Coagulant Dose, and Preliminary
Oxidation on Non-ionic
Detergent Removal
A high nonionic detergent content equaling
from several to more than 10 g/m3 was found in
the wastewaters after biological treatment.
Mean values from the obtained test investi-
gation data for five wastewater samples are
shown in Figure 5. From the shape of the
curves the resulting speculation is that in-
creased doses of the coagulant at a wastewater
pH of between 7.4 and 6.5 cause an increase in
detergent removal. At a pH of about 5.8,
removal of nonionic detergents is highest, but
does not depend on the dose of the coagulant
when exceeding 100 mg/liter. A lower waste-
water pH during coagulation causes a con-
siderable increase in nonionic detergent
removal. Investigations were also carried out
with six samples of wastewaters coagulated
with a dose of 200 g/m3 at a constant pH of 6.6
and previous oxidation with various doses of
NaOCl. The obtained results are represented
in Figure 6. Preliminary oxidation with doses
of about 5 mg Cl2/liter gave a considerable in-
crease in nonionic detergent removal. Increase
in the oxidant doses to a value above 5 mg
Cl2/liter caused a decrease in detergent
removal.
Influence of the Coagulant Dose,
Wastewater pH, and Preliminary
Oxidation on COD Removal
The data from the investigations presented
in Figure 7 show that increase in the coagulant
dose causes increase in COD removal. On the
other hand, a decrease in the pH from 7.4 to 6.5
causes an increase in COD removal by about 20
percent. A further decrease in the pH to 5.8 in-
creases the COD removal in relation to the
result at a pH of 6.5 by about 10 percent.
Figure 8 shows the influence of of preliminary
oxidation on COD removal under a pH of 6.6
and a coagulent dose of 200 mgAiter of
A12(S04)3 x 18 H20. In the majority of sam-
ples, contrary to the color, the initial decrease
in the COD removal was found at oxidant doses
up to 10 mg Cl2/liter. Bigger doses of the oxi-
dant caused an increase in the COD removal.
The pilot plant data have shown the positive in-
fluence of post chlorination on COD removal
and the positive effect on residual aluminum in
the effluent (Figure 9).
PILOT INVESTIGATIONS AND
DISCUSSION OF RESULTS
Technological investigations on coagulation
of a filtration bed were carried out for a num-
ber of systems that differed in construction of
27
-------
100mm
500 mm
flow contrcj
valve
Dosing
tank
y 3QOmm-,
effluent
_ JSOmm
^inlet pipp
head
degasification pipe
f wash water
Figure 3. Experimental contact filter for tertiary wastewater treatment at Zywiec pilot plant (ref. 7).
28
-------
^h
I
\
o
1,
a
o
u
sludge disposal
biological
•ffluent
Y
1
1
Figure 4. Schematic diagram of the pilot coagulation installation.
100
90-
^ 80-
| 70-
i so
0>
O)
0)
T3
so-
§ 40^
•«M
§
c 30-
20-
10-
pH-5.8
100 200 300 400
coagulant dose mg Al2(S04)3X 18
Figure 5. Average removal of nonionic detergents in relation to the coagulant
dose and the pH values in wastewater.
29
-------
Initial concentration of detergents
. 16.0, 14.3, 13.0mg/l
X 10.3, 8.2, 6.3mg/l
initial detergent
concentration
pH: 6.6
initial alum dose: 200 mg/l
20 30
dose NaOCLmgCi2/l
Figure 6. Influence of preliminary oxidation on nonionic detergent removal
in wastewater after coagulation.
100
90
*80
2
I 70
860
o
50
AO
30
20
10
pH-5
TOO 200
coagulant dose mg
300
WX)
Figure 7. Average COD removal in relation to coagulant and pH value.
30
-------
S 10 20 30
dose NaOCI mg CI2/I
Figure 8. Influence of preliminary oxidation before coagulation on COD removal.
(Tests for six wastewater samples.)
pH: 6.6
initial alum dose: 200 mg/l
dose NaOCI mg CI2/I
Figure 9. Influence of oxidation with NaOCI on removal of residual
aluminum from coagulated wastewater.
31
-------
filter media. The bed systems that gave posi-
tive results during investigations are charac-
terized below:
System I — Structure of the filtration bed
and grain size of particular layers in the
system were as follows:
• height - 0.15 m, gravel depth - 2
to 4 m
• height - 1.5 m, gravel depth — 1.2
to 2 m
• height — 0.15 m, sand depth — 0.4
to 0.8 m.
The mean dose of the coagulant was up to
200 mg/liter of aluminum and the time of
wastewater storage after addition of chemi-
cal compounds was about 4 minutes. The
process was carried out in parallel in two
columns, with the pH being lowered in one
of them by means of hydrochloric acid. The
average filtration cycle was about 24 hours.
Table 4 presents mean results of the opera-
tion of a pilot installation for 24 days at a pH
of under 6.8. The obtained results of con-
taminant removal were satisfactory. Re-
moval of biodegradable organic substances
was fairly great and equaled, on the
average, 87 percent. Organic refractory
compounds determined by COD accom-
plished 65 percent removal.
The influence of preliminary oxidation on
the obtained treatment effectiveness is evi-
dent in Table 5. An increase higher than 10
percent was also found in color removal
after preliminary oxidation of the waste-
water with a dose of 10 to 14 mg Cl2/liter at a
contact time of 15 minutes.
System H — The structure of bed medium
and the granular size of particle layers in
this system were as follows:
• height — 0.4 m, gravel depth — 3 to
6 mm
• height — 1.3 m, sand depth — 1.2 to
2 mm
• height — 0.2 m, sand depth — 0.4 to
0.8 mm.
The applied aluminum dose equaled 250 to
270 mg/liter. In spite of a higher filtration
rate the cycle was about 1 day. In this sys-
tem a series of investigations was carried
out by using hydrochloric acid at a de-
creased wastewater reaction. The waste-
waters, after coagulation on contact filter,
were oxidized with NaOCl with a dose of 15
to 20 mg C^Aiter and subsequently filtrated
through an anthracite sand bed at a rate of 3
m/h. The results are presented in Table 6.
A decrease in the wastewater reaction
from a pH of 7.3 to a pH of 6.8 caused a par-
ticularly evident increase in the removal of
red color measured by extinction at a wave-
length equal to 400 mm. Oxidation of postco-
agulation wastewaters with secondary
filtration caused an increase in the removal
percent of all parameters.
TABLE 4. CONTAMINANT REMOVAL DURING FILTRATION WITH SIMULTANEOUS
COAGULATION FOR LOWER pH
(AVERAGE RESULTS BASED ON 24 DAYS OF OPERATION )
Biologically
treated effluent
Parameters
PH
Alkalinity, meg/liter
Permanganate COD,
mg 02/liter
COD, mg 02/liter
BODs, m9 02/liter
Suspended solids.
mg/liter
Color, mg pt/liter
Color threshold
Average
concentration
7.8
7.8
18.5
72
13.6
32
57
19:100
Standard
deviation
0.2
2.2
7
26
4.6
16
23
6:100
Coagulated
effluent
Average
concentration
6.5
3.4
9.2
25
1.7
13
21
60:100
Standard
deviation
0.3
1.6
3.6
17
1.1
10
12
30:100
Removal
(percent)
50
65
87
56
60
32
-------
The combined treatment system, which in-
cluded the use of contact filters, was examined
in order to check the combined operation and
effectiveness of selected tertiary processes
evaluated as optimum (Figure 10). Average
effectiveness of this system for COD para-
meters is given in Table 7.
TABLE 5. CONTAMINANT REMOVAL DURING FILTRATION WITH SIMULTANEOUS
COAGULATION APPLIED TO WASTE WATER PRELIMINARILY OXIDIZED
WITH NaOCI (AVERAGE RESULTS BASED ON 28 DAYS OF OPERATION)
Biologically
treated effluent
Parameters
PH
Alkalinity, meq/liter
Permanganate COD,
mg 02/liter
COD, mg 02/liter
Suspended solids.
mg/liter
Color, mg pt/liter
Color thresh old
Average light
absorption, E
Average
concentration
7.7
8.3
19.0
80
50
56
20:100
0.093
Standard
deviation
0.2
3.2
3.5
29
22
15
7:100
0.040
Coagulated
effluent
Average
concentration
6.4
3.3
7.9
22
9
17
69:100
0.030
Standard
deviation
0.4
1.6
1.3
6
7
8
31:100
0.026
Removal
(percent)
58
72
82
70
68
TABLE 6. CONTAMINANT REMOVAL DURING FILTRATION WITH SIMULTANEOUS COAGULATION,
FOLLOWED BY OXIDATION WITH NaOCI AND SUBSEQUENT FILTERING (AVERAGE RESULTS
BASED ON 14 DAYS OF OPERATION )
Biologically
treated effluent
Parameters
PH
Alkalinity, meq/liter
Permanganate COD,
mg 02/liter
COD, mg 02/liter
Suspended solids,
mg/liter
Average light absorp-
tion, E X=400 nm
Average light absorp-
tion, E X=550 nm
Average
concentration
8
8.6
24
100
20
0.350
0.250
Standard
deviation
0.2
1.8
6
25
15
0.120
0.100
Coagulated effluent
Average
concentration
6.8
4
13
40
10
0.150
0.160
Standard
deviation
0.3
2.2
3
13
7.5
0.070
0.090
Coagulated, oxidized
and secondarily filtered
Average
concentration
6.8
4
9
30
2
0.040
0.010
Standard
deviation
0.3
2.2
3
10
1.5
0.030
0.005
Removal
during
coagulation
process
(percent)
45
60
50
57
36
Total
removal
(percent)
62
70
90
88
98
33
-------
TABLE 7. AVERAGE EFFECTIVENESS OF THE COMBINED SYSTEM FOR COD
Kind of sample
Concentration
mg 02/liter
Removal in
unit processes,
in percent
Cumulative
removal,
in percent
Biological
Contact coagulation
effluent
After filtration and
preliminary oxidation
with NaOCI
After adsorption
After ozonation
100
40
30
6
5
60
25
80
16
60
70
94
95
Figure 10. Schematic diagram of the pilot installation for combined treatment processes.
CONCLUSIONS
• The method of surface coagulation on con-
tact filters lends itself more to coagulation
of biologically treated wastewater than
conventional coagulation performed in a
reactor, with suspended floes. This is an
effective and economical method of ter-
tiary treatment.
• Preliminary oxidation before the coagula-
tion process increases the removal of color
and nonionic detergents.
• Decrease in pH produces better removal
from refractory pollutants.
• Aluminum salts should be used for
coagulation of wastewater, since use of
ferric salts caused secondary wastewater
color.
• Before the coagulation process, prelim-
inary treatment of textile wastewater is
necessary for pH control and efficient
removal of biodegradable organics in the
activated sludge process.
REFERENCES
1. D. M. Mine, W. P. Krisztul, and A. M.
Pierlina, Mietod Kontaletnoyo Osbietlenia
Wody, "Method of Contact Clarification of
Water," Sanitarnaja Technika, Moskva,
1954.
2. G. Alpe and A. Barret, "U.K. Developments
34
-------
in Upflow Titration," Fifth Federal Conven-
tion of the Australian Water and Waste-
water Association, Adelaide, 1972.
3. D. M. Mine, "Modern Theory of Filtration,"
International Water Supply Congress, Spe-
cial Subject No. 10, Barcelona, 1966.
4. K. J. Ives, "Theory of Filtration," Interna-
tional Water Supply Congress, Special Sub-
ject No. 7, Wien, 1969.
5. G. J. Iranijuszin, Primienienie pieszczanych
filtroir na stancjoch alracji, "Application of
Sand Filters at the Sewage Treatment
Plants," Wodosnabzenie, Sanitarnaja
Technika, No. 6, 1966.
6. N. A. Lukinych and B. L. Lipman, Dooc-
zistka Gorodskich Stocznych vod na filtrach
z voschodieszczim potokom, "Tertiary
Municipal Wastewater Treatment on Up-
flow Contact Filters," Naucznye Trudy, No.
105, Ak. Ch., Moskva, 1976.
7. H. Kloss-Trebaczkiewicz, Zastosowenie
filtroir kontaktozych do doczyszczania
scieleow biologiernie ocryszcronych, "Ap-
plication of Contact Filters to Tertiary
Treatment of Biological Effluents," Doc-
toral Thesis, Warsaw Technical University,
Warsaw, 1978.
35
-------
USE OF POWDERED ACTIVATED CARBON FOR TEXTILE
WASTEWATER POLLUTION CONTROL
David G. Button, Francis L. Robertaccio*
Abstract
Plant-scale performance tests show that the
addition of powdered activated carbon to ac-
tivated sludge plants treating textile wastes
significantly improves effluent quality. The
combined process has also been found to signif-
icantly reduce certain priority pollutants. Ex-
perience shows that the combined activated
carbon-activated sludge process can be easily
scaled up. Studies suggest that carbon regen-
eration is not required for an economically
practical process, but may be suggested for im-
proved economics and as an efficient means of
sludge disposal Performance data from Du
Font's 1.75 m3/s (40 MOD) wastewater treat-
ment plant treating waste from dye and textile
chemicals manufacture are presented. Carbon
regeneration performance will also be dis-
cussed.
INTRODUCTION
Best practicable treatment (BPT) is now well
over a year behind most of us. The EPA
reports that 90 percent of industry met their
goal of July 1, 1977 for operation of BPT
facilities. While surely there have been permit
violations from time to time, the literature
does not suggest any widespread problem of
meeting NPDES permits for the conventional
pollutants required under BPT. Most of the
textile industry utilizes biological treatment,
which tends to give log-normal distribution of
effluent quality data for any particular plant.
Hence, most plants can expect to periodically
have some permit violations.
Now our thoughts turn toward coping with
the provisions of the Clean Water Act of 1977.
While the requirements are still in the formula-
tion stage, it seems clear that three types of
technology based effluent standards will be re-
quired on or before July 1, 1984. These are:
• Conventional pollutants — BOD, suspended
solids, etc., that are now in our permits;
plus, perhaps, something like total organic
carbon (TOO, dissolved organic carbon
(DOC), or chemical oxygen demand (COD).
• Toxic compounds — Any of the 129 specific
compounds derived from the 65 classes of
toxic compounds incorporated in the Clean
Water Act of 1977. These are listed in
Table 1, for reference.
• Nontoxic nonconventional pollutants — A
third group of pollutants, which are not
specifically listed as either toxic or con-
ventional. Such pollutants may include col-
or, turbidity, chlorides, TDS, etc.
This paper will compare the effectiveness of
a combined powdered activated carbon-
activated sludge treatment (Du Pont PACTt
Process) with conventional activated sludge
for all three classes of pollutants. Economics
will also be discussed.
CONVENTIONAL AND
NONCONVENTIONAL POLLUTANTS
Du Font's Chambers Works
While Du Font's Chambers Works is not a
textile mill, it is a dye and textile chemical
manufacturing site and has many, if not most,
of the pollutants in the wastewater that one
would find at a textile mill. In June 1977,
Chambers Works began operation of its waste-
water treatment plant to comply with best
practicable technology (BPT) regulations. This
wastewater treatment plant consists of neu-
tralization and settling for primary treatment
followed by combined powdered activated
carbon-activated sludge (Du Pont PACT Proc-
ess) for secondary/tertiary treatment. The
wastewater, varying from 1.0 to 2.2 m3/s (23 to
50 MGD) during the first year of operation, is
fed to the process without equalization (Figure
1). Wastewater is pumped into the neutralizers
where lime is added for neutralization. Poly-
*E. I. du Pont de Nemours, Inc., Wilmington, DE.
tDu Pont service mark for powdered activated carbon
treatment.
37
-------
TABLE 1. PRIORITY POLLUTANTS
Chlorinated alkanes
Methyl chloride
Methylene chloride
Methyl bromide
Chloroform
Bromoform
Carbon tetrachloride
Dichlorobromomethane
Trichlorofluoromethane
Dichlorodifluro methane
Chlorodibromomethane
Chloroethane
1,1-Dichloroethane
1,2-Dichloroethane
1,1,1-Trichloroethane
1,1,2-Trichloroethane
1,1,2,2-Tetrachloroethane
Hexachloroethane
1,1-Dichloroethylene
1,2-trans-Dichloroethylene
1,2-Dichloropropane
1,2-Dichloropropylene
Tetrachloroethylene
Vinyl chloride
Hexachlorobutadiene
Hexachlorocyclopentadiene
Chlorinated aromatics
1,2,4-Trichlorobenzene
Chlorobenzene
Hexachlorobenzene
2-Chloronaphthalene
1,2-Dichlorobenzene
1,3-Dichlorobenzene
1,4-Dichlorobenzene
Chlorinated ethers
bis(Chloromethyl) ether
2-Chloroethyl vinyl ether
4-Bromophenyl phenyl ether
bis(2-Chloroethoxy) methane
bis(2-Chloroethyl) ether
4-Chlorophenyl phenyl ether
bis(2-Chloroisopropyl) ether
Aromatics
Benzene
Toluene
Ethylbenzene
Naphthalene
Fluoranthene
Aromatics (con.)
Acenaphthene
Benzo(a)anthracene
Benzo(a)pyrene
Chrysene
lndeno(1,2,3-c,d)pyrene
3,4-Benzofluoranthene
Benzo(K)fluoranthene
Acenaphthylene
Anthracene
Benzo(ghi)perylene
Fluorene
Phenanthene
Dibenzofa, h)anthracene
Pyrene
Phthalate esters
bis(2-ethylhexyl)phthalate
Butyl benzyl phthalate
Di-n-butyl phthalate
Di-n-octyl phthalate
Diethyl phthalate
Dimethyl phthalate
Phenols
Phenol
2-Chlorophenol
2,4-Dichlorophenol
Pentachlorophenol
2-Nitrophenol
2,4-Dimethylphenol
4-Nitrophenol
2,4-Dinitrophenol
4,6-Dinitro-o-cresol
2,4,6-Trichlorophenol
para-Chloro-meta-cresol
Substituted aromatics
Nitrobenzene
2,4-Dinitrotoluene
2,6-Oinitrotoluene
2,3,7,8-Tetrachiorodibenzo-p-dioxin
Benzidine
3,3'-Dichlorobenzidene
1,2-Diphenyl hydrazine
Polychlorinated biphenyls
Miscellaneous
Acrolein
Acrylonitrile
Asbestos
Cyanide
Isophorone
N-Nitrosodimethylamine
N-Nitrosodiphenylamine
N-Nitrosodi-n-propylamine
Pesticides
Aldrin
Oieldrin
Chlordane
4,4-DDT
4,4-DDE
4,4-DDD
a-Endosulfan-alpha
b-Endosulfan-beta
Endosulfan sulfate
Endrin
Endrin aldehyde
Heptachlor
Heptachlor epoxide
a-BHC- alpha
b-BHC-beta
r-B H C( Lindane) -gamma
g-BHC-delta
Toxaphene
Metals
Antimony
Arsenic
Beryllium
Cadmium
Chromium
Copper
Lead
Mercury
Nickel
Selenium
Silver
Thallium
Zinc
PCB-1242
PCB-1254
PCB-1248
PCB-1221
PCB-1232
PCB-1260
PCB-1016
38
-------
electrolyte is added to the neutralizer effluent
to aid in coagulating the fines prior to clarifica-
tion. Primary sludge settles and thickens in the
clarifiers, is pumped to storage, and finally is
pumped to a filter press for concentration
before disposal in a lined landfill.
Powdered activated carbon and return
PACT sludge are added to the primary effluent
as it is fed to the aeration tanks. Polyelec-
trolyte is added to the aerator effluent to aid in
coagulating fines in the final clarifier. Treated
effluent combines with other plant streams in a
settling lagoon and is discharged to the Dela-
ware River.
Excess PACT sludge is wasted from the
aerator effluent into a thickening and storage
tank. From here, the sludge is pumped to filter
presses. The 40 percent solids cake obtained by
filtration is then conveyed to a multiple hearth
furnace where the bacteria are burned and the
powdered carbon regenerated. Regenerated
carbon is acid washed to prevent inert solids
buildup, and then returned to the aerators.
Makeup carbon is added as required.
Typically, the feed to the Chambers Works
wastewater treatment plant has the composi-
tion shown in Table 2. The wastewater can be
characterized as a strongly acidic, highly col-
ored, medium strength waste. It has a low
BOD/DOC ratio, suggesting a high ratio of non-
biodegradable compounds. The wastewater is
somewhat brackish and has a relatively low
suspended solids content.
For a portion of that year, December 14,
1977 through May 30, 1978, a laboratory acti-
vated sludge unit and a plant PACT tracking
unit were run in parallel with the wastewater
treatment plant (WWTP). The plant tracking
unit operated at the same temperature, carbon
type, and dose as the plant. For this time frame
not only can plant performance be observed,
but scaleup information and a comparison of
PACT versus activated sludge can be obtained.
A schematic of the operation during this time
frame is shown in Figure 2.
BOD5 performance is shown in Table 3 and
Figure 3. The feed BOD5 averaged 174
mg/liter, virtually the same as the primary ef-
fluent. However, as noted in Figure 3, the feed
BOD5 distribution was very peculiar, at-
tributed to the somewhat biologically inhibi-
tory nature of this waste, and to the inability to
Primary
Lime Polymer
r»tv,n & " •'" '*'• * PfCJ ?r(XS3B •-'. • •
:, ,:,«ili
Chambers
\Afcrks
Solids
Handling
Hydrochloric
Add
Figure 1. Schematic diagram of Chambers Works wastewater treatment system.
39
-------
TABLE 2. CHAMBERS WORKS WASTEWATER TREATMENT
PLANT FEED
Flow, liters/sec
PH
Acidity, mg/liter
Total suspended solids, mg/liter
Total dissolved solids, mg/liter
Total BODg, mg/liter
DOC, mg/liter
Color, APHA units
December 1977-May 1978
1,530 (24,200 gpm)
1.98
1,050
111
3,000
174
194
1,400
Feed
w
Plant
Primary
Treatment
-*
Full scale
PACT
Laboratory
Activated
Sludge
Laboratory
PACT: Plant
Tracking
^ ^t*f\ »
> Effluent
^Effluent
Figure 2. Parallel operation of units at Chambers Works.
40
-------
TABLE 3. CHAMBERS WORKS SOLUBLE BOD PERFORMANCE
(AT -100 PPM CARBON OOSE)
Full Scale
Treatment Plant
No. of data points
Average BOD 5, mg/liter
Standard deviation,
mg/liter
Feed
79
174
93
Primary
effluent
115
175
74
PACT
effluent
121
5.1
3.8
Lab
PACT:plant
tracking
effluent
76
5.0
7.5
Lab
activated
sludge
effluent
78
10.3
9.9
01
O)
&
600
500
400
300
T§ 200
100
Bio Effluent = Laboratory unit Activated
~ " "•" Sludge Effluent
Lab Tracking = Lab Unit PACT: Plant
r~i"p%- Tracking Effluent
WWTP Effluent = Full scale Treatment
«-"-*« plant Effluent
>
,-f
/ Feed
X primary
/ /
X
X
X
X
X
X
Bio -•»«- LabTracking
assrjrLT^-*1*^^, WWTP Effluent
.1 .5 1.0 2.55.010.0 25.0 50.0 75.0 90.095.097.599.099.5 99.9
Cumulative Percent
Figure 3. Chambers Works soluble BOD performance.
always get meaningful BOD5 data. Effluent
BOD5 from the plant averaged 5.1 mg/liter ver-
sus 5.0 mg/liter from the lab tracking unit. As
you can see, this represents virtually perfect
scaleup from the laboratory 7-liter aerator to
the plant's 45,400 m3 (12 million gallons) of
aeration. The activated sludge unit also did
well during this 6-month period, but not as well
as the PACT units.
Chambers Works has a stringent DOC limit
in its permit, so there is a lot of interest in this
parameter. As with the BOD5 data, matched
laboratory and plant data for a greater than
6-month period can be compared. Average
DOC feed to the plant was 194 mg/liter, with an
average effluent of 34.5 mg/liter (Table 4). Only
a minor amount of DOC was removed in pri-
mary treatment. Figure 4 shows the histogram
41
-------
plots of the data, with the envelops showing
the best log normal fit of the data. While there
is some difference between the laboratory and
plant results, this difference is well within one
standard deviation, and is small compared to
the difference between PACT and the acti-
vated sludge unit. With activated sludge alone,
Chambers Works could not meet its DOC per-
mit limit.
Chambers Works already has a permit limit
for color, a nonconventional parameter. The
matched laboratory and plant samples show a
color increase over primary treatment and no
removal of color by the activated sludge pro-
cess (Table 5). The carbon addition in the
PACT units effects about 60 percent color
removal at the carbon dose used, which is
enough for Chambers Works to meet its per-
mit. Again the log normal distribution of data
is seen in Figure 5.
Acidic Dye and
Fine Chemical Manufacturer
Similar results to the above Chambers
Works case were reported by ICI, United
States, for a small acidic dye and fine chemical
plant. Following lime neutralization, the waste
was treated by a conventional extended aera-
tion activated sludge process 817 m3/day (0.216
MGD) design flow plant. Two carbons were
tested at 85 ppm dosage levels, which resulted
in equilibrium aerator carbon levels of 1,700
mg/liter. The carbon addition improved both
COD and phosphorus removal (Table 6), com-
pared to the control (ref. 1).
Eastern Pennsylvania Municipality (ref. 6)
Tests of carbon addition were conducted at
an eastern Pennsylvania municipality that
TABLE 4. CHAMBERS WORKS DOC PERFORMANCE
(AT ~100 PPM CARBON DOSE)
\lo. of data points
Average DOC, mg/liter
Full Scale
Treatment Plant
Primary PACT
Feed effluent effluent
183 192 198
194 186 34.5
Lab
PACTiplant
tracking
effluent
177
29.0
Lab
activated
sludge
effluent
175
61.5
Standard deviation,
mg/liter
35.2
30.5
9.4
8.7
15.2
TABLE 5. CHAMBERS WORKS COLOR PERFORMANCE
(AT -100 PPM CARBON DOSE)
Full Scale
Treatment Plant
Feed
\lo.ofdatapoints 119
\verage color, APHA 1,130
Primary
effluent
142
1,410
PACT
effluent
153
460
Lab
PACT:plant
tracking
effluent
165
420
Lab
activated
sludge
effluent
160
1,410
Standard deviation,
APHA 410
450
180
190
490
42
-------
WWTP Feed
80
80
Primary Effluent
c:
o>
CT
LL
/
/
/
1
1 WWTP Effluent
1
160
160
240
320
400
c
QJ
CT
UL
/„
h
\ Lab Tracking Effluc
1
IL.
240
320
400
OJ
I
.
\ Lab Bio Effluent
nnjjc
. Ill JnirTT^a . r_ , 1 I
80
160
240
320
400
Figure 4. Chambers Works dissolved organic carbon performance (concentrations in mg/l).
43
-------
WWTP Feed
1600 2400 3200 4000
Primary Effluent
1600 2400 3200 4000
1
CD
LJ_
f
I
M
I
WWTP Effluent .
OS- -
0
800 1600 2400 3200 4000
Lab Tracking Effluent
1600 2400 3200
4000
o
Lab Bio Effluent
1600 2400 3200
4000
Figure 5. Chambers Works color performance (concentrations in mg/l).
44
-------
TABLE 6. EFFECT OF CARBON ADDITION TO ACIDIC DYE AND FINE
CHEMICAL MANUFACTURER WASTE*
"Hydrodarco C"
'Hydrodarco H"
Control
Average flow, m^/day
Average food/microorganism
Aerator detention, hours
Average influent COD, mg/liter
Average effluent COD, mg/liter
Average percent COD removed
Average effluent SS, mg/liter
Average influent P, mg/liter
Average effluent P, mg/liter
Average percent P removed
530
.42
69
1,106
672
39
39
13
2.7
79
318
.15
114
774
332
57
30
4
0.6
85
530
.42
68
1,585
1,155
27
76
21
12
43
'Reference 1.
TABLE 7. COMPARISON OF METALS REMOVAL AT
CHAMBERS WORKS*
Cadmium, jug/liter
Chromium, mg/liter
Copper, mg/liter
Lead.jug/liter
Mercury, jug/liter
Nickel, mg/liter
Zinc, mg/liter
WWTP
feed
19.5
0.31
0.39
252
0.51
0.12
1.38
Primary
effluent
7.4
0.04
0.16
32
<0.1
0.08
0.52
Plant
PACT
effluent
6.5
0.02
0.05
21
<0.1
0.06
0.32
•Full-scale treatment plant.
received 70 percent of its feed from a textile
dyeing and finishing mill. A principal problem
at this plant was overloading, with daily
average flows from 0.03 to 0.08 m3/s (0.75 to 1.8
MOD) and peaks to over 0.09 m3/s (2 MGD) in a
plant designed for 0.045 m3/s (1 MGD).
Powdered carbon additions of from 20 to 25
ppm, resulting in aerator concentrations of 900
mg/liter, increased BOD removals from 70 per-
cent to 90 percent (Figure 6).
PRIORITY POLLUTANTS
Metals
The studies of Dyes Environmental and Tox-
icology Organization, Inc. suggest that a num-
ber of metals are found in the textile industry.
Du Font's Chambers Works has some of these
in its wastewater, where removal by primary
treatment and the PACT process can be ob-
served. As seen in Table 7, good removal of all
metals was obtained. It is to be noted, how-
ever, that most of the metals removal is accom-
plished by lime neutralization to pH 7.0. In all
cases the metal removal is governed by the
final solubility, with neither activated sludge
nor carbon being particularly effective for ad-
sorbing metals.
Nonmetals
Chambers Works has a lot of data on removal
of phenols from wastewater. This is not
C6H5OH, as measured by gas chromatography/
mass spectrometry (GCMS), but phenolics as
45
-------
100
90
o
70
60
50
S 40
30
20
10
i i
Hydrodarco HJ^**
Control
1 5 10 30 50 70 90 95 99
Cumulative Percent
Figure 6. Effect of powdered carbon on BOD removals.
measured by ASTM Method 510, a class test.
As seen in Figure 7, the Chambers Works
Wastewater Treatment Plant gives 95 percent
removal of phenolics, with a residual phenolics
concentration about 20 percent less than that
seen by activated sludge. Phenolics are basical-
ly quite biodegradable, and the difference bet-
ween PACT and activated sludge may not be
significant.
For other organic priority pollutants Cham-
bers Works has some data on removal, as indi-
cated in Table 8. Excellent removals by PACT
have been observed. In addition, a 99 percent
removal of acrylonitrile by the PACT process
was observed in a laboratory study.
Berndt and Polkowski (ref. 3) report on the
comparison of activated sludge and PACT with
air oxidation for carbon regeneration (Table 9).
The improved removal of PACT is readily
seen. In this scheme it is reasonable to expect
that individual priority pollutants can be col-
lected at the source, co-fed with PACT sludge
to the regeneration unit, and destroyed.
46
-------
WWTP Feed
Avg. = 3.99
Primary Effluent
Avg. = 4.21
WWTP Effluent
Avg. = 0.22
Lab Tracking Effluent
Avg. = 0.16
Lab Bio Effluent
Avg. = 0.27
Figure 7. Chambers Works phenolics performance (concentrations in mg/l).
47
-------
TABLE 8. REMOVAL OF ORGANIC PRIORITY POLLUTANTS
BY PACT PROCESS AT CHAMBERS WORKS
Concentration in ppb
4-Nitrophenol
Phenol
Toluene
Ethylbenzene
Chlorobenzene
WWTP
feed
56
440
680
29
1,900
WWTP
effluent
N.D.*
17.7
4.1
6.5
12
Percent
removal
>99
96
99
78
99
*N.D. = not detectable.
TABLE 9. COMPARISON OF PACT AND ACTIVATED SLUDGE
FOR ORGANICS REMOVAL*
Effluent concentration
Organic
Chlorinated pesticides, mg/liter
Organo-sulfur pesticides, mg/liter
Organo-phosphate pesticides, mg/liter
PCB's, mg/liter
PACT
0.017
0
1.23
0.008
Activated sludge
0.35
15.0
3.03
0.13
* Reference 3.
ECONOMICS
Gulp and Shuckrow (ref. 4) did an extensive
economic study that compared, among other
processes, the cost of a 0.44-m3/s (10-MGD) acti-
vated sludge and a PACT process with wet air
oxidation for carbon regeneration. While these
economics were for a municipal plant with ex-
tensive primary treatment, the primary treat-
ment investment can be split out to give the
comparison shown in Table 10. For reference,
the Chambers Works PACT process cost about
$26 million ($159 per m3/day at 151,400 m3/day
capacity or $602,000/MGD at 43.2 MOD capaci-
ty). Thus the PACT process cost, including car-
bon regeneration, is 25 percent above a conven-
tional activated sludge process, and 7 percent
above a nitrifying activated sludge process.
They projected the cost of the carbon feed
system at $207,000 for the 0.44-m3/s (10-MGD)
plant, including engineering, interest, etc.
Operating costs were also projected for a
0.44-m3/s (10-MGD) municipal wastewater
treatment plant (Table 11). In this case it was
not possible to split out the cost of primary
treatment. On an overall basis, the PACT pro-
cess costs were 15 percent higher than conven-
tional activated sludge and 6 percent higher
than nitrifying activated sludge.
CARBON REGENERATION
A primary question often asked about the
PACT process is: Can you regenerate the car-
bon? The immediate response to this is: Do the
economics justify it? Nearly everyone has a
wastewater treatment system already, most
likely a biological treatment. Anyone with a
biological system has a sludge disposal system.
Adding carbon very likely will not add to the
volume of sludge, only the dry weight. Hence,
the PACT sludge can be processed in the same
manner as the activated sludge. In fact, ex-
perience shows that PACT can be processed
more easily.
In a study Chambers Works did for the pe-
troleum industry, carbon addition to an ex-
isting activated process at a dosage of 75 ppm
48
-------
TABLE 10. INVESTMENT FOR 10 MGD MUNICIPAL SECONDARY
WASTEWATER TREATMENT PLANTS (SECONDARY ONLY)*
PACT with
wet air
oxidation
Activated
sludge
(conventional)
Activated
sludge
(nitrification)
Secondary treatment,
$ thousands
6,388
5,124
5,986
Investment
$ Per m3/day
$ Per MGD
169
639,000
135
512,000
158
599,000
TABLE 11. OPERATING COST FOR 10 MGD MUNICIPAL WASTEWATER
TREATMENT PLANT (INCLUDING PRIMARY TREATMENT)*
PACT with
wet air
oxidation
Total annual operating
cost, (in thousands of dollars) 1,580
Dollars/m^ treated 0.114
Dollars/million gallons
treated 433
Activated
sludge
1,380
0.100
378
Nitrifying
activated
sludge
1,500
0.108
410
^Reference 4.
(75 pounds of carbon added per million pounds
of wastewater treated) was assumed. At a 30
percent carbon yield loss during regeneration,
the carbon value alone could not justify carbon
regeneration at plant scales below 0.88 m3/s (20
MGD). Even then, only a 10 percent net return
on investment was achieved, or a payout time
of 5 years.
On the other hand, carbon regeneration can
be very attractive if it also serves as a means
of ultimate disposal of bacterial sludge. Any of
several methods can be utilized to regenerate
carbon from waste PACT sludge. Chambers
Works has evaluated four methods: wet air ox-
idation, multiple hearth furnace, transport
reactor, and fluidized bed. The high tempera-
ture methods appeared to be more rigorous
when the Chambers Works plant was de-
signed. However, recently laboratory tests
with wet air oxidation have produced a satis-
factory regenerated carbon.
Chambers Works chose multiple hearth fur-
nace regeneration, and installed a 5-hearth fur-
nace 7.6 meters (25 ft) in diameter. While a lot
of problems with sludge handling, particularly
feeding, have hampered demonstration of what
the multiple hearth furnace can do, one thing is
clear; powdered carbon from the PACT proc-
ess can be regenerated. Chambers Works is
able to dispose of its waste biological solids,
and has been able to recycle regenerated car-
bon. Data from the period February through
June 1978 from the Chambers Works furnace
are shown in Table 12. As the furnace in-time
increases, the carbon recovery is expected to
increase. The carbon properties from the
mixed carbons being regenerated are very
near one of the carbons being purchased (Table
13). Thus, the regenerated carbon quality is
satisfactory.
From February through June 1978, carbon
purchases averaged 111 percent of that pro-
jected in the technical manual for the waste-
water treatment plant. A lower carbon dosage
requirement compensates for the low regener-
ated carbon yield. Within the next year
49
-------
TABLE 12. CHAMBERS WORKS CARBONS
REGENERATION EXPERIENCE
Regeneration furnace in time, hours/day 16.6
Average feed rate, tons/day of dry solids 18.5
Carbon recovery, in percent 58
TABLE 13. CHAMBERS WORKS REGENERATED CARBON PROPERTIES
Decolorization index
Iodine number
Regenerated
carbon
12.9
514
Virqin carbons
Type A
21.5
983
TypeB
15
504
Type C
13.6
581
Chambers Works may be purchasing less car-
bon than expected when the plant was de-
signed.
CONCLUSION
In conclusion, full-scale plant tests show that
the combined powdered activated carbon-
activated sludge process (PACT process) not
only does better than activated sludge on
removing conventional pollutants (BOD, DOC,
etc.), but is also more effective at removing
nonconventional pollutants such as color. The
PACT process can remove many metals to low
levels. Preliminary results show that the
PACT process is also very effective for remov-
ing some organic priority pollutants. PACT
process carbon can be regenerated if econom-
ics justify the investment.
REFERENCES
1. "Powdered Hydrodarco Activated Carbons
Improve Activated Sludge Treatment,"
ICI United States Bulletin PC-4, October
1972.
2. Alan D. Adams, "Carbon Adsorption Aids
Activated Sludge Treatment at Dyeing
and Finishing Plants," American Dyestuff
Reporter, Vol. 65, No. 4, pp. 32, 34, 36, and
38, 1976.
3. Craig L. Berndt, and L. B. Polkowski, "Pact
Upgrades Wastewater Treatment," Water
& Wastes Engineering, pp. 48-50, May
1978.
4. Gordon L. Gulp, and A. L. Schuckrow,
"What Lies Ahead for PAC," Water &
Wastes Engineering, pp. 67-74, February
1977.
50
-------
ACTIVATED CARBON ADSORPTION FOR TEXTILE
WASTEWATER POLLUTION CONTROL
E. J. Schroeder, A. W. Loven*
Abstract
Carbon adsorption technology may be uti-
lized for advanced wastewater treatment to im-
prove the effluent quality beyond what can
generally be achieved with biological treat-
ment and to treat wastewaters not amenable to
biological treatment. Carbon adsorption is pri-
marily effective for the removal of dissolved
organic materials and, therefore, will reduce
the concentration of such parameters as bio-
chemical oxygen demand (BOD), chemical oxy-
gen demand (COD), total organic carbon (TOC),
and specific organic constituents. Carbon ad-
sorption processes will likely play an impor-
tant role in future treatment systems in the
removal of toxic pollutants as well as conven-
tional pollutants.
INTRODUCTION
Full-scale carbon adsorption systems are
now being utilized in several municipal and in-
dustrial wastewater pollution control applica-
tions. Carbon adsorption technology has been
successfully utilized and proven in the waste-
water pollution control field, and although it is
not effective in all cases, it is capable of re-
moving a wide range of organic pollutants. Its
major disadvantage has been cost (dollars per
pound of pollutant removed) as compared to
biological treatment. Therefore, carbon proc-
esses are generally used only when specific ef-
fluent requirements justify the greater ex-
penditure, biological treatment does not work,
or if it is economical to reuse the treated ef-
fluent.
Currently there are few waste treatment in-
stallations reported with full-scale carbon ad-
sorption processes in the textile industry as
compared with such industries as organic
chemicals, dye manufacturing, and pesticides.
However, there have been several extensive
studies and experimental programs on carbon
*Engineering-Science, Inc., Atlanta, GA.
adsorption processes in the treatment of tex-
tile wastewaters and a good understanding has
emerged concerning application and effective-
ness. It should be understood that the textile
industry has a wide variety of manufacturing
processes that contribute to significantly dif-
ferent wastewater characteristics among the
plants even in a similar category. Also the
large variety of dyes, sizing agents, dispersing
agents, finishing materials, etc. have an in-
fluence on wastewater characteristics. These
waste characteristics definitely impact on the
treatment effectiveness and economic feasibil-
ity of carbon adsorption systems.
CARBON ADSORPTION PROCESSES FOR
WASTEWATER TREATMENT
With the present state of the technology and
economic factors, there are two basic types of
carbon adsorption treatment processes that
may be considered for end-of-pipe application
with textile waste treatment; granular acti-
vated carbon adsorption (GAC), and powdered
activated carbon (PAC) with activated sludge.
Each of these processes has certain advan-
tages which should be tested and evaluated for
a particular application.
GRANULAR ACTIVATED CARBON
ADSORPTION
A typical treatment scheme utilizing a GAC
tertiary treatment system (i.e., following acti-
vated sludge secondary treatment) is shown in
Figure 1. A multimedia filter (MMF) is recom-
mended prior to the GAC columns to reduce
the concentration of total suspended solids in
the influent to the GAC system. The lower
total suspended solids concentration allows the
GAC system to perform more efficiently by not
"blinding" the carbon surface. The MMF also
provides protection for the GAC unit in the
event of an upset in the biological treatment
system which might result in a shock loading to
the GAC system of total suspended solids
51
-------
being discharged from the secondary clarifier.
For illustration purposes a conceptual proc-
ess flow diagram of a 2-stage, countercurrent
pressure downflow GAC system is shown in
Figure 2. This system consists of three ad-
sorber units, a backwash sump, and influent
and backwash pumps. In this scheme the ad-
sorbers are operated with two units in series
and the third adsorber as a standby unit. When
the carbon is exhausted in the first adsorber it
is taken out of service and the standby unit is
placed in the number two series position. The
exhausted adsorber is charged with regener-
ated or virgin carbon and then assumes the
standby position. The GAC effluent is dis-
charged to a backwash sump before final dis-
charge. The backwash sump serves as a reser-
voir of water for backwashing the adsorbers.
Backwash water containing carbon fines and
other suspended solids is discharged into the
biological system.
Carbon regeneration can be provided offsite
through a regeneration contract service or
may be performed with onsite regeneration
furnace facilities. Large carbon usages
(approximately 2,000 Ib/day) are required to
economically justify onsite regeneration facil-
ities.
Powdered Activated Carbon
With Activated Sludge
One of the most promising applications of
PAC is in the activated sludge process. The
PAC with activated sludge process is a recent-
ly developed modification to the activated
sludge process that has been demonstrated to
significantly upgrade the effluent quality of
the conventional activated sludge treatment
system. In this process, powdered activated
carbon is added to the activated sludge aera-
tion basin thus enhancing the organic removal
through adsorption and also by some interac-
tion with the biological system. Compared to
other tertiary treatment processes, PAC with
activated sludge may have considerable eco-
nomic advantages. The primary economic ad-
vantage is the low capital expenditure re-
quired to convert an existing activated sludge
treatment plant to a PAC with an activated
sludge treatment process. However, operating
costs may be high depending on the desired ef-
fluent quality and the required carbon dosage.
Although carbon regeneration for the PAC
with activated sludge process has been demon-
strated to be technically feasible, it may be
several years before regeneration equipment
develops to the point of making regeneration
economically attractive.
A typical treatment scheme utilizing a PAC
treatment process is shown in Figure 3. The
aeration basin should have sufficient mixing to
maintain the carbon and mixed liquor in sus-
pension. The solids loading on the secondary
clarifier is increased because of the additional
carbon in the mixed liquor; therefore, coagu-
lant usage may be required to enhance solids
settling and, hence, prevent overloading the
COLUMN BACKWASH
RAM
WASTEWATER
FROM
TEXTILE'
PLANT
FILTER BACKWASH
1
BIOLOGICAL
TREATMENT
(AERATION BASIN)
i
RETURN
s
/SECONDARY^
H CLARIFIER I •
SLUDGE
I
WAS
SLU
MULTIMEDIA
FILTER
GRANULAR
ACTIVATED
CARBON
ADSORPTION
SYSTEM
IM^— ^^BV
i I
1 1 CARBON
1 L MAKEUP
* i
TE
DGE
CARBON
REGENERATION
TREATED
EFFLUENT TO
RECEIVING
'STREAM OR
RECYCLED
FOR REUSE
(ONSITE OR OFFSITE OPTION)
Figure 1. Typical wastewater treatment scheme utilizing a GAC tertiary system.
52
-------
miivt u*u> FOR mnaai
IKFUEHT FROf
MULTI-MEDIA FILTER
TO PLANT JNFLUEKT
CARBOK CHARGE
(SLURRY)
CARBOS DISCHARGE
BACKWASH SUMP
Figure 2. Conceptual process flow diagram activated carbon adsorption.
ADDITI
OF ROW
ACTIVA
RAM
WASTEWATER
FROM ,
ON
DERE
TED
TEXTILE
PLANT
D
CARBON
AERATION
BASIN
POLYM
ADDIT
(OPTI
ER
ION
ONAL)
/SECONDARY'S
/ /•>( noTi-tr-n \ ~.
MULTIMEDIA
FILTER
(OPTIONAL)
* RETURN SLUDGE
WASTE SLUDGE
FILTER BACKWASH
TREATED
EFFLUENT TO
RECEIVING
"STREAM OR
RECYCLED
FOR REUSE
Figure 3. Typical wastewater treatment scheme utilizing a PAC
with activated sludge treatment process.
secondary clarifier. Powdered activated car-
bon -'.oded to the mixed liquor means more
waste sludge is produced for disposal. How-
ever, the powdered carbon may improve de-
watering characteristics of the sludge although
sometimes the reverse has been shown to be
'rue. The addition of multimedia filtration
after secondary clarification is an option that is
dependent on the desired effluent total sus-
pended solids level.
PAC has also been applied with limited suc-
cess on a tertiary treatment process either in a
53
-------
single-stage rapid mix-clarification step, as in a
reactor clarifier, or in a multistage countercur-
rent application. This approach has not been
proven economically competitive with GAC
due to poorer inherent adsorption efficiency
and problems in regeneration.
EXPERIMENTAL TESTING OF CARBON
ADSORPTION WITH TEXTILE
WASTEWATER
Extensive indepth carbon adsorption ex-
perimental studies have been conducted over
the last 2 years with textile wastewaters in a
project jointly sponsored by the American
Textile Manufacturers Institute, Northern
Textile Association, Carpet and Rug Institute,
and the U.S. Environmental Protection Agen-
cy. As part of this project, GAC was evaluated
for tertiary treatment with pilot-scale equip-
ment at 19 textile plants that were chosen as
representatives of the various subcategories in
the industry. The PAC with activated sludge
process was also evaluated with bench-scale
equipment using wastewaters from 10 repre-
sentative textile plants. The results from these
studies indicate potential effectiveness of car-
bon adsorption for textile wastewater treat-
ment.
Granular Activated Carbon
Experimental Studies
The GAC tests were conducted in two
mobile pilot plant trailers that were trans-
ported to the 19 textile plant sites during a
17-month period from May 1977 to September
1978. Effluents from the existing biological
treatment plants were used as influents to the
pilot plant units. Hence, variability of best
practical treatment (BPT) plant effluent qual-
ity was incorporated into the technical evalua-
tion of pilot-scale performance.
The GAC pilot equipment consisted of a mul-
timedia filter unit followed by three carbon col-
umns operated in series. A sketch of the filter
unit is presented in Figure 4. The flow rate to
the 13.5-in. diameter pilot filter was varied be-
tween 1 and 7 gpm, which represented a sur-
face loading rate of 1 to 7 gpm/ft2. The op-
timum surface loading rate for the filter (based
on head loss, effluent quality, and filter run
time) was determined during an initial 2- to
3-week screening period at each plant site, ihe
filter was operated at a selected optimum rate
(if secondary effluent characteristics remained
relatively constant) during the remainder of
the experiment at that site.
A sketch of a single carbon column is shown
in Figure 5. The carbon selected for the in-
dividual experiments was based on a prelim-
inary comparison of TOC and color adsorption
isotherms for each wastewater with both a lig-
nite- and bituminous-based carbon. Three pilot
carbon columns were operated in series with
an empty-bed retention time of 45 minutes
(total) unless individual wastewater charac-
teristics dictated a longer or shorter retention
time. Samples were collected before and after
each column in order to monitor breakthrough
and exhaustion by developing breakthrough
curves. Single-cycle carbon regeneration tests
were performed on the spent carbon samples
from 12 different sites.
The COD reduction performance with the
GAC pilot units is summarized by a bar chart
in Figure 6. COD is one of seven promulgated
effluent guideline parameters and it is the
parameter that can generally be effectively
reduced with carbon adsorption. The GAC
process obtained the highest percent COD
reduction in Subcategories II (Wool Finishing),
V (Knit Finishing), and VI (Carpet). The GAC
process was not as effective for overall per-
centage COD reductions in Subcategories I
(Wool Scouring), IV (Woven Fabric Finishing),
and VII (Stock and Yarn Dyeing). With the ad-
dition of the GAC process for tertiary treat-
ment it can be predicted that the presently
promulgated best available technology eco-
nomically achievable (BATEA) guideline lim-
itations for conventional pollutants can be
achieved on a technical basis in Subcategories
I, II, V, and VI, but could not be achieved in
Subcategories IV and VII. This does not mean
that the GAC process is the most cost-effective
treatment alternative, indeed there may be
other less expensive processes that are suit-
able, only that it is technically feasible in the
Subcategories mentioned above (I, II, V, and
VI). The economic effectiveness and economic
impacts of the application of these technologies
is still under review by participating organiza-
tions.
The results from the carbon regeneration
54
-------
SURFACE WASH
PO
12"
12"
16"
DRAIN
INFLUENT I
X
SIGHT GLASS
O-
(PRESSURE
INDICATOR
ANTHRACITE COAL
0.9 - 1.5 mm
13.5"
. • -SAND . -
-0.4 •- 0.8-mm
.. •••/••
v • •
/-• °f GRAVEL
t/4" x 5/8"
AIR SCOUR
BACKWASH DRAIN
r
-STEEL COLUMN
BACKWASH
-|X]——EFFLUENT
Figure 4. Sketch of pilot multimedia filter unit.
55
-------
INFLUENT
1X1 -1
GRAB SAMPLE
PORT
CARBON DRAIN
AIR SCOURIX}
BACKWASH [X]
SCHEDULE 80 PVC
EFFLUENT
Figure 5. Sketch of pilot carbon column unit.
tests, after one adsorption-regeneration cycle,
indicated the carbon was successfully regener-
ated in all cases. No unusual or major problems
are indicated in regenerating the granular car-
bon, based on this preliminary data.
Wastewaters from Subcategory IV (Woven
Fabric Finishing) and Subcategory VII (Stock
and Yarn Dyeing) were not particularly ame-
nable to carbon adsorption for COD removal.
Throughout both subcategories, problems re-
lated to nonadsorbability of certain COD frac-
tions as well as COD "stripping" were ob-
served. Identification of specific compounds
and wastewater characteristics contributing to
this effect demands further detailed study.
Powdered Activated Carbon With
Activated Sludge Experimental Studies
The PAC with activated sludge experiments
were conducted on a bench-scale basis as a pre-
liminary screening evaluation to determine the
56
-------
TOO
90
80
§ 70
i— i
1—
l_>
g 60
£
g 50
0
z 40
UJ
g
£ 30
20
10
TEST °
PLANT
SUBCATEGORY
•
1
I
- - ; - ,
2
I
:::::
3
I
^_
5
.••
6
—
7
8
p
9
10
11
::::;
:::::
12,
Y"
IV
r*n
:::::
13
<
14
~^~
V
JJLi
EFI
GU
EFI
BA1
16
VI
:LUENT ACHIEVED BATEA
DELINE VALUES
:LUENT COULD NOT ACHIE1
fEA GUIDELINE VALUES
17
VII
V
18
III
19
IX
WOOL WOOL
SCOURING FINISHING
WOVEN FABRIC FINISHING
KNIT FABRIC CARPET STOCK COM. MULTI-
FINISHING AND YARN FIN. OPS.
DYEING
Figure 6. Summary of COD reduction performance with GAS pilot units.
technical feasibility of the process as applied to
textile wastewaters. Ten textile plants repre-
senting the various subcategories were se-
lected for the study. Raw waste samples were
shipped weekly to the Engineering Science,
Inc. laboratory in Atlanta, Georgia during the
study period. Several different test conditions,
primarily with regards to carbon dosage, were
evaluated simultaneously in the lab.
Each plant experiment was conducted utiliz-
ing three 10-liter plexiglas bioreactors (similar
to the one shown in Figure 7). The three bio-
reactors were seeded with activated sludges
from the textile mill and a municipal/industrial
wastewater treatment plant and subsequently
acclimated to the textile wastewater. After ac-
climation was achieved, TOC isotherms were
performed on the bioreactor effluent to select a
powdered carbon (lignite or bituminous base)
and carbon dosage range to evaluate. The car-
bon dosages were typically in the range of
1,000 to 10,000 mg/liter (in the mixed liquor) in
ordei to bracket the expected technically effec-
tive range. Carbon was added at the selected
high and low carbon dosages to two of the bio-
reactors and the system was allowed further
time to acclimate. The third bioreactor was
operated as a control unit. Data was collected
and evaluated after steady-state was achieved
in the three bioreactors.
The results of the PAC with activated
sludge screening tests with the ten textile
plant wastewaters are summarized in the bar
charts in Figure 8. The PAC with activated
sludge treatment process was generally suc-
cessful in improving the effluent quality in
bench-scale evaluations of wastes from Subca-
tegories II, IV, V, VI, and VII, but was not ef-
fective with waste from Subcategory I. The
treatment effectiveness in Subcategory I
(Wool Scouring) was probably reduced by the
higher oil and grease levels in the raw waste-
water. Based on the favorable results from the
screening studies, a second level of onsite pilot-
scale testing at selected plants from Subcate-
gories IV and V has been recommended.
SUMMARY
Recent studies have been completed eval-
uating carbon adsorption technology for ad-
vanced waste treatment of textile waste-
waters. Granular carbon adsorption pilot tests
were conducted with secondary effluents from
57
-------
AIR SUPPLY
NEEDLE VALVE
AIR DIFFUSER LINES
ING BAFFLE
OVERFLOW DRAIN
FEED CONTAINER
PUMP
Figure 7. Experimental biological reactor.
19 representative textile wastewaters. Granu-
lar carbon adsorption pilot tests were con-
ducted with secondary effluents from 19 repre-
sentative textile plants, and bench-scale
powdered activated carbon with activated
sludge screening tests were conducted with
ten textile wastewaters. The following obser-
vations were developed from evaluating these
experimental studies.
The granular activated carbon pilot studies
demonstrated reasonable effectiveness for
reducing secondary effluent COD concen-
trations (63 to 84 percent) of textile plants
in Subcategories II, V, and VI, but only
marginal COD removal effectiveness (14 to
80 percent) was achieved in Subcategories
I, IV, and VII.
From the results of the granular activated
58
-------
NT OF COD
PAC TREATMENT
% IMPROV
REDUCTION WI
100
90
80
70
60
50
TEST PLANT
SUBCATEGORY
LOW CARBON DOSAGE
10
o
1 23 457 14
15
16
17
1 II IV V VI VII
o
-------
BIOLOGICAL TREATMENT OF TEXTILE WASTE BY AERATION
T. A. Alspaugh, John Hodges, Arthur Toompas*
Abstract
Biological treatment systems have shown
the adaptability and flexibility to handle
almost any textile waste mixture. This blend-
ing of prehandling, treatment, waste character-
istics, and effluent results, necessary with
space requirements and operation costs, has
produced the numerous textile wastes treat-
ment systems found in the industry. These sys-
tems have been shown to handle shock loads
and vary wastes characteristics, while pro-
ducing good results at a reasonable construc-
tion and operating cost.
INTRODUCTION
The words "textile wastes" mean the
wastewater from the manufacturing and proc-
essing of textile fibers or fabrics containing
natural and/or synthetic fibers. In more
specific terms lint, trash, grease, and other
residue from fiber cleaning; wastes from the
manufacturing or greige operations, including
size material; the wastes from desizing, scour-
ing, bleaching, and mercerizing before dyeing
and finishing; all dye bath dumps, overflows,
and rinses; finish bath dumps, overflows, and
rinses; machine and equipment spills, cleanups,
and the like. The mixture will be warm with
the temperature depending on the use of heat
exchangers. The chemical content will depend
upon the type fiber, fibers, or fabric processed.
The natural fibers usually encountered in
the United States are wool and cotton. Waste-
water from wool preparation will be very
strong, high in grease, BOD, COD, total sus-
pended solids (TSS), heavy metals, and color,
but usually low in pH. The contaminant mate-
rial removed from wool by scouring is the ma-
jor problem. Wool finishing wastes will be
much weaker in all aspects with a near neutral
pH. Cotton, on the other hand, produces a
wastewater high in pH due to caustic use in
mercerizing and/or cleaning the fibers. The
BOD will be moderate, the TSS usually on the
low side with trace heavy metals and with
varied color, but high COD coming from the
processing chemicals. Synthetic fiber process
wastewater will be near neutral in pH, low in
TSS, moderate BOD and COD and, with a weak
color. The major problems are the materials
used to aid the dye in penetrating and staying
on the fibers.
A total mixture then can have a low to high
pH depending on the processed fiber, will be
Colored, and can contain grease, fibers, lint,
and trash from the fiber. The mixture will con-
tain processing chemical residues, size, pene-
trants, wetting agents, detergents, carriers,
dyes, finishes, etc., all mixed in the water used
in processing to rinse or wash the fabric or
fiber, to dye it, and to finish it. (Tables 1, 2, and
3)
TABLE 1. RAW WASTES ANALYSIS •
BLEND OR COTTON
pH
BOD,mg/l
COD, mg/l
TSS, mg/l
Color
Cr, mg/l
Zn, mg/l
TKN, mg/l
TP04, mg/l
10-12
600-1,000
1,000-2,000
70-100
Varied (dark)
0-0.50
0-1.0
10-20
10-15
"Chemical Engineer, Engineering Department, Cone
Mills Corporation, Greensboro, NC.
METHODS OF BIOLOGICAL TREATMENT
There are a number of treatment alternates
available for the biological treatment of textile
wastes. Based on work carried out within the
industry, any of these waste mixtures can be
treated satisfactorily either by aeration sys-
tems or biological filters if the necessary
prehandling is done. This prehandling involves
61
-------
pH adjustment in some cases and the addition
of nutrients (nitrogen and/or phosphorus). In
others, possible removal of lint, grease, or
trash, equalization if needed, and removal or
replacement of processing chemicals when tox-
ic to the biological treatment process.
Except for oxidization ponds, most of the dif-
ferent aeration type biological systems for
treatment of textile wastes can be found listed
in the Metcalf & Eddy, Inc. book entitled
Wastewater Engineering: Collection, Treat-
ment, Disposal under the heading "The Ac-
tivated Sludge Process." These include conven-
tional, complete-mix, step-aeration, modified-
aeration, contact stabilization, extended-
aeration, Kraus process, high-rate aeration,
and pure oxygen systems.
TABLE 2. RAW WASTES ANALYSIS - SYNTHETIC
TABLE 3. MIXED WOOL WASTE
PH
BOD, mg/l
COD, mg/l
TSS, mg/l
Color
Cr, mg/l
Zn, mg/l
TKN, mg/l
TP04, mg/l
6-9
200-400
400-800
40-80
Varied (lighter)
0-0.3
0-1.0
5-15
5-10
Scouring
Finishing
pH
BOD, mg/l
COD, mg/l
TSS, mg/l
Cr, mg/l
Zn, mg/l
P04, mg/l
TKN, mg/l, mg/l
Phenol
2-5
2,300
6,000
2,500
—
—
—
—
—
5-7
180
740
200
0.6
—
—
—
.04
The information included in Table 4 indi-
cates the various systems that have been used
for textile waste treatment. (Figures 1 through
9)
Aerobic biological treatment sounds simple
enough but has required extensive laboratory,
pilot work, and full-scale systems to convince
regulatory agencies that it would work, and
was both feasible and practical. This early
work developed the minimum maintenance,
long aeration time period activated sludge
type waste treatment system generally found
throughout the textile industry today. This
system uses the aeration time period to buffer
the waste, blend it, absorb the shock loads, and
allow acclimation of biological organisms to the
TABLE 4. OPERATING CHARACTERISTICS OF WASTE TREATMENT SYSTEMS FOR TEXTILES
Trickling filter
Bio-disc
Oxidization ponds
Conventional AS
Completely mixed AS
Step aeration
Contact stabilization
Extended aeration
Oxygen/activated sludge
Detention
period
10-20 mm
Depends on
tank size
5-50 days
6-8 h
2448 h
6-8 h
1-h contact
3-8 stab.
24-120 h
8-24 h
Flow
model
Pluj
Completely
mixed
Plug flow
Plug flow
Completely
mixed
Plug flow
Plug flow
Completely
mixed
Completely
mixed
Recycle
1/1 to 4/1
.3/1 to 1/1
0-1/1
.3/1 to 1/1
1/1 to 3/1
.3/1 to 1/1
.3/1 to 1/1
1/1 to 3/1
.3/1 to 1/1
Aeration
type
NA
NA
Wind currents
or mech. aerators
Diffused air,
mech. aerators
Diffused air.
mech. aerators
Diffused air
Diffused air.
mech. aerators
Diffused air.
mech. aerators
Mechanical
aerators
BOD removal
efficiency %
50-85
85-95
6085
80-90
85-99
70-85
80-90
75-95
85-95
Application
Susceptible to shock loads; used for weak
strength textile wastes, mixed domestic
sewage, and textile wastes
Useful for plant expansions, medium strength
textile wastes, in combination with activated
sludge for strong wastes
Weak strength textile wastes, susceptible to
shock loads
Susceptible to shock loads, weaker textile
wastes, and/or domestic wastes (not used
frequently)
General application, resistant to shock loads
General application (not used frequently)
Flexible; useful for plant expanstons,small
to medium plant, package plants (not used .
frequently)
Variable strength wastes, small and medium
sized plants
General application on all textile wastes, use
where land space is limited
Source: Reference 1.
62
-------
material. The method and/or methods used
depends on the type and strength of the waste,
land availability, power cost, labor cost, effi-
ciency required, etc. In general, the processes
used for textile wastes treatment have tended
to fall into the completely mixed or extended
aeration category or a hybrid of these two,
with some scattered biological filter systems.
Completely mixed systems are most often
used for strong, highly colored wastes, usually
with detention aeration periods of from 24 to
48 hours. They tend to carry higher mixed liq-
uor suspended solids (MLSS) than the other
systems and usually require sludge disposal.
They most often use mechanical aeration but if
not too large and designed properly they can
use diffused air. Extended aeration systems on
the other hand can be either completely mixed
or plug flow, depending upon the type aeration
and raw waste distribution devices. Extended
aeration uses low organic loadings and much
longer aeration time periods; up to 5 days in
some cases to handle the wastes. Since the
space required to treat a large volume with
high or moderate organic concentrations by
this method is large, extended aeration is most
often used for small to medium plants or
weaker wastes unless land is available, (ref. 1)
Influent
Aeration Sasin
Recycle Sludge
Clan'f ier
Effluent
Waste Sludge
Figure 1. Conventional activated sludge system.
Influent
Aeration Basin
o o b
Mechanical Aerators
O O O
Clarifier
_-. Effluent
Recycle Sludge
_»- Waste Sludge
Figure 2. Completely mixed activated sludge system.
63
-------
Influent.
Clarifier
Aeration Tank
Recycle Sludge
Effluent
*. Waste Sludge
Figure 3. Step aeration activated sludge system.
Influent.
Stabilization Tank
Clarifier
Contact
Tank
Recycle Sludge
Effluent
*- Waste Sludge
Figure 4. Contact stabilization system.
Influent
Aeration Basin
o o
Mechanical Aerators
O O
Recycle Sludge
Clarifier
Effluent
-•-Waste Sludge
Figure 5. Extended aeration activated sludge system.
64
-------
Bio Disc Units
Irf ,
'erati-n ark
riarifie--
DTD DO
moo
-*• refluent
Re--y~le Sludge
_». Waste "ludie
Figure 9. Bio-disc activated sludge system.
Oxidation ponds are shallow lagoons or
basins, usually not over 6- to 7-ft deep and even
as shallow as 4 ft, and use natural processes to
treat the waste with or without the use of me-
chanical means for mixing or aeration. Solids
will settle to the bottom of the lagoon or pond
and if so, are decomposed anaerobically. The
suspended organic solids are decomposed by
aerobic bacteria that use surface aeration.
Algae can grow on the surface of lagoons of
this sort using the nitrogen and phosphorus in
the incoming waste for nourishment. A system
of this type can be a dual aerobic/anaerobic
system; aerobic on top, anaerobic on the bot-
tom. This type of system has a limited use in
particular situations for weak, easily oxidized
wastes where land is available and the climate
is warm. (ref. 2)
The pure oxygen aeration system is becom-
ing more popular as a means of treating highly
concentrated waste. The aeration basin, in
general, will be smaller, possibly covered, and
contain oxygen mixed with the waste. This
process results in a more rapid production of a
biological floe, which reportedly has better set-
tling characteristics, and is said to be more effi-
cient than a system using air as its oxygen
source. The molecular sieve is one device being
used to generate the pure oxygen onsite. As
electric power costs increase, oxygen systems
are forecast to become more prominent.
The trickling filter is another biological
treatment system that has found some use in
certain situations. Trickling filters are normal-
ly beds of media 5- to 20-ft deep with influent
spray distribution systems and underdrain sys-
tems for effluent collection. The sprays dis-
charge onto the bed of filter medium which con-
tains the bacteria or other biota. This slime ab-
sorbs and breaks down organic matter in the
wastewater. The concentration of the influent
waste is diluted by recirculating a portion of
the filter effluent back to the incoming waste.
A major difference between activated sludge
and trickling filters is that the trickling filter
uses a series of biological communities at vari-
ous levels in the unit and each series has its
own type of life and its own degrees of
removal.
Strong waste, however, can cause sloughing
off of matter and biomass from the media.
Although recirculation is widely practiced and
forced air used in some cases, the trickling
filter normally lacks the inherent buffering
capacity of activated sludge. Waste that is not
readily biodegradable will not be removed ef-
fectively by trickling filters due to the shorter
detention period in contact with the active
organisms.
The rotating biological disc is an effort to
overcome some of the disadvantages of the bio-
logical trickling filter. It is a series of closely
spaced perforated discs rotating partially sub-
merged in waste and the dripping, turning disc
aerates and mixes the waste and the active
organisms. This process tends to overcome the
lack of long-term contact inherent with the con-
ventional filter. In some cases rotating discs
have been added to activated sludge plants as a
means to increase treatment. Many textile
plants that have been operating for some
period of time can become overloaded as pro-
duction increases. Rotating discs may be added
to upgrade an existing plant as an improve-
66
-------
ment step and provide a large amount of sur-
face area in a small space for bacteria to attach
to and thus increase removal. Bio-discs usually
have to be covered as they are affected by air
temperatures and cold weather.
Numerous relationships have been devel-
oped to explain the actions of the various aero-
bic biological systems. For activated sludge
these relationships may include:
• Waste strength vs. mixed liquor suspended
solids concentration,
• Waste strength vs. detention time,
• pH reduction as function of C02 produc-
tion by aeration,
• Color removal vs. mixed liquor suspended
solids concentration,
• Organic removal vs. mixed liquor suspend-
ed solids concentration,
• Effluent heavy metal concentration vs.
sludge concentration,
• Effluent heavy metal concentration vs.
sludge removal rate,
• Efficiency of removal vs. aeration deten-
tion time,
• Efficiency of removal vs. cost of power,
and
• Removal rate as a function of recircula-
tion.
And for biological filters and rotating discs,
they may include:
• Removal rate as a function of recircula-
tion,
• Removal rate as a function of media
volume or surface area,
• Waste strength vs. detention time (Figure
10),
• Removal rates vs. disc speed, and
• pH reduction vs. recirculation rate.
CONTROL PARAMETERS FOR PLANT
DESIGN & OPERATION (Figures 10 and 11)
Food-to-Mass Ratio
The waste loading or food-to-mass (F/M)
ratio is used to control the rate of biological ac-
tivity in an activated sludge system. The F/M
ratio is the amount of BOD applied per day
a
o
a
DC
Detention Time, Days
Figure 10. Effect of increased detention time on cost and removal.
67
-------
9 c
._ m
l!
Ul O
Detention Time, Days
Figure 11. Effect of increased detention time on treatment efficiency.
divided by the amount of volatile suspended
solids in the aeration basin. As a general rule,
high F/M ratios tend to give better settling
solids. However, if F/M ratios are too high,
some hard to break down BOD may not be
broken down and removal efficiencies may suf-
fer. If low F/M ratios are used, one may expect
better organic removal efficiencies, although a
poor settling sludge sometimes occurs. In
general, textile waste can accept lower F/M
loadings than municipal waste (0.05 to 0.15),
however, ideal levels should be set on a case-
by-case basis.
Recycle Rate
The recycle rate is the volume of liquid
returned sludge divided by the incoming
volume of wastewater to be treated. This rate
can be varied in order to prevent solids from
leaving the plant in the effluent. If the return
rate is too low then septic solids along with
solids in the effluent sometimes can be ex-
pected. If the return rate is too high, problems
with hydraulic loadings and difficulties in
maintaining an adequate sludge blanket can oc-
cur. Textile wastes normally require return
rates of 100 to 300 percent while municipal
waste treatment plants use lower recycle
rates.
Solids Level
The solids level refers to the concentration
of MLSS in the aeration system. At low levels
it is difficult to achieve high removals of some
organics and high pH or hard-to-treat wastes
often cause problems due to a limited buffering
capacity. At high solids levels increased color
and COD removal is often found along with
some problems in suspended solids removal.
Again, the MLSS level can be higher when
treating textile waste than municipal waste,
levels of 2,500 to 4,000 ppm are typical, but
4,000 to 8,000 are not uncommon.
Sludge Volume Index
The sludge volume index (SVI) utilizes infor-
mation on sludge volume and dry basis solids
to give a useful operating parameter. The SVI
is the percent volume of sludge settled in 30
68
-------
minutes divided by the percent of suspended
solids in the aeration basin. The sludge density
index is 100 divided by the SVI. This para-
meter is not too useful in very high MLSS.
Sludge Age
Sludge age is the pounds of TSS in the aera-
tion tank per day per pounds of suspended
solids added. Sludge age is intended to be an in-
dicator of the activity of the sludge. When
sludge is too old it may be considered inactive
and may also concentrate undesirable products
such as heavy metals and/or toxicants. Low
sludge age operation normally causes a more
rapid growth of solids which in turn can re-
quire more disposal of solids. Again, very long
aeration periods with low production rates of
solids can tend automatically to have high
sludge ages.
Detention Time
Detention time in an aeration basin or filter
is the volume of the basin or filter divided by
the daily flow rate. This parameter is extreme-
ly difficult to increase once a plant has been
built, unless the designer happened to sec-
tionalize the system and build in capacity. Most
textile wastes require longer detention times
than municipal wastes. Aeration basins with 1
to 3 days detention are typical.
METHODS OF PROVIDING OXYGEN
There exist two primary means of aeration
to provide oxygen. One is mechanical aeration
which includes floating or fixed surface aera-
tors, biological filters, and rotating discs, and
the other is diffused air. Mechanical aeration
throws the liquid waste up into the air, drips it,
or waterfalls it, producing more surface area
and allowing more oxygen to be absorbed. Dif-
fused aeration, on the other hand, is inducing
compressed air into the waste through air dif-
fusers usually at the bottom of the tank for
complete mixing and absorption. There are
also some hybrid devices including mixers with
diffused air, and air ejectors or jets.
Diffused aeration is usually considered the
more expensive method of the two, as the capi-
tal costs of installation are larger. It is also con-
sidered the more efficient aeration method as
most of the horsepower goes to maintain the
dissolved oxygen level and less of it goes to
strictly mixing. It is most often used in deep
basins with vertical walls, normally longer
than wide due to the air pattern, and con-
structed of concrete. Maintenance costs for
operation of diffusers in textile waste can be a
real problem because they are so subject to
plugging.
Surface aerators are of two types: slow
speed and high speed and are usually more
flexible and easier to maintain than diffused air
because individual units can be cut off or
removed from the aeration basin without shut-
ting down the system and causing a defficiency
in dissolved oxygen. They do, however, trans-
fer somewhat less oxygen per horsepower and
often use more horsepower for mixing than
aeration. This is due, in the main, to the type of
installations in ponds and lagoons. Large
shallow ponds and lagoons have a very large
surface area when compared to vertical wall
concrete basins. In colder climates aerators
tend to cool the waste thus slowing down bio-
logical activity; diffused air would do the op-
posite.
Slow speed fixed surface aerators require
the construction of a bridge structure and use
a speed reduction gear unit. They are less flexi-
ble than floating units because the water or
waste level cannot be varied to any extent and
the individual units are stationary. Mainte-
nance and capital costs tend to be high when
compared to a high speed surface unit because
of the cost of the speed reduction gear unit and
the support platform. Slow speed units are said
to have the advantage over high speed units
when it comes to floe breakdown. High speed
units are said to shear the biological material
and to produce a less easily settled floe.
Floating surface aerators require little in
the way of structural supports and produce a
very flexible aeration system as they can be
moved around and are not affected by water
level changes. There are again two types: slow
speed and high speed. The slow speed unit is
just a fixed surface aerator on floats and has
most of the same advantages and disadvan-
tages of the fixed unit. The high speed (direct
drive) floating surface aerator is the most flexi-
ble system of all. It is easy to maintain, but not
quite as efficient as either the slow speed me-
chanical aerator or diffused air. It is, however,
69
-------
usually the least expensive aeration/pound of
air delivered if all costs are taken into con-
sideration.
Trickling filters are not used as often on tex-
tile wastes as the other methods but can work
well on some wastes. The construction costs
will be high, but operation costs can be lower
unless a mechanical means is used to increase
air supply through the unit (fans). They are
also subject to weather, especially if above
ground in cold climates. Freezing problems can
also be encountered with the rotating element
during cold v/eather.
Rotating discs or rotating biological discs
are a combination of a biological filter and an
activated sludge system and are a successful
treatment device. As the cost of power in-
creases, they will become more competitive
and their initial cost will possibly overcome the
flexibility and ease of handling of the floating
high speed (direct drive) surface aerator.
SOLIDS SEPARATION
In order to operate an activated sludge type
treatment system efficiently, an optimum bio-
logical population must be maintained in the
aeration system by settling the solids and
returning them to the aeration portion. This
settling takes place in clarifiers. The settled
sludge is collected by any one of a number of
drawoff systems and is picked up by the return
sludge pumps for recycling. As discussed pre-
viously, the return sludge rate can be a critical
factor in the successful operation of this sys-
tem on textile wastes.
The biological floe from textile waste treat-
ment is often difficult to settle and light sludge
is one of the most common problems. This pro-
blem can be partially controlled by making
sure that the control parameters used and dis-
cussed previously are properly regulated; the
most effective F/M ratio, sludge age, and
sludge concentration must be maintained. Nor-
mally if these conditions are met a fairly good
settling sludge will result. However, the floe
may still tend to be very small and extremely
light. Sludge bulking can also be a problem.
Since the incoming wastes often are strong and
not always consistent, a change in the plant
process, if not controlled, can cause temporary
rises and falls of highly concentrated waste
flows that may upset the system temporarily
killing off bacteria that can cause bulking
sludge. The growth of filamentous bacteria or a
change in water temperature both may also
cause bulking or floating sludge. Floating
sludge, caused by a change in temperature, is
often noted when the air temperature changes
rapidly for several days from high to low or low
to high in relationship to the water tempera-
ture. Bulking from filamentous sludge is more
noticeable in cold weather with rapid with-
drawal clarifiers, long aeration detention
times, and in waste with high carbohydrate
fractions.
To reduce the solids content in the clarifier
effluent resulting from the carryover of these
small fines or when bulking occurs, a weighting
agent can be used to aid the settling. A number
of organic polymers can be used to coagulate
the fine floe and produce a larger, heavier one.
Inorganic coagulants such as alum and ferric
chloride may also be used. If organic polymers
are added they will be oxidized and neutralized
when returned to the aeration system, thus a
constant feed must be maintained. Inorganic
coagulants, on the other hand, tend to be
recycled. If the use is large and sludge wasting
is low, they will in time concentrate in the
system and can cause problems themselves.
They may reduce the biological activity and
result in poor treatment and/or bulking. Pow-
dered and granular activated carbon have also
been added to biological treatment plants to
aid the biological process and increase settling
by weighting the sludge.
Sludge disposal is required for solids that
are no longer needed in the system; these
solids may contain refractories, nonbiodegrad-
able produced solids, and trace heavy metals.
Sludge wasting is determined by the desired
F/M ratio, solids level, sludge age, and treat-
ment expected.
Biological treatment sludge has a structure
that enables it to absorb color bodies, organic
materials, and trace heavy metals. This serves
to concentrate these materials and allows them
to be removed from the liquid phase in a form
that can be concentrated for disposal. Con-
tinuous in-plant efforts can serve to remove or
replace many of these substances that will end
up in this sludge. Chromium use for dye ox-
idization has been successfully replaced at a
higher cost. Zinc, which is everywhere, in-
cluding the raw water, can be reduced by
replacement as a corrosion inhibitor, chemical
70
-------
type changes, replacement by piping, etc.; but
many dyes are made with zinc. Other products
containing heavy metals that are determined
to be hazardous must be replaced or removed.
At this time, little long-term data on textile
wastes is available to determine levels to be
expected or required on other problem heavy
metals. The few analyses made to date in tex-
tiles indicate that few, if any, other heavy
metals will be a problem.
Biological activity and thus sludge produc-
tion will vary and can be related to F/M ratio.
Generally at high F/M ratios (0.15 to 0.20) pro-
ductivity of sludge is high and is reduced at
lower F/M ratios (less than 0.10). However, bet-
ter settling solids can usually be expected at
these higher loading rates, thereby reducing
effluent suspended solids and clarifier size.
Solids level and sludge age also affect treat-
ment expected, with higher MLSS levels gen-
erally giving better color and COD removals,
however, high effluent suspended solids may
be expected at these high solids levels. Very
long aeration detention times also can result
in poorly settling sludge. This can often be
handled with larger clarifiers and lower over-
flow rates. Nutrient removal, however, is en-
hanced by this long detention time.
OPERATION
Experience treating various types of textile
waste including greige operations, dyeing
waste from all types of dyes on cottons, wools,
blends, synthetics, and mixtures; finishing
wastes, including wash-and-wear treatments,
mothproofing, resin treatment, fire retardant
treatment; printing wastes including roller,
flatbed, and screen; will indicate that no two
wastes are exactly alike. General relationships
hold true and reasonable efficiencies can be ob-
tained without pilot work, but when high effi-
ciencies are required and strict conditions
must be met, each waste must be considered to
have individual characteristics requiring
evaluation. Due to this premise, off-the-shelf
treatments cannot be expected to produce ex-
tremely desirable results at the least cost with-
out some prior waste characterization and/or
pilot plant data. Common sense will indicate
that screening, lint removal, grease removal,
and the like will be required in many cases.
Wool waste cannot be treated if the grease
from scouring is not removed prior to the
biological step. Lint from cotton operations
will clog diffusers, hang on mechanical aera-
tors, and otherwise reduce efficiencies.
In some situations, segregation of the wastes
will reduce the volume to treat. In others, the
piping changes are so expensive that it may be
more economical to handle the volume or re-
duce the amount of segregation used. It has
been found that in many plants air can be used
to provide neutralization of the alkaline waste
through the production of C02 from biological
activity and from C02 in the air. In other cases
with certain dye mixtures, acid neutralization
is required to produce acceptable color re-
moval.
What we mean to say is that conditions must
be maintained as optimum as is cost-effective
to allow the biological matter to treat the many
colored, strong textile wastes. Time has been
used in place of chemicals, space, or area in
place of concrete, and both in place of labor to
produce the same or similar degree of treat-
ment.
In selecting a treatment scheme, experience
is valuable but the decision should be tem-
pered by actual data. This data will aid in pro-
viding a much more sound and reasonable deci-
sion.
Consider the following four options as ex-
amples.
Medium strength waste Pretreatment (in-
direct discharge)
Medium strength waste Direct discharge
Concentrated waste Pretreatment (in-
direct discharge)
Concentrated waste Direct discharge
A reasonable evaluation of the wastes and
sources of wastes will indicate that for most
textile operations primary clarification is not
required, however, screening for lint and rag
removal is usually desirable and necessary.
Pilot plant data will determine the method of
treatment that is expected to have the efficien-
cy to produce the required effluent at an accept-
able cost.
For pretreatment of a medium strength
waste, an oxidation pond, extended aeration,
bio-disc, or trickling filter should be given con-
sideration. For direct discharge of this waste
an oxygen system, extended aeration, or com-
pletely mixed activated sludge system should
71
-------
be considered. In all activated sludge evalua-
tions prior to construction, detention time,
solids level, sludge age, and F/M ratio should
be varied and evaluated.
Pretreatment of a strong waste would prob-
ably omit the trickling filters and maybe the
bio-disc as separate systems (depending on the
waste's strength) and would consider oxidation
ponds with mechanical aeration and recycle
and either extended aeration or completely
mixed activated sludge. Depending on whether
solids removal is or is not provided will help ad-
just the detention time of pretreatment if ac-
tivated sludge is desired. Direct discharge of
strong wastes would normally limit the selec-
tion to include only an oxygen system, a com-
pletely mixed activated sludge system and an
oxidation pond, or a bio-disc/activated sludge
system or some other combination of these.
CONCLUSION
In summary, the use of biological treatment
for textile wastes indicates that this process
can be adapted to almost any textile wastes.
There are various parameters or relationships
that control this treatability, including F/M
ratio, nutrient balance, MLSS concentrations,
settleability of the sludge produced, wastes
strength and makeup, recycle rate, and sludge
age. These things all have levels that can be
measured. The optimum blend then is deter-
mined for the wastes and results required. Ex-
ternal factors such as power cost, changing re-
quirements, and the so-called toxic limits all
now affect the engineering decisions and must
be evaluated as another variable.
Sludge removal aids in the increased effi-
ciency of treatment by taking refractory or-
ganics, trace heavy metals, some residual col-
or, hard to degrade organics, and degraded
biomass from the biological system. This
serves to enhance an active biological popula-
tion by removing concentrations to potentially
undesirable constituents which in turn in-
creases the efficiency of the biological
organisms.
To date, biological treatment has been the
most successful treatment method developed
for textile wastes and it should not be aban-
doned.
REFERENCES
1. Metcalf & Eddy, Inc., Wastewater Engi-
neering, p. 503.
2. Sacramento State College, Department of
Civil Engineering, Operation of Waste-
water Treatment Plants, Chapter 9, p. 1.
72
-------
A FUNDAMENTAL APPROACH TO APPLICATION OF
REVERSE OSMOSIS FOR WATER POLLUTION CONTROL*
S. Sourirajan, Takeshi Matsuuraf
Abstract
Reverse osmosis separation is the combined
result of preferential sorption of one of the con-
stituents of the feed solution at the membrane-
solution interface, and mass transport by fluid
permeation under pressure through the capil-
laries in the microporous membrane. Preferen-
tial sorption at the interface is a function of
solute-solvent-membrane material interactions
arising from polar (including ionic), steric, and
nonpolar characters of each one of the compo-
nents involved in the reverse osmosis system-
Basic reverse osmosis transport equations, and
available quantitative data on parameters
characterizing polar, steric, and nonpolar inter-
facial forces applicable to reverse osmosis sys-
tems involving dilute aqueous solutions, pref-
erential sorption of water at the membrane-
solution interface, and a cellulose acetate
(CA-398) or an aromatic polyamide-hydrazide
copolymer (PPPH-8273) membrane material
are presented. By integrating the latter param-
eters into the basic transport equations, one
can predict the reverse osmosis performance of
a membrane for a large number of feed solution
systems involving inorganic and organic (ion-
ized, nonionized, or partially ionized) solutes in
aqueous solutions from only a single set of ex-
perimental data for a reference sodium chlo-
ride-water feed solution. This prediction tech-
nique is illustrated with respect to two spec-
ified membranes, one made of CA-398 material
and the other made of PPPH-8273 material.
The application of the above transport equa-
tions for predicting the performance of com-
mercial reverse osmosis modules is also illus-
trated. Brief discussions on solute preferential
sorption at membrane-solution interfaces, tem-
perature effect, pH effect, membrane compac-
tion effect, fouling problems, and ultrafiltra-
tion are also included with particular reference
to application of reverse osmosis for water pol-
lution control.
INTRODUCTION
Reverse osmosis in industry is a general
process for the separation of substances in
fluid (liquid or gaseous) solution by permeation
under pressure through an appropriate syn-
thetic membrane. Since the original develop-
ment in 1960 of the now well known asymmet-
ric porous cellulose acetate membranes for
water desalination applications (ref. 1), many
synthetic membranes have been successfully
developed and tested for a variety of water pol-
lution control applications (ref. 2). The opera-
tional simplicity and the obvious potential of
reverse osmosis for many separation applica-
tions have generated a fast growing worldwide
activity on the subject in recent years. In spite
of such activity, reverse osmosis has not yet
gained a major entry into the industrial world
of water pollution control. This is not due to
lack of sufficient interest in the application of
reverse osmosis for water pollution control;
but this is due to, at least, an apparent lack of
such interest in a fundamental understanding
of reverse osmosis itself, and in the assiduous
development of an overall science of reverse
osmosis (based on such understanding) neces-
sary for its effective utilization with respect to
any application, including water pollution con-
trol. Hence the object of this paper is to call at-
tention to a fundamental approach to reverse
osmosis research and development, and the
science of reverse osmosis arising from that ap-
proach (ref. 3), and illustrate this approach by
some quantitative data predicted by the ap-
plication of this science. Such predictive
capability is useful for the industrial develop-
ment of reverse osmosis for water pollution
control applications.
•Issued as N.R.C. No. 17022.
' National Research Council of Canada, Division of Chem-
i>.trv. Ottawa. Canada.
73
-------
THE APPROACH AND THE SCIENCE
The fundamental approach referred to above
is the preferential sorption-capillary flow
mechanism for reverse osmosis, which led to
the development of the original porous cellu-
lose acetate reverse osmosis membranes (ref.
1). According to this mechanism, which is
discussed in detail in the literature (refs.
1,3,4,5), reverse osmosis separation is the com-
bined result of preferential sorption of one of
the constituents of the feed solution at the
membrane-solution interface, and mass trans-
port by fluid permeation under pressure
through the capillaries of the microporous
membrane. An appropriate chemical nature of
the membrane surface on the feed solution
side, and the existence of pores of appropriate
size and number on the area of the membrane
at the interface on that side together consti-
tute the indispensable twin requirement for
the practical success of this separation process.
Thus for reverse osmosis separation to take
place, one of the constituents of the feed solu-
tion must be preferentially sorbed at the mem-
brane-solution interface; further, to be indus-
trially useful, the reverse osmosis membrane
must have a microporous and heterogeneous
surface layer at all levels of solute separation,
its entire porous structure must be asymmet-
ric, and there should be no chemical reaction
between the constituents of the feed solution
and the material of the membrane surface.
With an appropriate material for the mem-
brane surface, practically any degree of solute
separation is possible in reverse osmosis by
simply changing the average pore size on the
membrane surface and the operating condi-
tions of the experiment. Aside from such proc-
ess requirements, reverse osmosis is funda-
mentally not limited to any particular solvent,
solute, membrane material, level of solute
separation or that of solvent flux, or operating
conditions of the experiment. Consequently,
the overall science of reverse osmosis arising
from the above approach unfolds itself through
proper integration of the physicochemical fac-
tors governing preferential sorption of solvent
or solute at membrane-solution interfaces, ma-
terials science of reverse osmosis membranes,
and the engineering science of reverse osmosis
transport, system analysis, and process design.
While there is still a long way to go towards
the full development of the science of reverse
osmosis in all its aspects, significant progress
has already been made, all of which is practical-
ly useful, which should be evident from the fol-
lowing discussion.
PREFERENTIAL SORPTION AT
INTERFACES
Preferential sorption at the membrane-
solution interface is a function of solute-
solvent-membrane material (i.e., material of
the membrane surface) interactions. These in-
teractions arise in general from polar (hydro-
gen bonding), steric, nonpolar (hydrophobic),
and/or ionic character of each one of the above
components. The overall result of these inter-
actions determines whether solvent or solute
(or neither) is preferentially sorbed at the
membrane-solution interface. Application of
reverse osmosis to water pollution control may
involve preferential sorption of water or solute
at the membrane-solution interface depending
on the chemical nature of solute and that of
membrane material. When water is preferen-
tially sorbed, reverse osmosis results in
positive solute separation (i.e., solute concen-
tration in the membrane permeated solution is
less than that in the feed solution); when solute
is preferentially sorbed, solute separation in
reverse osmosis can be positive, negative, or
zero, depending on the relative mobility of the
solute species (compared to water) through the
membrane pores under the conditions of the
experiment. Both cases are illustrated in the
literature (ref. 6).
MEMBRANE MATERIALS AND
MEMBRANES
Though many membrane materials and
membranes have been successfully tested for
reverse osmosis applications (ref. 2), cellulose
acetate (acetyl content -39.8 percent, desig-
nated here as CA-398) membranes are still by
far the most widely used ones in current indus-
trial practice; these membranes are also the
ones for which extensive scientific data have
been generated. For this reason this paper is
concerned mainly with such cellulose acetate
membranes. Du Pont reverse osmosis modules
are also widely used in industry; the mem-
branes in these modules are made of an aro-
74
-------
matic polyamide-hydrazide copolymer material
(ref. 7), which is not commercially available for
outside independent researchers. However, a
chemically similar polymer (designated here as
PPPH-8273) synthesized in the laboratory, has
been studied as reverse osmosis membrane
material, and a limited amount of scientific
data has been generated for membranes made
from the latter material (ref. 8) (which, for prac-
tical purposes, may be considered to be analo-
gous to the Du Pont material); this paper also
includes the latter data, and illustrates their
utility.
The polymer materials CA-398 and PPPH-
8273 have been chemically characterized in
terms of their ap (polar), an (nonpolar), and 0
(affinity for nonionized polar organic solutes
relative to that for an ionized inorganic solute)
parameters generated from liquid-solid chro-
matography (LSC) data (refs. 9,10). The rel-
evance of such characterization to reverse os-
mosis stems from the principle that the solute-
solvent-polymer material interactions govern-
ing the relative retention times of solutes in
LSC are analogous to the interactions prevail-
ing at the membrane-solution interface under
reverse osmosis conditions. On arbitrarily
defined relative scales (refs. 9,10), the values of
ap, an, and 0 are -1.00, -0.017, and 1.37
respectively for CA-398, and -0.31, 0.144, and
0.17 respectively for PPPH-8273 polymers. A
higher value of ap indicates stronger hydrogen
bonding capacity of polymer material and
hence its higher sorption capacity for water. A
higher value of an for the polymer material
results in greater mobility of the preferentially
sorbed water through the membrane pores
when water is preferentially sorbed, and lower
mobility of the solute through the membrane
pores when solute is preferentially sorbed at
the membrane-solution interface; both these ef-
fects contribute to higher solute separation in
reverse osmosis. A higher value of /3 denotes
stronger affinity of the polymer for nonionized
polar organic solutes (and consequently lesser
preferential sorption of water at the mem-
brane-solution interface) relative to ionized
solutes. Thus the numerical values of ap, an,
and /3 direct attention to the inherent potential
of polymer materials with relatively higher ap
and «„ values together with relatively lower /?
values (such as the case for PPPH-8273) for
possible use in making high flux reverse
osmosis membranes especially for applications
requiring high separations of both ionized
solutes and nonionized polar organic solutes in
wastewaters.
Futher, with a given membrane material, an
infinite number of membranes can be made
with different porous structures and hence dif-
ferent reverse osmosis properties. The specific
details of film casting techniques used in mak-
ing the membranes decide the overall porous
structure of the resulting membranes. For ex-
ample, the aromatic polyamide membranes are
usually cast at high temperatures (90° to
130° C) whereas the cellulose acetate mem-
branes are usually cast at relatively low tem-
peratures (0° to 30° C). The supermolecular
polymer aggregates in a given casting solution
are generally smaller in size at higher tem-
peratures, which results ultimately in cor-
respondingly smaller size pores on the mem-
brane surface in the as-cast condition (ref. 11).
Consequently, the use of higher temperatures
for casting solution and/or casting atmosphere
offers one of the several means of creating
smaller average size of pores on the surface of
resulting membranes. For some water pollu-
tion control applications, the practical possibil-
ity of using the above means of pore size gener-
ation on the membrane surface may be the
deciding factor for the choice of membrane
material, and the film casting conditions for
making the required membranes.
PHYSICOCHEMICAL PARAMETERS
CHARACTERIZING SOLUTES IN THE
VICINITY OF INTERFACES
For reverse osmosis systems involving di-
lute aqueous feed solutions, three physico-
chemical parameters have been developed for
characterizing ions and nonionized solutes in
the vicinity of different membrane-solution in-
terfaces. These parameters are: polar free
energy parameter -AAG/RT, steric parameter
5*-Es, and nonpolar parameter w*:Cs*. These
parameters may be described briefly as fol-
lows.
Polar Free Energy Parameter -AAG/RT
When water is preferentially sorbed at the
membrane-solution interface, AAG/RT gives a
quantitative measure of solute repulsion at the
75
-------
interface on a relative scale. This parameter
was initially developed for inorganic ions, and
later extended to nonionized polar organic sol-
utes, organic ions, and inorganic ion pairs (refs.
8,12,13,14,15,16). The basis for this parameter
arises from (1) Born equation for ion-solvent in-
teraction (free energy of hydration, AG) as ap-
plied to the bulk feed solution phase and the
membrane-solution interface, (2) the thermody-
namic basis of Hammett and Taft equations
representing the effect of structure on reaction
rates and equilibria (ref. 17), (3) the relationship
between the parameters of Taft eq. (16) in ref-
erence 17 and reverse osmosis data on solute
transport parameter DAM/^ (defined by eq. (8)
later in this discussion) for different solutes
and membranes (refs. 18,19,20), and (4) the func-
tional similarity of the thermodynamic quanti-
ty AAF* representing the transition state free
energy change (ref. 17) and the quantity AAG
defined as
AAG = AGi-AGB,
(1)
which represents the energy needed to bring
an ion, or the nonionized solute molecule, from
the bulk solution phase (subscript B) to the
membrane-solution interface (subscript I). The
quantity -AAG/RT, which is an interfacial
property, is a function of the chemical nature of
the solute, the solvent, and the membrane ma-
terial, and it is independent of the porous
structure of the membrane used. A lower value
of -AAG/RT for the solute yields a lower value
for DAM/KS for the solute, and hence higher
solute separation in reverse osmosis.
The values of AAG, and hence -AAG/RT, for
different ions and nonionized polar organic sol-
ute molecules can be computed from data on
AGs (which arises from the physical chemistry
of the bulk solution phase) and AGj calculated
from experimental reverse osmosis data on
DAM/KS for different solutes (refs. 8,12,13,14,
15,16). Table 1 gives the available data on
-AAG/RT for several inorganic and organic
ions, and inorganic ion-pairs for interfaces in-
volving aqueous solutions and CA-398 or
PPPH-8273 polymer membrane material. The
value of -AAG/RT for a completely ionized in-
organic salt is simply the sum of the -AAG/RT
values for the ions involved. For example,
AAGJ
AAGJ = < AAG)
RT (NaC1 " 1 RT |Nm+
i AAG(
+ {-^RT~f
I RT )cr
j_AAG/
= r RT )Mg-
Mg(N03)2
2-
AAG/
For nonionized polar organic solutes, the val-
ues of AGB and AGi are simply the additive
functions of the contributions <7B and j^ respec-
tively) of the structural groups in the molecule,
so they can be expressed as:
= ETB (structural group ) + yEO and (2)
AGi = 2-yj (structural group) + >j 0. (3)
For example, for s-butyl alcohol whose
molecular formula is CH3CH2CH(OH)CH3
j AAGJ =
RT
Ti(CH2)
for cyclohexanol whose molecular structure is
CH2- CH2
AAG
~RT~
+ 7i(OH) + 7X (cyclic,
6-membered) + 7i,o(-
+ 7B(CH) + 7B(OH) +7B (cyclic,
6-membered) + 7BO|i ;
76
-------
TABLE 1. DATA ON FREE ENERGY PARAMETER (-AAG/RT) FOR SOME IONS AND ION-PAIRS AT 25° C,
APPLICABLE FOR INTERFACES INVOLVING DILUTE AQUEOUS SOLUTIONS AND
CA-398 OR PPPH-8273 POLYMER MEMBRANE MATERIAL
(-AAG/RT) (-AAG/RT)
Species SDCCJC:
PPPH- PPPH-
CA-398 8273 CA-398 8273
Inorganic cations Inorganic anions (con.)
H+ 6.34 I" -3.98 1.33
Li+ 5.77 -1.20 103~ -5.69
Na* 5.79 -1.35 H2P04" -6.16
K+ 5.91 -1.28 Br03" -4.89
Rb+ 5.86 -1.27 N02" -3.85
Cs+ 5.72 -1.23 N03" -3.66 2.54
NH4+ 5.97 ClOg- -4.10
Mg2+ 8.72 -2.41 CI04" -3.60
Ca2"1" 8.88 HC03" -5.32
Sr2"1" 8.76 HS04" -6.21
Ba2+ 8.50 S042~ -13.20 2.18
Mn2+ 8.58 S2°32~ -14-03
Co2+ 8.76 S032~ -13.12
Ni2+ 8.47 Cr042" -13.69
Cu2"1" 8.41 Cr2072~ -11.16
Zn2"1" 8.76 C032" -13.22
Cd2+ 8.71 Fe(CN)g3~ -20.87
Pb2+ 8.40 Fe(CN)g4" -26.83
Fe2+ 9.33
fe 9.82 Inorganic ion-pairs
Al3* 10.41 MgS04 3.45
Ce3* 10.62 CoS04 3.41
Cr3+ 11.28 ZnS04 2.46
La3+ 12.89 MnS04 2.48
Th4"1" 12.42 CuS04 2.85
CdS04 3.04
Inorganic anions NiS04 2.18
OH" -6.18 KFe(CN)62~ -2.53
r -4.91 1.03 KFe(CN)63- -17.18
OP -4.42 1.35
Br~ -4.25 1.35
* Applicable data on dissociation equilibrium constant are given below:
Dissociation
Solute equilibrium constant (ref. 21) Solute
MgS04 4.36 x 10~3 CuS04
CoS04 3.38 x 10r3 CdS04
ZnSO, 4.89 x 10"3 NiS04
4 . *
MnS04 5.20 x 10"15 K3Fe(CN)g
K4Fe(CN)g
Species -
Organic ions
HCOQ-
H Phthalate"
c2o42-
t-C4HgCOO"
i-c3H7coo~
cyclo-CgH^COQ-
n-C4H9COO~
n-C3H7COO~
C2H5COO~
CH3COO"
CgH5(CH3)2COO""
C6H5(CH2)2COO-
C6H5(CH2)COQ-
CgHgCOQ-
p-CH3OC6H4COO"
m-CH3CgH4COO"
m-OHCgH4COO~
p-CIC6H4COO"
m-N02C6H4COQ-
p-N02CgH4COO~
o-CIC6H4COO"
o-N02CgH4COO""
HOOCCOO"
HOOCCH2COO~"
HOOC(CH2)2COO~
CHjCHOHCOQ-
HOOCCH(OH)CH2COO"
HOOCCH(OH)CH(OH)COQ-
HOOCCH2C(OH)(COOH)CH2COO-
Dissociation
equilibrium constant
2.62 x 10~4
4.89 x 10~3
3.98 x 10~3
3.98 x 10"2
5.01 x 10"~3
(-AAG/RT)
PPPH-
CA-398 8273
-4.78
-4.63
-14.06
-6.90
-6.11
-6.24
-6.11
-6.06
-6.14
-5.95 1.33
-5.93
-5.86
-5.69
-5.66
-5.74
-5.67
-5.64
-5.63
-5.92
-5.93
-6.41
-6.61
-6.60
-6.46
-5.65
-6.30
-5.97
-6.40
-6.24
77
-------
for D-glucose whose molecular structure is
CH2OH
01 OH
+ 57j (OH)
+ 7 j (cyclic, 6-membered)
(cyclic, 6-membered)
and for sucrose whose molecular structure is
O H CH2OH
r i
HO %C C
H
I/A \l l>°\l
C V1 r1^ ^*r"
• - r>« H /V V\ H OH r
l/L^\i_i/(
L
OH H
8Tl(CH)
+ 7j (C) + 7j (cyclic, 6-membered)
+ 7j(cyclic, 5-membered)
7B (cyclic, 6-membered)
TB (cyclic, 5-membered)
The available values of TB and TI for different
structural groups and the applicable constants
7B>0 and 7j „ for interfaces involving aqueous
solutions and CA-398 or PPPH-8273 polymer
membrane material are listed in Table 2.
TABLE 2. STRUCTURAL GROUP CONTRIBUTIONS FOR
AGB AND AGj AT25° C, APPLICABLE FOR INTERFACES
INVOLVING DILUTE AQUEOUS SOLUTIONS AND
CA-398 OR PPPH-8273 POLYMER MEMBRANE MATERIAL
7B Q=-12.04; 7| 0=~41.21 for CA-398,
and 44.01 for PPPH-8273
Structural
group
-CH
>CH
>CH
>C<
Cyclic (5-membereu)
Cyclic (6-membered)
-OH
>0
-CN
>C=0
-CHO
-COOH
-N02
- CONH2
~ ceHs
-coo-
r»
11.07
0.17
-10.62
-21.50
20.60
20.49
3.99
-4.03
5.19
-5.80
5.80
2.49
6.56
-0.21
8.41
-5.14
CA-398
24.23
0.24
-23.65
-47.39
46.99
47.12
17.04
-4.59
17.17
-6.32
18.84
15.13
18.95
12.75
21.43
-6.32
*l
PPPH-8273
- 17.74
0.32
18.41
36.26
-
-37.37
-25.21
-4.21
-
-6.11
-23.36
—
-
-
-20.09
-5.49
Steric Parameter £*EES
The electrostatic repulsion of ions in aque-
ous solution in the vicinity of membrane mate-
rials of low dielectric constant (ref. 22) is much
larger in magnitude than repulsions or attrac-
tions resulting from interfacial forces due to
steric and nonpolar (hydrophobic) effects;
hence the latter two effects are significant only
for reverse osmosis systems involving nonion-
ized solutes. Steric hindrance to reactivity of
molecules arises from repulsions between non-
bonded atoms, and also from interference of
groups or atoms with each other's motions:
consequently steric hindrance (which also
parallels the effective size of the molecules con-
cerned) is always a repulsive force at the inter-
face. The parameter 5*DES represents the total
steric effect affecting solute transport through
the porous membrane in reverse osmosis (ref.
19); steric hindrance to such solute transport
results in a negative value for the parameter
6*rEs, which contributes to higher solute
78
-------
separation in reverse osmosis. Unlike the polar
free energy parameter (-AAG/RT), the steric
parameter 6*1!ES is a function not only of the
chemical nature of the solute, the solvent and
the membrane material, but it is also a function
of the porous structure of the membrane sur-
face.
With respect to solute molecules involving
monofunctional groups, the value of the steric
constant EES for the substituent group in the
polar organic molecule can be computed as the
summation of the steric substituent constants
for the structural units involved from the data
on Es given by Taft (listed in Table 3) (refs.
17,19). For example, 2ES for the substituent
group in t-butyl alcohol = Es for t-C^g; that
for cyclohexanpl = Es for cyclo-CeHn; that for
t-butyl isopropyl ether = Es for t-C^g + Es
for i-CaH?; and that for diisopropyl ketone
= 2 x Es for i-CaH?, etc. It must be noted that
the value of EES for the substituent group in
the solute molecule depends only on the chem-
ical nature of the group, and represents the
property of the solute in the bulk solution
phase. The steric coefficient 8* associated with
EES, however, depends on the chemical nature
of the functional group in the solute molecule,
as well as the chemical nature and porous
structure of the membrane surface. The avail-
able values of 6* for alcohol-, aldehyde-,
ketone-, and ether solutes (refs. 8,13) for
CA-398 and PPPH-8273 membrane materials
are given in Figure 1 as a function of average
pore size on the membrane surface as repre-
sented by the quantity ln(CNaci/A). Using the
values of Es given by Taft (ref. 17) and suitable
interpolation techniques, additional such data
have been generated for practical use in re-
verse osmosis (refs. 19,23); some of these data
are also included in Table 3.
For solute molecules involving polyfunc-
tional groups (for example, glucose, sucrose,
etc.), there is no simple way by which the
values of EES can be computed from data given
TABLE 3. TAFT'SSTERIC SUBSTITUENT CONSTANTS, Es at 25° C APPLICABLE FOR
AQUEOUS SOLUTIONS AND SOLUTE MOLECULES INVOLVING MONOFUNCTIONAL GROUPS (refs. 17,19)
Substituent
group
Aliphatic compounds
CH3
C2H5
Cyclo-C4H7
n-C3H7
n-C4H9
n-CgH,,
i-CsHn
n-C8H17
t-C4HgCH2CH2
C6H5CH2
C6H5CH2CH2
C6H5CH2CH2CH2
i-C3H7
Cyclo-C5Hg
cyclo-CgH^
i-C4Hg
cyclo-C6HnCH2
-------
CA-398 MEMBRANES
PPPH-8273 MEMBRANES
3.0r
2.0
1.0
1.0
2.0
ETHERS a
NITRO
COMPOUNDS
ESTERS
KETONES
I (A) I
KETONES
ETHERS
ALDEHYDES
3.0
4.0
1.0
2.0
3.0
4.0
5.0
Figure 1. Correlation of 6 * versus ln(CNaC,/A) for CA-398 and PPPH-8273 membranes at 25° C.
Line A in Figure is applicable to alcohol-, aldehyde-, amide-, and undissociated carbox-
ylic acid solutes.
in Table 3. For such solutes, an empirical
method has been established (ref. 13) for esti-
mating the value of the parameter 5*£ES on
the basis that the latter reaches a limiting
value, designated as (S*EEs)iiin, when the
average pore size on the membrane surface be-
comes sufficiently small. For a given mem-
brane material, this quantity (6*£Es)ijm is simp-
ly an additive function of the contribution of
each of the structural units involved in the
solute molecule, so that
20(structural unit) + 0. (4)
Table 4 lists the available data (refs. 13,24) on <£
and 0 applicable for aqueous solutions involv-
ing polyfunctional solutes and CA-398 mem-
brane material. For illustration, from the
molecular structures of D-glucose and sucrose
given above,
(6*EEs)um for D-glucose
= (MCH2) + 5
-------
TABLE 4. STRUCTURAL GROUP CONTRIBUTIONS FOR
(5*2Es),jm FOR SOLUTE MOLECULES WITH POLY-
FUNCTIONAL GROUPS AT 25° C, APPLICABLE FOR
INTERFACES INVOLVING DILUTE AQUEOUS SOLUTIONS
AND CA-398 POLYMER MEMBRANE MATERIAL
(t>Q= -4.83 for CA-398
Structural unit 0
Es*(CH = CHCH2CH2CH2)
= 2s*(-CH = ) + 3(>CH2)
-CH3
XH2
>CH
4
Cyclic (5-membered
Cyclic (6-membered
Cyclic (5-membered
Cyclic (6-membered
-OH
>0
>C=0
-CHO
and single ring)
and single ring)
and two rings)
and two rings)
2.90
-0.02
-3.19
-6.88
5.22
4.70
5.99
5.47
2.01
0.63
-1.15
2.14
hydrocarbon. The Small's number versus loga-
rithm of solubility is a straight line which is dif-
ferent for different reactive series of com-
pounds of similar chemical nature including
paraffins, cycloparaffins, olefins, cycloolefins,
diolefins, acetylenes, and aromatics. By ad-
justing the data on Small's number for various
structural groups such that the correlation of
Small's number for the paraffin hydrocarbon
(taken as reference) versus logarithm of its
molar solubility in water is valid for hydrocar-
bons in all the above reactive series, a modified
number for the molar attraction constant,
called the "modified Small's number" repre-
sented by s*, has been generated for different
structural groups (ref. 26) as given in Table 5.
The modified Small's number Es* for a
hydrocarbon molecule, or the substituent
group in a polar organic molecule, is obtained
from its chemical structure using the additive
property of s* for different structural groups.
For example,
CH - CH2)
= 6s*( - CH = )
+ s*(>CH2) + s*(conjugation)
Es* for the substituent group in CH3OH
= s*(CH3)
Es* for the substituent group in
CH3(CH2)2CH(CH3)CH2OH
= 2s*(CH3) + 3s*(>CH2) + s*(>CH)
Es* for the substituent group
in (CH3)3COH
= 3s*(CH3) + s*(>C<)
Es* for the substituent group
s*(C6H5) + s*(H)
£s*(C6H4(CH3)2)
= s*(C6H4) + 2s*(CH3)
= 5s*(>CH2) + s*(>CH)
etc. The scale of Es* is consistent within each
of the groups of aromatic, cyclic, and noncyclic
hydrocarbon structure. Just as EES, Es*
represents the property of the solute in the
bulk solution phase.
The coefficient w* associated with Es* in the
nonpolar parameter is a function of the chem-
ical nature of the interface and independent of
the porous structure of the membrane surface.
For interfaces involving aqueous solutions, d
to Cg monohydric alcohol-solutes and CA-398 or
PPPH-8273 polymer membrane material, the
available values of w* obtained experimentally
(ref. 27) are listed in Table 6.
Referring to Table 6, the value of w* is a
function of both the number of carbon atoms
and degree of branching in the molecular struc-
ture of the alcohol-solute. The quantity As*,
defined as
As*(alcohol) = Es*(alcohol) - Es*
(straight chain primary alcohol with the
same number of carbon atoms) (5)
serves as a quantitative measure of the latter
variable. As the degree of branching increases,
the value of As* becomes more negative. An in-
crease in the value of «* denotes greater non-
polar attraction of the solute towards the mem-
brane surface, whereas a decrease in w* de-
notes lesser such attraction. This is under-
81
-------
CA-398 MEMBRANES
LOr—
0.9
(8*SEs)lim
0.7
0.6
1.0
2.0
3.0
4.0
Figure 2. Correlation of 8*2Es/(5*2Es)|jm versus ln(CNaC!/A) for CA-398 and PPPH-8273
membranes at 25° C.
standable on the basis of the solubility of
alcohol in water, which increases with increase
in the degree of branching in the molecular
structure of the alcohol. An increase in solubili-
ty in water means greater solute-solvent at-
traction, which tends to decrease the solute-
polymer attraction, and this is expressed by a
decrease in o>*. Thus the data given in Table 6
show that while an increase in carbon number
tends to increase the value of «*, the opposite
effect is the case as the As* value becomes
more negative. When these opposing values
TABLE 5. MODIFIED SMALL'S NUMBER s* (IN cal^cc^g-mokf1) FOR SOME ORGANIC STRUCTURAL GROUPS AT
25° C APPLICABLE FOR DILUTE AQUEOUS SOLUTIONS (ref. 26)
Structural
group
-CH3,
>CH2 angle
>CH J bonded
>C<
H
$»
214
133
28
-93
90
Structural
group
CH2 =
-CH= I double
>C = ) bonded
CH=C-
-CSC-
Phenyl
s*
140
61
-31
-15
-78
335
Structural
group
Phenylene (o,m,p)
Ring (5-membered)
Ring (6-membered)
Conjugation
W~
=C=
s»
258
0
0
25
181
-31
82
-------
TABLE 6. DATA ON co* (IN THE NONPOLAR PARAMETER co*Zs*) AT 25° C APPLICABLE FOR INTERFACES
INVOLVING Ci-CgMONOHYDRIC ALCOHOL-SOLUTES, DILUTE AQUEOUS SOLUTIONS
AND CA-398 OR PPPH-8273 POLYMER MEMBRANE MATERIAL
CA-398
As*(alcohol)
co*x
103
NOTES:
co* in cal~"cc ^g-mole.
nc = number of carbon atoms in the alcohol molecule.
PPPH-8273
As* (alcohol) = Zs*(alcohol) - Ss* (straight chain primary alcohol with the same nc). For example,
As*(CH3CHOHCH3) = 2s* (CH3CHOHCH3) - 2s*(CH3CH2CH2OH)
= 456-480 =-24
CO* X
103
0
-24
-48
0
-24
64r72
88
1to3
3 to 5
6 and 7
4to8
6 to 9
4to7
6 and 7
0
0
0
2.1
1.2
-1.0
-2.2
1to4
3 and 4
Band?
5 to 9
5to7
7
Band 7
0
0
-0.84
0.55
-0.28
-1.35
-1.35
cancel each other w* becomes zero. An explana-
tion for the negative values for co* is in order;
this is simply a matter of scale which arbitrari-
ly sets «* = 0 for solute molecules containing
no more than three straight chain carbon
atoms not associated with a polar functional
group.
TRANSPORT EQUATIONS FOR REVERSE
OSMOSIS SYSTEMS INVOLVING
AQUEOUS SOLUTIONS AND
PREFERENTIAL SORPTION OF WATER
AT THE MEMBRANE-SOLUTION
INTERFACE
Basic Equations
At any given operating temperature and
pressure, the experimental reverse osmosis
data on pure water permeation rate (PWP),
product rate (PR), and fraction solute separa-
tion / at any point (position or time) in the
reverse osmosis system can be analyzed on the
basis that PWP is directly proportional to the
operating pressure, the solvent water trans-
port (NB) through the membrane is propor-
tional to the effective pressure, the solute
transport (NA) through the membrane is due to
pore diffusion through the membrane capil-
laries and hence proportional to the concentra-
tion difference across the membrane, and the
mass transfer coefficient k on the high
pressure side of the membrane is given by the
"film" theory on mass transport. This analysis,
which is applicable to all membrane materials
and membranes at all levels of solute separa-
tion, gives rise to the following basic reverse
osmosis transport equations (ref. 1):
A = (PWP)/Mx X S X 3600 X P
NB =
I (C2XA2~ C3XA3)
'XA2 ~ XA3l
kCl(l-XA3)ln
-------
DAM/K<5 is treated as a single quantity for pur-
poses of analysis (ref. 1).
It must be emphasized that neither any one
equation in the set of equations (6) to (9), nor
any part of this set of equations, is adequate
representation of reverse osmosis transport;
the latter is governed simultaneously by the
entire set of equations (6) to (9).
Correlations of A, DAM/K§, and k
The correlations of these quantities with
operating pressure, temperature, feed concen-
tration, feed flow rate, nature of solute and
membrane compaction during continuous op-
eration are of practical interest from the point
of view of specifying membranes and predict-
ing membrane performance under different ex-
perimental conditions. Such correlations are
best established experimentally for different
membrane materials, membranes, and solutes.
The following correlations so established are
used in obtaining the results presented in this
paper.
1. For CA-398 membranes, DAM/K£
for a given solute (whether ion-
ized or nonionized) is independ-
ent of feed concentration and
feed flow rate.
2. For PPPH-8273 membranes,
DAM/KS for any nonionic solute is
independent of feed concentra-
tion and feed flow rate; in limited
range of concentrations, DAM/KS
is also independent of XAS for
ionized solutes involving diva-
lent ions; and for ionized solutes
involving monovalent ions only,
3.
(10)
With respect to any given mem-
brane, k for any solute (k^te)
can be obtained from the corre-
sponding value of k for NaCl
&NaCi) tar a reference NaCl-HaO
feed solution using the relation:
ksalute = kNaCl
NaCl] %
(ID
diffusivities of the solute in the feed solu-
tion under consideration, and NaCl in ref-
erence feed solution respectively.
Membrane Specifications
At a specified operating temperature and
pressure, a reverse osmosis membrane can be
specified in terms of its performance data with
respect to a reference solute in aqueous solu-
tion. For the purpose of this paper, NaCl is
taken as the reference solute. Any CA-398
membrane can be completely specified in
terms of its A and (DAM/KS^CI values; any
PPPH-8273 membrane can be completely
specified in terms of its values for A and
ff>AM/K%faCl corresponding to a given XAB
value. A single set of experimental reverse
osmosis data on PWP, PR, and f for a
reference NaCl-H20 feed solution under known
operating conditions is enough for use in equa-
tions (6) to (9) to obtain the required specifying
data on A, ff>AM/K$NaCl. m& XA2- When a
membrane is so specified, its reverse osmosis
performance with respect to a large number of
inorganic and organic solutes, and different
feed concentrations and feed flow rates (as ex-
pressed by different k values) at the specified
pressure and temperature can be predicted
from the basic transport equations or by com-
bining them with the polar free energy, steric,
and nonpolar parameters given above. This
predicion technique is outlined and illustrated
below with respect to a CA-398 membrane and
a PPPH-8273 membrane specified as in Table 7.
Predictive Capability of Basic Transport
Equations
Combining equations (7), (8), and (9),
(12)
(c2XA2-c3XA3)
(13)
AIP-TrCX^) +
JDAM(U-XA3
XA3
(C2XA2 ~ C3XA3)
where (DAB) solute and (DAB) NaCl are the
84
-------
TABLE 7. MEMBRANE SPECIFICATIONS
Operating pressure: 405 psig (27.56 atm)
Operating temperature: 25° C
CA-398
Membrane
PPPH-8273
Membrane
A, g-mole H20/cm2s atm 1.660 x10~6 1.300 x 1IT6
(DAM/K8)NaC|, cm/s 1.790x10-5 1.470
XA2 1.386:
' f \ I w
Ix 10-5)
ix ID'3 /
For a membrane specified in terms of A and
DAM/KS, equations (12) and (13) enable one to
predict membrane performance (solute separa-
tion and membrane flux) for any feed concen-
tration XAI, and any chosen feed flow condition
specified in terms of k as follows. First deter-
mine, by trial and error, the particular com-
bination of XA£ and XAS which simultaneously
satisfies equations (12) and (13), and also equa-
tion (10) when applicable. Using that combina-
tion of XA2 and XAS values, determine NB from
equation (7). The above values of XAS and NB
give the obtainable solute separation and mem-
brane flux which may be expressed in terms of
/"and PR respectively as follows (ref. 1):
These results also illustrate that while data on
A and (DAM/K%aC1 (along with data needed for
equation (10)) are enough to specify a mem-
brane, they are not enough to predict mem-
brane performance with respect to any solute,
any feed concentration, and any feed flow rate;
for such prediction, the relationship between
(DAM/K6)NaC1 and DAM/K6 for other solutes, and
the applicable value of k are needed in addition
to data on A and (DAM/K%aCi; variations in k
can have significant effects on / and PR; and,
consequently, reverse osmosis experimental or
performance data on f and PR are always in-
complete without the corresponding k value or
its equivalent.
Considering the problem of predicting the
reverse osmosis performance of a specified
membrane at any point in the membrane-
solution-operating system, the quantity A is in-
dependent of any solute under consideration,
the quantity k is fixed either by choice or by
operating conditions, or can be calculated from
equation (11) or by other means (ref. 28), and
hence the problem is primarily one of gener-
ating the applicable value of DAM/KS for the
solute under consideration from data on
(DAM/K^NaCl given in membrane specifica-
tions.
ml
X
= 1—(
A3
- X
A3
1-X
Al
X
•Al
X
* 1-
A3
(14)
X
Al
PR = -
NB X MB X S X 3600
I {-
1
1000
mi (1 — /^MA ,
(15)
Using the above prediction procedure, one can
calculate, for example, the effect of feed con-
centration and feed flow rate (as expressed by
different values of k) on /and PR for a specified
membrane.
A set of results of such calculations is il-
lustrated in Figure 3 for the CA-398 and
PPPH-8273 membranes specified in Table 7
with respect to NaCl-H20 feed solutions in the
solute concentration range 200 to 10,000 ppm.
Relationships Between (DAM/K5)NaCi and
DAM/K6 for Other Solutes
For completely ionized inorganic and (sim-
ple) organic solutes (refs. 12, 14, 15, 16),
(DAM/K5)soIute oc expnc
AAG
AAG^
(16)
na i RT
anion
where nc and na represent the number of moles
of cation and anion respectively in one mole of
ionized solute. Applying equation (16) to
*
MDAM/K6)NaC1 = lnCNaC]
AAG
\
RT
Na
AAG
c
-------
CA-398 MEMBRANE
PPPH-8273 MEMBRANE
e
o
.c
X.
o>
of
a.
z- 98
g
h-
8 96
CO
_ 94
o
o
z
92
J
I
I
2000
6000 10000 2000 6000
NaCI CONCENTRATION IN FEED, ppm
10000
Figure 3. Effect of feed concentration and mass transfer coefficient k on the performance of
the CA-398 and PPPH-8273 membranes specified in Table 7 for NaCI-H2O feed
solutions. Operating pressure - 405 psig.
where In CN«CI is a constant representing the
porous structure of the membrane surface ex-
pressed in terms of (DAM/KS)NaCl- Using the
data on - AAG/RT for Na + and Cl - ions (Table
1) which are independent of In CNBCI»tne value
of the latter can be calculated correspondingly
to any given value of (DAM/K5)NaCl- Using the
value of In CN«CI so obtained, the correspond-
ing value for DAM/K& for any completely ion-
ized inorganic or (simple) organic solute can be
obtained from the relation
aC1
anionl
Thus, for any specified value of
the corresponding values of DAM/KS for a large
number of completely ionized solutes can be
obtained from equation 18 using data on
-AAG/RT for different ions given in Table 1.
For the PPPH-8273 membrane considered
for illustration in this work, (DAM/K^hfeCl at
XA2 = 1.386 x 10-3 was used to calculate
In C&aci for use in obtaining (DAM/KSUnne
values for nonionized organic solutes, and
(DAM/K%aCl at XAZ - 0.830 x 10-3 was used
to calculate In CNaCl for use in obtaining
(DAM/K5)solute values for ionized inorganic
solutes.
With respect to electrolytic inorganic
solutes, a few special cases are of particular in-
terest to water pollution control. Some solutes
are subject to partial dissociation and ion-pair
formation (for example, sulfates of divalent
86
-------
metals particularly at high concentrations) or
partial hydrolysis (for example, Na2C03, FeCl2
etc.). For a solution system involving ions and
ion pairs, equation (18) can be written as:
ln(DAM/KS)sohlte =
cation
where ap represents the degree of dissocia-
tion, and the subscript ip refers to the ion-pair
formed; for the particular case where the ion-
pair itself is an ion, equation (18) assumes the
more general form
ln(DAM/K6)solute =
n -
(20)
AAG\
RT ) t.
' cation
RT /
anion
M AAGN|
-aon- RT;_
/
AAG\
RT )
' i
cation
(l-aD)(na-nipc)
AAG\
RT /
anion
where the subscripts nipc and nipa are number
of moles of cation and anion respectively in-
volved in one mole of ion-pair. For example,
when K3Fe(CN)6 dissociates to give K+ and
Fe(CN)63- ions, and the ion-pair KFe(CN)62-,
nc = 3, na = 1, njpc = 1. and njpa = 1; when
K4Fe(CN)g dissociates to give K+ and
Fe(CN)64- ions, and the ion-pair KFe(CN)63-,
nc = 4, na = 1, nipc = 1. and nipa = 1. Further,
for the case of a feed solution which is subject
to partial hydrolysis, equation (18) becomes
ln(DAM/K5)solute =
(21)
AAG
RT /
AAG
RT )
anion
cation
1
«H
/ AAG\
\ RT /
hy
AAG
[\
RT /
OH- or H+
where «H represents the degree of hydrolysis,
and the subscript hy refers to the hydrolyzed
species resulting from the hydrolysis reaction
and the subscripts OH - and H + represent the
hydroxyl and hydrogen ions respectively. In
equations (19), (20), and (21), the applicable
values of «D and «H are those corresponding to
the boundary concentration XA2-
Using equation (18) (and equation (19) when
necessary) and data given in Table 1 for
calculating the DAM/KS values for MgClg,
NaaSO,!, and MgS04 solutes, the corresponding
data on solute separation and product rate
were calculated at a constant value of
k( = 20 x 10 - 4 cm/s) as a function of feed solu-
tion concentration (in the range 200 to 10,000
ppm of solute) for the CA-398 and PPPH-8273
membranes specified in Table 7. The results
obtained are given in Figure 4, which il-
lustrates the effect of chemical nature of solute
and that of membrane material on reverse
osmosis performance, and also the utility of
equations (18) and (19) together with the basic
transport equations (6) to (9) for predicting
such performance data. The latter is fui t-her il-
lustrated by data on solute separation pre-
sented in Appendix 1.
For a completely nonionized polar organic
solute (refs. 13,27),
ln(DAM/K5)soiute =
lnA<
(22)
RT /
6*2E
87
-------
CA-398 MEMBRANE
PPPH-8273 MEMBRANE
OJ
o 2
o»
of
a.
MgS04
Na2S04
MgS04
SOLUTE SEPARATION, %
-------
-1.0
-2.0
CA-398
MEMBRANES
PPPH-8273
MEMBRANES
\
\
1.0
2.0
Figure 5. Correlation of In A* versus
3.0
4.0
5.0
for CA-398 and PPPH-8273 membranes at 25° C.
(vs) of the product solution through the mem-
brane may be expressed as
AP
c
(23)
Applying the basic transport equations (6) to (9)
to such dilute feed solutions, DAM/K6 for the
solute involved can be expressed as (ref. 6):
(DAM/K8)solute -
and hence
(1
f =
(DAM/K5)solute
exp<
-1 (24)
(25)
Equation (25) is an extremely useful relation-
ship; it enables one to calculate solute separa-
tion f as a function of (DAM/K6)soiute for any
given value of k for any specified membrane.
The. results of such calculations for the CA-398
and PPPH-8273 membranes specified in Table
7 are presented in Figure 6 as f vs.
In (DAM/K£)solute for different k values. These
correlations are applicable to both inorganic
and organic solutes (whether ionized or non-
ionized) so long as the feed solution involved
satisfies the criteria stated above for dilute
solutions, and (DAM/K6)solute ig independent of
XA2- For such feed solution systems, one can
calculate ln(DAM/K<5)solute f°r anv solute from
one of the applicable relations from the set of
equations (18) to (22), and then directly read the
corresponding solute separation f from Figure
6 for a wide range of k-values for the mem-
branes specified in Table 7; thus the practical
utility of the correlations of the type shown in
Figure 6 is obvious.
In view of the practical importance of dilute
feed solutions in the application of reverse
89
-------
CA-398 MEMBRANE
PPPH-8273 MEMBRANE
lOOi-
-12 -10
-6
-4 -2 -12 -10
in(DAM/K8)so|ufe
Figure 6. Correlation of solute separation versus ln(DA,vi/k8)Solute for tne CA-398 and PPPH-8273
membranes specified in Table 7 for dilute aqueous feed solutions. Operating pressure =
405 psig.
osmosis to water pollution control, solute
separation data for the membranes specified in
Table 7 were calculated for a large number of
inorganic and organic solutes using the ap-
plicable relationships from the set of equations
(18) to (25), and the physicochemical data given
in Tables 1 to 6; the results obtained are given
in Appendix 1.
The effect of operating pressure P on solute
separation /is of interest in many water pollu-
tion control applications. According to equa-
tion (25), when the values of (DAM/K5)soiute and
k remain constants and vs 1. These predictions of
equation (25), which are both of fundamental
and practical interest, have been experimental-
ly verified (ref. 29).
Many polluted waters contain organic acids
which are dissociated to different extents
depending upon their concentrations. Reverse
osmosis separations of such acids in dilute solu-
tions can be calculated by determining sepa-
rately the applicable (DAM/KS) value for the
ionized part from equation (18), and that for the
nonionized acid from equation (22), and then
calculating solute flux NA for each part
separately from the general relation
NA = (DAM/KS)c(XA2-XA3>.
(26)
From the values of NB (= AP) and total solute
flux (NA)total ( = NA for ionized part + NA for
nonionized acid), XAS can be calculated from
the relation
XA3 = (NA)total/ {NB + (NA)t0tal} (27)
and then the solute separation. / can be cal-
culated from equation (14). Further computa-
tional details of this method of calculation are
given in reference 14. Using this method, sol-
ute separation data for several organic acids
were calculated as a function of feed concentra-
tion at a specified k value for the CA-398 mem-
brane specified in Table 7, and the results ob-
tained are given in Figure 7; additional data on
f for a number of such organic acids are in-
cluded in Appendix 1.
90
-------
CA-398 MEMBRANE
TARTARIC ACID
CITRIC ACID
LACTIC ACID
D,L-MALIC ACID
BENZOIC
ACID
SUCCINIC
ACID
ACETIC
ACID
CITRIC ACID
TARTARIC ACID/
. .
D,L- MALIC ACID
LACTIC ACID
SUCCINIC ACID
r—BENZOIC ACID
\ ACETIC ACID
5 10 50 100 500
SOLUTE CONCENTRATION IN FEED, ppm
1000
Figure 7. Effect of feed concentration on the performance of the CA-398 membrane specified in
Table 7 for aqueous feed solutions of partially dissociated organic acids. Operating
pressure = 405 psig; k = 20 X 10" 4 cm/s.
Predictability of Membrane Performance
for Aqueous Feed Solution Systems
Involving Mixed-Solutes
The foregoing reverse osmosis transport
equations refer specifically to single-solute
aqueous feed solution systems. Most polluted
waters however contain more than one solute.
Hence the predictability of membrane per-
formance for feed solution systems involving
mixed-solutes is of real interest in the applica-
tion of reverse osmosis for water pollution con-
trol. Even though a comprehensive analytical
technique for predicting such membrane per-
formance in reverse osmosis for all kinds of
mixed-solute systems from data on membrane
specification only is yet to be developed, con-
siderable progress has been made at least with
respect to certain kinds of mixed-solute
systems (refs. 1,18,30,31). The latter include (1)
mixture of any number of electrolytic solutes
involving a common ion (ref. 1), (2) mixture of
any number of nonionized organic solutes with
no mutual solute-solute interactions (refs. 18,
30) and (3) mixture of two uni-univalent elec-
trolytic solutes without a common ion (ref. 31).
The techniques involved for predicting mem-
brane performance (i.e., both solute separation
and rate of membrane permeated product) for
the above kinds of mixed-solute systems are
91
-------
described fully in the indicated references.
Briefly, the techniques involved for the mixed-
solute systems (1) and (2) use the reverse
osmosis transport equations given above
treating each solute independently, so that the
net result is simply the additive effect of each
individual component in the mixed-solute
system. In the prediction technique for the
mixed-solute system (3), the basic reverse
osmosis transport equations are written for
each ion along with the necessary additional
equations for overall electroneutrality for the
system; these equations together with equa-
tion (18) written for each possible electrolytic
solute combining the cations and the anions
present in the system, yield a set of equations
which can be solved to give the necessary data
on ion separations and product rates from data
on membrane specifications only. This latter
technique has recently been extended to
mixed-solute systems involving one mono-
valent-monovalent electrolyte and one
divalent-monovalent electrolyte (ref. 32). These
techniques (refs. 31, 32) are simple in principle
and fundamental in approach but require con-
siderable effort in solving the computational
complexities involved. For the purpose of il-
lustration, the mixed solute system sodium
acetate + calcium nitrate was analyzed with
respect to the CA-398 membrane specified in
Table 7, and the results obtained are shown in
Figure 8. These results are significant from the
o
O>
of
CL
2.0
1.0
J_
TOTAL MOLALITY : 0.2
NO:
0.2 0.4 0.6 0.8
mol CH3COONa/{mol CH3COONa + mol
.0
Figure 8. Effect of relative concentration of CH3COONa on ion flux and product rate for 0.2
molal feed solution CH3COONa-Ca(NO3)2-H2O during reverse osmosis operation
us.ng the CA-398 membrane specified in Table 7. Operating pressure = 405 psig;
92
-------
points of view of both water pollution control
and fractionation of dissolved solutes by
reverse osmosis.
Predictability of Performance of
Commercial Reverse Osmosis Modules
The basic transport equations given above
(equations (6) to (9)) apply to any point in the
reverse osmosis system. With reference to
commercial reverse osmosis modules, this
means that solute separation and product rate
values predicted by the above equations cor-
respond to values obtainable at module en-
trance or at zero product water recovery. The
true performance of an entire module, how-
ever, involves a finite fraction recovery (A) of
product water whose concentration changes
continuously from the entrance end to the exit
end of the module. Thus the module as a whole
has an average solute separation (/) and an
average product rate. This performance of the
module as a whole can be predicted by apply-
ing the basic transport equations to the entire
reverse osmosis system and analyzing the
various relationships applicable to the system.
Such system analysis is simple in principle,
fundamental in approach, and often complex in
computation. The technique for such analysis
has been developed in detail with particular
reference to water treatment applications and
CA-398 membranes (refs. 33,34). The following
is a brief outline of this technique.
System Specification
Any reverse osmosis system may be
specified in terms of three nondimensional
parameters y, 6, and X defined as:
7 =
(28)
osmotic pressure of initial feed solution
operating pressure
(DAM/K5)
w
_ solute transport parameter (29)
pure water permeation velocity '
(DAM/K6)
mass transfer coefficient on
the high pressure side of membrane
solute transport parameter
(30)
The quantity v£ is obtained from the relation
AP
c
(31)
and the quantity iKX^) refers to the osmotic
pressure of the feed solution at membrane en-
trance in a flow process or start of operation in
a batch process. The quantities y, 6, and
X0 (= k/Vw) may be described as the osmotic
pressure characteristic, membrane charact-
eristic, and the mass transfer coefficient
characteristic respectively for the reverse
osmosis system under consideration. The
significance of system specification is that a
single set of numerical parameters can repre-
sent an infinite number of membrane-solution-
operating systems; conversely, any two
membrane-solution-operating systems can be
simply and precisely differentiated in terms of
unique combinations of numerical parameters.
System Analysis
Figure 9 indicates the performance
parameters of a reverse osmosis system. The
dimensionless concentration C is defined in
terms of actual concentration ratios as
so that
C = XA/XA1
GI = XAI/XAI
C2 = XA2/X&!
C3 =
C3 =
(32)
(33)
(34)
(35)
(36)
(37)
(38)
(39)
-------
Thus the quantities GI, €2, and CB represent
the concentrations of the bulk solution and the
concentrated boundary solution on the high
pressure side of the membrane and the mem-
brane permeated product solution on the low
pressure side of the membrane respectively at
any point in the system; the quantities C°, C%,
and €3 are the values of Ci, C& and Cs respec-
tively at membrane entrance in flow process,
or at start of_pperation in batch process; and,
the quantity Ca is the average value of €3 cor-
responding to a specified fraction A of product
water recovery. Further, a dimensionless
longitudinal length parameter X for a flow
process, and a dimensionless time parameter r
for a batch process are defined as:
and
VW X
X = =o ~
u h
T =
(40)
(41)
When (1/h) • (S/V°), and (x/Tr0)st, the
numerical values of X and T are identical. The
quantities Ci, C2, C3, 63, r or X, and A repre-
sent the main performance data for any
specified reverse osmosis system.
On the basis of the transport equations (6) to
(9), and the foregoing definitions, and a few
reasonable assumptions applicable for most
water treatment applications of reverse
osmosis, it has been shown that in any reverse
osmosis system specified in terms of 7,6, and X,
the quantities vw( = local permeation velocity
of product water) and v£, C% and Ca, Ci and €3,
Ci and GS, €3 andr or X, and 63 and A are
uniquely related as follows:
= C
(42)
(43)
(44)
cV = i =
(45)
exp (46)
+ CsA = 1 (47)
A = 1 - exp(-Z) (48)
where Z =
X(7C3 + 0)
and
- 7 +
(7C3 + 6) (50)
(7C3 + 0)exp
exp (-Z)
sM'
' ^
0
dC,
The above relationships show that with
respect to a specified reverse osmosis system,
if any 4 of the 12 quantities of 7,6, X (or X0), Z, A,
Ci, Cz, C§, €3, Cjj, €3, T or X are fixed, the re-
maining 8 quantities can be obtained by
simultaneous solution of equations (42) to (50).
Illustrative Calculations
A commercial ROGA-4000 reverse osmosis
module with a cellulose acetate (assumed to be
CA-398) membrane has the following dimen-
94
-------
C|
0
u°
lllllllllllllimmilllllllllllllllllllllllimimilimUllllMlmillllfllllllllllll
Figure 9. Performance parameters for a reverse osmosis system.
sional characteristics:
effective area of film
surface = 5.763 x 104 cm2,
longitudinal length for feed flow,
x = 73.7 cm,
membrane area per unit volume of fluid
space, h"1 = 18.19 cm"1, and
cross sectional area of module = 44.15
cm2.
A single reverse osmosis experiment with the
above module gave the following data for a
3,500 ppm NaCl-H20 reference feed solution
whose osmotic pressure is 40 psi:
operating pressure
pure water permeation
rate
feed flow rate at module
entrance
product rate
= 400 psig
= 0.75 gal/min
= 3.45 gal/min
= 0.656 gal/min
flow rate of concentrate = 2.794 gal/min
concentration of NaCl
in product = 140 ppm.
From the above data, the values for the four
quantities 7, A, Ca, and X can be calculated to
be as follows:
7
A
Uu
X
x(3,500 ppm NaCD/400 = 0.1
product rate/feed flow rate = 0.19
NaCl concentration in
product/Nad concentration in
feed = 0.04
pure water permeation rate/
effective area of film
surface = 8.242 x 10 - 4 cm/s
feed flow rate/cross sectional area
of module = 4.95 cm/s
x/u°h = 0.2232.
v:
Solving equations (42) to (50) simultaneously,
the values of 6 and X0 corresponding to the
above values of 7, A, C& and X are obtained as
6 = 0.02176, and
X0 = 2.233
from which the values of (DAM/K6)NaCl f°r the
membrane in the module and k]staci for the
95
-------
operating conditions of the module can be cal-
culated to be
(DAM/K6)NaCi = 1.793 x 10 - 5 cm/s, and
kNaCl = 18.4 x 10 ~4 cm/s.
From the above values of (DAM/K5)fjaci and
kNaCi,thecorrespondingvaluesof(DAM/K5)solute
and ksoiute for different solutes can be
calculated using equation (11) and equations
(18) to (22) discussed earlier. Using these values
of ODAM/K6)solute and ksoiute, and the value of v£
obtained already for the module, the reverse
osmosis system can be respecified in terms of
y, 6, and X for any desired feed solution system.
These latter three parameters, together with
the already calculated value of X for the
module, yield the_pther performance data A,
Ci, Cz, 63, and €3 for the solution system
under consideration.
For further illustration, a few numerical
data for the above ROGA module were calcu-
lated by system analysis with particular refer-
ence to very dilute feed solutions for which 7
may be assumed to be equal to zero. For such a
case, the numerical values of A and X become
identical, and Figure 10 shows the effects of
variations of (1) ktfaCl and (2) nature of solute on
20
SODIUM CHLORIDE
kxlO
cm/s
7
0
<0.0
'0.02176
18.4,
100,00
f
1.00-
•
0.96="
•
0.92-
20
10
Cl
i I i I
kNoCI =
18.4 x 1
0~4 cm/s
y=o.o
©NoCl=0.02176
i-PROPYL ALCOHOL—i
GLYCEROL-i
Figure 10. Some results of system analysis for a reverse osmosis module for very dilute
feed solutions.
96
-------
average solute separation / (= 1 - Ca), and the
ratio, solute concentration in the concentrate
leaving the system/solute concentration in the
feed entering the system = Ci, as a function of
fraction product water recovery. For the par-
ticular ROGA module analyzed, when 7 = 0,
0 = 0.02176 and X0 = 2.233, A = 0.2232; the values
of /and Ci corresponding to this value of A are
indicated in Figure 10 by dotted lines.
The foregoing calculations illustrate the ap-
plication of basic transport equations and
system analysis for predicting the perfor-
mance of reverse osmosis membranes and
modules; conversely, one can also perform
similar calculations to specify the membranes
and modules needed to obtain a given reverse
osmosis performance.
A FUNDAMENTAL APPROACH TO SOME
OF THE OTHER ASPECTS OF REVERSE
OSMOSIS RELEVANT TO WATER
POLLUTION CONTROL
Preferential Sorption of Solute at
Membrane-Solution Interfaces
The transport equations and predictability
techniques discussed above apply specifically
to reverse osmosis systems where water is
preferentially sorbed at the membrane-solu-
tion interface. There are a large number of
practically important reverse osmosis systems
where solute is preferentially sorbed at the in-
terface. For example, with respect to cellulose
acetate membranes and aqueous solutions,
such systems involve solutes such as phenols,
chlorinated and/or aromatic hydrocarbons,
mineral oils, higher alcohols, and many un-
dissociated organic compounds. For such
reverse osmosis systems, comprehensive
precise criteria for preferential sorption of
solute at membrane-solution interface, and
transport equations and predictability tech-
niques similar to those described above are yet
to be fully developed. However, certain
characteristics of such systems are known
(refs. 26,35,36,37). When solute is preferentially
sorbed at the membrane-solution interface,
solute separation in reverse osmosis can be
positive, negative, or zero, depending on the
nature of solute, membrane, and experimental
conditions; under otherwise identical ex-
perimental conditions, solute separation can
pass through a maximum and minimum with
decrease in average pore size on the membrane
surface, and solute separation generally
decreases with increase in operating pressure.
Solute preferential sorption results in pore-
blocking effect, so that even when the osmotic
pressure effects involved are negligible, the
PR/PWP ratio is significantly less than unity.
The quantities 14PR/PWP) (which is a measure
of pore-blocking effect) and NA (solute flux) are
both unique functions of XAZ for the calculation
of which equation (9) is still applicable.
Temperature Effects
The effects of operating temperature on (1)
A, (DAM/K5)Soiute. and ksoiute. (2) the rate of any
chemical reactions involved (such as hydrol-
ysis, dissociation, etc.), and (3) the physical
properties of the membrane (compaction,
swelling, etc.), and the result of such effects on
the overall performance of the reverse osmosis
system are all yet to be studied in detail.
Effects of pH
The pH of the feed solution may change the
chemical nature of the solute species and/or
that of the membrane surface due to dissocia-
tion and/or hydrolysis reactions. Such changes
must naturally affect the performance of the
membrane during reverse osmosis. The effects
of such changes must be distinguished from the
effects of other parameters governing reverse
osmosis separations in the choice of ap-
propriate reverse osmosis systems for specific
applications.
Membrane Compaction
The factors governing membrane compac-
tion in reverse osmosis include (1) molecular
structure and physical properties of the mem-
brane material, (2) overall porous structure of
the membrane arising from specific membrane
making conditions, and (3) reverse osmosis ex-
perimental conditions. Precise quantitative
relationships between membrane compaction
and the above factors have not yet been
studied in detail.
97
-------
Membrane Fouling
The term "fouling" represents no definite
scientific phenomenon; it is just an incoherent
description of the occurrence of flux decline of
product water during reverse osmosis due to a
variety of reasons which include (1) pore-
blocking by suspended or colloidal matter pre-
sent in the feed solution, (2) pore-blocking due
to precipitation of some solid from the feed
solution during reverse osmosis, (3) pore-
blocking due to solute preferential sorption at
the membrane-solution interface, (4) high con-
centration polarization on the high pressure
side of the membrane, (5) low solute diffusivity
and low mass transfer coefficient on the high
pressure side of the membrane, (6) high viscosi-
ty of the boundary solution on the high pres-
sure side of the membrane, (7) pre-gel or gel
formation on the high pressure side of the
membrane, (8) membrane compaction, and (9)
chemical and physicochemical processes affect-
ing the chemical nature and porous structure
of the membrane. Often several of the above
reasons contribute simultaneously to flux
decline during reverse osmosis. Therefore
practical solutions to the problem of membrane
fouling must be based on an understanding of
the causes of such fouling in the particular
reverse osmosis application concerned.
Ultrafiltration (UF) and Reverse
Osmosis (RO)
The distinction between UF and RO is not
well defined, and it will continue to be so
because of historical and practical reasons.
This need not prevent a basic understanding of
these two processes which are extremely im-
portant for water pollution control applica-
tions. Simply because the osmotic pressure of
the solution involved is low or negligible, or
the molecular weight of the solute involved is
high, or the operating pressure for the system
is low, or the resulting water flux in the system
is high, the process involved does not become
UF nor does it cease to be RO. If UF is re-
stricted to size separation by mechanical sieve
filtration, and RO is restricted to solute separa-
tion by virtue of preferential sorption of sol-
vent or solute at the membrane-solution inter-
face, the two processes are mutually exclusive
and can be treated as such, even though both
processes may occur simultaneously in a given
system. Further, even when the process in-
volved is clearly a mechanical sieving opera-
tion, the chemical nature and porous structure
of the membrane surface may have significant
effects on the filtration characteristics of the
system. These considerations point to the need
for a critical study of the UF and RO processes
involved in practical membrane separation
systems nominally described as ultrafiltration.
CONCLUSION
The field of reverse osmosis is new, open,
visible, and versatile. In the present context of
widespread industrial pollution and public con-
cern on the quality of our environment, the ef-
fective utilization of reverse osmosis for water
pollution control would make the social rele-
vance of reverse osmosis second to none. For
such utilization, one must understand and
develop the science of reverse osmosis. The
physicochemical basis for reverse osmosis
separations, the materials science of reverse
osmosis membranes, and the engineering
science of reverse osmosis transport are the
major components of the science of reverse
osmosis-. All applications of reverse osmosis
arise from this science. The development and
practical utilization of this science calls for
dedicated efforts. This paper is part of such ef-
forts.
REFERENCES
1. S. Sourirajan, Reverse Osmosis, Chapters
1, 2, 3, and 6, and Appendix 1, Academic
Press, New York, 1970.
2. Reverse Osmosis and Synthetic Mem-
branes, S. Sourirajan, ed., National
Research Council Canada, Ottawa, 1977.
3. S. Sourirajan, Pure Appl. Chem., Vol. 50,
No. 7, p. 593,1978.
4. S. Sourirajan, Reverse Osmosis and Syn-
thetic Membranes, S. Sourirajan, ed.,
Chap. 1, National Research Council
Canada, Ottawa, 1977.
5. S. Sourirajan, Ind. Eng. Chem. Funda-
mentals, Vol. 2, p. 51,1963.
6. S. Sourirajan and T. Matsuura, Reverse
Osmosis and Synthetic Membranes, S.
98
-------
Sourirajan, ed., Chapters 2 and 3, Nation-
al Research Council Canada, Ottawa, 1977.
7. L. E. Applegate and C. R. Antonson,
Reverse Osmosis Membrane Research, H.
K. Lonsdale and H. E. Podall, eds., p. 243
Plenum, New York, 1972.
8. T. Matsuura, P. Blais, L. Pageau, and S.
Sourirajan, Ind. Eng. Chem. Process Des.
Dev., Vol. 16, p. 510, 1977.
9. T. Matsuura, P. Blais, and S. Sourirajan,
J. AppL Polym. Sci., Vol. 20, p. 1515,1976.
10. T. Matsuura and S. Sourirajan, Ind. Eng.
Chem. Process Des. Dev., Vol. 17, p. 419,
1978.
11. S. Sourirajan and B. Kunst, Reverse Os-
mosis and Synthetic Membranes, S. Sour-
irajan, ed., Chapter 7, National Research
Council Canada, Ottawa, 1977.
12. T. Matsuura, L. Pageau, and S. Sourira-
jan, J. AppL Polym. Sci., Vol. 19, p. 179,
1975.
13. T. Matsuura, J. M. Dickson, and S.
Sourirajan, Ind. Eng. Chem. Process Des.
Dev., Vol. 15, p. 149, 1976.
14. T. Matsuura, J. M. Dickson, and S.
Sourirajan, Ind. Eng. Chem. Process Des.
Dev., Vol. 15, p. 350, 1976.
15. R. Rangarajan, T. Matsuura, E. C. Good-
hue, and S. Sourirajan, Ind. Eng. Chem.
Process Des. Dev., Vol. 15, p. 529, 1976.
16. R. Rangarajan, T. Matsuura, E. C. Good-
hue, and S. Sourirajan, Ind. Eng. Chem.
Process Des. Dev., Vol. 17, p. 71, 1978.
17. R. W. Taft, Jr., Steric Effects in Organic
Chemistry, M. S. Newman, ed., pp.
556-675, Wiley, New York, 1956.
18. T. Matsuura, M. E. Bednas, and S.
Sourirajan, J. Appl. Polym. Sci., Vol. 18,
p. 567,1974.
19. T. Matsuura, M. E. Bednas, J. M. Dickson,
and S. Sourirajan, J. Appl. Polym. Sci.,
Vol. 18, p. 2829, 1974.
20. T. Matsuura, M. E. Bednas, J. M. Dickson,
and S. Sourirajan, J. Appl. Polym. Sci.,
Vol. 19, p. 2473, 1975.
21. L. G. Sillen and A. E. Martell, "Stability
Constants of Metal Ion Complexes," Spe-
cial Publication No. 17, The Chemical So-
ciety, London,1964.
22. C. P. Bean, Research and Development
Progress Report, No. 465, OSW, U.S.
Dept. Interior, Washington, DC, 1969.
23. T. Matsuura and S. Sourirajan, J. Appl.
Polym. Sci., Vol. 18, p. 3593, 1974.
24. E. N. Pereira, T. Matsuura, and S.
Sourirajan, J. Food Sci., Vol. 41, p. 672,
1976.
25. P. A. Small, J. Appl. Chem., Vol. 3, p. 71,
1953.
26. T. Matsuura and S. Sourirajan, J. AppL
Polym. Sci., Vol. 17, p. 3683,1973.
27. T. Matsuura, A. G. Baxter, and S. Sourira-
jan, Ind. Eng. Chem. Process Des. Dev.,
Vol. 16, p. 82, 1977.
28. S. Kimura and S. Sourirajan, Ind. Eng.
Chem. Process Des. Dev., Vol. 7, p. 539,
1968.
29. F. Hsieh, T. Matsuura, and S. Sourirajan,
J. AppL Polym. Sci., in press.
30. T. Matsuura and S. Sourirajan, Ind Eng.
Chem. Process Des. Dev., Vol. 10, p. 102,
1971.
31. R. Rangarajan, T. Matsuura, E. C. Good-
hue, and S. Sourirajan, Ind. Eng. Chem.
Process Des. Dev., Vol. 17, p. 46,1978.
32. R. Rangarajan, T. Matsuura, E. C.
Goodhue, and S. Sourirajan, Ind. Eng.
Cheni. Process Des. Dev., Vol. 23, p. 561,
1979.
33. H. Ohya and S. Sourirajan, Reverse
Osmosis System Specification and Per-
formance Data for Water Treatment Ap-
plications, The Thayer School of Engi-
neering, Dartmouth College, Hanover,
NH, 1971.
34. S. Sourirajan and H. Ohya, Reverse Os-
mosis and Synthetic Membranes, S.
Sourirajan, ed., Chapter 4, National Re-
search Council Canada, Ottawa, 1977.
35. T. Matsuura and S. Sourirajan, J. AppL
Polym. Sci., Vol. 16, p. 2531, 1972.
36. T. Matsuura and S. Sourirajan, J. AppL
Polym. Sci., Vol. 17, p. 3661,1973.
37. J. M. Dickson, T. Matsuura, and S.
Sourirajan, "Transport Characteristics in
the Reverse Osmosis System p-Chloro-
phenol-Water-Cellulose Acetate Mem-
brane," Ind. Eng. Chem. Process Des.
Dev., in press.
99
-------
APPENDIX 1.
(a) CALCULATED DATA ON REVERSE OSMOSIS SEPARATIONS OF DIFFERENT SOLUTES
FOR THE CA-398 MEMBRANE SPECIFIED IN TABLE 7
Operating conditions: very dilute feed solutions; Operating pressure: 405 psig; k|yjag| = 20 x 10" cm/s
Solute
NaCI
Inorganic solutes
AI(N03)3
NH4Br
NH4CI
NH4I
NH4N03
BaBr2
BaCI2
Bal2
Ba(N03)2
Cd(N03)2
CdS04*
CaBr2
CaCI2
Cal2
Ca(N03)2
Ce(N03)2
CsBr
CsCI
Csl
CsN03
Cr(N03)3
Co(N03)2
CoS04*
Cu(N03)2
CuS04*
FeCI2
Fe(N03)2
Fe(N03)3
LaCI3
La(N03)3
Pb{N03)2
i • f\
LiBr
1 " ^1
LiCI
LiF
Lil
LiN03
* • •*
MgBr2
kxlO4
cm/s
20
14.5
23.3
23.0
23.1
22.5
18.3
18.2
18.2
17.7
17.2
13.1
17.8
17.6
17.7
17.2
16.9
23.6
23.5
23.5
22.9
16.7
16.8
12.6
16.6
17.1
'16.9
16.7
15.4
17.3
16.7
18.3
18.0
17.9
16.5
18.0
17.6
17.0
Sepn.
%
96.8
99.5
95.8
96.4
94.5
92.6
99.1
99.4
98.5
97.2
96.5
>99.9
98.7
99.1
97.8
95.9
99.4
96.7
97.2
95.7
94.2
98.8
96.3
>99.9
97.4
99.1
98.6
93.7
99.7
99.4
94.3
97.5
96.2
96.7
97.9
95.0
93.2
98.9
Solute
MgCI2
Mgl2
Mg(N03)2
MgS04*
Mn(N03)2
MnS04*
Ni(N03)2
NiS04*
K acetate
K anisate
K benzoate
KBr
KBr03
K butyrate
K o-ctilorobenzoate
K p-chlorobenzoate
KCI
KCI03
KCI04
K2C03*
K2C204
K2Cr04
K2Cr207
K cyclohexylcarboxylate
KF
KHCOO
KHC03
KH phthalate
KH2P04
KHS04
K hydroxybenzoate
Kl
KI03
K3Fe(CN)6»
K4Fe(CN)6*
K isobutyrate
K m-nitrobenzoate
K o-nitrobenzoate
K p-nitrobenzoate
KN02
kx104
cm/s
16.9
16.9
16.5
13.0
16.6
13.1
16.6
13.1
18.2
15.6
16.4
23.3
20.6
16.8
16.5
16.4
23.0
21.7
22.1
18.5
11.7
19.5
16.7
16.3
20.6
20.5
18.9
20.1
18.0
19.8
16.4
23.1
18.2
19.0
18.8
16.8
16.5
16.5
16.5
20.9
Sepn.
%
99.2
98.1
96.5
>99.9
96.9
>99.9
97.2
>99.9
99.2
98.9
98.8
96.0
97.8
99.2
99.5
98.8
96.6
95.3
92.5
>99.9
99.9
99.9
98.3
99.4
97.8
97.5
98.5
97.1
99.3
• 99.4
98.8
94.8
98.9
>99.9
>99.9
99.3
99.1
99.6
99.1
94.0
100
(continued)
-------
APPENDIX 1 (continued).
Solute
KN03
KOH
K phenylacetate
K 4-phenylbutyrate
K (3-phenylpropionate
K pivalate
K propionate
K2S03
K2S04
K2S203
K tofuate
K valerate
RbBr
RbCI
Rbl
RbN03
Na acetate
Na anlsate
Na benzoate
NaBr
NaBr03
Na butyrate
Na o-chlorobenzoate
Na p-chlorobenzoate
NaCI03
NaCI04
Na2C03*
NaiCoOi
Na2Cr04
NaiCroOi
Na cyclohexylcarboxylate
NaP
NaHCOO
NaHCO,
O
NaH phthalate
NaH2P04
NaHS04
Na hydroxy-benzoate
Nal
Nal03
Na3Fe(CN)6*
Na4Fe(CN)6*
Na isobutyrate
Na m-nitrobenzoate
kxlO4
cm/s
22.5
29.3
16.0
14.2
14.5
16.3
17.0
18.7
19.4
19.7
16.4
16.6
23.7
23.5
23.5
23.0
16.5
14.4
15.0
20.1
18.3
15.4
15.1
15.0
19.1
19.3
16.1
10.9
16.9
14.8
14.9
18.3
18.2
16.9
17.9
16.3
17.8
15.0
20.0
16.4
16.3
16.0
15.4
15.0
Sepn.
%
93.0
99.4
98.8
99.0
99.0
99.7
99.3
99.8
99.8
>99.9
98.9
99.3
96.2
96.8
95.1
93.3
99.2
99.0
98.9
96.3
97.9
99.3
99.5
98.9
95.6
93.0
>99.9
99.9
99.9
98.6
99.4
98.0
97.7
98.6
97.3
99.4
99.4
98.9
95.2
99.0
>99.9
>99.9
99.3
99.2
Solute
Na o-nitrobenzoate
Na p-nitrobenzoate
NaN02
NaN03
NaOH
Na phenylacetate
Na 4-phenylbutyrate
Na /3-phenylpropionate
Na pivalate
Na propionate
Na2S03
Na2S04
Na2S203
Na toluate
Na valerate
SrBr2
SrCI2
Srl2
Sr(N03)2
Th(N03)4
Zn(N03)2
ZnS04*
Alcohols
Methanol
Ethanol
1-Propanol
2-Propanol
1-Butanol
2-Butanol
2-Methyl-1-propanol
2-Methyl-2-propanol
1-Pentanol
2-Pentanol
3-Pentanol
2-Methyl-l-butanol
3-Methyl-1-butanol
2,2-Dimethyl-1-propanol
1-Hexanol
2-Hexanpl
3-Hexanol
2-Methyl-1-pentanol
3-Methyl-l-pentanol
4-Methyl-1-pentanol
N104
cm/s
15.0
15.0
18.5
19.6
15.1
14.6
13.2
13.5
14.9
15.5
16.6
16.7
17.1
12.0
15.2
17.8
17.6
17.7
17.3
11.1
16.5
13.0
22.6
18.7
16.5
16.5
15.0
15.0
15.0
15.0
13.9
13.9
13.9
13.9
13.9
13.9
13.0
13.0
13.0
13.0
13.0
13.0
Sepn.
%
99.6
99.2
94.4
93.5
99.4
98.9
99.1
99.1
99.7
99.3
99.8
99.8
>99.9
98.8
99.3
98.9
99.2
98.1
96.4
99.9
96.3
>99.9
9.4
18.2
29.3
48.0
21.8
49.7
49.7
86.8
18.5
51.7
51.7
51.7
51.7
75.7
15.5
29.2
29.2
29.2
29.2
f\f\ A
29.2
(continued)
-------
APPENDIX 1 (continued).
===^=33=========
Solute
3-Methyl-2-pentanol
2-Methyl-3-pentanol
3,3-Dimethyl-2-butanol
1-Heptanol
2-Heptanol
3-Heptanol
4-Heptanol
2-Methyl-3-hexanol
4-Methyl-3-hexanol
3-Methyl-3-hexanol
2,2-Dimethyl-1-pentanol
2,4-Dimethyl-3-pentanol
2,2-Dimethyl-3-pentanol
1-Octanol
2-Octanol
4-Octanol
5-Nonanol
Cyclohexanol
Aldehydes
Acetaldehyde
Propionaldehyde
n-Butyraldehyde
i-Butyraldehyde
i-Valeraldehyde
Benzaldehyde
Hexanal
Ketones
Acetone
Methyl ethyl ketone
2-Pentanone
Methyl isopropyl ketone
3-Pentanone
Cyclopentanone
Methyl n-butyl ketone
Methyl s-butyl ketone
Methyl i-butyl ketone
Methyl t-butyl ketone
Ethyl propyl ketone
Ethyl i-propyl ketone
Cyclohexanone
Methyl amyl ketone
Methyl i-amyl ketone
-I^M «•!! —1 '
kxlO4
cm/s
13.0
13.0
13.0
12.3
12.3
113
12.3
12.3
12.3
12.3
12.3
12.3
12.3
11.7
11.7
11.7
11.2
13.9
19.5
17.0
15.3
15.3
14.0
14.2
13.3
17.0
15.3
14.0
14.0
14.0
14.3
13.1
13.1
13.1
13.1
13.1
13.1
13.4
12.4
12.4
Sepn.
%
53.1
53.1
90.5
13.1
27.6
27.6
27.6
55.2
55.2 .
82.6
82.6
74.8
90.2
11.2
26.0
26.0
24.8
71.0
44.5
45.9
47.7
46.7
47.8
34.7
17.8
27.8
29.5
34.1
46.0
31.7
53.3
36.1
44.3
42.1
58.9
36.6
40.1
39.5
21.5
37.1
Solute
Dipropyl ketone
Diisopropyl ketone
Acetophenone
Dibutyl ketone
Diisobutyl ketone
Benzyl methyl ketone
Ftfpre
L-dlCi «
Methyl acetate
Ethyl acetate
Methyl propionate
n-Propyl acetate
2-Propyl acetate
Ethyl propionate
Methyl butyrate
Methyl 2-methyl-propionate
n-Butyl acetate
2-Methyl-1-propyl acetate
2-Propyl propionate
Ethyl butyrate
Ethyl 2-methyl-propionate
Methyl pentanoate
Methyl 2-methyl-butyrate
Methyl 3-methyl-butyrate
2-Methyl-1 -butyl acetate
3-Methyl-l-butyl acetate
n-Butyl propionate
2-Propyl butyrate
2-Propyl 2-methyl-propionate
Ethyl pentanoate
Ethyl 2-methyl-butyrate
Ethyl 3-methyl-butyrate
Methyl caproate
Methyl 4-methyl-pentanoate
n-Hexyl acetate
n-Butyl butyrate
Ethyl caproate
Methyl heptanoate
2-Propyl caproate
Ethers
Methyl propyl ether
Methyl isopropyl ether
Methyl butyl ether
kxlO4
cm/s
12.4
12.4
12.5
11.4
11.4
11.9
16.3
14.8
14.8
13.7
13.7
13.7
13.7
13.7
12.9
12.9
12.9
12.9
12.9
12.9
12.9
12.9
12.2
12.2
12.2
12.2
12.2
12.2
12.2
12.2
12.2
12.2
11.6
11.6
11.6
11.6
11.2
14.9
14.9
13.8
Sepn.
%
41.6
4 rt 1
49.3
Aft f*
20.6
46.8
59.2
24.7
12.4
16.0
14.3
20.4
27.7
16.2
20.5
27.7
22.1
35.4
31.0
23.0
31.0
22.1
41.4
35.4
47.1
22.5
25.2
41.2
51.3
25.2
45.6
39.5
12.3
22.6
6.4
34.5
14.2
6.4
27.9
42.9
56.3
46.0
(continued)
-------
APPENDIX 1 (continued).
Solute
Ethyl propyl ether
Ethyl isopropyl ether
Methyl amyl ether
Ethyl butyl ether
Ethyl t-butyl ether
Oipropyl ether
Oiisopropyl ether
Ethyl amyl ether
Isopropyl t-butyl ether
Anisole
Dibutyl ether
Diisobutyl ether
Phenetole
Carboxylic acids
Acetic acid*
Propionic acid*
n-Butyric acid*
i-Butyricacid*
Valeric acid*
Pivalkacid*
Cyclohexane-carboxylic acid*
Benzoic acid*
Phenylacetic acid *
m-Toluic acid*
Anisic acid*
(3-Phenylpropionic acid*
4-Phenylbutyric acid*
Oxalic acid *
Malonic acid*
Succinic acid*
Lactic acid*
D,L-Malicacid*
kxlO4
em/s
13.8
13.8
12.9
12.9
12.9
12.9
12.9
12.2
12.2
13.6
11.6
11.6
12.7
17.0
15.1
13.7
13.7
12.7
12.7
12.0
12.8
12.0
12.0
11.7
11.4
10.8
15.8
14.4
14.0
14.4
12.6
Sepn.
%
47.7
60.8
29.6
51.3
90.9
62.3
82.4
33.7
96.7
21.5
67.2
911
24.8
28.2
29.4
30.9
30.5
32.3
65.2
53.9
22.2
23.2
22.3
23.8
24.1
25.7
95.6
66.3
56.9
79.0
89.1
Solute
Tartaric acid*
Citric acid*
Polyfunctional compounds
1,2-Ethanedoil
1,3-Propanediol
Propylene glycol
Gtycerol
1,3-Butanediol
2,3-Butanediol
i-Erythritol
3-Hydroxy-2-butanone
1,5-Pentanediol
Arabitol
Xylitol
Adonitol
L-Arabinose
1,6-Hexanediol
cis- and trans-1,2-cyclohexanediol
trans-1,2-Cyclohexanediol
1,2,6-Hexanetriol
Pinacole
D-Sorbitol
Dulcitol
4-Hydroxy-4-methyl-2-pentanone
Acetonyl acetone
D-Glucose
D-Mannose
D-Galactose
D-Fructose
Sucrose
Maltose
Lactose
kxlO4
cm/s
12.3
10.9
17.9
16.0
16.0
15.5
14.5
14.5
13.9
15.0
13.5
12.7
12.7
12.7
13.0
12.7
13.5
13.5
12.5
12.7
11.9
11.9
13.0
13.3
12.1
12.1
12.1
12.1
9.0
9.0
9.0
Sepn.
%
97.0
98.3
66.1
68.2
72.4
85.5
74.1
77.8
95.0
58.2
71.9
98.4
98.4
98.4
97.9
55.5
89.2
89.2
89.0
97.8
99.5
99.5
84.4
39.6
99.3
99.3
99.3
99.3
99.9
99.9
99.9
(continued)
•Separations data correspond to solute concentrations in feed = 200 ppm. See also NOTE at the end of Appendix.
-------
APPENDIX 1 (continued).
(b) CALCULATED DATA ON REVERSE OSMOSIS SEPARATIONS OF DIFFERENT SOLUTES
FOR THE PPPH-8273 MEMBRANE SPECIFIED IN TABLE 7
Operating conditions: very dilute feed solutions; Operating pressure: 405 psig; kNaCJ = 20 x 10 cm/s
Solute
NaCI
Inorganic Solutes
CsBr
CsCI
Csl
LiBr
LiCI
UF
Li I
KBr
KCI
KF
Kl
RbBr
RbCI
Rbl
NaBr
NaF
Nal
Alcohols
Methanol
Ethanol
1-Propanol
2-Propanol
1-Butanol
2-Butanol
2-Methyl-1-propanol
2-Methyl-2-propanol
1-Pentanol
3-Methyl-1-butanol
3-Hexanol
2-Methyl-l-pentanol
3-Methyl-1-pentanol
4-Methyl-1-pentanol
3-Methyl-2-pentanol
2-Methyl-3-pentanol
3,3-Dimethyl-2-butanol
1-Heptanol
kxlO4
cm/s
20
23.6
23.5
23.5
18.0
17.9
16.5
18.0
23.3
23.0
20.6
23.1
23.7
23.5
23.5
20.1
18.3
20.0
22.6
18.7
16.5
16.5
15.0
15.0
15.0
15.0
13.9
13.9
13.0
13.0
13.0
13.0
13.0
13.0
13.0
12.3
Sepn.
%
98.4
95.6
95.6
95.6
98.1
98.1
98.5
98.1
95.8
95.7
98.8
95.8
95.7
95.7
95.8
98.4
98.8
98.4
25.3
70.0
80.2
87.0
83.9
91.5
86.9
98.7
89.4
94.6
94.7
94.8
94.5
95.9
97.1
98.4
97.9
92.6
Solute
2-Heptanol
3-Heptanol
4-Heptanol
2-Methyl-3-hexanol
4-Methyl-3-hexanol
3-Methyl-3-hexanoi
2,4-Dimethyl-3-pentanol
2,2-Dimethyi-3-pentanol
1-Octanol
Aldehydes
Acetaldehyde
Propionaldehyde
n-Buty raided yde
i-Butyraldehyde
i-Valeraldehyde
Benzaldehyde
Ketones
Acetone
Methyl ethyl ketone
?* Ppnt ft n n n P
c. rciiioiiuiic
Methyl isopropyl ketone
3-Pentanone
Cyclopentanone
Methyl isobutyl ketone
Cyclohexanone
Diisopropyl ketone
Acetophenone
Diisobutyl ketone
Benzyl methyl ketone
Esters
Methyl acetate
Ethyl acetate
Methyl propionate
n-Propyl acetate
2-Propyl acetate
Ethyl propionate
Methyl butyrate
kxlO4
cm/s
12.3
12.3
12.3
12.3
12.3
12.3
12.3
12.3
11.7
19.5
17.0
15.3
15.3
14.0
14.2
17.0
15.3
id n
l*t.U
14.0
14.0
14.3
13.1
13.4
12.4
12.5
11.4
11.9
16.3
14.8
14.8
13.7
13.7
13.7
13.7
Sepn.
%
97.8
97.2
96.2
98.6
97.2
99.1
97.6
98.4
90.8
64.3
72.2
87.0
93.2
96.8
75.0
64.6
71.9
RR S
Ou.3
93.0
78.4
89.9
94.2
95.3
98.4
74.5
99.6
87.8
81.6
86.3
86.4
93.3
96.1
89.8
93.3
(continued)
-------
APPENDIX 1 (continued).
Solute
Methyl 2-methyl propionate
n-Butyl acetate
2-Methyl-1-propyl acetate
2-Propyl propionate
Ethyl butyrate
Ethyl 2-methyl-propionate
Methyl pentanoate
Methyl 2-methyl-butyrate
Methyl 3-m ethyl butyrate
2-MethyM-butyl acetate
3-Methyl-1-butyl acetate
n-Butyl propionate
2-Propyl butyrate
2-Propyl 2-methyl-propionate
kxlO4
cm/s
13.7
12.9
12.9
12.9
12.9
12.9
12.9
12.9
12.9
12.2
12.2
12.2
12.2
12.2
Sepn.
%
96.1
94.7
97.9
97.2
95.1
97.2
94.7
98.6
97.9
99.1
94.8
96.3
98.7
99.3
Solute
Ethyl pentanoate
Ethyl 2-methyl-butyrate
Ethyl 3-methyl-butyrate
Methyl 4-methyl-pentanoate
n-Butyl butyrate
Ethers
Ethyl butyl ether
Ethyl t-butyl ether
Dipropyl ether
Diisopropyl ether
Isopropyl t-butyl ether
Anisole
Phenetole
kx104
cm/s
12.2
12.2
12.2
12.2
11.6
12.9
12.9
12.9
12.9
12.2
13.6
12.7
Sepn.
%
96.3
99.0
98.6
93.5
98.3
93.5
99.0
96.1
98.7
99.7
79.2
84.7
NOTE: Data on solute separations calculated above assume that, In each case, water is preferentially sorbed at the membrane-solution
interface during reverse osomosis. Available knowledge on precise physicochemical criteria on preferential sorption at interfaces
is still incomplete (3).
Nomenclature
C, Cf, Cj, Cy ^
D
AB
DAM/K6
AAF*
f
f
AG
AGB,AG,
AAG
-AAG/RT =
n ss
k
MA, MB
m
na, nc
nipa, nipc
P
PR
PWP
pure water permeability constant, g-mol of H20/cm2s atm
, C§ and C3 = concentration ratios defined by eq. (32), (33), (34), (35), (36), (37), (38), and (39) respectively
quantity defined by eq. (17)
molar density of solution, g-mol/cm3
diffusivity of solute in water, cm2/s
solute transport parameter (treated as a single quantity), cm/s
Taft's steric constant for the substituent group in the organic molecule
transition state free energy change
fraction solute separation defined by eq. (14)
average value of f
free energy of hydration of solute, kcal/g-mol
value of AG for the bulk solution phase and membrane-solution interface respectively
quantity defined by eq. (1)
polar free energy parameter
membrane area per unit volume of fluid space, cm ~J
mass transfer coefficient for the solute on the high pressure side of the membrane, cm/s
molecular weights of solute and water respectively
solute molality
solute flux and solvent flux respectively through membrane, g-mol/cm2s
number of moles of anion and cation respectively in one mole of ionized solute
number of moles of hydrolyzed species arising from the hydrolysis of one mole of solute
number of moles of anion and cation respectively in one mole of ion-pair
operating pressure, atm
product rate through given area of membrane surface, g/h
pure water permeation rate through given area of membrane surface, g/h
-------
R = gas constant
S = effective membrane area, cm2
s*. Es* = modified Small's number, cal'^cc'^g-mol"1
As* = quantity defined by eq. (5)
T = absolute temperature
t = time, s
F° - average fluid velocity at membrane entrance, cm/s
V? - initial volume of feed solution on the high pressure side of membrane, cm3
vs «= permeation velocity of product solution, cm/s
vw = permeation velocity of solvent water, cm/s
vj = pure water permeation velocity, cm/s
X = quantity defined by eq. (40)
XA = mole fraction of solute
x = longitudinal length of feed flow, cm
Z = quantity defined by eq. (49)
Greek Letters
aD
aH
&
y
A
mA*
8*
e
degree of dissociation
degree of hydrolysis
nonpolar and polar parameters respectively characterizing polymer membrane materials
a parameter characterizing polymer membrane material, representing affinity of polymer material for nonion-
ized polar organic solutes relative to that for an ionized inorganic solute
quantity defined by eq. (28)
structural group contributions to AGB and AG, respectively
constants in eq. (2) and (3) respectively
fraction product water recovery
quantity defined by eq. (22) when polar, steric and nonpolar parameters are each set equal to zero
coefficient associated with steric constant EEs
quantity defined by eq. (29)
osmotic pressure of solution corresponding to mole fraction XA of solute, atm
quantity defined by eq. (41)
structural group contribution to (S*£Es),jm
constant in eq. (4)
coefficient associated with modified Small's number £s*, cal"l/! cc~ 1/!g-mol
Subscripts
1 = bulk feed solution
2 = concentrated boundary solution on the high pressure side of membrane
3 = membrane permeated product solution on the low pressure side of membrane
Superscript
o = initial value, or value at membrane entrance
106
-------
HYPERFILTRATION OF N ON ELECTROLYTES: DEPENDENCE OF
REJECTION ON SOLUBILITY PARAMETERS
H. G. Spencer, J. L. Gaddis*
Abstract
The dependence of rejection of nonelectrolyte
solutes in single-solute water solutions on
solubility parameters is demonstrated using
data from previously reported hyperfiltration
studies. Solubility parameters characterizing
the hyperfiltration systems are evaluated and a
procedure for estimating the rejection of
solutes from their solubility parameters is
described.
SUMMARY
The dependence of hyperfiltration rejection
of nonelectrolyte solutes in single-solute water
solutions on solubility parameters is demon-
strated using hyperfiltration results reported
in the literature. The hyperfiltration systems
are characterized by a solubility parameter
derived empirically from the rejection-solu-
bility parameter dependence. A criterion for
high rejection follows.
INTRODUCTION
Hyperfiltration possesses high potential for
separating toxic solutes in industrial unit oper-
ation effluents (ref. 1). Some of the nonelectro-
lytes of concern are quite volatile and many are
only slightly soluble in water. Thus, the direct
experimental measurement of the salt rejec-
tion R. of the approximately 100 nonelectrolyte
priority pollutants would be difficult and a reli-
able method for predicting R. in a hyperfiltra-
tion system from a few reference measure-
ments and molecular properties of the solutes i
would be valuable.
The most detailed model developed for this
purpose has been provided by Sourirajan and
coworkers (ref. 2). Using one or more molecular
*Departments of Chemistry and Mechanical Engineer-
ing, respectively, Clemson University, Clemson, SC.
properties (acidity, basicity, Hammett and Taft
numbers, steric parameters, and Small's num-
ber) and the permeability and rejection of a
reference solution, one can relate these prop-
erties to R.. Other models use flux equations to
relate the measurable properties of a hyper-
filtration system (refs. 3,4). All approaches in-
clude both a transport property of the hyper-
filtration system and a coefficient for the
distribution of the solute between the bulk
solution and the barrier.
We have previously pointed out the value of
solute molecular weights in predicting R. of
nonelectrolytes (ref. 1). Most high rejection
hyperfiltration membranes effectively reject
nonelectrolytes with molecular weights
greater than about 80. Cellulose acetate is an
exception to this generalization. Although scat-
ter can be large in plots of rejection vs. molec-
ular weights, when the molecular weight is the
most reliable or perhaps the only molecular
property available it can be used to estimate R.
in systems characterized by a few measure-
ments.
This report demonstrates a dependence of R.
for nonelectrolytes on the solute solubility
parameter introduced by Hildebrand and Scott
(ref. 5) and characterizes the hyperfiltration
systems by solubility parameters. It also pro-
vides an empirical method for the rough esti-
mation of R. for a solute of known solubility
parameter in a hyperfiltration system from
values of R. obtained for a few reference
solutes without explicitly considering a trans-
port property for the solute, providing its
molecular volume is not vastly larger than
those of the reference solutes.
Chian and Fang (ref. 6) proposed the dif-
ference between the solubility parameters of
the solute and membrane plays the major role
in determining R. of nonelectrolyte solutes.
Klein et al. (ref. 3)'qualitatively related solute
permeabilities in the absence of hydraulic flux
with 2-dimensional solubility parameters and
used the experimental permeabilities to pre-
107
-------
diet specific separations of organic solutes
under hyperfiltration conditions. A quan-
titative correlation of R. with solubility param-
eters was not attempted in either report.
DEFINITIONS, CONCEPTS, AND
CALCULATIONS
The solubility parameter is defined by <5 =
(AE/V)I/2 where AE/V is the energy of
evaporation per unit volume, called the
cohesive energy density. The units of <5 are
(J/m3)1^. Solubility parameter theory predicts
that the best solvent for a given solute, e.g., a
polymer, is one whose solubility parameter is
equal to or close to that of the solute (ref. 5).
The rejection R. of a solute is defined as
1 - Cp/Cb, where Cp and Cb are concentrations
in the permeate and bulk feed solutions respec-
tively. An intrinsic rejection, 1 - CP/CW) based
on the concentrations of permeate and that oc-
curring at the feed-membrane interface Cw is
commonly defined. Normally the intrinsic re-
jection is projected as the infinite-velocity
asymptote of the rejection R. , and this intrin-
sic rejection is the property logically ad-
dressed in this study. Because most investiga-
tions are conducted so as to preclude large dif-
ferences in the two rejections, and virtually no
data exist projecting the intrinsic rejection,
the observed rejection is used throughout. Er-
rors produced by this simplification may be
substantial and are largest near rejections of
0.5.
In hyperfiltration the distribution of solute
between the bulk' solution and the barrier is
assumed to be important in determining R..
Further, assuming the concentration of solute
available for transport across the hyperfilter
depends on Ajm = 6j - 6m, where <5m char-
acterizes the hyperfilter, R. should be a func-
tion of A;m. Of course, an attempt to relate R. to
A;m alone is incomplete because a transport
property characterizing the hyperfiltration
system is not explicitly included.
The group contribution method of Konstam
and Fairheller (ref. 5) was used to calculate
Small's number (ref. 8) S, and the 6; were ob-
tained by 5j = S{/v"j where vj is the molar
volume. This method is used although tables
containing 6j for many solutes are available
(refs. 5,9). It is desirable to use a consistent
method for as many compounds as possible.
Even with the use of this general approach
some values of 6; are not available, especially
those for polyfunctional molecules. Values of vj
TABLE 1. CHARACTERISTICS OF HYPERFILTRATION SYSTEMS
Membrane
Pressure
(MPa)
Temperature
ID"3 6m
(J/m3)%
Classes of solutes
Reference
Aromatic polyamide 1.72
Aromatic polyamide,
Permasep B-9, hollow
fiber 3.10'
Cellulose acetate 1.72
Cellulose acetate 1.72
NS-100 5.52
Polyfether/amide)
PA-300 6.89
25
20
25
23-25
25
25
29.0 Alcohols, aldehydes, 11
ethers, ketones
30.0 Alcohols, acids 12
25.0 Alcohols, aldehydes 11
25.0 Alcohols, aldehydes, 15
esters, ethers, hydro-
carbons
34.5 Acids, alcohols, aide- 13
hydes, esters, ketones,
amines
(~34) * Acid, alcohols, aide- 14
hydes, esters, ketone,
chlorohydrocarbons
tpjot enough data for extrapolation.
108
-------
were calculated by dividing molecular weight
of the solute M; by its density ei in the liquid
state at the temperature of the experiment,
with Mj and Q{ obtained from commonly used
tables (ref. 10).
DEPENDENCE OF REJECTION ON
SOLUBILITY PARAMETERS
Figures 1 through 5 show the dependence of
R. on 6; for the six hyperfiltration systems
described in Table 1 (ret's. 11,12,13,14,15). The
8j occurring at R. = 0 is assumed to be the
solubility parameter characterizing the mem-
brane system and is designated 6m. The 6m
values were obtained by a visual linear extrap-
olation of the plots. A slope of -10 x 10'3
(m3/J)'/z was satisfactory for all graphs in the
region A;m < 0. Insufficient data are provided
to determine 5m by extrapolation in the
polylether/amide) thin film composite (PA-300)
system, Figure 4. Using the slope observed in
the remainder of the graphs, 6m should be in
the interval 34 x 103 < dm < 36 x 103(J/m3)'«.
It should be noted that the scatter is very large
in the cellulose acetate systems.
In the Permasep B-9 and cellulose acetate
systems (see Figures 2 and 5), several values of
R. at Aim > 0 are available. It is clear in Figure
2 that R. increases monotonically with increas-
ing Jkim in the region \m > 0. The dependence
is not well defined in the cellulose acetate case
where in this region several of the solutes
listed contain two functional groups and the
calculation of 6; is less reliable.
The molecular weight, or vj, may be used to
estimate R. in many systems, however this
dependence appears to be absent in the
cellulose acetate system. Figure 6 is provided
to illustrate this observation and it should be
compared with Figure 5, where the R. values
are plotted vs. o,.
DISCUSSION
The dependence of R. on fy has been il-
I.O,-//-
.8
o
s
3.4
.2-
0°0
oL//
^
18 20 22 24 26 28 30 32
SOLUBILITY PARAMETER I0~3 S\ (J/m3)l/2
Polyamide Membrane, 1.72 MPa, 25°C
Figure 1. Solute rejection vs. solubility parameter: Polyamide membrane, 1.72 MPa, 25° C (ref. 11).
109
-------
1.0
.6
z
o
UJ .
(T .4
.2
OL//
n/'
O O
18 20 ' 22 24 26 28 30 32
SOLUBILITY PARAMETER 10*
-------
.8
.6
z
a:
.4
.2
O
O O
o
O
Jf—L
_L
_L
±
J_
J_
16
18 20 22 24 26 28 30 32 34
SOLUBILITY PARAMETER IO"3 S\ (J/m3)l/2
36 38
PA - 300, 6.89 MPa, 25°C
Figure 4. Solute rejection vs. solubility parameter: poly(ether/amide),
PA-300, 6.89 MPa, 25° C (ref. 14).
lustrated in several hyperfiltration systems. It
is evident that high rejection occurs when
\m is large. Figure 7 provides a graph of R;
vs. Ajm for all the membranes other than the
cellulose acetate membranes and Figure 5 pro-
vides a graph of R. vs. 6j for the cellulose
acetate systems, where the scatter is greater.
This empirical treatment of rejection of non-
electrolytes in hyperfiltration systems also
provides a method for estimating R. for any
solute of known 6; from a few reference ex-
periments. The value of hm is determined from
a graph of R. vs. fy obtained for the reference
solutes. Solutes having .Ajm < -10 x IO3
(J/m3)1/z should have R. > 0.90. In the region
-10 x 103 < Aim
-------
•Or//
.8
.6
o
ui
-3
UI „
or .4
.2
14 16 18 20 22 24 26 28 30
SOLUBILITY PARAMETER IO"3 S\ (J/m3)l/2
Cellulose Acetate, 1.72 MPa, 23-25°C
Figure 5.Solute rejection vs. solubility parameter: cellulose acetate,
1.72 MPa, 23-25° C (refs. 11, 15).
32 34
are much smaller. Hyperfiltration systems
with 6m greater than 36 x 1Q3 or 38 x 103
(J/m3)l/z should provide high rejections of these
solutes.
Refinement of this approach using solubility
parameters as a measure of the membrane-
solute interaction will likely require incorpora-
tion in a flux model.
ACKNOWLEDGMENT
The authors wish to acknowledge the gen-
erous financial support of this work by the U.S.
Environmental Protection Agency, Industrial
Environmental Research Laboratory, Re-
search Triangle Park, NC, EPA Grant Number
R805777-1.
112
-------
1.0
.8
.6
.4
(£
.2
_L
e
o
o
o
o
o
0 C
o o o
00 o
00 °
o
0°
°0
o o
o
o
D° O O
o o
JL
0 20 40 60 80 [00120 I40
MOLECULAR WEIGHT Mj
Cellulose Acetate, 1.72 MPa, 23-25°C
Figure 6. Solute rejection vs. molecular weight: cellulose acetate, 1.72 MPa, 23-25° C (ref. 15).
REFERENCES
1. H. G. Spencer, J. L. Gaddis, and C. A. Bran-
don, "Membranes for Toxic Control," pre-
sented at the Membrane Separation Tech-
nology Seminar, Clemson University,
Clemson, SC, 1977.
2. S. Sourirajan and T. Matsuura, "Reverse
Osmosis and Synthetic Membranes," S.
Sourirajan, ed., National Research Coun-
cil of Canada Publications, Ottawa, Cana-
da, Chapter 2, 1977.
3. K. S. Spiegler and 0. Kedem, Desalina-
tion, Vol. 1, p. 311, 1966; H. K. Lonsdale,
U. Merten, and R. L. Riley, J. Appl.
Polymer Sci, Vol. 9, p. 1341, 1965; and L.
Dresner and J. S. Johnson, Jr., Principles
of Desalination, 2nd ed., K. S. Spiegler
4.
6.
7.
and A. D. K. Laird, eds., Academic Press,
New York, in press.
E. Klein, J. Eichelberger, C. Eyer, and J.
Smith, Water Res., Vol. 9, p. 807, 1975.
J. H. Hildebrand and R. L. Scott, "The
Solubility of Nonelectrolytes," Rheinhold,
New York, 1950.
E. S. K. Chian and H. H. P. Fang, AIChE
Symposium Ser., Vol. 70, p. 497, 1973.
H. H. Konstam and W. R. Fairheller, Jr.,
AIChE J., Vol. 16, p. 837, 1970.
8. P. A. Small, J. Appl. Chem., Vol. 3, p. 71,
1953.
9. J. L. Gordon, Encyclopedia of Polymer
Science and Technology, Vol. 3, p. 833,
1965; H. Burrell, J. Paint Technol.,Vo\. 27,
p. 726, 1955; C. M. Hansen, Ind. Eng.
Chem., Prod. Res. Dev.,Vo\. 8, No. 2,1960.
113
-------
0
-18
-16 -14 -12 -10 -8 -6 -4 -2 0 +2 +4 + 6
RELATIVE SOLUBILITY PARAMETER I0~3 A im (J/m3)172
Noncellulosic Membranes
Figure 7. Solute rejection vs. relative solubility parameter Aim = 8i - 5m
for the noncellulosic membranes.
+ 8
10. Handbook of Chemistry and Physics,
R. E. Weast, ed., CRC Press, Inc.,
Cleveland, OH. 14.
11. J. M. Dickson, T. Matsuura, P. Blais, and
S. Sourirajan, J. Appl. Polymer Sci., Vol.
19, p. 801, 1975.
12. V. B. Caraceiolo, N. W. Rosenblatt, and V.
J. Tomsic, "Reverse Osmosis and Syn-
thetic Membranes," S. Sourirajan, ed.,
National Research Council of Canada, Ot- 15.
tawa, Canada, Chapter 16,1977.
13. L. T. Rozelle, J. E. Cadotte, K. E. Cobian,
and C. V. Kopp, Jr., "Reverse Osmosis
and Synthetic Membranes," S. Sourira-
jan, ed., National Research Council of
Canada, Ottawa, Canada, Chapter 12,
1977.
R. L. Riley, R. L. Fox, C. R. Lyons, C. E.
Milstead, M. W. Seroy, and M. Togami,
"Spiral-wound Poly (ether/amide) Thin-
film Composite Membrane Systems," pre-
sented at the Membrane Separation Tech-
nology Seminar, Clemson University,
Clemson, SC, 1976.
T. Matsuura and S. Sourirajan, J. Appl.
Polymer Sci., Vol. 15, p. 2905, 1971; J.
Appl. Polymer Sci., Vol. 16, pp. 1663 and
2531, 1972; J. Appl. Polymer Sci., Vol. 16,
pp. 1043 and 3683, 1973.
114
-------
SEPARATION OF TOXIC SUBSTANCES BY HYPERFILTRATION*
J. Leo Gaddis, H. Garth Spencerf
Abstract
Experiments have been run to determine the
separation of toxic substances in textile water
streams when subjected to hyperfiltration. The
data indicate a high level of separation of toxic
substances by hyperfiltration. Three mem-
branes of commercial significance were tested:
cellulose acetate, poly (ether/amide) thin film
composite, and the dynamic membrane. The
membranes were exposed to textile streams
from mild caustic scouring and from dyeing of
cotton. Data regarding the separation of priori-
ty pollutants are presented according to chemi-
cal analysis of samples. Also presented are
data from biological testing with fathead min-
nows and daphnia. Acute rat toxicity, hamster
ovary clonal assay, and Ames tests are in-
cluded. Separation in LC50 from a level of 55 to
100 percent is produced in actual textile test
fluids.
INTRODUCTION
Hyperfiltration membranes may have the
capability of separating toxic substances in
textile process water. The literature contains
measurements of membrane rejections of ionic
and nonionic materials, but there are only a
few data on the rejection of toxic pollutants ex-
cept for metals. This program was initiated to
determine the separation characteristics of
membranes for process stream toxicity (deter-
mined from bioassay) and for toxic pollutants
(determined by analysis).
Toxic materials that enter a process stream
may have various origins. Field-applied agri-
cultural chemicals on cotton, spinning oils on
synthetics, warp size chemicals on cotton
woven goods, and preparation and dyeing solu-
tions are examples. The fluids selected for
*Work sponsored by EPA Grant 805777, Dr. Max Sam-
Afield, Project Officer.
'Departments of Mechanical Engineering and Chem-
istry, respectively, Clemson University, Clemson, SC.
study were from preparation and dyeing of cot-
ton goods. Three membranes were selected:
cellulose acetate (CA), poly ether/amide (PEA),
and a dynamic zirconium oxide/poly acrylic
acid membrane (DM). The basis for selection
was reasonable expectation of applicability to
industrial streams and commercial availability.
TEST DESCRIPTION
Figure 1 shows a schematic of the fluid col-
lection and test apparatus. Fluid from the sec-
ond washer of the range was pumped to the
test apparatus. Prior experience indicated the
fluid to be representative of the average fluid
from the range, not the highest nor lowest con-
centration of the individual streams emanating
from the range. The fluid was prefiltered using
a i-(i polypropylene cartridge filter for removal
of suspended materials, except for the dynamic
membrane runs where prefiltration is not re-
quired.
An initial feed volume of about 600 liters was
pressurized in a plunger pump to feed the
membranes. The membrane feed was concen-
trated slightly by the membrane as it passed
through the assembly. The concentrated feed
was returned to the feed reservoir, and the
permeate(s) collected in separate containers.
The original feed was thus gradually dimin-
ished in volume until 10 to 20 percent of its
original volume was remaining. This residual
comprised the concentrate fluid. A straightfor-
ward excercise indicates that this procedure
produces permeate and concentrate samples
equivalent to those expected in a prototype
unit.
Four tests were run: (1) PEA and CA simul-
taneously operating on dye fluid, (2) DM
operating on dye fluid, (3) PEA and CA simulta-
neously operating on scour fluid, and (4) DM
operating on scour fluid. Feed, permeate, and
concentrate samples were taken on each run.
In addition, samples were taken for chemical
analysis of the plant supply water and for
water passed through the range and circulated
in the apparatus.
The scour-fluid known chemicals are the
115
-------
CLOTH FLOW
DYE
I PAD
JET WASHER
COLLECTION
PAIL
CARTRIDGE FILTER
SKID
MOUNTED
PUMP STATION
APPROXIMATELY
40 METERS
FEED TANK
_L
FEED
OR
CONCENTRATE
SAMPLE
PERMEATE
BARREL
JPERMEATE
E f SIPHON
PERMEATE
FANK
PERMEATE DRAIN
Figure 1. Schematic of fluid acquisition and operations.
116
-------
warp size removed in the scour, a detergent/
wetting agent, hydrogen peroxide, and sodium
carbonate. Residuals of field chemicals, motes
on the cotton fiber, and natural oils and waxes
may also be present. The dye fluids contain a
thickener (gum), two detergent/wetting agents,
and a particular mixture of direct dyes.
RESULTS
In general, all membranes performed accord-
ing to expectations. Rejection (fraction re-
moved) of color was nearly perfect (>99 per-
cent); rejection of total solids averaged 94 per-
cent (88 to 98 percent); and rejection of conduc-
tivity averaged 95 percent (85 to 99 percent).
Analyses for specific toxic organic and metal
compounds were performed by Monsanto Re-
search Corporation under a separate grant
with EPA. The detailed results are contained
in the project final report (ref. 1) and are sum-
marized here. Only 19 organic compounds were
sensed; the highest concentration was below
300 mg/m3. Analysis in this low concentration
range is not expected to be more than modest-
ly accurate, effectively precluding a mean-
ingful calculation of rejection.
However, it appears that chloroform, toluene,
trichloroethylene, and methylene chloride
were not effectively rejected. Phenol and di-n-
Butyl phthalate showed a mixed tendency to
be rejected. The other compounds were shown
to be rejected and are: bis(2-ethylhexl)
phthalate, dimethyl phthalate, butylbenzyl
phthalate, diethyl phthalate, acenaphthene, an-
thracene, fluoranthene, pyrene, naphthalene,
phenanthrene, chlorobenzene, and ethylben
zene. The basis for rejection, or no rejection,
for all the above is relatively weak. Some of tho
compounds may have vaporized, others may
have been coated on the apparatus surfaces,
and still others may have been rendered nonex
tractable for analysis. Mass balances of solutes
were mostly poor.
Toxic metals were also present only in low
concentration. Adequate concentrations tor
analysis of only arsenic, copper, and zinc oc-
curred. All these metals were rejected well.
Arsenic (one point has rej>69) rejection
averaged 89 percent, copper averaged 97 per
cent, and zinc 100 percent. Other nontoxic
metals were analyzed and showed high rejec-
tion.
Rat acute toxicity and microbial mutagen-
TABLE 1. LETHAL CONCENTRATION AND IMPLIED TOXICANT CONCENTRATIONS
Fluid and type
Dye-feed
Dye-pea permeate
Dye-CA permeate
Dye-concentrate
Dye-feed
Dye-DM permeate
Dye-concentrate
Scour-feed
Scour-pea permeate
Scour-CA permeate
Scour-concentrate
Scour-feed
Scour-DM permeate
Scour-concentrate
Sample
no.
10
11
12
13
14
15
16
3
4
5
6
7
8
9
96-h minnows
I-CBO
% solution
9.7
82
>100
1.6
25
NAT
5.3
16
28
>100
1.5
13
NAT
2.0
48-h daphnia
Implied
concentration
no. units
10
1.2
<1
62
4
0
19
6
3.6
<1
67
7.7
0.0
50
LC50
% solution
33.5
60 to 100
60 to 100
4.1
49
80
17
26
53
42
5.1
25
>100
9.9
Implied
concentration
no. units
3.0
1 to 1.7
1 to 1.7
24
2.0
1.2
5.9
3.8
1.9
2.4
20
4
1
<10
117
-------
TABLE 2. REJECTION OF TOXICITY BY HYPERFILTRATION
Membrane
Scour fluid
Dynamic ZrO/PAA
Cellulose acetate
Poly ether/amide
Dye fluid
Dynamic ZrO/PAA
Cellulose actate
Poly ether/amide
Daphnia toxicant
>88
55
68
60
62 to 82
62 to 82
% Rejection
Fathead minnow toxicant
100
>92
60
100
>96
95
SCOUR DYE
o ° 48 HOUR
• • 24 HOUR
40 50 60 70 80 90 100
CONCENTRATION (rjP) TO DAPHNIDS
LC50
Figure 2. Correlation of concentration toxic to fathead minnows with concentration
toxic to daphnids.
118
-------
icity bioassays showed no effect. Hamster
ovary clone cytotoxicity showed effects on the
concentrate samples only. Results of acute
studies with daphniids and fathead minnows
showed consistent, interesting results. Table 1
shows the results obtained for LC5o(concentra-
tion of sample in water was lethal to half the
test subjects) at 96 hours for minnows and 48
hours for daphniids. Also shown in the table is
the corresponding value of 100/LC50, which is
the number proportional to the concentration
for the toxicant in the sample. Use of this im-
plied concentration allows calculation of mass
balance of toxicant and rejection of toxicant by
regular procedures. It is acknowledged that
such a calculation may oversimplify the prob-
lem.
Using the implied concentrations for the
four test runs allows mass balances of toxicant
to be calculated as 0.95, 1.28, 1.55, and 1.17
(minnow data), and 1.16, 1.23, 1.08, 0.65
(daphniid data). These values are highly con-
sistent compared to expectations and are
almost as good as those obtained for total
solids. Table 2 shows the rejection data cor-
responding to the bioassay. The average rejec-
tion of daphniid toxicant is 69 percent, and min-
now toxicants is 90 percent. The relative mem-
brane effectiveness for removal of toxicant is
DM > CA > PEA, while for salt removal the
relative order is PEA > CA > DM.
There is a high correlation of the data ob-
tained for daphniid and minnow toxicity.
Figure 2 shows the results of both 24- and 48-h
responses as reciprocal LC50 for each animal.
The correlation coefficient for the data is 0.94,
suggesting an excellent empirical relationship.
CONCLUDING REMARKS
Hyperfiltration has been shown to be an ef-
fective separator for toxic materials in textile
process streams. The bioassay data for
daphniids and minnows provide consistent
evidence of high separation of about 70 percent
for daphniids and 90 percent for minnows.
Because of analytical difficulties associated
with low concentrations, most of the organic
and metal toxic pollutant rejections were not
determined. There is good evidence, in agree-
ment with prior experience, for high rejection
of metals. Some organics were rejected badly
or not at all, but most were separated. Further
studies under controlled conditions should be
undertaken to determine the separation of
these materials.
REFERENCES
1. J. L. Gaddis and H. G. Spencer, Evaluation
of Hyperfiltration for Separation of Toxic
Substances in Textile Water Process
Water, final report, Grant R805777, to be
printed by EPA.
119
-------
Session III: AIR POLLUTION CONTROL TECHNOLOGY
Donald F. Walters, Session Chairman
121
-------
INCINERATION AND HEAT RECOVERY IN THE TEXTILE INDUSTRY
William H. Hebrank*
Abstract
The paper discusses ways to solve air pollu-
tion and excess fuel uses caused by textile
dryer exhausts. Heat recovery with incinera-
tion minimizes heat exchanger cleaning main-
tenance and provides net fuel savings, even
without exhaust reduction. Five purposes for
exhaust size are presented, along with param-
eters to evaluate the size of a dryer's exhaust.
The three mechanisms of incineration described
are flame exposure, thermal, and catalytic. In
increasing order of complexity, five different
types of incinerators in industrial use are
described. These are the dryer counterflow in-
cinerator, the afterburner incinerator, the in-
cinerator with a preheater, the recycling loop
incinerator, and the high-temperature thermal
incinerator.
The subject of this paper is incineration with
heat recovery applied to exhaust gases from
textile finishing dryers. Textile finishing dryer
exhausts are a problem for two reasons. First
the exhaust is a relatively large flow, usually
about 9,000 to 18,000 kg of air per hour, and it
is hot enough to cause a large heat loss. The
reduction of large heat losses from dryer ex-
hausts is the most important fuel conservation
task in the textile industry. The second prob-
lem with textile dryer exhaust relates to air
pollution. The blue haze and odor from many
finishing dryer exhausts takes first place as
the most important air pollution problem in the
textile industry.
The amount of air pollution from dryer ex-
haust stacks depends on the finishing process.
Various kinds and amounts of oils and finishing
resins are vaporized from the cloth in dryers at
high drying and cloth setting temperatures.
Some exhausts also carry quite a bit of lint.
There are process means to reduce some of the
air pollution. Nowadays, most of the oils are
"Hebrank Incorporated, Greenville, SC.
removed by a prescour process and, some-
times, the use of a vacuum slice before the
dryer decreases the amount of resin on cloth
entering the dryer. Alternate selection of
finishing chemicals may also help, but cloth
finish quality and economics limit the ability of
processes to correct the problem.
Most dryers use natural gas for fuel. If the
dryer exhaust is clean exhaust, heat recovery
should be used to save fuel. A good exhaust
heat recovery application can pay for itself in 1
to 2 years. With gas prices going up, it is
economically foolish to put off heat recovery
plans. However, in most cases, unless heat
recovery equipment can be easily cleaned, con-
densing smoke from the exhaust stream soon
plugs heat transfer surfaces and diminishes
economic gains.
If the exhaust can be cleaned of dirt and
smoke from lint and condensing oils before
passing heat-recovery heat exchangers, the
prospects for success are greatly improved.
Unfortunately, air-cleaning devices such as
scrubbers or electrostatic precipitators re-
quire cooling the exhaust before cleaning so we
are back to problems with heat exchanger
plugging. Incinerators have an advantage. The
sequence is reversed; exhaust is cleaned before
going to heat recovery. That point, plus the
fact that burning is the most convenient way to
get rid of air polluting smoke, makes incinera-
tion very attractive.
You may have heard that incineration re-
quires additional fuel use. That just is not so; a
properly designed incineration system with
heat recovery can reduce the total fuel used for
drying. Depending on circumstances, the sav-
ings can run as high as 60 to 65 percent. It is
possible to get 15 to 25 percent savings, even
without any exhaust reduction.
It is important to review the size of the ex-
isting exhaust flow rate before planning to in-
stall exhaust air pollution and heat recovery
equipment. Size and cost of equipment depends
on the amount of exhaust flow. Many exhausts
can be reduced, sometimes by as much as 50
percent. Exhaust reduction also concentrates
123
-------
smoke and vapors in the exhaust steam, thus
making them easier to remove. The amount of
exhaust needed depends on the cloth drying-
finishing-setting process, the rate of cloth flow,
the dryer's tightness against smoke leakage,
and plant operating-room ventilation condi-
tions.
In considering the amount of a textile dryer
exhaust, remember that there are at least five
purposes for that exhaust, the first of which is
to decrease evaporated water vapor concentra-
tion in the dryer housing. Exhaust should be
thought of in terms of decreasing concentra-
tion; not in terms of removing water vapor. It
can not remove all of the evaporated water; it
serves to bring water vapor concentration
down to levels consistent with good operation.
Many dryers use more than 30 kg of exhaust to
carry off 1 kg of evaporated water. This can be
reduced to the 16- to 18-kg range. I have heard
it said that with extensive effort to seal the
dryer and control plant conditions on some
dryers, one can reach 10 to 12 kg without no-
ticeably affecting drying rates. The paper in-
dustry has known this with their high speed
dryers for years.
The second purpose for exhaust is to
decrease the concentration of process smoke
and vapors in the dryer. Again all smoke can-
not be removed. Fresh makeup air to replace
air leaving as exhaust dilutes the concentra-
tion. As concentration goes down, additional
exhaust has less and less benefit.
The third purpose of exhaust is to reduce
dryer pressure below room air pressure to
maintain a slight inflow to all dryer openings
from the room. Water vapor and smoke gases
generated from the cloth add their partial pres-
sures to the total inside the dryer and must be
accounted for in exhaust calculations. Local
high pressure areas due to circulation flows in-
side the dryer make it more difficult to main-
tain a slight inflow at all dryer openings. The
inflow of air aspirates and cleans air near the
openings, not unlike a ventilation hood over a
shop welding or grinding operation. Also, the
partial pressure of fumes inside the dryer is
higher than the partial pressure of fumes in
the room. This difference of partial pressure
causes fumes to diffuse or flow towards the
room's lower concentration level. With a very
slow inflow air stream, the fume diffusion rate
may be fast enough to swim against the inflow-
ing air and, thus, cause smoke or haze to occur
in the room.
The fourth purpose for exhaust is to remove
products of combustion from the dryer burn
ers. The fifth purpose is to properly ventilate
the dryer during startup prior to the ignition.
Textile dryer exhaust dirt and smoke are
organic compounds. They can be destroyed by
burning, that is reduced to mostly carbon
dioxide and water vapor, with three different
incineration mechanisms. The first mechanism
is by direct exposure to flames. For efficient
operation, design of conditions around the com-
bustor such as control of flame geometry and
airflows are most important. The flame front
itself is a highly ionized and reactive region
.and should be used to best advantage..Tem-
perature level is not that important but should
be at least 320° to 340° C.
The second incineration mechanism is the
phenomena of thermal ignition. Flames are not
necessary for a substance to thermally ignite.
If a combustible substance in air is hot enough
it will go through the chemical reactions to ox-
idize to mostly carbon dioxide and water
vapor. Thermal ignition temperatures of var-
ious organic compounds in air are listed in
chemical handbooks. Most of the common sol-
vents have thermal ignition temperatures in
the 320° to 590° C range. If essentially all of the
substances in exhaust smoke thermally ignite
at 510° C or lower, it makes little sense to go to
650° or 820° C in the incinerator. A rule of
thumb for thermal incinerators is to heat ex-
haust to 650° to 820° C and maintain the tem-
perature for one-half second or so. Those high
temperatures require expensive high-tempera-
ture materials, and the incinerator uses more
fuel at high temperatures. In taking the overall
environment and available resources into con-
sideration, the aim should be to do an adequate
job of cleaning exhaust. It can be wrong and
wasteful to aim for a 100 percent cleaning
operation.
The third mechanism for incineration is to
accomplish chemical reactions of the dryer
smoke substances at the surface of a catalytic
material. Chemical reaction rates are related
to temperature and surface conditions. At the
catalytic interface, reaction rates go faster,
thus allowing a reduction of temperature with
attendant savings on materials and fuel usage.
Catalytic reactors are costly and, with certain
124
-------
340°
CLOTH
Figure 1. First incinerator category.
gases or elements, are subject to surface
poisoning, which can permanently lessen their
effectiveness.
In order of simplicity, there are about five
different kinds of incineration systems for
reducing the air pollution from textile dryers.
The first incinerator category (Figure 1)
removes air from the back end of a dryer,
passes it through a combustor to heat it to 320°
or 370° C, then reenters the incinerated stream
into the drier front end. Since oil and resins
evaporate at higher temperatures than water
does, most of the smoke in textile dryers is
created in the last zones of the dryer. This
method destroys the smoke at the point of ori-
gin and has been reported to make a noticeable
reduction of exhaust smoke. Actually, the ex-
haust from the dryer contains more smoke
than the incinerated flow from the incinerator
because the exhaust has whatever concen-
tration exists inside the dryer. Any fuel used
should be used on the exhaust stream leaving
the dryer, thus obtaining maximum cleaning
effect on the stream going to the atmosphere.
The second incinerator category (Figure 2A),
often called an afterburner, takes air to be ex-
hausted from the dryer, passes it through an
incinerating combustor which heats it to 320°
to 370° C, then either exhausts the incinerated
stream directly to the atmosphere or passes it
through a heat exchanger to recover some of
the heat energy before exhausting. A variation
of this, using the combustor of an indirectly
fired dryer as the exhaust incinerator (Figure
2B), ran successfully in Mexico in 1974. The
afterburner concept uses the presence of
flames rather than high temperature to ac-
complish incineration.
The third incinerator category (Figure 3) and
probably the most common in use in other in-
dustries has a heat exchanger before the in-
cinerator to preheat exhaust air to be in-
cinerated. High-temperature flow leaving the
incinerator passes back through the hot side of
the preheater exchanger to provide the pre-
heat energy, and in so doing it gives up an
equivalent amount of heat and is cooled. In
essence the preheat heat exchanger makes it
possible to raise the exhaust stream to higher
temperatures, using less fuel to do so. Earlier
it was mentioned that putting heat exchangers
into the exhaust stream before the stream had
been cleaned would cause dirt and smoke to
soon plug the heat transfer surfaces. In this ap-
plication, that is not so, because the heat ex-
changer is hotter than the exhaust gases; it is
being used to heat, not cool, the exhaust
stream. Oils, resins, and smoke will not con-
dense on it. If the transfer surfaces are design-
ed to pass lint on to the incinerator it will not
require cleaning maintenance. The incinerated
exhaust stream leaving the preheater-inciner-
ator combination is still very hot and usually is
passed through a heat recovery exchanger
125
-------
370° C
-VWW O
CLOTH
Figure 2A
Figure 2B
Figure 2. Second incinerator category.
before exhausting to the atmosphere. With
this system the first heat exchanger, the
preheater, is most often a shell and tube ar-
rangement. Incinerator chamber temperatures
will probably exceed 450° C with this system; it
does a good job of cleaning the exhaust stream,
and, if enough heat exchanger surface is used
for heat recovery to dryer makeup air, it can
save 10 to 15 percent of the finishing process
fuel bill.
The fourth incinerator category (Figures 4A
and 4B) uses a recycling loop to pass dryer ex-
haust through a flame field a number of times
before exhausting to recovery heat exchangers
and the atmosphere. The recycle principle was
first tried in 1974 on a modified indirect-fired
dryer (Figure 4A) using the burners in the
dryer as exhaust incinerators. The unit has
been in continuous production since then in
compliance with North Carolina code, and fuel
savings exceed 60 percent. The recycle loop
type incinerator (Figure 4B) incinerates at
temperatures of 320° to 340° C with retention
times of 2 seconds or longer. Retention is not
as important as flame front exposure with a
thin flat flame spread as wide as possible
across the flow. The recycle duct carries hot
gases back to premix with entering dryer ex-
126
-------
Figure 3. Third incinerator category.
Figure 4A
Figure 4B
>/wwv—Q
-»»|i
-------
560° C ||ll|lllllllUHHll|lh
vAAA/VV
www
CLOTH
Figure 5. Fifth incinerator category.
haust. The mixing of hot and cooler gas acts as
a preheater to preheat dryer exhaust to tem-
peratures around 260° C before entering the
combustor. The system uses a recovery heat
exchanger to heat makeup air for the dryer to
about 260° C. This hot makeup air carries
enough heat back to the dryer to allow dryer
operation with all of its burners turned off.
Depending on the installation, fuel savings can
run from 20 to 50 percent.
The fifth incinerator category (Figure 5)
utilizes a high efficiency heat exchanger, 80 to
90 percent, to preheat exhaust to about 560° C
before entering the incinerating chamber. The
incinerating chamber has sufficient volume to
hold the flow for about one-half second at
650° C. The method of heating from the enter-
ing temperature to the incinerator tem-
perature is not important for incineration; that
is, exposure to flame is not necessary. For
practical design, the preheating heat ex-
changer should be a regenerating type. The
regenerating exchanger reverses hot and cold
flow over heat-storage elements. The system is
very efficient, but 650° C temperature levels
require construction with expensive high-tern
perature materials: ceramics for heat storage
and refractory lining in the incinerator. Textile
dryer exhausts contain many hydrocarbons
which thermally ignite at temperatures below
540° C, probably all of them or what is left is of
such small quantity to be of no consequence. It
is very likely that 540° C peak levels will do an
adequate job of incineration. The costs of con-
struction materials for 650° C are very expen-
sive compared to the cost of materials for
540° C. Again we are back to the point of doing
that which is practical for the overall ecology
with available resources.
128
-------
FIXED BED CARBON ADSORPTION FOR CONTROL
OF ORGANIC EMISSIONS
W. Macon Sheppard*
Abstract
Activated carbon processes have long been
used for solvent recovery; however, these proc-
esses and variations of them are receiving in-
creased attention for organics considered to be
air pollution problems requiring emission con-
trol. In addition to the air pollution aspects,
many recovery schemes previously labeled as
uneconomical are now undergoing reconsidera-
tion due to increasing chemical costs. Acti-
vated carbon is one of several operations that
can be employed to reduce organic emissions
and in many cases recover valuable material or
heat value.
When selecting the appropriate type of pollu-
tion control process, the physical properties of
the contaminant and the level and frequency of
emission can be used to determine whether the
pollutant can be discarded, incinerated for heat
value, or recovered as product. By concentrat-
ing emission streams, activated carbon can be
used to produce fuels suitable for incineration,
thereby increasing pollution control with re-
duced energy drain.
In designing fixed bed adsorption systems,
pilot scale studies are the best method of ob-
taining the required design information. The
most important factors to be evaluated are the
breakthrough time, the length of the mass
transfer zone, and the required regeneration cy-
cle necessary to sustain an adequate carbon
working capacity. Overall design considera-
tions are governed by the required control effi-
ciency and the available plant facilities.
INTRODUCTION
Some States are imposing, in State Im-
plementations Plans, stricter limitations on
organic emissions from stationary sources in
order to meet Federal requirements for con-
trolling ozone. Also, for the first time many
•Environmental consultant.
States are beginning to require emission inven-
tories on organics.
The most important regulatory action con-
cerning organics is the increased attention
being placed on carcinogenic compounds. The
ruling of vinyl chloride as a hazardous sub-
stance due to increased incidents of angio-
sarcoma among workers was the first of what
will surely be many organic compounds to
undergo reduced emission levels and strict en-
forcement action. Since the vinyl chloride rul-
ing, benzene has now been singled out as a
health hazard and others are sure to follow. As
in the case of vinyl chloride, strict emission
limits set for health reasons will not be
lowered due to complaints of poor economics or
shortages of energy and technology. These
emission limits, imposed by either State, EPA,
or OSHA regulations, will be required of those
industries where the specified organic is
emitted from stacks, vents, or fugitive sources.
In addition to the air pollution aspects, in-
creased attention is also being given to many
recovery schemes previously labeled uneco-.
nomical, but now undergoing reconsideration
due to increasing chemical costs. Activated
carbon is one of several operations that can be
employed to reduce organic emissions, and in
many cases recover valuable material or heat
value.
THEORY OF ADSORPTION
Adsorption is a process involving the con-
centration of materials at a surface or inter-
face, whereas in absorption, the material goes
into solution. Adsorption is normally thought
of as a liquid or gas adhering to the surface of a
solid; such as activated carbon. Such a material
can either be physically or chemically ad-
sorbed. Physical adsorption is the result of in-
termolecular forces of attraction between
molecules of the solid and the substance being
adsorbed. The adsorbed substance does not
penetrate the lattice structure, but remains at
the surface. The phenomena is reversible with
129
-------
an increase in tertiperature or a decrease in
pressure removing the adsorbate in an un-
changed form. In almost all cases, organics are
physically adsorbed on activated carbon.
Chemical adsorption (chemisorption) is a
chemical reaction between the solid and adsorb-
ate with the formation of bonds much greater
than the forces of physical adsorption. This
phenomena is frequently irreversible; e.g.,
oxygen chemisorbed on carbon is desorbed as
CO or C02 (ref. 1).
The amount of a given material that will ad-
sorb on activated carbon is most commonly
determined by static equilibrium tests. The
carbon is contacted with a given concentration
at a constant temperature, and after equilib-
rium is attained, the mass increase of the car-
bon is measured. From these tests adsorption
isotherms are formed. A typical isotherm is
shown in Figure 1 (ref. 2). The adsorption
capacity increases with a decrease in temper-
ature and an increase in concentration and as a
general rule, a vapor is more readily adsorbed
the higher the molecular weight and the lower
the critical temperature.
FIXED BED OPERATION
Carbon adsorption systems normally employ
fixed beds of activated carbon. As shown in
Figure 1 (ref. 3), as the pollutant is adsorbed
from the gas stream onto the bed, an adsorp-
tion zone is formed which moves through the
bed at a velocity much slower than the gas
velocity. The breakpoint (or breakthrough) is
that time when the adsorption zone (or mass
transfer zone) reaches the end of the bed, and
the pollutant concentration of the bed effluent
begins to increase. The effluent concentration
increases with onstream time, and at complete
bed saturation will equal the influent concen-
tration. Static equilibrium measurements of
adsorption isotherms indicate the amount of
organic adsorbed at saturation, but these data
are not good design parameters since the on-
stream time of a fixed bed system is normally
terminated short of complete saturation, usual-
ly soon after the breakthrough occurs.
.0001
.001
.01 .1
Partial Pressure—PS!A
1.0
Figure 1. Toluene isotherms.
130
-------
DESIGN CONSIDERATIONS
Pilot-scale studies are the best method of ob-
taining design information. Some factors that
must be taken into account when designing an
adsorption system and that can be evaluated in
a pilot study are (ref. 4):
• The breakpoint and shape of the break-
through curve or length of the mass trans-
Progress of B stable mass-transfer
through an adsorbent bed
front
. T T
JL
T T T
L
o «, .,
Onstream time, 6
Figure 2. Stable mass transfer front.
Position of the stoichiometric front relative to the
stable mzss-lransfer front during dynamic adsorption
1
§8
= e v
•U^. *,
aeea
C 3
S 8
i£
l«y
Kff '«
tj3 1
1
X
^ ')
1 -
3 »* JL e.
•Onstream time, 9.
Figure 3. Stable front relative to
stoichiometric front.
ferzone. Breakpoint times decrease with:
— a decrease in bed depth,
— an increase in carbon particle size,
— an increase in flow rate, and
— an increase in initial concentration.
The length of the mass transfer zone is re-
quired to determine adequate bed depths
(Figures 2 and 3). Normal bed depths are
at least 2 to 3 times the length of the mass
transfer zone. If the mass transfer zone is
longer than the carbon bed, breakthrough
is immediate.
• The working capacity of the carbon, at
various temperatures, for the organic be-
ing removed from the stream. The work-
ing capacity is the capacity restored by
regeneration and cannot be determined
from saturation capacities as shown in
Figure 4 (ref. 4).
• The pressure and flow rate of the stream
being treated. The pressure drop is re-
quired for determining bed depth and the
particle size of the carbon to be used.
• The desired and/or economical onstream
time between regenerations. Factors that
may govern onstream time are:
— minimum effluent concentration,
— bed volume, and
— regeneration time required to restore
a desired working capacity.
• The need for pretreatment of the pollutant
stream prior to adsorption. Pretreatment
is sometimes necessary to remove heat,
moisture, particulates, or aerosols (Figure
5).
• The type of regeneration to be used. Alter-
natives for regeneration are steam, air,
vacuum, and thermal, as can be seen in
Figure 5 (ref. 5).
POLLUTION CONTROL PROCESSES
Four general types of adsorption schemes
can be defined as aids in selecting the ap-
propriate type of control process (Figure 6).
These are odor removal, pollution control, sol-
vent recovery, and tank vent emission control
(ref. 6).
Odor Removal (Figure 7}
Odor problems can occur when concentra-
tions are less than 1 ppm, even though emis-
131
-------
>z 15
H ^.
— o
OCQ
^ rf
3° 10
o
LU o
^^^\
^ O
Q3 W)
»^^
71
X
V
>
/^
>•
/
X
X
X
/
/j
^s
Ads
»\ —
Des
Pur
7^
v/
X
•^/
y
/
X
- ^«
Puri
1
'"j
O
X
X
X
X
^•^PIB
7H
\\\\\\\N
1
Adsorption Temperature 75°F
DesorptKm Temperature 75°F
Purge Rate, bed volumes/min 10
Purge Time, minutes 20
Saturation Capacity
Working Capacity
1234
CARBON NO.
Figure 4. Working capacity is independent of saturation capacity.
Adsorption Cycle
Pretreatment
(If Necessary)
A/C
Adsorbers
Polluted
Air
Source
- •
»J Absorption
»i Condensation
A. ^^%« « .^ • .^k
—
— ^
—
I
1
Clean
Air to
Atmos
s
Desorption Cycle
Heated air-
Ambient air-
Electrical heat
Vacuum »
Inert gas-
Steam —
Combustion gas •• »
Chemical ^
A/C
Adsorbers
Condensation
Absorption
Incineration
Figure 5. Adsorption/desorption alternatives.
132
Reuse
Disposal
Fuel
Heat
Recovery
-------
POLLUTION CONTROL PROCESSES
Concentration
Bed D«plh (Typical)
Regeneration Method
Disposition of
Collected
Materia!
Adsorption Cvcie
Odor
Removal
1 ppm
1/Jm.
Reactivate
or discard
None
1 yetr
Poituiton
Control
1 ppm-0.1%
9m.
Regenerate
m puce
Incinera-
tion or
disposal
1-8 hours
Solvent
Recovery
01V3*
24 in.
Steam Re-
generation
Reuw
30-60 mtn .
Tank Vent
Emission Control
Various
, Various
Discard Carbon
M2mon|ns
INLET
Figure 6. Pollution control processes.
INLET
EXHAUST
FAN ("ARTICULATE ODOR
REMOVAL REMOVAL
AIR FLOW
PRESSURE DROP
CARBON LIFE
NEW FILTER COST
REACTIVATION COST
2000CFM/FILTER
0.25" H-0
1-2 YEARS
S220-S250
S 80-S100
Figure 7. Odor removal system.
sion quantities are minimal. In this system the
air is filtered to remove particulates before be-
ing treated by the carbon. A thin bed of carbon
or multiple sets of thin beds can be designed to
handle large volumes of air and keep the con-
taminant below its threshold level for over a
year. The spent carbon is then removed from
the bed and either reactivated or discarded.
Pollution Control (Figure 8)
Emission problems where pollutant concen-
trations range from 1 to 1,000 ppm can be
caused by odor or emission quantities. In this
system, the stream may be filtered to remove
particulates and/or cooled to increase adsorp-
tion. Normally, multiple beds are employed in
cycles to allow one to be regenerated while
another is undergoing adsorption. Economics
normally dictate that the carbon be regener-
ated in place, but that the desorbed material
not be recovered for reuse. Incineration is the
normal disposal method.
Solvent Recovery (Figure 9)
A conventional solvent recovery system is
BLOWER AIR PARTICULATE
CONDITIONING REMOVAL
DISPOSAL
CONDENSER
AIR FLOW
ADSORPTION
PRESSURE DROP
f^ STEAM
REGENERATION
PURIFICATION
1000-100.000 CFM/BED
1 - 8 HOURS
10"-20" H20
Figure 8. Pollution control system.
BLOWER
AIR
CONDITIONING
PARTICULATE
REMOVAL
CARBON
CARBON
CONDENSER REGENERA-
DECANTER TION
I EXHAUST
PURIFl- "
CATION
AIRFLOW - 1.000-40.000 CFM/BED
ADSORPTION - 30-60 MIN.
PRESSURE DROP - 20"-30" HjO
Figure 9. Solvent recovery system.
usually employed when the pollutant concen-
tration ranges from 1,000 ppm to the lower ex-
plosive limit. This system is the same as the
pollution control system previously described,
except that the large pollutant loadings re-
quire deeper beds even for short adsorption
cycles, and organics are recovered for reuse in
the regeneration cycle.
Tank Vent Emission Control (Figure 10)
In many cases the conventional solvent
recovery process can be used to control emis-
sions from tank vents; however, the process as
described in Figure 10 is applicable where
regeneration facilities such as steam or heat
are not available. The process is simply a
replaceable carbon bed which adsorbs vapor on
tank filling and can be partially regenerated
133
-------
with inert gas as the tank is emptied.
The appropriate adsorption process can be
selected based on the pollutant's physical prop-
erities, concentration, and emission frequency
as shown in Figure 11. As a general rule,
organics with a molecular weight below 30 are
not adsorbed in quantities which make adsorp-
tion as practical as other control methods.
Another factor is that with increasing boiling
point, the pollutant is less readily desorbed
from the carbon. Applicable systems are,
therefore, those in which the carbon is not
regenerated in place. Continuous operations
are more conducive to cyclical adsorption
systems with disposal or recovery based on the
pollutant concentration in the stream.
INERT GAS
SOURCE
VENT
AIRFLOW -UptoZOOCFV
PRESSURE DROP - Negligible
CARBON LIFE -1-12 MONTHS
CARBON COST - S1-S2/L8. SOLVENT
(THROW AWAY BASIS!
Figure 10. Tank vent emission control.
NOVEL ADSORPTION SYSTEMS
By concentrating emission streams, ac-
tivated carbon can be used to produce fuels
suitable for incineration, thereby increasing
pollution control with reduced energy drain. In
a typical solvent recovery unit, two beds are
used to allow one to be regenerated with steam
while the other is adsorbing solvent. Figure 12
illustrates a modification to this process in that
the carbon is contained in annular beds at a
depth of 4 inches. As steam is used to regene-
rate the beds countercurrently, concentrated
airstreams containing steam can be sent direct-
ly to an incinerator (ref. 2).
Another adsorption system is used to con-
centrate a benzene emission stream from 85
ppm to 3,500 ppm, thereby reducing fuel re-
quirements for subsequent incineration by a
factor of 86 (ref. 2). As shown in Figure 13, the
85 ppm benzene stream is cooled to 150° F,
then adsorbed on the carbon bed in an 8.5-h cy-
cle. For regeneration a slip stream of the
original 300° F emission is introduced to the
carbon bed in a 4-h desorption cycle, which pro-
duces an effluent average concentration of
3,500 ppm. Whereas the 85 ppm stream would
require 1,530,000 Btu's to incinerate at
1,400°F, the 3,500 ppm stream requires only
17,800 Btu's. A similar adsorption/incineration
system patented as the Zorbcin Process (Fig-
ure 14) uses incinerator exhaust to desorb the
carbon beds (ref. 7).
In addition to the commonly used solvent
recovery process, recovery is also feasible with
vacuum desorption. In such a system, the pres-
sure is reduced below the partial pressure of
the adsorbed organic material. The organic is
then removed from the carbon in concentra-
tions as high as 90 percent, which makes con-
densation and recovery attractive. A vacuum
desorption system has been shown to be a
viable method for controlling emissions from
gasoline vapors during refueling operations.
As shown in Figure 15, vapor concentrations of
10 percent to 40 percent are adsorbed and
vacuum regenerated to 95 percent concentra-
tion levels (above the upper explosive limit).
Although successful under these conditions,
there is little data for vacuum desorption of
dilute streams and low carbon loadings.
CRITERIA FOR SELECTION OF
GAS CLEANING EQUIPMENT
Carbon adsorption is one of several opera-
tions that can be used to control organic emis-
sions. Other techniques are adsorption by var-
ious liquid scrubbers and direct flame, thermal,
and catalytic incineration. Criteria for selec-
tion of the proper processes are: desired collec-
tion efficiency, vapor stream characteristics,
plant facilities and available utilities, and cost
of control (Figure 16).
134
-------
TABLE II PROCESS SELECTION
BOILING
POINT 400" F
MOLECULAR
WEIGHT > 30
! MOLECULAR
WEIGHT <30
CONCENTRATION
>1 PPM
REMOTE
LOCATION
f
CONCENTRATION
<1 PPM
CONTINUOUS
EMISSION
SINGLE
OCCURRENCE
INTERMITTENT
EMISSION
INCINERATION
ADSORPTION OR
CHEMICAL REACTION
PERSONNEL
NEARBY
TANK VENT EMISSION CONTROL
POOR REMOVAL
ODOR REMOVAL
CONCENTRATION
3%
— SOLVENT RECOVERY
TANK VENT
EMISSION CONTROL
SOLVENT RECOVERY
CONCENTRATION
0.1%-3.0%
CONCENTRATION
1 PPM-0.1%
CONCENTRATION
1 PPM
SOLVENT RECOVERY
POLLUTION CONTROL
ODOR REMOVAL
TANK VENT EMISSION CONTROL
TANK VENT EMISSION CONTROL
HIGH
CONCENTRATION]
ODOR OR LOW
CONCENTRATION
ODOR REMOVAL
Figure 11. Selection process for adsorption control.
135
-------
Steam
Inlet
Air
p<
1
1
r —
C/inrffinsfir
«
*
Incinerator
Steam
Figure 12. Steam regeneration/incineration.
85 ppm Benzerv
300«F
4
20 SCFM
B
^^ / ,A
" Vv j
v^
200 SCFt
^
* M
1 1 N
Adsorber
t
ri
H ?
M 1 I
Adsorber
t
3500 ppm Benzene
30O°F L , _ _,_
Figure 13. Adsorption/incineration system.
136
-------
CONTAMINATED
AIR STREAM
AIR
*P
CONTAMINATED
AIR FAN
EXHAUST TO ATMOSPHERE
AOSOR8ER NO. 2
SLOWDOWN STREAM
I
REGEN.
AIR COOLER,
COOLING
WATER
REGENERATING
FAN
HEAT !NTER-
CHANGER
•
MAKE-UP
TO
EXHAUST
INCINERATOR INCINERATOR
FAN I
NAT. GAS
Figure 1.
ZORBCIN
Process
for
purifying
contaminated
air streams.
Figure 14. Zorbcin process.
10-40%
Vol. HC
? M
I !M
Adsorber
1 +
Adsorber
vacuum
Regeneration
System
^^ •^pv
r>
h±r
95V
HC
-*
ol°/<
Recovery
Storage
>
Figure 15. Gasoline vapor control system.
137
-------
EMISSIONS AND EMISSIONS
STANDARDS
DETERMINES COLLECTION EFFICIENCY
CONTROL EQUIPMENT ALTERNATIVES
* V 1
PACKED
TOWER
VENTURI
SCRUBBER
r t
ACTIVATED
CARBON
ADSORPTION
DIRECT
FLAME
CATALYTIC
VAPOR STREAM
CHARACTERISTICS
VOLUME
TEMPERATURE
MOISTURE CONTENT
CORROSIVENESS
ODOR
PROCESS
VAPOR STREAM
CHARACTERISTICS
EXPLOSIVENESS
VISCOSITY
IGNITION POINT
DENSITY
PHYSICAL PROPERTIES
WASTE TREATMENT
SPACE RESTRICTION
PRODUCT RECOVERY
PLANT
FACILITY
WATER AVAILABILITY
FORM OF HEAT RECOVERY
(GAS OR LIQUID)
ENGINEERING STUDIES
HARDWARE
AUXILIARY EQUIPMENT
LAND
STRUCTURES
INSTALLATION
START-UP
I
COST OF
CONTROL
POWER
WASTE DISPOSAL
WATER
MATERIALS
CAS CONDITIONING
TAXES
INSURANCE
RETURN ON INVESTMENT
«t •
SELECTED
GAS CLEANING SYSTEM
t
DESIRED EMISSION RATE
Figure 16. Criteria for selection of gas cleaning equipment.
138
-------
REFERENCES
1. W. M. Sheppard, "A Study of the Steam
Regeneration of a Fixed Bed of Activated
Carbon Used to Adsorb Hydrogen
Sulfide," Master's Thesis, Clemson Univer-
sity, 1970.
2. W. D. Lovett and F- T. Cunniff, "Activated
Carbon Methods for Controlling Sta-
tionary Source Air Pollutants," presented
at the Calgon Adsorption Conference,
Pittsburgh, PA, 1975.
3. G. M. Lukchis, "Adsorption Systems,"
Union Carbide Corp. Pamphlet, reprinted
from Chemical Engineering, 1973.
4. W. D. Faulkner, W. G. Shuliger, and J. E.
Urbanic, "Odor Control Methods Using
Granular Activated Carbon," presented at
74th National AIChE Meeting, New Or-
leans, 1973.
5. B. Grandjacques, "Air Pollution Control
and Energy Savings With Carbon Adsorp-
tion Systems," Calgon Corporation Report
No. APC 12-A, 1975.
6. J. C. Enneking, "Control of Vapor Emis-
sions by Adsorption," ASHRAE Bulletin
CH 73-2,1973.
7. M. M. Mattia, "Processs for Solvent Pollu-
tion Control," Chemical Engineering Prog-
ress, Vol. 66, No. 12,1970.
139
-------
FLUIDIZED BED ACTIVATED CARBON ADSORPTION
Ram Chandrasekhar*, Carmen M. Yonf
Abstract
Activated carbon adsorption and recovery of
hydrocarbons has long been practiced in in-
dustry using fixed beds. With the development
of special beaded carbon it is now possible to
recover these hydrocarbons continuously in a
fluidized bed. Fluidized beds are compact,
easier to operate and maintain, and offer im-
proved life cycle costs. PuraSiv® HR is a pro-
prietary fluidized bed activated carbon adsorp-
tion/desorption system licensed by Union Car-
bide Corporation from Kureha Chemical Co.,
Ltd.* Details of the operation of fluidized bed
activated carbon adsorbers, their application
to textile plant exhausts, and relative merits
over fixed beds are discussed in this paper.
ACTIVATED CARBON ADSORPTION
The control of environmental pollution has
become one of the industrialized world's major
social goals. The textile industry is but one of
many faced with a need for new technological
innovations in dealing with the control of unde-
sirable emissions. In the area of air pollution
there are many opportunities for reducing the
quantities of organic solvents, trace toxic sub-
stances, and malodorous compounds released
to the atmosphere.
The processes that have emerged as the
most successful in treating gaseous emissions
are adsorption, absorption, incineration, and
catalytic oxidation. Among these, only the first
two offer the opportunity to recover sub-
stances for subsequent reuse. Although highly
effective, the latter two methods require the
consumption of much more valuable fuel, espe-
cially in applications where the pollutant is at
low ppm levels. The use of adsorption, prin-
cipally with activated carbons, has gained wide
*Foster-Miller Associates, Inc., Waltham, MA.
^Union Carbide Corporation, Tarrytown, NY.
•••Carbon Products Department (formerly Taiyo Kaken
Co., Ltd.).
acceptance due to its versatility in removing a
wide range of pollutants, its effectiveness of
low concentrations, and its ability to yield a
concentrated recovered substance for reuse.
Currently, nearly all activated carbon sys-
tems are of the fixed bed type. Granular or
pelleted activated carbon of the desired size is
packed in a vessel through which the solvent-
(or pollutant-) laden gas is passed to effect
removal. At some point before the activated
carbon's adsorptive capacity is exhausted, the
contaminated airstream must be switched to
another vessel containing a bed of regenerated
adsorbent. Steam is then passed through the
spent bed of carbon to heat the adsorbent and
strip the contaminant. In the case of solvent
recovery the steam-vapor mixture goes to heat
exchanger where it is condensed. If the solvent
and water are immiscible, the two phases are
separated in a decanter and each is further
refined as necessary. If the solvent is miscible
with water the stream must go to distillation in
order to achieve a separation. For applications
where recovery of the contaminant is not de-
sired, the steam-vapor mixture can be in-
cinerated since it is now more economically
feasible due to the much higher fuel value of
the stream. After the activated carbon has
been regenerated, a stream of cool air is passed
through the bed to cool and dry the adsorbent,
preparatory to using it again for adsorption.
There are several limitations and disadvan-
tages to the use of fixed beds. When a large
volume of gas is to be treated, adsorption effi-
ciency drops off due to the difficulty in obtain-
ing uniform flow distribution across the face of
the bed. In order to minimize the pressure drop
across the bed (and thus power consumption),
the adsorbent layer is usually very thin, some-
times only a few inches. Unfortunately, this
leads to very poor utilization of the volume
within a vessel and subsequent increase in
equipment costs. In cases where high concen-
trations in the air must be treated, the heat
generated by adsorption cannot easily be re-
moved. Other problems arise from the batch-
wise nature of fixed beds. Each vessel must
141
-------
have several valves in order to properly se-
quence through the cycle steps: adsorption,
desorption, cooling, and drying. These valves
lead to leakage, operating, and maintenance
problems. The containment of the activated
carbon in a steel vessel requires that the metal
as well as the adsorbent be heated up and cooled
down every cycle. This results in an unneces-
sary use of utilities. When adsorption is car-
ried out in fixed beds, there is always a portion
of the bed that is not completely exhausted due
to mass transfer resistance. This unused bed
must still be regenerated each cycle, however,
and this leads to increased utility require-
ments. The use of steam as the direct heating
and desorbing medium leads to more difficult
solvent recovery steps than if water were not
present. The condensed steam can also present
a very bothersome water pollution problem.
Some years ago it became apparent that a con-
tinuous process based on fluidized bed adsorp-
tion offered a means of improving upon acti-
vated carbon adsorption.
FLUIDIZED BED ADSORPTION
If a bed of activated carbon is maintained in
a fluidized state, there are many advantageous
properties of the system. Particles in a fluid-
ized condition behave much like a liquid. This
means that the regenerated carbon can be con-
tinuously brought into contact with the con-
taminated air and a portion of spent carbon
constantly withdrawn. The result is that only a
small quantity of adsorbent need be main-
tained in the adsorption zone and very low
pressure drops are achievable. The fluidity
allows the adsorption zone to be designed as a
series of stages analagous to a distillation
tower with activated carbon introduced coun-
tercurrent to the gas onto the top stage and
withdrawn from a bottom stage. When the ad-
sorption is staged in this manner, it is possible
to completely exhaust every bit of adsorbent
before it is regenerated and regeneration heat
and stripping gas are conserved. Because the
velocities of gas needed to fluidize a bed are
much higher than those allowable in a fixed
bed, the gas phase mass transfer resistance is
drastically reduced. A considerable amount of
heat is generated when solvent is adsorbed at
high concentrations. Fluidized beds transfer
the heat away well, preventing "hot spots"
that have been implicated in fixed bed fires.
The ability to move the carbon into and out of a
desorption zone eliminates the need to cyclic-
ally heat and cool the vessel walls with the ac-
companying waste of energy. Several attempts
to apply fluid bed adsorption principles had
previously been attempted with very limited
success.
In the early 1930's, there was an attempt to
develop moving beds of activated carbon with-
out success (refs. 1,2). Around 1950, Union Oil
Company developed the Hypersorption mov-,
ing bed process for the recovery of hydrocar-
bons from natural gas and for other hydrocar-
bon fractionations (refs. 3,4). It was not until
the mid 1960's that Cortaulds, Ltd. developed a
fluid bed adsorber for the recovery of carbon
disulfide from air in a rayon fiber plant and ap-
plied it to several solvents (refs. 5,6,7). All of
the processes were plagued by excessive
losses of activated carbon due to breakage and
attrition. The capital costs were high and there
were economic disadvantages in scaledown.
The other major difficulty was breakdown of
mechanical equipment, primarily that asso-
ciated with transporting the carbon from sec-
tion to section. One reason was the erosive
nature of the adsorbent on steel parts. The suc-
cessful use of fluid bed adsorption had to wait
for the introduction of a dramatically different
activated carbon and the development of a new
process system to take full advantage of that
product's unique properties.
PURASIV® HR ADSORBENT
PuraSiv® HR is a proprietary fluidized bed
activated carbon adsorption/desorption system
licensed from Kureha Chemical Company, Ltd.
and sold by Union Carbide Corporation in the
United States and Canada (refs. 8,9,10,11). The
adsorbent around which the system has evolved
is a beaded activated carbon (BAC) which is
produced by a patented forming process (ref.
12). The raw material for the BAC is a molten
petroleum pitch. Being thermoplastic, the mol-
ten material can be formed into nearly perfect
spheres (see Figure 1) by the action of surface
tension. Since the spheres are still thermo-
plastic, oversize and undersize particles can be
recycled, yielding as narrow a size distribution
as is needed without expensive losses. The
beads are then carbonized and steam activated
142
-------
Figure 1. Beaded activated carbon.
under conditions similar to those used for con-
ventional carbon activation. Careful control of
the activation yields a PuraSiv* HR adsorbent
with equilibrium loadings comparable to those
of the highest grades of granular activated car-
bon.
The good mass transfer characteristics of
BAG are attributable to its superior activation,
very small size, and spherical shape. The com-
bination of small size and spherical shape re-
sults in a high ratio of external surface area
per unit volume, and this is conducive to high
mass transfer rates. Also, the small diameter
particles which characterize the PuraSiv® HR
adsorbent have relatively small diffusion paths
from the adsorbent surface to the adsorption
sites and this results in higher diffusion rates.
The activation given to the product results in a
very large percentage of micropores in the
range needed for solvent recovery. The result-
ing high pore volume and narrow pore size
distribution also improve the mass transfer
characteristics.
The fluidity of BAG, especially in the fluid-
ized state, is nearly that of a liquid as would be
expected from its particle shape. The spheric-
ity and the narrow size distribution contribute
to very stable fluidization and homogeneous
flow patterns. This fluidity also allows for very
easy transfer by a carrier gas and ease in
charging and discharging of equipment. These
same properties result in the ability to trans-
port the BAG in a dense phase by gravity with
little tendency to bridge in narrow passages.
Another result of the forming process is the
excellent attrition resistance exhibited by the
PuraSiv® HR adsorbent. As contrasted with
products manufactured by binding powdery
raw material or granulating nonhomogeneous
substances, BAG has a structure that is uni-
form, homogeneous, and strong. This strength
and the sphericity result in an adsorbent with
attrition resistance that is significantly
superior to conventional activated carbon. This
higher attrition resistance not only reduces
costs associated with adsorbent replacement,
143
-------
but also adds significantly to on-line operating
time and minimizes particulate emissions.
All of these properties make BAG an ideal
adsorbent for a fluidized bed adsorption .proc-
ess. It is around these unique qualities that the
PuraSiv® HR process has been developed.
PURASIV® HR ADSORPTION SYSTEM
The heart of the PuraSiv® HR system is the
adsorber/desorber vessel shown in Figure 2.
The particular unit depicted is for a system uti-
lizing nitrogen as the stripping gas and steam
as the heating medium. There are many other
means of stripping and heating that will be dis-
cussed later.
The adsorption section consists of multi-
staged fluidized beds. An important considera-
tion is that the influent gas be introduced into
the fluidizing section as uniformly as possible
to insure uniform contact between the fluid-
ized adsorbent and influent stream. The Pura-
Siv® HR system incorporates gas distribution
structures which reduce impact pressure and
allow proper distribution with a minimal pres-
sure drop. Good gas distribution not only as-
sures efficient adsorption but also eliminates
"hot spots" believed to be one cause of fixed
bed fires. The particles of adsorbent are fluid-
ized by the upward moving gas and flow
smoothly across the tray to the downcomers.
The position of the downcomers are alternated
on each tray, thereby eliminating dead space
and allowing for uniform contact between the
adsorbent and gas through the adsorption sec-
tion. The gas velocities are several times those
achievable in fixed beds and since there is only
a single vessel the overall unit is more com-
pact. The fluidized bed height is kept at a
specified level by the weir'. Particles from an
upper stage fall onto a tray and overflow the
weir into the downcomer. This patented tray
design (ref. 8) results in efficient countercur-
rent contact of the PuraSiv® HR adsorbent
with the upward moving gas. The adsorbent
becomes progressively saturated with the sol-
vent as it falls downward through successive
trays of the adsorber and, after passing
through the seal zone, falls into the secondary
adsorption zone.
The purpose of the secondary adsorption
section is to remove from the stripping nitro-
gen most of the residual solvent not condensed
in the condenser. In this section the adsorbent
falls as a dense moving bed countercurrent to
the upward flowing stripping gas. The nitro-
gen leaving at the top has a solvent level low
enough that it can be recycled as fresh desorp-
tion gas. The BAG has enough adsorptive ca-
pacity to accomplish this since the concentra-
tion of solvent in the nitrogen leaving the con-
denser is easily an order of magnitude higher
than that in the adsorber. The adsorbent exits
through another seal zone into the desorption
section.
Desorption is usually carried out by raising
the temperature and/or by reducing the partial
pressure of the adsorbate (solvent that is ad-
sorbed). This is often accomplished by desorb-
ing with gases such as steam, nitrogen, or air,
while simultaneously supplying sufficient heat
to bring the material to the desired desorption
temperature and to provide the heat of desorp-
tion. This heating can be accomplished directly
or indirectly. In the PuraSiv® HR desorption
section of Figure 2, the adsorbent, falling as a
dense moving bed phase, is countercurrently
contacted with the upward flowing stripping
nitrogen. The heat of desorption is supplied in-
directly by steam on the shell side of a shell
and tube heat exchanger. The nitrogen for
stripping is recycled from the secondary ad-
sorber to the desorber bottom by a recycle
blower. The solvent-laden stripping gas exits
the top of the desorber to the condenser. Most
of the solvent in the nitrogen is condensed and
recovered while the nitrogen goes to the sec-
ondary adsorber. One of the advantages of us-
ing nitrogen in this recycle manner is that only
a small amount of fresh nitrogen is required —
that to make up losses. Since nitrogen does not
condense as steam from a fixed bed would in
the condenser, the condenser size and cooling
water requirements are significantly reduced.
The use of nitrogen allows solvents to be
recovered with much lower water content,
thus reducing the amount of post treatment
before reuse. Because the steam for heating is
applied indirectly, the steam condensate can be
returned directly to the boiler system without
expensive treatment to remove the dissolved
solvent. In many cases with nitrogen stripping
gas there will be no water condensate from the
condenser to creat a secondary pollution prob-
lem.
After the BAG flows down to the bottom of
144
-------
CLEAN
GAS
ADSORPTION
SECTION
SECONDARY
ADSORPTION
SECTION
DESORPTION
SECTION
N_ RECYCLE
BLOWER
^ BAG FLOW
<^ 1 GAS FLOW
WEIR
DOWNCOMER
ADSORBER TRAY
BAG LIFT PIPE
SEAL ZONE
SECONDARY
ADSORBER TUBE
CONDENSER
RECOVERED
SOLVENT
DESORBER TUBE
CONDENSATE
BAG LIFT AIR
Figure 2. PuraSiv® HR adsorber/desorber.
145
-------
the column past the desorption section, it is
recycled to the top of the adsorption section
through the airlift line. The carbon and air are
blended in a special nozzle and the BAG is
lifted by the airstream. No mechanically mov-
ing devices are involved in the system. The
airlift nozzle also acts as a seal zone between
the desorber and the airlift line. Due to the ex-
cellent flow properties and attrition resistance
possessed by the PuraSiv® HR absorbent,
these transfers are accomplished with very
small adsorbent attrition losses and relatively
low power consumption.
As mentioned above, the PuraSiv® HR
system has wide versatility in design and
operation. The adsorber section can have any
number of stages and any desired BAC-to-gas
ratio. For some applications special formula-
tions of BAG have been developed for difficult
separations. However, the broadest flexibility
is in the desorption section. Here the stripping
gas and the heating medium can be custom de-
signed for the application. In addition to the
use of nitrogen discussed above, the desorbing
gas can be air, steam, carbon dioxide, combus-
tion gas, or fuel gas. Figure 3 illustrates the
use of air for stripping perchloroethylene. The
use of inexpensive air is possible for solvents
that are not combustible. Steam can be used
just as it is in conventional fixed beds. COe and
combustion gas would be used in the same way
as nitrogen wherever they are more available.
Fuel gas might be the choice for desorbing if
the solvent or contaminant is to be incinerated
or used as a fuel supplement.
The process also has a variety of heating
means available. The indirect heating medium
can be combustion gas, high temperature heat
transfer fluids, or electric resistance heaters as
well as steam. This allows the temperature of
desorption to be selected for optimal perform-
ance. This is in contrast to fixed beds, which
are generally limited to about 100° C. The
higher temperatures are especially useful in
recovering high boiling point materials and for
deodorization applications to accomplish near-
ly complete desorption. Figure 4 is a schematic
of a PuraSiv® HR system for recovering
trichloroethylene, which uses direct steam for
heating and desorption. Here, the steam is
used directly because the trichloroethylene
hydrolyzes with heat and moisture to form cor-
rosive hydrochloric acid. To make a shell and
tube desorber of material exotic enough to
withstand such an environment can be prohib-
itively expensive. The use of direct steam re-
quires only that the desorber walls be corro-
sion resistant and there can be a nonmetallic
Raw gas
Steam
Recovered ,ank
solvent
Figure 3. PuraSiv® HR for perchloroethylene recovery.
146
-------
Treated gas
Raw gas ,>.
0 . . , Safety valve
Reducing valve 7
r—tf
Steam
Compressed air
Cooling
water
Solvent
Figure 4. PuraSiv® HR for trichloroethylene recovery.
coating since there is no need for high heat
transfer. Although the desorber is now similar
to a fixed bed, the adsorber section still offers
all of the fluidized bed advantages. Since the
desorption section has a large impact on the
capital and operating costs of the unit, the
variations and flexibilities available in the
PuraSiv® HR system allow an optimization of
processing techniques and a minimum of costs.
The range of emission control problems that
can be solved with this process are innumer-
able. Table 1 lists the solvents and contami-
nants that have been commercially treated by
fluidized bed activated carbon adsorption. The
range of applications treated include 6- to
4,000-nm3/min flows and concentrations of 100
to 10,000 ppm by volume. A representative list
of the commercial experience of the Pura-
Siv HR system to date is shown in Table 2.
Figure 5 is a photograph of one of these in-
stallations. The unit is used to recover an
TABLE 1. FLUID BED APPLICATIONS
Acetone
Ammonia
Butyl alcohol
Carbon disulfide
Carbon tetrachloride
Ethyl acetate
Ethyl alcohol
Ethylene dichloride
Formaldehyde
Formalin
Isopropanol
Methyl cellosolve acetate
Methyl ethyl ketone
Naphtha
n-Hexane
Perchloroethylene
Phenol
Toluene
Trichloroethylene
Trimethyl benzene
Vinyl chloride
Xylene
aromatic solvent from an adhesive coating
vent stream. It is quite apparent that many of
these, if not actually in textile finishing plants,
are very closely related to the problems en-
countered there.
147
-------
Figure 5. Commercial PuraSiv© HR unit (820 nm/min).
148
-------
TABLE 2. PURASIV® HR COMMERCIAL EXPERIENCE
Application
Cloth washing
Degreasing
Spray coating
Chemical plant
.Film coating
Printing
Degreasing
Printing
Chemical plant
Degreasing
Adhesive coating
Foundry
Lamination
Degreasing
Magnetic tape
Film coating
Paper coating
Metal coating
Solvent
Chlorinated solvent
Chlorinated solvent
Thinner
Chlorinated solvent
Aromatic solvent
Thinner
Mixed solvent
Aromatic solvent
Alcohol
Chlorinated solvent
Aromatic solvent
Odor
Mixed solvent
Chlorinated solvent
Mixed solvent
Aromatic solvent
Mixed solvent
Mixed solvent
Flow
(Nm3/min)
20
30
120
50
250
40
40
180
30
40
820
4,000
120
50
200
100
200
300
Influent
(ppm)
4,600
3,500
100
3,560
1,680
1,450
390
1,800
1,200
600
2,710
20
2,000
1,500
2,800
1,000
550
2,800
Effluent
(ppm)
30
50
_
70
30
50
40
80
50
50
50
0.4
50
50
50
100
80
100
APPLICATION OF TEXTILE PLANT
EXHAUSTS
Use of activated carbon adsorption and/or
solvent recovery from textile plant exhausts is
complicated by the variety of fabrics and asso-
ciated treatments involved. The fabric itself or
the colors and patterns are changed very fre-
quently. Common textile finishing operations
include fabric preparation (including singeing,
desizing, scouring, bleaching, and drying),
printing/dyeing (including ageing, washing,
and drying), and final finishing (application of
special finishes and curing/heatsetting).
The principal variable in fabric preparation
is the inclusion or exclusion of scouring and the
need for chlorinated solvent scouring, as with
some polyester fabrics, versus the use of water
and detergents, as with most other fabrics. .
In printing, combinations of certain dyes
with acetate or acrylic fabrics causes emis-
sions. Most other operations only evolve water
vapor (ref. 13). If print pastes used with cotton
or rayon fabrics include significant quantities
of urea, again substantial emissions can be ex-
pected. In dyeing, the dye carriers change with
colors used and fabric. All emissions occur dur-
ing dye fixations (ageing).
The final finishing step basically involves
the application of resins, most based on for-
maldehyde polymers, to bestow properties
such as antistat, crease resistance and water-
proofing, and heat setting the resin in a tenter
frame at high temperatures. The resin formula-
tions change with both the fabric as well as the
kind of property imparted to the fabric.
The exhausts could be further diluted by
area ventilations.
EFFECTS OF PROCESS VARIABLES
Various parameters significantly affect the
performance of fluidized bed activated carbon
adsorber. These include temperature, species
of hydrocarbons and their concentrations,
quantity of moisture present, and dust loading.
The sizing of the adsorber, the choice of con-
figurations, mode of desorption, materials of
construction, and selection of ancillary equip-
ment for pre- or post-treatment of the exhaust
gases are highly dependent on these param-
eters as well as the presence of toxic or odor-
ous compounds.
Activated carbon, in general, has greater af-
finity to higher molecular weight compounds
and is effective with C4 to Ci4 (boiling points in
the range 65° C [149° FJ/1760 C [349° F]) species.
Changes in adsorber/desorber are required be-
149
-------
yond this range, for example, higher boiling
organics will increase the severity of desorp-
tion. Certain chlorinated compounds break
down forming inorganic acids necessitating
proper selection of materials of construction.
The adsorption of hydrocarbons by activated
carbon is concentration dependent at a given
temperature. Hence, lower inlet concentra-
tions will require more severe desorption to
achieve a preset degree of emission control.
Textile plant exhaust temperature ap-
proaches or exceeds normal desorption tem-
perature of activated carbon. These gases,
therefore, will require precooling prior to
entering the adsorber. The degree of precool-
ing usually prescribed can easily be carried out
either with tower cooling water or with gas/
gas heat exchangers depending on the reuse
economics of the recovered heat. Close atten-
tion will have to be paid to undesirable conden-
sation of vapors in the heat exchanger. Lower
inlet concentrations may require lower inlet
temperatures if the hydrocarbon recovery re-
quired is a very high percentage of the inlet
concentration.
Relative humidities of up to 50 percent at
38° C (100° F) have been easily handled in car-
bon bed adsorbers. Depending on the mode of
desorption, products end up with 2 to 20 per-
cent water. Higher moisture content of the ex-
haust gases result in more water in the pro-
duct. Special dehumidification may be required
if the humidity far exceeds 50 percent.
Dust, in finishing plants, appears in the form
of lint. Unless sticky or rendered sticky prior
to entering the adsorber and likely to clog the
distributor plates, the system will allow dust
to pass right through without affecting per-
formance.
Levels of toxic and odorous hydrocarbons
permitted in the exhaust gases are usually
lower than those for other hydrocarbons.
Often, the inlet concentrations are also low. If
it is not possible to adsorb them along with the
rest of the hydrocarbons, a separate finishing
stage is added to the fluidized bed and the ac-
tivated carbon stream is handled separately
with more severe desorption conditions and
the desorbed gases are usually incinerated to
render them innocuous using less fuel than if
TABLE 3. APPLICATION EXAMPLES OF FLUIDIZED BED ADSORBERS
Solvent
Perchloroethylene Trichloroethylene Toluene
Inlet gas
Flow rate (nm^/min)
Temperature (° C)
Concentration (ppm/vol)
15
40
4,600
9.3
20
2,600
240
47
1,680
Outlet gas
Concentration (ppm/vol) 3040 20-30 <50
Desorption
Gas flow rate (nm3/min) 36'air' 4y(steam) 80(N2>
Temperature (° C) 145 110 160
Carbon
Inventory (kg) 300 90 1,500
Circulation rate (kg/h) 120 30 600
Utilities
Electricity (kwh) 37 2.3 22
Steam (kg/h) 20 50 100
Cooling water (ton/h) 0.6 1 3
Nitrogen (nmfyh) - _ 15
Solvent recovered (kg/h) 20 8 100
150
-------
the incineration were carried out without ad-
sorption/desorption.
PERFORMANCE CHARACTERISTICS OF
FLUIDIZED BED
Table 3 illustrates the performance capabili-
ties of some typical fluidized bed activated car-
bon adsorbers. It must be noted that these are
not the performance limits.
Fluidized Bed Versus Fixed Bed
• Adsorption and desorption can be carried
out continuously in a single vessel.
— There are no valves to be manipulated
to switch vessels.
— The vessels are not alternately
heated and cooled, only the carbon is,
effecting significant energy savings.
Capital and installation costs are lower
because of the single vessel configuration,
elimination of timers, valves, duplicate
heating and cooling systems, emissions
monitoring system to check bed satura-
tion, etc., and smaller, compact equipment
which results from high gas flow rates per
unit area.
Homogeneous fluidization and continuous
flow of carbon avoids maldistribution of
TABLE 4. ECONOMICS OF FLUIDIZED BED ADSORBERS
Operating conditions
Solvent
Inlet gas
Flow rate (nmfymin)
Temperature (° C)
Concentration (ppm)
Casel
Toluene
240
47
1,680
Case 2
Naphtha
940
38
105
Outlet gas
Concentration (ppm)
Solvent recovery (kg/h)
Operating time (h/yr)
Costs
Equipment cost ($)
Installation cost ($)
Space requirements (m*)
Utilities
Steam (kg/h)
Electricity (kwh)
Cooling water (ton/h)
Nitrogen (nmfyh)
Annual operating cost ($)
Annual maintenance cost ($)
Annual carbon replacement cost ($)
Payback period (yr)
<50
100
8,000
Fluidized bed Fixed bed
200,000
50,000
37
100
22
3
15
14,400
8,000
800
2.1
180,000
75,000
75
365
44
22
38,000
15,000
1,000
2.9
8
Fluidized bed
510,000
130,000
150
33
80
0.7
0.25
21,900
20,000
200
_
26
,000
Fixed bed
500,000
200,000
200
135
140
8
44,000
40,000
2,000
_
151
-------
gases improving adsorption efficiency.
• For the same reason above, hot spots are
not formed, making fluidized beds less
prone to fires.
Economics of Fluidized Bed Adsorbers
Table 4 shows cost comparisons for two dif-
ferent applications of fixed and fluidized bed
activated carbon adsorbers.
CONCLUSIONS
With the advent of stronger, fluidizable ac-
tivated carbon, fluidized bed carbon adsorption
of hydrocarbon emission control and solvent
recovery has become a reality and shows dis-
tinct advantages over traditional fixed bed
operations. Maldistribution of gases, asso-
ciated with fixed beds, is minimized resulting
in the most efficient use of the carbon. Both hot
spots created by nonuniform gas distribution
and high bed temperatures, which result if the
inlet solvent concentration is very high, are
also eliminated. Design flexibilities permit
fluidized beds to cope with ever-changing com-
position of exhaust gases from textile finishing
plants. Compactness and continuous operation
in a single vessel improve the overall eco-
nomics of fluidized beds over fixed beds.
REFERENCES
1. V. Pantenburg, "Process for Regeneration
of Adsorption Material," U.S. Patent
1,784,536, issued December 9,1930.
2. F. J. Bechtold, "Regenerating Granular
Adsorbents," U.S. Patent 1,836,301,
issued December 15,1931.
3. W. F. Bland, "Design .Details, Operating
Results of First Commercial Hyper-
sorber," Petroleum Processing, July 1948.
4. C. Berg, "Hypersorption in Modern Gas
Processing Plants," Petroleum Refiner,
Vol. 30 (No. 9), September 1951.
5. H. M. Rowson, "Fluid Bed Adsorption of
Carbon Disulfide," British Chemical Engi-
neering, Vol. 8 (No. 3), March 1963.
6. D. A. Avery and D. H. Tracey, "The Ap-
plication of Fluidized Beds of Activated
Carbon to Solvent Recovery from Air or
Gas Streams," Institution of Chemical En-
gineers Symposium Series No. 30,1968.
7. D. A. Avery and D. A. Borston, "The
Recovery of Solvents From Gaseous Ef-
fluents," Chemical Engineer, January/
February, 1969.
8. H. Murakami, et al., "Method for the
Purification of Waste Gas Containing
Gaseous Pollutants," U.S. Patent
4,047,906, issued September 13,1977.
9. H. Murakami, et al., "Method and Ap-
paratus for the Purification of Waste Gas
Containing Gaseous Pollutants," U.S. Pa-
tent 4,061,477, issued December 6,1977.
10. Y. Sakaguchi, "Development of Solvent
Recovery Technology Using Activated
Carbon," Chemical Economy & Engineer-
ing Review, Vol. 8 (No. 12), December
1976.
11. Anon, "Beaded Carbon Ups Solvent
Recovery," Chemical Engineering, Vol 84
(No 18), August 29,1977.
12. Y. Amagi, et al., "Method for the Prepara-
tion of Carbon Moldings and Activated
Carbon Moldings Therefrom," U.S. Pa-
tent 3,917,806, issued November 4,1975.
13. H. H. Northup and W. F. Turner, "Air Pol-
lution Control in Textile Finishing," Tex-
tile Chemist & Colorist, Vol. 7 (No. 4),
April 1975.
152
-------
CATALYTIC OXIDATION OF HYDROCARBON FUMES
Richard E. Githens, Donald M. Sowards*
Abstract
The efficiency of fixed-bed vapor/solids
catalyst beds has been determined and correla-
tions published to permit accurate design of
catalyst beds. These designs have been used
successfully in textile and other applications.
All existing EPA guidelines can be met for the
oxidation of hydrocarbon emissions if these
design procedures are followed.
The operating costs for fuel are substantially
less than for thermal incineration. Both ther-
mal and catalytic afterburners utilize heat re-
covery, recirculation, and stack gas econo-
mizers in the design of systems.
INTRODUCTION
The utility of heterogeneous catalysis in air
pollution control is well established in such
fields as nitric acid manufacture, chemical
plant emissions, and paint bake ovens. It has
been equally successful, but less well known, in
the textile fibers field. This article will discuss
some of the applications of particular interest
to the textile fibers industry.
DESIGN BASIS
There are many published reports on
catalytic reactions in the gas phase using a
solid catalyst. The manufacture of many chemi-
cals is dependent upon a detailed knowledge of
the reaction kinetics, heat and mass transfer,
and pressure drop. Nitric acid, hydrogen
cyanide, and formaldehyde are but a few of
such commercial heterogeneous catalytic
operations.
Less publicized, but equally well known to
those in the field, are the design methods for
oxidation of hydrocarbons to carbon dioxide
and water (refs. 1,2 and Table 1). Each supplier
*E. I. du Pont dc Nemours & Company, Int.
of catalysts has his own research data and cal-
culation methods but, for reference purposes,
the report prepared by Shell Development
Company for the U.S. ^n-ironmental Protec-
tion Agency will be used (ref. 3). Referring to
Figure 1, it has been found that most of the
practical commercial installations of catalytic
afterburners are operated in the range of flow
and temperature where mass transfer is con-
trolling. There are at least six different forms
for the catalyst substrate and as many catalyst
compositions. For brevity, this paper will show
correlations only for three substrates and a
single precious metal catalyst (Figures 1
through 7).
A monolithic honeycomb structure made
from various grades of alumina or other
ceramics is one type of substrate. Cylindrical
pellets (1/8 in. diameter x 1/8 in. high) are
another type of substrate or support. The third
is a wire mesh pad made of appropriate high-
temperature metallic ribbon. All can be coated
with platinum, palladium, or other precious
metals. Not considered here are the various
metal-oxide catalysts that are also used in
some applications. Miller and Wilhoyte (ref. 2)
showed that when the flow velocity and tem-
perature are in the range where mass transfer
is controlling, the honeycomb ceramic supports
are more efficient than pellets or metal foil
supports. Note that both pellet beds and metal-
ribbon beds must be larger in diameter than
honeycomb ceramic beds to run at the same
pressure drop. This is important because the
size of the catalyst bed affects the size and cost
of the system containing the catalyst.
CAPITAL COST
Each system is custom designed and there
are many variables that can affect the cost of
the system. Consequently two systems with
identical gas flow rates can have widely dif-
ferent costs. Some of the variables that have
significant effects on cost are:
153
-------
• Type and efficiency of the heat ex-
changer, if used at all;
• Degree of complexity of the instrument
control panel and the controls;
• The number and type of dampers used to
control recirculation;
• The type of burner train used; and
• The amount and type of catalyst.
In spite of these differences, it seems useful
to have some guidelines as to the relative cost
versus the flow rate. Several recent examples
.from our sales files attempt to break them
down into broad categories. (These are shown
in Figure 5 as unit costs per SCFM versus flow
rate in SCFM.)
• The main incinerator body including
catalyst, plus all the blowers, burners,
controls, and instruments.
• The recuperative heat exchanger and its
associated duct work.
Limited personal experience with thermal
oxidation systems indicates that the total cost
of the system is very similar to the total cost
for a catalytic oxidation system but the in-
dividual breakdown is, of course, different
(refs. 3,4). The thermal system must ue de-
signed for higher operating temperatures.
The spread of our data around the curves
shown is probably ± 20 percent due to the dif-
ferences previously noted and should be in-
dexed from June 1978.
OPERATING COST
The main advantage of a catalytic system
versus a thermal system is the energy savings
due to reduced fuel consumption (refs. 5,6,7).
The catalyst replacement cost must be con-
sidered as an offset to the fuel savings.
Although suppliers usually warrant the cata-
lytic life for only 1 year, our experience in
hydrocarbon applications is that the catalyst
can last well beyond this with proper care and
maintenance. A good average figure would be 4
years, but we have catalysts still in use after 6
years.
Examples chosen to illustrate the effect of
solvent load on fuel consumption are shown in
Tables 2, 3, and 4. At any concentration above
about 5 percent of the lower explosive limit
(LED, the catalyst system can run self-
sustaining and no fuel is required if a 70 per-
cent effective heat exchanger is used. Above
about 30 percent of the LEL, the solvent con-
centration will cause enough temperature rise
that dilution must be used to prevent over-
heating of the catalyst. For high concentra-
tions like this, it is freqently more economic to
use thermal oxidation. However, in most tex-
tile operations we are operating catalytic
abatement equipment well below 30 percent of
the LEL. In most tenter frames the concentra-
tion is below 3 percent of the LEL.
INSTALLATION COST
For a typical textile plant installation, labor
cost for tie-ins is about 40 man-days of time
($1.50/SCFM). However, each case must be de-
termined individually because of site dif-
ferences. All of the units up to 20,000 SCFM
are skid-mounted on one skid and need only be
connected to the duct work and utilities after
mounting on the roof support steel. Above
20,000 SCFM, the units may be on one skid or
may be broken into two or more skids for ease
of shipping. In this case, the pieces may be
mounted on the roof and then Connected by
field contractors (Photographs 6, 7, and 8).
EXAMPLES OF EXISTING
INSTALLATIONS OF
CATALYTIC AFTERBURNERS
• A printworks in New England using a
Thermosol oven to cure polyester/cotton
printed fabric had problems with visibil-
ity and odor complaints from residents. A
catalytic incinerator was installed, which
solved these problems. Approximately 80
percent of the clean, hot air is recycled to
the oven. The net reduction in fuel costs
was approximately $5,000/yr at 1976 fuel
values. A thermal incinerator required to
do the same job was more expensive and
would have consumed far more fuel. The
plant engineer stated that the catalytic
reactor "is the only control system I saw
that is not energy intensive — the only
one that does not require more energy
than we normally use in our printing
operation." (See Photograph 5.)
• A finishing plant in southern Penn-
154
-------
PHOTOGRAPH 1
PHOTOGRAPH 2
PHOTOGRAPH 3
PHOTOGRAPH 4
155
-------
PHOTOGRAPH 5
PHOTOGRAPH 6
PHOTOGRAPH 7
PHOTOGRAPH 8
156
-------
TABLE 1. RECOMMENDED CORRELATIONS FOR TRANSPORT PROPERTIES
Catalyst typo
Torvex 2B
(1/8 in. Hexcell
Honeycomb)
S/V
(frl) e
268 0.61
L
(ft)
0.0091
X M M <
Iff) "Nu^Sh,1
0.833 NNu = 3.66 (1 + 0.095 NR Np Is) 0-45
M . = 7 cfi /i 4. n noc u u .L\0.45
f = 64/NRe(1+ 0.0445 N,,,,
5
Thermocomb
(8 dm. Honeycomb)
695
0.6
0.00345
Normally
0.167-0.333
N Nu = 2.35(1 +0.095 NReNprjj)
N oh= 2.35 (1 + 0.095 N noNCr t)°'4
Oil nc oC A
f = 53.3/N Re (1 +.0.0445 NRe^|D'5
Oxycals
36.5
0.515
0.0133
N -051
NNu-U.Sl
D'56 °'333
Nsh-0.51NRe
0.56 0.333
f = 0.5/NRB
0.15
Metal ribbon
("D" Series)
336
0.93
0.004
M - n EG u 0-5 u 0.333
NNu-0.55NBe Npr
N -DEEM
Nsh-0.55N
°-5
Re
f=(1-e)/e(5.2 + 3600(1-e)/NRe)
-AP = Z G 2/2e2p-gc'/L dZ + 2.1 G
o g
Spherical catalyst pellets
6e
dp
Normally
0.35-0.4
dp
f=(1-e)/
Note: Reproduced with permission from Shell Development Company
sylvania was having similar problems
with visibility complaints from two
tenter frames. A catalytic afterburner
was installed on one frame and the
before/after photographs are shown in
Figure 8. The energy recovery is higher
than in the first example because a
recuperative heat exchanger is used as
well as recycling of the hot, clean air.
Operation since the first quarter of 1978
has been satisfactory (Photographs 1, 2,
3, and 4).
An automotive assembly plant presently
being constructed will have catalytic in-
cineration in all of the paint curing ovens.
They will be using waterbase paints,
which are quite low in solvents but con-
tain plasticizers, surfactants, and biocide.
The purpose is to eliminate any visibility
and odor problems and recover the
energy value of the solvents in the vent
gas. A large portion of the hot, clean air
will be returned to the ovens to reduce
fuel consumption. The smaller remaining
portion is heat exchanged with ambient
makeup air before exhausting to the at-
mosphere. Because the exhaust gases are
free of the high molecular weight hydro-
157
-------
TABLE 2. EFFECT OF SOLVENT LOAD ON FUEL CONSUMPTION
Basis: 10,000 SCFM Total Flow
Toluene
AH = 17,613 Btu/lb =4,196 Btu/SCF
No Heat Exchanger
Fuel Cost = $2.25/mm Btu
Oven Temperature = 250° F
Preheat required:
Catalytic - 550° F, Thermal - 1,250°
%
LEL
1
2
5
10
20
50
ppm
V/V
120
240
600
1,200
2,400
6,000
Meters Btu/h
AH
302.1
604.3
1,510.7
3,021.3
6,042.6
15,106.6
Theoretical
AT-°F
26.5
53.0
132.5
264.9
529.9
1,324.6
mm Btu/h
required to preheat
Catalytic Thermal*
3.1
2.8
1.9
0.4
0
0
11.1
10.8
9.9
8.4
5.4
0
'Allows credit for heat released by combustion of solvent and no heat exchanger.
Note: In both catalytic and thermal incinerators, heat exchangers can be used to provide the preheat. Catalytic after-
burners can run self-sustaining (zero fuel consumption) at any concentration above about 4 percent of the LEL.
TABLE 3.
BASIS: $2.50/mn) Btu
Preheat 250° F •
Preheat 250° F-
500° F operating temperature for catalyst
1,300° F operating temperature for thermal
Fuel = SCFM x 1.085 X (To - 250)/i^L/ 8,000 h
/ UP/
Btu/h
/ $ /JL
/ mm Btu/ yr
SCFM
2,000
10,000
20,000
50,000
100,000
Fuel-catalytic
13,020
65,100
130,200
325,500
651,000
Fuel-thermal
45,570
227,850
455,700
1,139,250
2,278,500 .
Annual $
savings vs. thermal
32,550
162,750
325,500
813,750
1,627,500
Note: Annual fuel costs: 8,000 h/yr if no recirculation or recovery of heat.
158
-------
BASIS: $2.50/mm Btu
Preheat 250° F
TABLE 4.
550° F operating temperature for catalyst
Preheat 250 F— » 1,300 F operating temperature for thermal
Fuel = SCFM x 1.085 x (To - 250)
10
8,000 h
SCFM
2,000
10,000
20,000
50,000
100,000
. Btu/h — * -IL
mm Btu yr
$/yr
Fuel-catalytic
3,906
19,530
39,060
97,650
195,300
= $/yr
$/yr
Fuel-thermal
13,671
68,355
136,710
341,775
683,550
Annual $
savings vs. thermal
9,765
48,825
97,650
244,125
488,250
Notes: Annual fuel costs: 8,000 h/yr if 70 percent of heat exit incinerator is recovered in some form,
e.g. recirculation, steam generation, air preheat.
Electrical costs for blower motor will average about $467/yr per 1,000 ACFM; based on $140/
hp-yrandSOOACFM/hp.
* 10 r
1000
200 300 400 500 600 "700 800 900 1000 1100
TEMPERATURE,°F
i.o
400 800 1200 1600
SCFM/SO FT OF FACE AREA
2000
Figure 1. Effective first order rate constant and
chemical and mass transfer rates
contributing to it.
Figure 2. Pressure drop vs. flow.
159
-------
1000
o 100 -
1.0
Fig 3
400 800 1200 1600
SCFM/SO FT OF FACE AREA
2000
100 r
1.0 r
O.I
Fi,4
400 800 1200 1600
SCFM/SO FT. OF FACE AREA
2000
Figure 3. Heat transfer coefficient vs. flow. Figure 4. Ratio of heat transfer and mass transfer
to pressure drop.
10000
SCFM (I ATM,70°F)
100000
2.0
15
10
0.5
0
10
Note ; Bused on 7000 operating hours/year
and fuel at S225/MM Btu.
Etectnca! energy is $ 140 per HP-yr
70% HX ONLY
75% RECIRCULATION ONLY
70% HX t 75% RECIRCULATION
Fig 6
30 10000 100000
SCFMd ATM, 70°F)
FigureS. Equipment costs.
Figure 6. Annual operating costs.
160
-------
1.0—
1.3-
2.3
HONEYCOMB
WIRE RIBBON
PELLETS
Figure 7. Vessel diameter vs. support type.
carbons the heat exchangers should re-
main at their design effectiveness and
not require frequent cleaning.
A 2-piece can plant in central New Jersey
had an existing thermal incinerator
designed to handle 12,000 SCFM and was
using approximately $250,000/yr of fuel
at 1977 prices. After retrofitting with a
catalyst and reducing temperatures to
600° F, the plant engineer says, "the
energy savings are very gratifying." The
fuel bill for the first 9 months of opera-
tion is $100,000 less than during the 9
months when thermal incineration was
used.
An aluminum coil coater in northern New
Jersey had a thermal incinerator but gas
consumption exceeded his allocation,
forcing him to use propane for fuel. The
unit was retrofitted with a catalyst and
the temperature reduced to 600° F. The
fuel consumption dropped by more than
50 percent. As a result, the use of natural
gas was resumed and the fuel cost re-
duced by more than 75 percent.
A 2-piece can line in northern New
Jersey utilized a thermal fume incin-
erator with a recuperative heat ex-
changer for approximately 1 year before
retrofitting the unit with $21,000 worth
of catalyst. Propane was the only fuel
available to the plant. After approx-
imately 6 months of catalytic operation,
the plant manager reported his propane
consumption had been reduced by 75
gal/h. Stack tests indicated more than 90
percent reduction of solvent hydrocarbon
at a catalyst inlet temperature of 600° F.
The catalyst cost was payed back in ap-
proximately 2 months.
REFERENCES
1. M. R. Miller and D. M. Sowards, "Solvent
Fume Abatement by Ceramic Honey-
comb Catalyst Systems," Franklin In-
stitute First National Symposium on
Heterogeneous Catalysis for Control of
Air Pollution, November 22, 1968.
2. M. R. Miller and H. J. Wilhoyte, "A Study
of Catalyst Support Systems for Fume
Abatement of Hydrocarbon Solvents,"
APCA Journal, Vol. 17, No. 12, December
1967.
3. R. D. Hawthorn et al., "Afterburner
Systems Study," Environmental Protec-
tion Agency Contract EHS-D-71-3, Shell
Development Company, Emeryville, CA.
4. D. M. Sowards, Fixed-Bed Vapor/Solids
Contacting Device, U.S. Patent 3,977,090,
issued August 31,1976.
5. H. L. Breckenridge and W. P. Jensen,
"Survey of Energy Use in Metal Coil
Coating," Contract EY-76-C-07-1570, U.S.
161
-------
Dept. of Energy, EG&G Idaho, Inc., 7. U.S. Environmental Protection Agency,
published February 1978. "Control of Volatile Organic Emissions
6. "Catalyst Saves a Printer's New From Existing Stationary Sources,"
Market," Textile World, August 1976. EPA-450/2-76-028, Vol. 1, November 1976.
162
-------
RADIATION CURING FOR TEXTILE COATING
William K. Walsh, B. S. Gupta*
Abstract
This paper briefly outlines the technology of
solvent-based fabric coating and the magni-
tude of the solvent emissions problem.
INTRODUCTION
The technological changes that occur in the
fabric coating industry in the next 5 to 10 years
will almost all be developments directed at
reducing solvent emissions (ref. 1). The shape
and nature of these developments will be
determined to a large degree by the time
frame imposed by government regulations (ref.
2). A short time frame will favor short-term
solutions that will tend to become permanent,
while a longer period should favor more techni-
cally complex answers. The technology of
solvent-based fabric coating and the magnitude
of the solvent emissions problem will be ad-
dressed further is this paper. Solvent-free,
radiation curable coatings will be described
and their potential for use in fabric coating will
be compared to other possible technical solu-
tions to the emissions problem that are being
examined by the coatings industry. As is true
with other industries with strong environ-
mental concerns, the decisions made on these
uncertain options will affect the fabric coating
manufacturers' future for years to come.
Regional and national economic well being will
also be affected to some extent, as well as our
trade balance, since imports from developing
countries are already making significant
penetration into this market.
Coated fabrics are textile materials covered
with a film-forming substance. The fabric
serves mainly to provide strength, and the film
provides a functional purpose that the raw
fabric is unable to perform. The variety of dif-
ferent textile materials, coating materials, and
coating methods result in a large number of
*School of Textiles, N.C. State University, Raleigh, NC.
marketable products. The earliest coated fab-
rics were linen or jute impregnated with dry-
ing oil, leading to the name "oilcloth," which
was first produced in volume in the 18th cen-
tury. Nitrocellulose coatings appeared soon
after, as did latex rubber, which produced syn-
thetic leather and rainwear that was so satis-
factory that it remained the standard for more
than 50 years (ref. 3). Synthetic polymers were
introduced in the 1930's and are used almost
exclusively today. Acrylates and styrene-
butadiene rubbers are presently consumed in
large volume for back coating of upholstery
fabrics and carpets. These are water-based
latices and do not constitute a major emissions
problem. The largest uses of vinyl coatings
(plasticized polyvinyl chloride) are in auto-
mobile tops, upholstery, and low-cost apparel.
These coatings, for the most part, do not re-
quire solvent evaporation.
Polyurethane-coateu fabrics are of the high-
est cost and best quality in the field, due to a
number of desirable attributes which set them
apart from other coatings (ref. 3). Such at-
tributes are:
• warm, dry, pleasant handle,
• stability to washing and dry cleaning,
• high cold flexibility (without plasticizers),
• good wear properties and abrasion
resistance,
• high breaking strength, elongation, and
toughness,
• stability to oils and fats, and
• low density.
Urethane-coated textiles are finding applica-
tions in an increasing variety of apparel,
domestic, and industrial goods. The annual con-
sumption of fabric for coating exceeds 110
million square yards and for chemicals, over 20
million pounds, figures which are on an increas-
ing trend. The two types of coating used are
direct and transfer, illustrated in Figure 1. In
the direct method, the coating collapses around
the yarns, leaving a surface with the general
appearance of the fabric. Substantial penetra-
tion of the coating into the fabric results in a
163
-------
DIRECT COATING
TRANSFER COATING
Figure 1. Direct and transfer coating.
stiffer end product, depending on the coating
modulus. This type of coating is generally
given to fabrics to stabilize them against yarn
raveling and distortion (back coating), and to
reduce penetration by wind and water (face
coating). Some of the major end uses of the
face-coated urethane fabrics are protective
clothing, camping fabrics, military fabrics, and
industrial fabrics.
In transfer coating, the penetration is highly
minimized by either precuring the coating to
some extent before applying it to the fabric,
napping the fabric so that the coating adheres
to the nap rather than to the yarns, or using a
highly flexible tie coat such as foamed poly-
urethane. All of these methods lead to a fabric
that is soft and flexible. This kind of coating is
carried out in order to produce fabrics with
leather-like appearance, soft hand, and durable
properties. Applications include apparel, shoe
uppers, handbags, luggage, upholstery, and
numerous other domestic and industrial ap-
plications. The coating is usually cured in con-
tact with a release paper, and the leather-like
surface is formed using the pretextured
release paper as a mold.
The excellent properties attainable with
polyurethane coating are accompanied by two
negative parts that have limited the extent of
their use —high chemical cost and the solvent
emission problem. These are not independent,
since the cost of the solvents, which are
evaporated, contributes to the high cost of the
coatings. The urethane fabric coating section
of the U.S. textile industry is the industry's
largest air polluter, on a per pound (of fabric)
basis, and is a major contributor on the basis of
total emissions. The current production is
about 65 million square yards of leather-like,
transfer-coated (from release paper) fabrics for
handbags, shoe uppers, and apparel; and about
50 million square yards of direct-coated (mostly
rip-stop nylon) fabrics for camping and general
outdoor use. The coatings on all these fabrics
are applied at 15 to 40 percent in solvent, the
great majority of which is released into the at-
mosphere. From these production figures, total
emissions from the two processes can be esti-
mated at around 107 Ib/yr. A number of
solvents are mixed in various proportions for
different coatings, but toluene, dimethyl for-
mamide, ethyl acetate, and methyl ethyl
ketone comprise the bulk of the emissions.
Fabric coaters are quite concerned about
this problem and are currently considering
possible solutions. Incineration of stack gases
164
-------
is a partial solution, but is being used only
minimally at present due to its high energy
consumption. Recovery has not been deemed
economical, mainly due to the mixed nature of
the solvents ("randomly" mixed recovered
solvents currently sell for only about 20
-------
TABLE 1. ENERGY COST COMPARISON ELECTRON BEAM CURING VS. THERMAL CURING (ref. 5)
Operation criteria
Thermal curing
1. Coating application - white coating on aluminum foil
2. Solvent type coating - 35 percent solids
3. Coating laydown - 6 Ib/ream
4. Web width - 60 in.
5. Line speed - 500 FPM
Electron beam curing
1. Coating application - white coating on aluminum foil
2. Coating type - 100 percent solids
3. Coating laydown - 2 Ib/ream
4. Web width - 60 in.
5. Megarad dosage - 2 megarads (could be lower)
6. Line speed - 500 FPM
7. Processor voltage - 175 kV
8. Processor output - 200 watts/in. = 12 kW
9. Power input - 17.5 kW
Cost per hour
Thermal curing
Heat: 6,000,000 Btu/h (natural gas) @ $2.EO/mm Btu's = $15.00
hp (fans) 50 hp = 38 kW @ $0.02 kWh
'0.76
Chill water - 25 tons/h
6,600 GPH @ $0.20/m =
Cooling® $0.10 per ton =
Subtotal = $19.58
Exhaust treatment
1.32
2.50
Gas incineration or precipitation to meet EPA requirements. Costs
normally 100 percent of processing heat costs = $15.00
Total = $34.58
Electron beam curing
Power for curing: (electric)
17.6 kWh@ 0.02 =$0.35
Water -1 Ton/h
300GPH@$0.20/m =0.06
Cooling @$0.10/ton =0.10
Inert ing gas- (using own generator)
Fuel: 244 SCFH natural gas @ $2.50/MCF
Water: 1,920 GPH @$0.20/m
Power: 3.6 kW@ $0.02
Total
= 0.61
= 0.38
= 0.07
$1.57
Energy savings/year - Assuming 2 shift or 4,000 h/year basis:
($34.58 - $1.57) x 4,000 = $132,040
gas availability. The cost of this energy per
square yard of coated fabric varies from 0.5c to
2c and is comparable to the amortization cost
for curing equipment, either thermal or radia-
tion. It is essential, however, to avoid neglec-
ting chemical costs, which are usually much
higher. These costs overwhelmingly dominate
the others and this fact makes it necessary
that radiation curable resins either require
less chemical add-on, be either cheaper or very
close in price to the thermally curable resins,
or offer a product or some other advantage
great enough that the increased cost can be
recovered in the marketplace through in-
creased sales and/or unit prices. Fabric back
coating, for example, currently uses acrylic
latexes, at 45c to 55c/dry pound. Since radia-
tion curable compositions that are being
developed cost considerably more than this
($1.00 to $2.50/lb), it seems doubtful that they
will be able to replace latexes (or vinyls) in any
process where they are performing satisfac-
torily unless a clear product advantage can be
shown, or until the narrowing of the chemical
cost gap and the increase in energy prices
reverse this imbalance. Thermally curable
urethane coatings on the other hand, however,
cost from $1.50 to $3.50/lb. Since many of the
radiation curable compositions have a similar
chemical basis, prospects seem quite good for
the elimination of chemical costs as a factor
and chemical savings are probable in some
cases. Introductory prices for 100 percent
solids, radiation curable, light-stable formula-
tions suitable for fabric coating are being
quoted in the range of $2.25/lb.
166
-------
Film Properties
Lack of chemicals available that were
suitable for high rates of polymerization has
been another barrier in the utilization of radia-
tion technology in the coatings of textiles. The
research on textile radiation processes in the
late sixties and early seventies relied almost
completely on the versatile N-methylol acryl-
amide, which has a high polymerization rate
constant, gives gel on polymerization, and has
cellulose reactivity through the methylol
group.
The current and rapid development of a
number of fast curing compositions by sup-
pliers to the coating industry has substantially
changed this picture. Most of the systems
designed for application in the particle board,
can, and printing industries, however, general-
ly give cured products that are too stiff for
more than a limited number of textile applica-
tions. The reasons for the above compositions
tending to be stiff can be found by examining
the chemistry of radiation curing, excellent
reviews of which are available (refs. 6,7). The
basic equations (ref. 8) for the polymerization
rate (Rp) and the chain length (Dp) in a chain
growth, addition polymerization, illustrate
some of the problems
DP =
(kt)y
(1)
Rf1/2 [M] (2)
where kp is the propagation rate constant for
the growing chain, kt is the termination rate
constant for stopping the chain growth, R; is
the rate of initiation of growing chains, which
is controlled by the radiation intensity, and M
is the concentration of monomer. As Rs is in-
creased (which must be done to achieve high
line speeds), equation (2) shows that the Dp
decreases. At the high radiation intensities
necessary, the Dp of the polymers from most
common monomers is so low that the material
is useless as a binder or coating. This problem
is solved by using monomers with high kp (such
as acrylates or vinyl pyrrolidone), monomers
which form gel on polymerization (acrylates
again) or, most commonly, monomers with two
or more double bonds to give crosslinking and
build up molecular weight. As polymer begins
to form, kt decreases due to the viscosity in-
crease (diffusion of the growing chain ends is
restricted) and rapid curing takes place. All of
these techniques, however, lead to increased
crosslinking in the cured film. As a result, most
of the resin formulations that have been
developed result in cured binder films that are
much stiffer and less extensible than those nor-
mally used in pigment printing. To increase
flexibility, reactive polymers (actually low
molecular weight [1,000], multifunctional
oligomers) are used. The double bonds are
usually located on the ends of the molecule, and
increasing the molecular weight moves the
double bonds further apart, reducing the
tightness of the network formed when the dou-
ble bonds are polymerized.
Several approaches (refs. 6,9,10) have lead to
compositions that, on curing, form elastomers
with properties approaching those normally
used in textiles. In particular, oligomers of
flexible polyesters end-capped with di-
isocyanates and hydroxyethyl acrylate can be
polymerized with monofunctional acrylates or
vinyl compounds to give fairly soft,
elastomeric films. The difunctional oligomer
leads to crosslinking, the degree of which
varies inversely with the oligomer molecular
weight (or distance along the chain between
the double bonds). It has been found, for exam-
ple, that elongation at the break of the cured
films varies inversely with the square root of
the oligomer molecular weight. Inclusion of
small amounts of low molecular weight
multifunctional acrylic monomers lowers the
elongation considerably (ref. 9). The film pro-
duced by these techniques has many of the
characteristics of urethanes, since the
urethane linkage is used in the oligomer
backbone.
Using these approaches, several formula-
tions have been developed that give cured film
properties adequate for replacing solvent-
based urethanes in fabric coatings. However,
these are not sufficient. The fabric coating in-
dustry requires a wide variety of products
with a convenient range of properties such as
stiffness, elasticity, adhesion, water and sol-
167
-------
vent resistance, and flame retardancy. New ap-
plication techniques must be explored to meet
the extra demands of 100 percent solids
coating. Preliminary work has been done in
these areas (refs. 12,13,14,15) but much re-
mains to be done.
OTHER ALTERNATIVES
Besides the approaches of solvent recovery
or incineration, several other coating techni-
ques are in the preliminary stages of develop-
ment and could offer potentially viable solu-
tions to the problem.
Water-Based Urethanes—
Aqueous dispersions of polyurethanes are
now available and are being used by a few
direct coaters for camping and outdoor fabrics.
Techniques for transfer coating have not been
developed for these materials. Their major
drawback is their high energy consumption
during drying and the accompanying bottle-
neck in drying speeds due to the higher latent
heat of vaporization of the water.
Laminating With Extruded
or Calendered Films-
Extrusion of polyurethane films or forming
of the films by calendering of polymer chips is
a common solvent-free technique in the vinyl
coating industry. The equipment is very expen-
sive and only suited for large production runs
of similar materials and probably not for the
multitude of special coatings manufactured by
urethane fabric coaters.
100 Percent Solids,
Thermoreactive Systems—
These are similar in principle to the radia-
tion cured coatings in that polymerization
takes place on the fabric. Generally two com-
ponents are mixed before curing. Polymeriza-
tion begins right after mixing, so it is generally
done in-line immediately before the coating
head, a process which is somewhat awkward
and difficult to control.
High Solids, Thermoreactive Systems—
These are two component systems with
some solvent added to control viscosity and
minimize premature reaction. The technique is
a compromise that would not eliminate solvent
emissions, but may make the thermoreactive
systems more practical (ref. 16).
Powder Coatings—
Powdered polymer is applied to the sub-
strate in a spray and is melted and cooled to
form the coating. This technique has been suc-
cessful in some types of metal coating and is
currently under investigation for transfer
coating of fabrics (ref. 17).
SUMMARY
The use of radiation for curing of coatings
and inks is still a relatively new technology,
but it has become widely accepted in such
areas as metal decorating, particle board filler
coatings, package printing, and clear coatings
on magazine covers and, more recently, on
vinyl floor tile. The practice, briefly, involves
substituting a polymerizable monomer for the
solvent, and a low molecular weight reactive
polymer (oligomer) for the polymer. The coat-
ing is cured by room temperature, in situ
polymerization, instead of by solvent evapora-
tion. The free-radical polymerization is induced
by high intensity ultraviolet light (using an ad-
ded photoinitiator), or less commonly, electron
beam.
The technique has been, therefore, suc-
cessfully introduced primarily in areas where
traditionally coatings have been applied from
solvent, because the advantages are the
greatest. Elimination of pollution, energy sav-
ings, higher line speeds, and^greatly reduced
floor space are significant advantages, but
eliminating the cost of the solvent is often the
deciding factor.
Studies of the possibilities of transferring
this technology into the field of fabric coating
have been initiated in this laboratory and by
others. These preliminary studies demonstrate
the feasibility and potential economic attrac-
tiveness of the method and point out the re-
maining work to be done. This work involves
development of the radiation-curable coating
formulations that meet the needs of fabric
coaters and demonstration by fabrication and
testing of coated fabrics using these formula-
tions.
All of the commercial radiation curing ap-
plications to date have involved coatings on
168
-------
relatively inflexible substrates. Fabric coating,
on the other hand, requires soft, highly elastic
films after curing—properties not heretofore
obtainable by radiation curing due to the high
degree of crosslinking in the films. A great deal
of development in this area has already taken
place, but still more is needed for producing at
minimal cost, flexible films with a wide variety
of properties suitable for the various types of
fabric uses.
If radiation curing could be demonstrated as
a technically and economically successful way
of eliminating solvent emissions in urethane
fabric coating operations, it is probable that it
would be adopted by other processors of flexi-
ble substrates with similar pollution problems.
The other techniques described here are also
deserving of study. It is possible that a com-
bination of radiation and one or more of the
other techniques would be the answer. It is
clear that more time is needed to study these
alternatives. There is much work to be done,
however, and the effort must be begun in
earnest.
REFERENCES
1. F. Schore, Daily New Rews Record, p. 13,
March 6,1978.
2. R. Dorchies, Sources and Resources, Vol.
11, No. 6, p. 25, 1978.
3. H.-J. Koch, "Dyeing/Printing/Finishing,"
/rest. Textile Bulletin, Vol. 1, p. 11, 1978.
4. First International Meeting of Radiation
Processing, J. Silverman and A. R. Van
Dyken eds., and Radiation Phys. and
Chem., Vols. 8 and 9, 1977, translation.
5. Data furnished by Energy Sciences, Inc.,
Bedford, MA.
6. S. H. Schroeter, "Annual Review,"
Material Sci., Vol. 5, p. 115, 1975.
7. A. J. Vrancken, Chem. Peintures, Vol. 73,
No. 1, p. 3, 1973.
8. A. Chapiro, "Radiation Chemistry of
Polymeric Systems," Interscience, p. 132,
New York, 1962.
9. W. Oraby and W. K. Walsh, "Coatings and
Plastics Preprints," J. Appl. Polymer Sci.,
Vol. 32, No. 2, p. 263, 1977, in press.
10. D. D. Howard and B. Martin, Radiation
Curing, Vol. 4, No. 2, p. 18, 1977.
11. A. D. Fussa and S. V. Nablo, J. Coated
Fabrics, Vol. 4, p. 170, 1975.
12. W. K. Walsh and B. S. Gupta, J. Coated
Fabrics, Vol. 7, No. 4, p. 253, 1978.
13. W. K. Walsh and B. S. Gupta, J. Coated
Fabrics, Vol. 8, No. 1, p. 30, 1978.
14. W. K. Walsh, and B. S. Gupta, J. Coated
Fabrics, in press.
15. W. K. Walsh, et al., High Energy Radiation
for Textiles: Assessment of a New
Technology," Final Report to NSF, Grant
GI 43105, October 1977.
16. C. Bluestein, Sources and Resources, Vol.
11, No. 6, p. 27, 1978.
17. C. Evans, Sources and Resources, Vol. 11,
No. 26, 1978.
169
-------
CONTROL OF AIRBORNE PARTICULATES IN TEXTILE PLANTS
Mansour H. Mohamed, Arthur C. Bullerwell*
Abstract
It is agreed that high concentrations of dust
level in the working environment in textile plants
represents a health hazard. There is a number of
ways by which the dust concentration could be
maintained below a certain level for every process.
New machinery is designed with more emphasis
on decreasing dust levels around them than was
done in the past With existing machines, engi-
neering controls have been developed, which in
most cases, are successful in meeting present
Federal regulations.
Developments in processing machinery, in ex-
haust ventilation systems, and in filtration
technology are reviewed from the standpoint of
dust generation and removal.
INTRODUCTION
Knowledge of respiratory problems of textile
workers has been traced back as early as 1713 in
connection with flax and hemp mills (ref. 1).
References to problems associated with cotton
dust go back to 1863 (ref. 2). The respiratory
disease resulting from the exposure to cotton
dust, which occurs in a certain percentage of the
work force, is called "byssinosis." Since cotton
dust regulations are the most serious particulate
control problems facing the textile industry at the
present time, this paper will deal only with this
subject.
From 1964 to 1971, a total dust concentration
level of 1 mg/m3 was established as the ceiling
limit of exposure by the American Conference of
Governmental Industrial Hygienists (ACGIH). A
dust concentration level of 1 mg/m3 became the
Federal standard under the Occupational Safety
and Health Acts of 1970 and 1972. In 1974, the Na-
tional Institute for Occupational Safety and
Health (NIOSH) recommended that the lint-free
dust concentration be set at the lowest level feasi-
ble, but not to exceed 200 uglm3 of vertical
*School of Textiles, N.C. State University, Raleigh, NC.
elutriator cotton dust for all segments of the cot-
ton industry.
METHODS OF REDUCING DUST LEVELS
Use of Cleaner Cotton
This means changes in harvesting and gin-
ning practices, leading to a reduction in trash
content of the cotton supplied to the textile in-
dustry. Almost all of the U.S. cotton crop is ma-
chine-harvested by two methods: machine-
stripped cotton, which is harvested by strip-
ping the entire cotton plant, and machine-
picked cotton, which is harvested by spindle-
type pickers. Machine-stripped cotton normal-
ly contains more trash than machine-picked
cotton, although reports indicate that true
comparisons are almost impossible and that
levels did not correlate well with harvesting
methods (refs. 4,5).
Ginning is performed to remove the seed
from the cotton lint. This is the most logical
step for trash removal. The amount of trash
removed depends on the ginning conditions
and the number of lint cleaners used. Ginners
are always concerned with the quality of the
lint cotton and fear damaging the cotton fibers
by excessive cleaning. Although fiber damage
was found to be greater with increased clean-
ing in ginning, no one investigated if increased
cleaning could be balanced by a reduction in
the number of cleaners in opening lines, which
would reduce fiber damage in opening.
The amount of trash removed at the gin is
controlled by the amount of seed cotton clean-
ing, the number of drying stages, and the
amount of gin lint cleaning. Two stages of dry-
ing at 250° F decreased the dust level in a
model card room by approximately 19 percent
(ref. 4). Increasing the number of saw-type lint
cleaners from zero to two was reported to
cause a maximum decrease in dust levels in a
model card room of 37 percent in one paper
(ref. 4) and 52 percent (ref. 6) in another. Cocke
and Hatcher (ref. 5) processed conventionally
cleaned cotton through a Shirley Analyzer and
171
-------
obtained a 43 percent reduction in dust levels.
Blending Cotton With Manmade Fibers
Although this would not be suitable for
every end-use application, it is a fact that the
amount of cotton dust generated is directly
proportional to the rate of cotton flow through
the plant. Therefore, a plant using a 50/50
polyester/cotton blend, for example, will gen-
erate 50 percent less dust compared to a plant
having to produce the same amount of 100 per-
cent cotton yarn.
Washing
This is known to remove the majority of
trash and dust in the cotton. This process is a
common practice in producing cotton for medi-
cal applications. However, to produce yarns
from washed cotton requires extensive re-
search to eliminate the problem of neps and the
increased friction due to the removal of the
natural waxes that exist on the surface of cot-
ton fibers.
Cotton has also been treated with steam (ref.
7). This process has been shown to decrease the
irritant level in the model card room, but at
present there are still questions about whether
this is an effective means to decrease byssino-
sis throughout a cotton textile plant. The
standard did not establish a higher allowable
dust level for steamed cotton; however, mills
processing washed cottons are exempted from
the standard.
Use of Additives
It has been demonstrated (refs. 8,9) that the
application of oils in the range of 0.4 to 1.5 per-
cent reduced the dust level by up to 94 percent.
This seems to be a very successful modification
and is worthy of consideration. It is certain
that the use of additives will require readjust-
ments in further processes, but reports seem
to indicate only minor problems.
Development of New Machinery
Old textile machinery was designed without
serious consideration to the dust problems. As
a matter of fact, some machines that were de-
signed for the purpose of separating trash
from cotton did not have any dust-collecting
devices. Over the last decade or two, various
machines were developed with much more em-
phasis on dust removal and collection. Ex-
amples of new machinery that help reduce dust
exposure are automated opening lines, chute-
feeding of cards, and open-end spinning. How-
ever, a large number of the machines in use to-
day are of the old design, and the most effec-
tive method of controlling dust levels is the use
of ventilation systems.
Use of Ventilation and Filtration Systems
Although there are general rules for the use
of ventilation and filtration systems, various
machines require different approaches to cov-
er the major points of dust emission. Varia-
tions in building size and structure present
problems. For example, in the case of looms the
best ventilation will be achieved when the air
vents are placed under the looms, which is not
always possible. Every mill and every process
has to be dealt with as a separate case.
Let carding be taken as a specific example,
since the card is considered the greatest
source of dust generation in a textile plant.
There are many ways by which dust is cap-
tured from the major emission points. Suction
plenums are added at the doffer, cylinder,
lickerin, and other points. In other systems,
orifices are added over the web takeoff area,
the coiler, and the doffer. In many systems, a
transparent cover is used over the web and
lickerin areas. Although some design Differ-
ences exist between systems, the control
points are basically the same. With high-speed
cards, undercard suction is necessary.
In most systems, the dust-laden air is passed
through two stages of filtration: primary and
secondary. The primary filter (sometimes re-
ferred to as the condenser) is a fine-mesh
screen which removes the lint in the form of a
mat that helps in filtering some of the dust.
The waste is doffed into a hopper and is re-
moved either manually or by means of a suc-
tion system to the waste room. The secondary
filter is generally a rotating drum covered with
a filter medium. Cleaning the dust off the filter
is continuously done by means of stationary
suction nozzles, and the dust is collected at a
172
-------
condenser. It is recommended that the clean
air be returned to the plant air conditioning
system, which acts as a third stage of filtration.
Some of the air could be discharged to the out-
side and replaced by fresh air. The percentage
of makeup air has a very important economical
significance. In other systems, the third stage
is an electrostatic precipitator followed by an
air washer through which the air passes before
it is returned to the plant atmosphere.
In other processes, the same control systems
are used. The major differences between proc-
esses are the collection points and the rate and
pattern of airflow. Variations in dust levels and
size distribution between processes have been
reported. The median particle size tends to
decrease in going from carding to warping, and
the particle size distribution tends to be
skewed to the small end. The decrease in me-
dian particle size is probably due to the
micronizing of trash by the cleaning operation.
The median particle size was reported to in-
crease in slashing and weaving areas. The
chemical nature of the dust in slashing and
weaving was found to be different, due to the
chemicals added on the yarn during slashing.
Based on these facts, the cotton dust standard
treats the slashing and weaving areas differ-
ently from other textile operations, as men-
tioned earlier.
In drawing, dust control consists of collect-
ing the air from the drafting roll cleaner
system and filtering it rather than blowing the
dust back into the room. The use of hoods over
the creel area is advised. In combing, the col-
lection of dust at the aspirator for further
filtration, control of dust from the drafting
zones and suction slot at the coiler are recom-
mended. Hoods might also be installed over the
combing area. In spinning, the broken ends col-
lection system is used to collect dust, which is
then removed in a central filtering system. In
most of these operations, it is preferred to
move the air downward through the machines.
Due to the large volumes of air handled, air
velocities at the filters are normally high.
Economically, it is advantageous to use high
velocities. However, filtration efficiency is
reduced and the pressure drop across the filter
is increased with the increase in air velocity.
This presents a definite need for research on
the development of high-efficiency economical
filters for fine cotton dust. The use of elec-
trostatic precipitators for the control of fine
cotton dust also requires further study. Wash-
ing the collecting plates might be turned into a
water treatment problem.
An experimental wet-wall electroinertial air
cleaner has been developed by the researchers
at the USDA Southern Regional Research Cen-
ter in New Orleans, Louisiana (ref. 10). Its
design consists of a vertical tube equipped
with coaxial charging wire and an air entrance
tangential to the periphery of the tube so as to
impart rotational flow. Dust entering the tube
is charged by corona from the wire and is at-
tracted to the tube wall by combination of the
electrical centrifugal'forces. A film of water
flows down the inner wall of the tube, washing
the dust down to a discharge sump. Dust re-
moval efficiency of up to 99.0 percent with a
10.16-cm (4-in.) diameter, 1.22-m (4-ft) long unit
operating at 5.66 m3/min (200 ft3/min) air flow
rate was reported.
CONCLUSIONS
1. Cotton dust concentrations could be re-
duced by many ways, such as improving
ginning conditions, use of additives, and
blending with synthetic fibers.
2. Control technology using local and
general ventilation systems with two or
three stages of filtration seems to be the
most successful method of lowering dust
concentrations to levels near or below the
levels required by the cotton dust stand-
ard. However, economic pressures on the
textile industry will be increased due to
the capital and running costs of such
equipment. In addition, every plant and
every process has to be dealt with on a
case-by-case basis, which makes it neces-
sary to custom design the air-cleaning
equipment.
3. The textile machinery industry has
placed more emphasis on dust cleaning in
the design of new machinery. Thus, the
severity of the dust problem in textile
plants is expected to reduce considerably
in the future.
4. There is still a need for the development
of high-efficiency, economical filters to
capture fine cotton dust at high air veloci-
173
-------
ties without excessive pressure drop.
5. The use of additives and washing of cot-
ton may require significant changes in
processing and have yet to be proven eco-
nomically feasible on a mill level.
REFERENCES
1. B. Ramazzini, "Diseases of Workers," De
Morbis Artificum of 1713, W. C. Wright,
translator, Hafner Publishing Co., New
York, 1964.
2. National Institute for Occupational Safe-
ty and Health, "Occupational Exposure to
Cotton Dust," NIOSH Report, U.S. De-
partment of Health, Education, and Wel-
fare, Washington, DC, 1974.
3. "Occupational Exposure to Cotton Dust,"
Federal Register, Vol. 43, Part III, June
23, 1978.
4. J. B. Cocke, R. A. Wesley, I. W. Kirt, and
R. V. Baker, Effects of Harvesting and
Ginning Practices on Mill Dust Levels,
The Cotton Gin and Oil Mill Press, Decem-
ber 18,1976.
5. J. B. Cocke and J. P. Hatcher, "Dust
Levels in Experimental Card Room as In-
fluenced by Cotton Variety and by Har-
vesting and Ginning Methods," Transac-
tion of the ASAE, 1975.
6. S. P. Hersh, R. E. Fornes, and E. V.
Caruolo, "Respirable Dust Levels Devel-
oped While Processing Cotton in a Model
Card Room," Cotton Dust, Proceedings of
ACGIH Symposium, Atlanta, GA, Novem-
ber 12-13, 1974.
7. C. Michaels and D. Whisnant, "Manufac-
turing Aspects of Steaming Raw Cotton
To Minimize Toxic Effects Upon the
Human Respiratory System," Engineer-
ing Control of Cotton Dust, proceedings
of a conference at Clemson University,
July 26, 1972.
8. J. B. Cocke, H. H. Perkins, and C. K.
Bragg, "Use of Additives to Reduce Cot-
ton Dust Levels," America's TextilesR/B,
March 1977.
9. R. E. Fornes, R. D. Gilbert, and S. P.
Hersh, "Collection and Characterization
of Cotton Dust in a Model Card Room,"
Beltwide Production Research Confer-
ence, Atlanta, GA, January 10,1977.
10. D. P. Thibodeaux, A. Banil Jr., and R. B.
Reif, "A Wet-Wall Electroinertial Precipi-
tator: A Highly Efficient Air Cleaner for
Cotton Dust," IAS 76 Annual.
174
-------
Session IV: SOLID WASTE POLLUTION CONTROL TECHNOLOGY
P. Aarne Vesilind, Session Chairman
175
-------
TEXTILE SLUDGE CHARACTERIZATION
Robert G. Shaver,* Deborah K. Guinant
Abstract
Analysis of sludge samples from six textile
industry segments has shown that metals and
organic wastewater pollutants concentrate in
the solid phase of the sludge. The six industry
segments sampled were wool scouring, wool
fabric dyeing and finishing, woven fabric dye-
ing and finishing, knit fabric dyeing and
finishing, carpet dyeing and finishing, and yarn
and stock dyeing and finishing. Approximately
2,000 of the more than 5,000 textile operations
in the United States fall into these segments.
The amount of sludge generated in 1974 on a
dry/weight basis was approximately 38,430
metric tons. This accounts for about 12 percent
of the total waste generated by the textile in-
dustry in 1974. Of the total sludge generated
approximately 0.5 percent was determined to
be heavy metals content with less than 0.1 per-
cent chlorinated organic material. The heavy
metals found included arsenic, cadmium, chro-
mium, cobalt, copper, lead, mercury, nickel,
and zinc. The concentrations of these heavy
metals found in the various textile industry
segments range from 3,600 ppm in yarn and
stock dyeing and finishing to 21,000 ppm in
wool fabric dyeing and finishing. The bulk of
these metals are washed or rinsed from fabric
into a mill's wastewater treatment system
from such operations as scouring incoming
greige goods, dyeing and printing cloth, yarn,
and carpet, and applying various finishes. The
chlorinated organics were not specifically iden-
tified, but such compounds as trichloro-
benzene, polyvinyl chloride, perchloroethyl-
ene, and others are commonly used as solvents,
dye carriers, and finishing agents. The total
chlorinated organics found in the sludges
analyzed range in value from 0.11 to 64.7ppm.
*Versar, Inc.
t
Association of American Railroads.
INTRODUCTION
The information presented in this paper is
the result of a study Versar performed for the
U.S. Environmental Protection Agency under
Contract No. 68-01-3178 to assess the industrial
hazardous waste practices of the textiles in-
dustry. The program began in April 1975 and a
final report was delivered in June 1976.
TEXTILE SLUDGE GENERATION
Processes
For the purposes of this study, the textiles
industry was classified into the seven cate-
gories that had been used earlier in an EPA ef-
fluent guidelines study for this industry. It was
agreed upon by the American Textiles Manu-
facturers Institute (ATMI), Versar, and EPA
that the best method for categorizing the in-
dustry was by process, not product. These pro-
cess categories and the Standard Industrial
Classification (SIC) codes are given in Table 1.
Six of these categories, excluding greige
goods, have wet processes such as scouring,
dyeing, printing, and finishing, which generate
wastewaters that are treated onsite or sent to
publicly owned treatment works. These treat-
ment systems generate various amounts of
sludge containing a variety of heavy metals
and other chemicals related to textile process-
ing.
Of the more than 5,000 textile plants in the
United States, 2,007 were identified as
performing wet operations and thereby gen-
erating wastewater. The remaining plants fall
into Category C and perform only such dry
operations as knitting and weaving. Greige
goods facilities also have yarn preparation
operations which include finishing yarn destin-
ed for knitting in an oil or wax emulsion and
slashing warp yarns used in weaving with car-
boxymethyl cellulose (CMC), polyvinyl alcohol
(PVA), or starch. But these operations gen-
erate no wastewater.
In the study, Versar visited a total of 80
177
-------
plants representing approximately 16 percent
of the total production in the seven textiles in-
dustry segments. The major effort was focused
in Categories D, Woven Fabric Dyeing and
Finishing, and E, Knit Fabric Dyeing and
Finishing, since over 69 percent of textile
plants are engaged in either woven or knit dye-
ing and finishing. Wastewater treatment
sludges from 14 selected plants were sampled
and analyzed to determine, in some measure,
the composition of this land-destined waste.
Each plant was composite-sampled once per
week for 4 weeks, taking samples from the
clarifier underflow. The sludges were analyzed
for heavy metals and chlorinated organics. The
total number of samples analyzed was 112.
Details on the sampling techniques, analytical
methods, and pollutant parameters measured
are presented in a later section of this report.
Location and Industry Breakdown by
Process
Table 1 shows the distribution of textile
plants by industry category. A breakdown of
the number of plants performing dyeing and
finishing operations by EPA region is given in
Table 2. As shown in the breakdown in Table 2,
91 percent of the plants that potentially
generate wastewater treatment sludges were
located in EPA Regions I, II, III, and IV with 50
percent located in Region IV alone. Also, the
textiles industry as a whole is heavily concen-
trated in the eastern part of the United States.
The 2,007 plants with wastewater loads are
approximately 37 percent of the total number
of textile operations in the United States. The
remaining 63 percent are dry operations.
Because the industry was categorized on a
process-oriented basis rather than by product,
Table 3 shows the distribution of the plants by
State and by process. Table 4 summarizes the
process distribution.
Over 62 percent of the textile plants are
engaged in greige goods manufacture. Of the
remaining plants that perform dyeing and/or
finishing operations, over 69 percent are in the
woven fabric and knit fabric dyeing and fin-
ishing categories.
Amount of Sludge Generated
The amount of sludge generated in each of
the six wet processing categories was deter-
mined for "typical" plants in each industry
category. The characteristics of a typical plant
are determined by industry survey in consul-
tation with industry personnel. In all cases the
sludges arose from wastewater treatment.
In looking at the sludge disposal problem, an
interesting phenomenon was noted. Sludge dis-
posal was found to be typically necessary in on-
ly two industry categories, Wool Scouring, arid
Woven Fabric Dyeing and Finishing. All plants
visited in the Wool Scouring category found it
necessary to dispose of sludge. About 40 per-
cent of the plants (56 percent of the category
production) engaged in Woven Fabric Dyeing
and Finishing category found it necessary to
dispose of excess sludge. Most facilities in the
other industry categories generate sludge at a
low enough rate to preclude the need to dis-
pose of any excess. Therefore, a distinction
was made between sludge requiring disposal
(wasted) versus that left accumulating in treat-
ment ponds (retained) (see Table 5).
Figure 1 summarizes the estimated State-by-
State distribution of total wastewater treat-
ment sludges. This illustration shows that 78
percent of sludges are generated in EPA
Regions I, II, III, and IV.
It was estimated that 488 plants of the 2,007
identified, which perform wet operations, have
wastewater treatment facilities (24 percent).
ATMI estimates that these 488 plants com-
prise 65 percent of the total production of the
2,007 plants. The remaining 1,519 plants use
municipal sewage treatment, which transfers
the problem of sludge disposal and its asso-
ciated costs to the municipalities. About 7 per-
cent of the plants visited contributed 50 to 95
percent of the total wastewater load to the
municipal sewage treatment system. Several
of the municipal treatment facilities were con-
structed with plant funds and turned over to
the municipality for operation and mainte-
nance.
Ninety-six percent of the 488 plants with
treatment systems are generating sludge in
unlined aeration basins with potential for per-
colation to underlying strata. Four percent
were found to have concrete lined aeration or
settling basins. The lined ponds were found to
be most prevalent in Category A, Wool Scour-
ing, where 67 percent of the plants use them.
Category E, Knit Fabric Dyeing and Finishing,
178
-------
was the only other industry category found to
be using lined ponds to the extent of 12 percent
of the plants in the category.
TEXTILE SLUDGE
CHARACTERIZATION
Analysis of sludge samples from the six tex-
tile industry segments found metals and
organic wastewater pollutants to be concen-
trated in the solid phase of the sludge. The
heavy metals found include arsenic, cadmium,
chromium, cobalt, copper, lead, mercury,
nickel, and zinc. The bulk of these metals are
washed or rinsed from fabric into a mill's
wastewater treatment system from operations
such as scouring incoming greige goods; dyeing
and printing cloth, yarn, and carpet; and apply-
ing various finishes. Total chlorinated organics
were determined but not further specifically
identified.
Sampling Techniques and Analytical
Methods
A representative number of dyeing and
finishing mills in each category except C,
Greige Goods, was sampled and the plants
were chosen to be representative in terms of
processing and fiber types used. Four-hour
composite wastewater treatment sludge
samples were collected from clarifier under-
flows returning to aeration ponds. Two
separate sets of samples were taken, one for
heavy metal and one for chlorinated organic
analyses. Sampling was repeated at each plant
once a week for 4 consecutive weeks.
Samples earmarked for heavy metal analysis
were acidified with nitric acid to a pH of 2 for
preservation. The samples for organics analy-
sis were not acidified and were handled very
carefully to avoid contamination. Total solids
content was measured on unpreserved
samples.
Sludge solids contents were below 2 percent
and atomic absorption was appropriate for
determining total trace metals. The samples
for atomic absorption analysis were digested
by heat and acid, the residue redissolved in
acid, and sample volume adjusted with distilled
water. The samples were then filtered to
remove insoluble materials and aspirated
directly into the flame source. The absorbance
was recorded and the corresponding metal con-
centration determined.
Alternate methods were used to determine
mercury and arsenic. Mercury concentration
was measured by the flameless atomic absorp-
tion method using a quartz lamp as a radiation
source to vaporize the mercury. Arsenic was
determined either by atomic absorption or by
the silver diethyldithiocarbamate method.
To verify that most of the heavy metals were
associated with the solid phase, the sludge
samples — nonacidified samples — were cen-
trifuged and the solids dried and digested.
Then the atomic absorption method, as men-
tioned above, was used.
The suspended solids measurement was
made by filtering, drying, and weighing a
known volume of sample.
Chlorinated organics were determined from
the nonacidified sludge samples by gas
chromatography. The samples were prepared
by adjusting the pH to between 6.5 and 7.5, ex-
tracting with methylene chloride in hexane,
concentrating the extract on an evaporating
hot water bath, and injecting the extract into a
gas chromatograph. The concentration of
chlorinated organics was then calculated by in-
tegrating the output absorption trace.
Results of Sample Analysis
Heavy metals are generally found in
municipal and industrial wastewater treat-
ment sludges, however, we found no published
data concerning the composition of textile
plant wastewater treatment sludge. During
this program selected samples were analyzed
for metal content to determine the distribution
of metals between the liquid and solid phases
of the sludge. This was accomplished by com-
paring the results of atomic absorption analy-
sis of digested samples of the entire sludge
(solid and liquid phases) and of the suspended
solids separated from the liquid phase in non-
preserved samples. Table 6 shows an example
comparison of these metal concentrations and
the metal concentration calculated from them
for the liquid phase (dissolved). Most of the
heavy metal content of the sludge is in the
solid phase and not dissolved. The exceptions
are lead, manganese, and zinc. Most of the lead
179
-------
and large parts of the manganese and zinc
were apparently dissolved.
Chlorinated organics determinations for all
samples were made on both the liquid and sus-
pended solids phases. Table 7 compares the
results from five woven fabric dyeing and fin-
ishing plants. Similar differentials between li-
quid and suspended solid phases of chlorinated
hydrocarbon concentrations were found from
other textile industry categories. Note that the
concentrations in the liquid phase are all below
the drinking water standard of 0.7 ppm for
total organics. Like the heavy metals, most
chlorinated hydrocarbons were found in the
suspended solids phase.
Table 8 summarizes the results obtained on
the five woven fabric dyeing and finishing
plants sampled for all measured parameters.
On a dry sludge basis, concentrations of
various heavy metals and chlorinated organics
are far above drinking water limits.
Table 9 summarizes the average total heavy
metal and chlorinated organics found from all
industry categories.
TREATMENT AND DISPOSAL PRACTICES
Treatment
A flowsheet of the prevalent wastewater
treatment technology is shown in Figure 2.
Plants in all textile industry categories were
found to retain sludge in unlined aeration
ponds, and Categories A and D were found to
dispose of excess sludge in general purpose
landfills, dumps, and/or landspreading on farm
land. A few plants in Categories A and E were
found to retain sludge in lined aeration ponds
and one plant in Category E disposed of
dewatered sludge in an approved landfill.
Disposal
There are various disposal and waste
management practices being used for sludges
by the textile industry. Discussions of five of
these practices follow.
Lagoon Storage or Retention—
Wastewater treatment sludges are being
stored and retained in the wastewater treat-
ment systems, either in disposal ponds or in
the bottom of ponds or lagoons that are used
for aeration and activated sludge treatment.
As this sludge builds up, it will eventually
reach the level where other storage or disposal
will be necessary.
Land Dumping —
Land dumping of wastewater sludges is
practiced by some textile plants. Some in-
stances of onsite dumping were found.
Land Spreading —
Wastewater treatment sludges have some
fertilizer value and in some instances are being
sprayed or spread on land for this value.
Various techniques are used, often utilizing
farm-type equipment or irrigation-type spray
units.
General Purpose Landfills—
Some plants send their wasted sludge to
general purpose landfills. These landfills are
characterized by their acceptance of a wide
variety of wastes, including garbage and other
organic materials, and usually by the absence
of special containment, monitoring, and
leachate treatment provisions.
Only one instance was found where an ap:
proved landfill site was used for disposal of
dewatered sludge. This was in EPA Region III.
Wet Oxidation of Sludge—
We found one plant that had installed a wet
oxidation process (Zimpro Process) for treat-
ment of wastewater sludge. This uses liquid
phase oxidation of wastes at high tempera-
tures and pressures and has the operational
flexibility of achieving either partial or nearly
complete oxidation, as needed. Wet oxidation
can reduce the amount of sludge and make the
remaining sludge easier to dewater. It also con-
verts much of the nonbiodegradable organic
material (measured as COD) to either oxidized
innocuous components or biologically de-
gradable material which can be recycled to the
plant wastewater treatment system for des-
truction. This operation had not been in use
since 1972 because the small amount of sludge
generated in the wastewater treatment
system made the equipment uneconomical to
operate.
All plants with wastewater treatment
systems retained some sludge in their treat-
ment ponds. Disposal of sludge is necessary
180
-------
only when the solids level in the wastewater
treatment system builds to a point where the
efficiency of the biological system would be af-
fected. In Table 10 is a summary of plants
visited that dispose of sludge. Local off site
disposal of sludge is usually within 16 km (10
miles) of the plant. Some of the plants that used
onsite disposal of wasted sludge preferred to
dispose of the sludges offsite in sanitary land-
fills and were seeking permission from local
authorities to do so.
CURRENT REGULATORY
STATUS
Wastewater treatment sludges from several
textile industry categories:
SIC 2231 - Wool Fabric Dyeing and
Finishing,
SIC 2261 - Woven Fabric Dyeing and
Finishing,
SIC 2250 - Knit Fabric Dyeing and
Finishing,
SIC 2269 - Yarn and Stock Dyeing and
Finishing,
SIC 2279 - Carpet Dyeing and Finishing,
and
SIC 2299 - Wool Scouring
have been incorporated by EPA into the
September 12,1978, draft of regulations under
Section 3001 of RCRA (PL 94-580), which deals
with criteria ami listing of hazardous wastes.
This draft is currently being circulated within
the agency and to interested parties for com-
ments.
181
-------
DEWATERING OF WASTE ACTIVATED SLUDGE
Eugene J. Donovan, Jr.*
Abstract
Sludge dewatering is often an integral step
in the ultimate disposal of solids removed from
wastewater treatment systems. Waste acti-
vated sludge from biological treatment con-
tains 95 to 98percent moisture and is relatively
difficult to dewater. There are a number of
alternative dewatering processes available,
each of which may have several variations.
Processes may include land spreading for solar
evaporation, sand bed drying with or without
chemicals, and various mechanical systems
which usually incorporate chemical addition.
These include vacuum filters, centrifuges,
pressure filters, and the relatively new belt
filter press devices.
The selection of an appropriate and cost-
effective dewatering system involves the eval-
uation of the process performance on the speci-
fic sludge, consideration of the ultimate dis-
posal selected for the sludge, and various site
specific factors. This paper reviews the applica-
tion of the various sludge dewatering processes
to waste activated sludges and outlines a step-
wise approach to the evaluation and com-
parison of alternatives. Typical results and
costs for dewatering waste activated sludge
are presented herein.
INTRODUCTION
Dewatering of waste activated sludge is an
element in the handling and ultimate disposal
of the excess sludge produced by biological
treatment of textile plant wastewater ef-
fluents. Dewatering processes remove water
from a sludge to the extent that its physical
form is changed essentially from that of a fluid
to that of a damp solid. Major dewatering
methods include vacuum filters, centrifuges,
pressure filters, belt filter presses, sand-
drying beds, lagoons, and land spreading. The
*Senior Engineer, Hydroscience, Inc., Westwood, NJ.
decision to incorporate a dewatering step into
a sludge handling scheme is based on the re-
quirements of the ultimate sludge disposal op-
tion selected, which is usually site specific.
Conversely, the ultimate disposal method
selected is a factor in the evaluation and selec-
tion of an appropriate sludge dewatering proc-
ess. Figure 1 shows the relationship of de-
watering alternatives in various sludge hand-
ling schemes. The application of the various
sludge dewatering processes to waste ac-
tivated sludges is reviewed, and an approach
to the evaluation and comparison of alter-
natives is presented, including data on perfor-
mance, sizing, and costs of the alternatives.
WASTE ACTIVATED SLUDGE
CHARACTERISTICS
Excess activated sludge is generated by the
conversion of a fraction of the organic matter
in wastewater to biological cells. The quantity
of excess sludge depends on the biological sys-
tem design (flood to microorganism ratio), raw
waste characteristics, and effluent solids. For
textile wastes, the solids production generally
ranges from 20 to 50 percent of the BOD re-
moved, plus the influent solids incorporated in-
to the sludge, less effluent solids. Typical ex-
tended aeration systems utilized in the textile
industry produce a relatively stable waste ac-
tivated sludge, whereas sludge from high-rate
systems may require further stabilization,
through digestion, prior to dewatering or dis-
posal.
Waste activated sludge from secondary clar-
ifiers is a liquid containing 98 to 99.5 percent
water. Gravity or dissolved air flotation may
be incorporated into sludge-handling systems
prior to dewatering or disposal.
The water contained in the sludge is distri-
buted into three general classifications, free or
between-cell water, bound or capillary water,
and intercellular water. Free water can largely
be removed by thickening or by drainage.
Dewatering devices are utilized to increase the
rate and amount of free water removal, and to
183
-------
Figure 1. Schematic diagram of sludge handling alternatives.
form a cake. Various organic and inorganic
chemicals are utilized to neutralize the bonding
forces by coagulating sludge particles and free-
ing the bound water for subsequent removal by
dewatering devices. Intercellular water, which
requires rupture of the cell wall to free the
water, is not removed to any great degree by
dewatering systems.
Removal of free water from activated sludge
produces a sludge of 4 to 6 percent solids, while
removal of the bound water by chemical condi-
tioning and dewatering results in sludge cakes
containing from 12 to 25 percent solids. Fur-
ther reduction in the moisture content involves
evaporation and heat drying.
MECHANICAL DEWATERING
ALTERNATIVES
Figure 2 is a schematic diagram of the major
elements of the plate and frame pressure filter
and of the rotary drum vacuum filter dewater-
ing systems. These systems generally require
substantial quantities of chemicals, typically
ferric chloride coagulant and lime, for dewater-
ing of waste activated sludges. Polymers have
been utilized in place of the coagulants, and
other materials such as pulverized ash or ce-
ment dust have been used for body feed. A pre-
coat may be applied to the filter media prior to
sludge dewatering to prevent blinding, to in-
crease solids capture, and to allow easier cake
discharge. In the case of the filter press, the
precoat would be applied in a thin cover on the
filter media prior to each filtration cycle, while
for a precoat vacuum filter, the precoat would
be applied as a relatively thick layer which
would be gradually shaved off along with de-
posited sludge cake as the rotating drum
moved past a discharge knife.
The filter press is operated on a batch basis.
Each cycle consists of load, pressure, drain,
and cake discharge steps, and ranges from 2 to
184
-------
FERRIC
CHLORIDE
STORAGE
TANK
Figure 2. Flow diagram — vacuum filter and pressure filter.
4 hours. The rotating vacuum filter operates
continuously. Sludge is picked up on the re-
volving drum from the vat, passes through a
drying zone, and is discharged from the media.
The pressure applied to the sludge in the
pressure filter by the feed pump is limited by
practical design considerations. However,
since increased pressure can reduce cycle time
and increase cake solids, pressures of about 7
atm for activated sludge dewatering have been
found to be satisfactory. The vacuum filter util-
izes a vacuum to draw water from the sludge
and is therefore limited by the vacuum which
can be applied, usually in the range of 35 to 60
cm of mercury.
Historically, the vacuum filter has been
widely used for dewatering wastewater treat-
ment sludges. The use of the pressure filter,
however, has received more attention in the
last several years due to potential economic ad-
vantages in situations where drier and less
sludge cake is advantageous for subsequent
sludge processing, such as incineration or long
hauls to limited landfill sites.
Bench-scale testing is conducted to deter-
mine chemical requirements. Optimum chemi-
cal dosages can be determined by the Buchner
funnel test or the capillary suction timer test,
in which sludge is dosed with varying amounts
of chemicals and the relative ease with which
free water can be removed is measured.
Small cells or pilot units are available to
determine the need for precoat, the volume of
sludge after dewatering, and cycle time re-
quired to size the pressure filter. A standard
leaf test in which a vacuum is applied on 93 cm2
(0.1 ft2) of filter media in a sludge sample for
various time periods can provide information
required for media selection and sizing a
vacuum filter. Appropriate scaleup factors are
applied to leaf test data when used for sizing
full-scale systems. Pilot scale testing to op-
timize design parameters and to obtain vari-
ability data on a sludge may, at times, be advis-
185
-------
able, particularly for large installations, or
where comparisons of specific types of equip-
ment or various dewatering systems are
desired. In general, for smaller sludge quan-
tities or in cases where only small quantities of
sludge from pilot treatability studies are avail-
able, design parameters are developed from
bench tests and from experience with similar
sludges. Figure 3 presents typical pilot test
data on dewatering activated sludges with the
pressure filter and vacuum filter. Final cake
solids are based on total solids including
chemicals.
Figure 4 is a schematic diagram showing the
major elements for dewatering sludges by cen-
trifuge and the relatively new belt filter press.
Both systems utilize polymers to agglomerate
sludge particles into large masses and to free
bound water.
There are various types of centrifuges, but
the solid bowl countercurrent centrifuge,
which is in effect a high-rate settling unit, is
the type most commonly selected for sludge de-
watering. This type of centrifuge consists of a
bowl and conveyor that rotate at slightly dif-
ferent speeds. A conical section in the bowl at
one end forms the dewatering beach on which
the conveyor screw pushes sludge solids to out-
let ports. The centrate discharges over a weir
that controls the depth of sludge in the cen-
trifuge. Polymer is usually injected into the in-
let line to the centrifuge. Major variables that
affect the performance of the centrifuge in-
clude the pond depth, rotation speed, differen-
tial scroll speed, feed rate, and sludge condi-
tioning. Full-scale designs should be based on
scaleup from pilot tests, where possible, since
bench tests do not provide the needed data on
throughput and cake solids performance.
The belt filter press has been quite exten-
sively used abroad and is rapidly growing in
popularity in the United States. In Europe
there are over 1,000 installations, while in the
United States there are predictions that up to
400 or more units will be installed in 1978-79.
There are some 8 to 10 manufacturers of vari-
ous types of belt filter devices, each of which
has different patented features and configura-
tions. Typically, a belt press will consist of
three distinct operations. The first is a simple
gravity drain section in which the bound water
freed by chemical addition is removed. The sec-
ond operation involves the gradual pressing of
the sludge between two converging parallel
moving belts under tension. The third section
is a shear zone, where the sludge between the
two belts is passed over a series of rollers,
causing the formed sludge cake to shear, fur-
ther releasing more bound water. Some equip-
ment incorporates an optional fourth stage,
which exerts higher pressures on the sludge to
maximize water removal. On activated
sludges, however, this stage only increases the
cake solids by a few percent at most. Bench-
scale tests can be used to determine approx-
imate polymer dosage requirements, and a
hand squeeze test of sludge between two
pieces of the press belts can provide an approx-
imation of the cake solids. The sizing of equip-
ment depends to an extent on the dewatering
rate in the gravity drain section. Laboratory-
scale belt press tests are conducted to develop
design sizing. There are no general standards
for measuring the actual throughput of sludge.
Manufacturers have been developing data
through pilot testing, however, the measure-
ment of actual throughput and performance of
a machine is best determined by pilot testing.
Various manufacturers have mobile units,
which can be used to develop a design sizing.
Figure 5 presents results of several pilot
dewatering tests using the centrifuge and belt
filter press. The lower polymer requirements
and higher cake solids generally achieved by
the belt filter press compared to the centrifuge
can be noted. Generally, it was found that dos-
ages above the minimum polymer dosage re-
quired to agglomerate the sludge particles
resulted in a slight increase in cake solids. The
range of results with different belt press
machines and different sludges indicate the
desirability of pilot tests for final system
scaleup and sizing.
LAND DEWATERING OF SLUDGES
Dewatering of sludges on land includes the
use of open or closed sand-drying beds, lagoon-
ing, and application to the land. Sludge that is
dewatered by a land method should be relative-
ly stable to achieve satisfactory results and to
minimize odors. For leachate and runoff drain-
age from land disposal sites, consideration of
potential effect on surface and groundwaters is
required. The necessity for under drainage,
normal for sand-drying beds but optional with
186
-------
100
RANGE OF TEST
RESULTS
PRESSURE FILTER
NOTES:
FEED SOLIDS- 2.0 -2. 5%
CHEMICALS - Fe C I3+ LIME - 30 -70% WT
FINAL CAKE SOLIDS- 28-34%
FINAL PRESSURE - 100 PSI
80 120
CYCLE TIME (WIN)
160
200
-LBS/HR/SF
01 C
Z 2
Q
Q
_l
tr I
UJ
b
U_
0.5
0.
\VACUUM FILTER
LEGEND:
%FEED
PLANT SOLIDS
• - A 2.2
• • - B 1.9
\ 0 - C 2.1
^P (11 FeCly+UMf
>.
\
\°
\
MINIMUM FORM \
TIME FOR CAKE >
npi rA^r ju 1 /A". . . - *.
I till
0.2 0.5 1.0 2 3
CHEM'"CAKE
25% 8%
42% 16%
20% 1 2%
I 1
4 5
FORM TIME (MIN)
Figure 3. Vacuum filter and pressure filter WAS dewatering test results.
187
-------
DILUTION
BELT FILTER PRESS
WASTE ACTIVATED SLUDGE
CAKE
DISCHARGE
CENTRATE
Figure 4. Flow diagram - belt filter press and centrifuge.
lagooning and land applications, requires eval-
uation. Many parameters affect the design and
selection of a land dewatering system. A prin-
cipal mechanism is drying through evapora-
tion; therefore local climatic conditions such as
amount and rate of precipitation, air
temperatures, humidity, wind velocity, and
percentage of sunshine are particularly rele-
vant to a design.
Sand bed designs generally consist of en-
closed beds with an underdrain' system,
overlaid with several layers of gravel and a
layer of sand. Sludge is applied to the sand
layer at depths consistent with design condi-
tions for the drying time and sludge cake con-
sistency desired. The design of the beds is in-
fluenced by the length of time required for the
sludge to dewater to a consistency at which it
can be removed from the bed and hauled to
final disposal. Use of covered beds will greatly
reduce the time required for dewatering, and
conditioning chemicals such as polymers will
also accelerate dewatering. Drying periods
range from 1 to 2 weeks with chemicals, and
from 2 to 6 weeks or more without chemicals,
depending on weather conditions. Sizing of
sand beds is based on local experience and
general design factors. Testing on test sand
beds can be conducted to obtain data on drying
time, cake consistency, and underdrain quality,
and factored with weather conditions to devel-
op design sizing.
A sludge drying lagoon is a low-cost, simple
system for sludge dewatering that has been
commonly used in the United States. Sludge
lagoons are generally not recommended since
they are particularly susceptible to odors and
nuisance conditions. They can stabilize and
consolidate sludge with very little attention
though, if properly sized and operated. Typical-
ly sludge is deposited in a lagoon at depths of 2
to 4 feet. Water drains into the soil or is evapo-
rated. Under drains may be used where drain-
age to groundwater is not acceptable, or where
permeable soils are not available. Two lagoons
are usually provided, one being filled, one be-
188
-------
14
h-
Z
UJ
o
or
UJ
a.
CO
o
CO
Ul
<
o
SOLID BOWL CENTRIFUGE
DIA.
NOTES:
FEED SOLIDS - 1.5 - 2.0 %
POLYMER 15-25 LBS/TON
I
10 15 20 25 30
FEED RATE GPM
35
40
45
18
IT
UJ
W f
O
JS'°
UJ
< 8
o
BELT FILTER PRESS
LEGEND:
UNIT
• -A
• — B
o-c
A—D
WIDTH %FEED RECOVERY
M SOLIDS AVG %
I
2
I
0.5
1.5
2.5
1.0
2.7
95
96
93
87
10 20 30 40 50 60
FEED RATE GPM
70
80
90
Figure 5. Belt filter press and centrifuge WAS pilot plant dewatering test results.
189
-------
ing allowed to dry out. Following drying
periods of 3 to 5 months or more, the de-
watered sludge is removed with a front-end
loader and hauled to disposal. Sizing drying
lagoons can be based on local experience and
small-scale tests, factoring in climatic condi-
tions. Where land is available, relatively con-
servative design factors can be used in the ab-
sence of local experience or testing. Area
should be reserved for expansion of the lagoon
system in the future, should actual results
necessitate more volume.
Sludge spreading on land is receiving some
attention in the last few years and has been
done at a number of locations. The effects of
various factors and the development of loading
parameters are in the developmental stage.
Groundwater contamination and the effect of
the sludge constituents on a crop grown on the
land are factors to be evaluated in a specific ap-
plication. Where land can be dedicated to
sludge disposal only, high rates of sludge
spreading may be feasible. The sludge is incor-
porated into the soil and there is limited to no
vegetation. Rainfall runoff and leachate must
be adequately considered in a design of a land
application system. Discing deposited sludge
can prevent surface blinding and minimize
odors; the need for and advantages of this pro-
cedure should be evaluated. Sizing of land ap-
plication systems depends upon a variety of
considerations, including sludge nutrients, salt
and heavy metal content, type crop, if any, to
be grown and harvested, soil type, land charac-
teristics, groundwater, and aesthetics. Each
site requires an evaluation of the various fac-
tors. Where feasible, particularly for large in-
stallations, test plots should be operated and
monitored for at least 6 months to a year.
EVALUATION AND COMPARISON OF
ALTERNATIVES
The selection of an appropriate and cost-
effective sludge dewatering alternative re-
quires the consideration and evaluation of the
various factors relating to a specific case. An
overall approach is suggested here, although
many of the factors involved are quite specific
to a given situation and their importance may
be weighted differently from plant to plant.
Table 1 outlines the procedures for develop-
ing and selecting an overall sludge dewatering
plan. The initial task is to establish the sludge
quantity and characteristics, either from data
from an existing system or using the design
basis for a yet-to-be-built treatment plant. Con-
sideration of future loads, present sludge
handling practices, if any, average and max-
imum sludge production rate, and the effect of
storage on sludge handling rates are included.
Potential sludge reduction alternatives such as
inplant recovery of waste constituents, or a
digestion step prior to dewatering would be
considered at this point to identify feasibility
on a preliminary basis.
An evaluation of potential ultimate disposal
alternatives will provide information on the re-
quired characteristics of the solids to be dis-
posed, and consequently, dewatering require-
ments. Availability, cost, and feasibility of the
use of land sites for possible land application or
landfill should be determined. The possible
need to reduce sludge or landfill should be de-
termined. The possible need to reduce sludge
quantities to a minimum by incineration due to
unavailability of landfills or high hauling costs
requires an early evaluation.
Bench-scale tests are conducted on potential
dewatering alternatives at this point, and
using these results along with general typical
information, as is shown in Tables 2 and 3, a
preliminary screening and consideration of
sludge dewatering alternatives can be made.
Available data from similar installations can be
evaluated to aid in developing sizing and costs,
and to evaluate extended operational perform-
ance and operating cost factors, particularly
for the newer and/or different types of specific
equipment available. State and local regula-
tions on land disposal and landfills are also con-
sidered in the screening process.
At this point a decision usually can be made,
and one or two sludge dewatering alternatives
selected for detailed evaluation. Although vari-
ous activated sludges generally have the same
characteristics, they also have differences
which can effect the performance of dewater-
ing systems. It is therefore advisable to per-
form pilot-scale tests, when possible, to
demonstrate performance and provide confir-
mation of preliminary bench-test sizing param-
eters. In some cases, side-by-side comparisons
of different sludge dewatering systems and dif-
ferent equipment may be desirable for com-
parison of performance. Extensive testing,
190
-------
TABLE 1. PROCEDURES FOR DEVELOPING AND SELECTING A SLUDGE DEWATERING PLAN
1. Define sludge quantity and characteristics
A. Average - maximum quantities
B. Effect of sludge storage on rates
C. Identify potential sludge reduction alternatives
2. Define ultimate disposal alternatives and sludge solids requirements
3. Bench scale evaluation of alternatives compatible with probable ultimate disposal
4. Preliminary sizing and capital and operating costs evaluation of alternatives
5. Pilot scale evaluations of indicated system(s) for final design parameters and performance confirmation
6. Final design sizing of dewatering system(s)
7. Final development of overall sludge handling plan(s)
Financial strategies
8. Selection of overall plan and required equipment
Capital cost estimate
Operating cost estimate
however, is not always necessary, particularly
for smaller installations where the effect of
conservative design parameters would have
only slight effect on the final cost of the
dewatering system. An example of this would
be the case where the minimum standard size
equipment capacity exceeds the design sludge
quantity.
With process design and performance param-
eters established, the selected dewatering
alternatives are sized, based on sludge quan-
tities, full-size equipment capabilities, and
operational requirements. At this point, with
characteristics of the sludge cake known, fur-
ther evaluation of the ultimate disposal alter-
natives available may be made. Additionally,
the cost of the dewatering step can be com-
pared with identified potential sludge reduc-
tion alternatives, and a decision on reduction
steps may be made. Following finalization, the
proposed total sludge handling plans are re-
viewed and capital and operating cost esti-
mates prepared. The type of dewatering equip-
ment can be selected and quotes for specific
equipment obtained for an t conomic analysis of
the sludge handling plans. Factored into these
cost analyses would be the company's financial
strategies for dealing with capital and
operating expenditures for pollution control.
Finally, with all relevant factors considered,
the overall sludge handling scheme and the
dewatering alternative can be selected.
Table 2 presents performance data for four
types of mechanical dewatering equipment.
The results shown are based on literature data
and actual test data developed in several
studies conducted by Hydroscience. The belt
filter press arid centrifuge use polymers for
sludge conditioning. Although polymers gen-
erally constitute a major part of the sludge
dewatering costs, they are very effective, and
eliminate the bulk chemical storage, handling,
and feed systems generally associated with the
filtration systems. The cake solids and recov-
ery are comparable for the belt filter press and
vacuum filter; somewhat lower cake solids and
capture are noted for the centrifuge, while the
pressure filter produces the highest cake
solids, and achieves virtually complete solids
capture. Table 2 shows the remaining moisture
content based on sludge solids, a significant
consideration particularly if the sludge is to be
incinerated. It also gives the approximate
volume of cake solids per ton of sludge solids,
which must be figured into sizing landfills and
determining haulage costs.
191
-------
TABLE 2. PERFORMANCE ESTIMATES FOR MECHANICAL DEWATEBING OF WASTE ACTIVATED SLUDGE
Dewatering
device
Belt filter
press
Solid bowl
centrifuge
Vacuum
filtration
Recessed
plate
Pressure
filter
Chemical
Polymer
Polymer
FeCIa
Ca(OH)2
FeCl9
4v
Ca(OH)2
Precoat
Dosages*
2.5-10
0-10
60-100
200400
60-200
200-500
25
Approx.t
chemical Unit sizing
cost criteria
12-45 115-180 kg
per hour
per meter of
belt width
045 135-225 kg
per hour per
56 cm diameter
x 1.7 m long
machine
20-35 5 to 15 kg
per hour/m^
28-66 0.5 to 1.0 kg
per hour/m?
25 mm thick cake
2 to 4 hours
cycle
Cake
solids
cone.
(percent)
10-20
6-10
12-18
24-30
Kg water*
Approx. remaining
% solids per kg dry
recovery sludge solids
90 6.7
50-90 11.5
93 7.8
99 4.1
Cubic meters
of cake
per ton dry
sludge solids
6.6
12.4
8.9
5.3
* Dosage = kg of chemical/metric ton of dry sludge solids.
* Cost = dollars/metric ton of dry sludge solids.
* Based on average cake solids
Feds - $165/ton, polymer - $4.40/kg
Lime - $55/ton, precoat -$220/ton.
Table 3 presents performance and general
design parameters for land alternatives for
sludge dewatering. These techniques typically
do not require chemical conditioning, and rely
primarily on draining and evaporation for de-
watering. The importance of general climatic
conditions for design, and local weather condi-
tions for operation is obvious. Normally use of
sand beds requires return of the drainage to
the treatment facility. With drying lagoons and
high-rate land applications, however, evapora-
tion rates, soil conditions, and leachate quality
will determine the necessity of returning
supernatant and leachate to treatment. Leach-
ate is usually not returned from low-rate land
application in which nutrients are incorporated
into a harvested crop and other constituents
such as Ijeavy metals are adsorbed and re-
tained in the soil. The 2-step operation asso-
ciated with drying beds and lagoons—sludge
application and sludge removal after dewater-
ing—is indicated by the range of time between
applications. Since with land application the
sludge is incorporated into the soil, that proc-
ess is also considered an ultimate disposal
scheme. Table 3 shows approximate cake solids
for each alternative and relative land require-
ments, based on an average loading for 1 ton
per day waste sludge solids, are presented.
Table 4 presents a cost comparison for
mechanical dewatering systems. The equip-
ment selected was based on the unit sizing
criteria, Table 2, and standard-sized manufac-
turers units. Equipment costs, which include
chemical storage and feed systems, represent
an approximate average estimate for the type
of equipment indicated, and are probably
within ±20 percent of current quoted costs.
192
-------
TABLE 3. PERFORMANCE AND GENERAL DESIGN CRITERIA FOR
LAND ALTERNATIVES FOR SLUDGE DEWATERING
System
Open sand
drying beds
Covered sand
drying beds
Drying
lagoons
Land
applications
Sizing criteria*
1. No chemicals - 50 to 100 kg/m2/yr, 4 to 6 weeks between
applications
2. Polymers - up to 400 kg/m2/yr, 1 to 3 weeks between
applications - drainage returned
No chemicals - 100 to 200 kg/m2/yr, 1 to 3 weeks between
applications - drainage returned
30-50 kg/mfyyr, 0.5 m to 1.25 m depth
3 to 5 months drying time, supernatent returned
lining may be required, 2 or more lagoons used
1. Crop-nutrient uptake-1.25 to 5.0 kg/m2/yr
2. Noncropland, drainage returned, up to 25 kg/m2/yr
storage during wet or cold seasons may be required
Land
required
acres/ton/d*
1.2
0.4
0.7
14.3
^
30
4.5
Cake sludge solids
40 to 50 percent
40 to 50 percent
40 to 50 percent
25 to 40 percent
Sludge
incorporated
into soil
*Dry sludge solids basis.
t Average sizing used.
TABLE 4. COMPARISON OF MECHANICAL DEWATERING ALTERNATIVES (ESTIMATED CAPITAL AND
OPERATING COSTS BASED ON 1,100 METRIC TONS/YR WASTE ACTIVATED SLUDGE SOLIDS
CONTINUOUS OPERATION - 300 DAYS/YR)
Equipment
One 1-meter belt
filter press
One-22"0x 66"
solid bowl centrifuge
One - 300 ft2
vacuum filter
One - 90 ft^ capacity
pressure filter
Equipment
costs „
($xifj3)*
100
130
135
200
Total
installation
costt($x103)
220
290
320
450
Operating
power
(hp)
10
50
50
20
0/M.
labor
(h/yr)
365
2,000
4,200
7,000
0/M costs §
($/ton dry
(sludge solids)
20-55
35-80
65-90
100-145
* Equipment costs approximate - vary with manufacture; includes major chem feed equipment.
* Total installation - 2 to 2.5 times equipment costs; building and utilities not included.
Labor requirements from reference 7.
§ 0/M costs include power at 3^/kWh, labor at $10/h, spare parts at 2 percent equipment cost, and chemical costs in Table 1.
Amortization of equipment and final disposal of sludge not included in 0/M costs.)
193
-------
TABLE 5. COMPARISON OF LAND DEWATERING ALTERNATIVES (ESTIMATED)
CAPITAL AND OPERATING COSTS BASED ON 1.100 METRIC TONS/YR WASTE
ACTIVATED SLUDGE SOLIDS)
System
Sand drying beds
4 acres open
Drying lagoons
1 million gallons, 7.5 acres
Land application/
crop -110 acres
High loading 20 acres1'
Capital*
costs
($x103)
420
120
300
175
0/M
labor
(h/yr)
6,600
150
3,100
2,100
0/M Costs*
($/ton/dry sludge solids)
45-80
2-6
3045
20-35
* Land costs do not include pumping and transmission to and recycle from land site.
t Includes underdrains.
* 0/M for pumping and transmission costs not included - variable site specific. (Costs for
ultimate disposal of sludge removed from drying beds or drying lagoons not included.)
Total installation cost is based on factoring the
equipment costs and includes engineering, con-
tractor costs, foundations, controls, and asso-
ciated piping and electrical work. The belt fil-
ter press, solid bowl centrifuge, and vacuum
filter capital costs are comparable, whereas
the pressure filter costs are somewhat higher.
The chemical feed systems required for the
bulk chemicals used with vacuum filtration and
pressure filtration affect capital costs.
Although advances in polymer chemistry and
use of polymers for conditioning in some cases
have replaced these bulk chemicals, the nature
of activated sludge has generally required the
bulk chemicals for filter application.
Chemical costs are the major part of the
operating and maintenance costs; however,
power and labor are significant particularly for
the filtration processes. It may be noted that
labor costs for pressure filters may vary con-
siderably with the type of equipment and
degree of automation. All of these estimates
would require final evaluation in a specific
case. As noted on the table, costs for final
sludge haulage and disposal are not included.
These costs are site specific and can be signifi-
cant to the extent that they will influence the
final selection of the dewatering alternative.
Table 5 compares relative costs for land de-
watering alternatives. Sand-drying beds re-
quire substantially less area, but they are
structured units. Consequently, capital costs
are significantly higher for these than for
earthen drying lagoons with liners or for the
preparation of relatively level, lightly vege-
tated land on which to spread or inject sludge.
The labor requirements reflect the frequency
of removing sludge from the drying beds and
lagoons, and the application of the sludge in the
land application schemes.
It should be noted that land costs, if any, or
special land preparation costs, as well as the
costs for construction and operation of pump-
ing and transmission facilities to transfer
sludge to the dewatering site and the return of
leachate or supernatant are not included, since
these costs are highly site specific and can vary
substantially. They should, however, be includ-
ed in the final cost evaluation of these alter-
natives for a specific case. The costs presented
for these alternatives are somewhat less pre-
cise than those developed for the mechanical
dewatering systems.
CONSIDERATIONS FOR SELECTION OF A
DEWATERING PROCESS FOR EXCESS
ACTIVATED SLUDGE
As previously mentioned, selection of a de-
watering alternative is site specific and to a
194
-------
large extent depends upon the ultimate dis-
posal technique. Nevertheless, certain general-
izations concerning the various techniques can
be made based on the information and data
generated and summarized in this paper.
Land application that includes ultimate dis-
posal appears to be a cost-effective solution to
dewatering and disposal, given two provisions.
Adequate suitable land must be available, and
the various concerns with respect to designing
for reasonable loading rates, i.e. the handling
of runoff and lechate and nonobjectionable
aesthetics, must be adequately addressed.
Lagoon drying is a low-cost dewatering solu-
tion, but is more susceptible to problems with
odors and unsightly conditions than land appli-
cations. Also periodic sludge removal to ulti-
mate disposal is required with this method.
Sand beds, while demanding less land than
the other land alternatives, are high in both
capital cost and labor cost due to frequent
sludge removal. Here, also, the sludge must be
hauled to ultimate disposal.
As with all equipment, the machinery for
mechanical dewatering alternatives is subject
to breakdown, and the need for repairs and
replacement requires consideration. Numerous
manufacturers offer various types of equip-
ment with different features and costs. Selec-
tion of the equipment best applicable to a
specific case requires an evaluation of these
aspects based on mechanical design and oper-
ating experience of the equipment being con-
sidered.
Of the mechanical dewatering systems, the
relatively new belt filter press equipment ap-
pears to be the most cost effective system to
produce a sludge cake suitable for disposal.
The pressure filter is a probable alternative in
the event incineration or long-distance hauling
is required. However, the desirability of charg-
ing an incinerator with the large quantities of
ferric chloride and lime used for conditioning,
and of subsequently disposing of significant
quantities of ash, requires consideration. Use
of the vacuum filter for dewatering activated
sludge appears to be less cost effective than
the belt filter press and therefore would prob-
ably no longer be considered in most cases for
this type application.
Dewatering with centrifuges does not pro-
duce as dry a cake as the other alternatives;
yet centrifugation may be applicable in a
sludge handling plan wherein chemical use is
minimized, and cost benefits are achieved by
producing a thickened, semifluid sludge for
subsequent handling by digestion, heat treat-
ment, or land disposal.
REFERENCES
1. National Council of the Paper Industry for
Air and Stream Improvement, "A Pilot
Plant Study of Mechanical Dewatering
Devices Operated on Waste Activated
Sludge," Technical Bulletin No. 288,
November 1976.
2. U.S. Environmental Protection Agency,
Process Design Manual for Sludge Treat-
ment and Disposal, EPA 625/1-74-006, Oc-
tober 1974.
3. W. E. Garrison, J. F. Stahl, L. Tortorici,
and R. P. Micle, "Pilot Plant Studies of
Waste Activated Sludge Processing," J.
Water Pollution Control Federation, Vol.
50, No. 10, p. 2374, October 1978.
4. G. T. O'Hara, S. K. Raksit, and D. R.
Olsen, "Sludge Dewatering Studies at
Hyperion Treatment Plant," J. Water
Pollution Control Federation, Vol. 50, No.
5, p. 912, May 1978.
5. E. Cole, S. Corr, and J. Albert, "Sludge
Dewatering in Textile Plants," Industrial
Wastes, January/February 1977.
6. National Council of the Paper Industry
for Air and Stream Improvement, "Full
Scale Operational Experience With Filter
Presses for Sludge Dewatering in the
North American Pulp and Paper In-
dustry," Technical Bulletin No. 299.
7. U.S. Environmental Protection Agency,
Operators Manual • Sludge Handling and
Conditioning, EPA 430/9-78002, February
1978.
8. R. V. Villiers and J. B. Farrell, "A Look at
Newer Methods for Dewatering Sewage
Sludges," Civil Engineering ASCE,
December 1977.
9. P. N. Cheremisinoff, "Sludge Handling
and Disposal —A Special Report," Pollu-
tion Engineering, January 1976.
10. J. A. Bell, R. Higgins, and D. G. Mason,
"Dewatering: A New Method Bows,"
Water and Wastes Engineering, April
1977.
195
-------
11. American Society of Civil Engineers, 14.
Wastewater Treatment Plant Design,
ASCE Manual No. 36,1977.
12. U.S. Environmental Protection Agency,
"Proceedings of the Joint Conference on 15.
Recycling Municipal Sludges and Ef-
fluents on Land," 1973.
13. -J. A. Mueller, "Land Disposal of Sewage 16.
Sludge — State of the Art," Hydroscience
Internal Research Report, September
1976.
G. J. Bisjak and A. E. Becher, "Wausau
Solves Dual Problems by Using Filter
Press," Water and Wastes Engineering,
February 1978.
A. E. Dembitz, "Belt Filter Presses: A
New Solution to Dewatering," Water and
Wastes Engineering, February 1978.
U.S. Environmental Protection Agency,
"Costs of Wastewater Treatment by Land
Application," EPA 430/9-75-003, June
1975.
196
-------
RATIONALE FOR SLUDGE MANAGEMENT IN A
WASTEWATER TREATMENT PLANT
P. Aarne Vesilind*
Abstract
Sludge management in wastewater is expen-
sive and can be troublesome. This paper pre-
sents three basic "laws" for sludge manage-
ment which, it is suggested, can reduce costs
and problems with sludge handling. These
three guidelines are:
1. Don't hold sludges,
2. Don't mix sludges, and
3. Don't recirculate sludges.
If these laws are followed, dewatering costs
can be reduced, odor problems alleviated, and
environmentally acceptable ultimate disposal
of the sludge facilitated.
INTRODUCTION
A wastewater treatment plant exists for the
purpose of producing an essentially clear ef-
fluent that can be discharged into the environ-
ment without adverse impact. The design engi-
neer and plant operator therefore are mostly
concerned with the liquid processing within
the plant, and solids management becomes im-
portant only when the plant malfunctions.
Often, these upsets can be avoided by following
a few fairly simple guidelines for sludge man-
agement within a plant. Malfunctions, from the
operator's viewpoint, are either effluent qual-
ity problems forcefully brought to his attention
by State or Federal inspectors, or sludge prob-
lems which usually consist of having too much
of a bad thing and nowhere to put it. The objec-
tive of this paper is to focus on the latter prob-
lem-sludge handling—and to present some
general guidelines, which in the author's opin-
ion will reduce the costs and headaches asso-
ciated with wastewater sludge management.
*Associate Professor of Civil Engineering, Duke Univer-
sity, Durham, NC.
THREE LAWS OF SLUDGE
MANAGEMENT
Stated concisely, these laws of sludge
management are:
1. Don't hold sludges,
2. Don't mix sludges, and
3. Don't recirculate sludges.
As is the case with any generalities, these
are not always applicable. However, they have
been found to be true at enough plants to war-
rant discussion.
Holding Sludge
Holding sludge (saving it, storing it, etc.) is
usually the doing of a plant operator, and the
procedure is often dictated by the vagaries of
the labor force. It should nevertheless be the
aim of an operator not to hold sludge before a
subsequent operation such as dewatering. Al-
though this principle seems to be especially ap-
plicable to raw primary sludges, the holding of
digested sludge (and the resulting cooling) has
been shown to cut the capacity of mechanical
dewatering to one third of the original (ref. 1).
One clear outward sign of impending trouble
is bubbling primary clarifiers, due to holding
raw sludge in the primary clarifiers, and the
subsequent gas formation within the settled
raw sludge. This will decrease the efficiency of
solids removal, as well as place an extra load on
the alkalinity of the anaerobic digesters. Sludge
should be pumped from the primary clarifiers
at sufficiently frequent intervals so as to avoid
septicity.
Mixing Sludge
The second axiom, "Don't mix sludges," is
violated by both design engineers and plant op-
erators. For example, experience seems to in-
dicate that the discharge of excess waste acti-
vated sludge to the head of the plant results in
deterioration of the primary clarifier perform-
ance, significantly increases the volume of
197
-------
sludge pumped, and necessitates more frequent
pumping due to increased septicity problems.
It has been known for some time that when a
raw primary sludge is mixed with waste acti-
vated sludge, the mixture assumes many of the
undesirable characteristics of waste activated
sludge. It is, for example, very difficult to
mechanically dewater a 1:1 mixture of raw pri-
mary and waste activated sludge. The excel-
lent dewaterability of raw primary sludges is
sacrificed when mixed with secondary sludges.
It is, however, necessary to achieve the
highest solids concentrations possible before a
process such as anerobic digestion, since the
governing operational parameter in anaerobic
digestion is solids retention time. One means of
achieving this has been to blend the raw pri-
mary and waste activated sludges in a gravity
thickener (ref. 2). This has at times caused
operational problems such as floating sludge
and odors, mainly due to the oxygen-starved
nature of the sludges. A more reasonable solu-
tion seems to be the gravitational thickening of
primary sludges and the use of floatation thick-
eners for aerobic secondary sludges, with
blending following the separate thickening
operations and then subsequent anaerobic di-
gestion. In plants where this type of scheme
has been instituted, the solids dewatering has
improved markedly and operational problems
have been reduced, (ref. 3)
Another example of improved operation by
not mixing sludges is the scheme of using
anaerobic digestion for raw primary sludge
only, and employing aerobic digestion for the
waste activated sludge. These stabilized
sludges can then be blended and dewatered at
a cost significantly lower than the alternative
of mixing prior to stabilization (ref. 3).
^circulating Sludge
The final rule, "Don't recirculate sludges" is
probably the most difficult of the three laws to
promote, since the recirculation of various sun-
dry process streams to the head of the plant is
almost a knee-jerk reaction for design engi-
neers. These streams can, however, impose a
significant excess solids load on the process.
Consider a 10 mgd plant which has a gravity
thickener for the primary and waste activated
sludges (5,080 kg/day at 4 percent solids and
7,130 kg/day at 1 percent solids, respectively)
and a centrifuge to dewater the anaerobically
digested sludge. The solids recovery of the
thickener and centrifuge is 90 percent and 80
percent respectively, and the supernatant
from the secondary digester has a solids con-
centration of 2 percent with digester influent
and effluent solids of 5 percent and 8 percent
respectively. What fraction of the total solids
load on the treatment plant is produced by the
return flows from the thickener, centrifuge,
and secondary digester?
1. Thickener overflow, with 90 percent solids
capture, is
0.10 (5,080 kg/day + 7,130 kg/day)
= 1,221 kg/day.
2. The total solids flow into the anaerobic di-
gester is thus
0.9 (5,080 kg/day + 7,130 kg/day)
= 10,989 kg/day.
Of this solids load, assume 20 percent is
lost (gas plus dissolved solids) so that the
digester effluent solids (supernatant +
underflow) is 8,791 kg/day, while 2,198
kg/day is lost in the digestion. Mass bal-
ance on the digester in terms of the
sludge solids is
inflow = underflow + supernatant
+ loss in digestion
10,989 kg/day
= x Ib/day + y Ib/day
+ (-2,198 kg/day).
The mass balance in terms of total liquid
is
inflow = underflow + supernatant
10,989 x
~
Solving the two simultaneous equations:
x - solids in underflow
= 2,930 kg/day
y = solids in supernatant
= 3,663 kg/day.
3. The centrifuge operates at 80 percent
solids capture on the underflow, so that
the concentrate contains
198
-------
0.2 (2,930 kg/day) = 586 kg/day.
The total contribution of recycled solids is
thus
1,221 kg/day + 3,663 kg/day
+ 586 kg/day = 5,470 kg/day.
Since the plant influent carries 8,527
kg/day, the recycled solids increase the
total solids load on the plant by
12,060
18,800
X 100 = 64 percent.
In addition to the total magnitude of the ad-
ditional solids load on the plant, the types of
solids recirculated are almost exclusively the
fine, difficult to handle solids. More than one
treatment plant operator has found himself
inundated with these fine solids, at the even-
tual expense in treatment efficiency. The only
solution in such cases is to clean the plant out.
In the design and operation of a plant, there-
fore, one should not recirculate solids if at all
possible. In the example above, the 2 percent
solids supernatant clearly has an impact on
operation, and the practice of supernatant
withdrawal is of highly questionable value.
Many operators prefer to dewater all of the
sludge, and to do this at a high rate of solids
recovery, recognizing that this investment is
highly profitable for future good plant opera-
tion.
CONCLUSION
Efficient and effective operation of waste-
water treatment facilities requires careful con-
sideration of sludge handling practices. It is
suggested that three practical guidelines
which could assist in the management of waste-
water sludges are: don't hold sludges; don't
mix sludges; and don't recirculate sludges.
REFERENCES
1. P. I. Vermehen, "Chemical Removal of Nu-
trient Salts from Plant Effluent," Sixth
Nordic Symposium on Water Research, Co-
penhagen, 1978.
2. W. N. Torpey, "Concentration of Combined
Primary and Activated Sludges in Sepa-
rate Thickening Tanks," Proc. ASCE 80,
No. 443, September 1954.
3 . U.S. Environmental Protection Agency,
Process Design Manual for Sludge Treat-
ment and Disposal, U.S. EPA Technology
Transfer, 625/1-74-006, Washington, DC,
1974.
199
-------
TEXTILE SLUDGE TREATMENT AND DISPOSAL
David H. Bauer,* John P. Woodyard.t Stephen P. Shelton*
Abstract
Increasing quantities of wastewater treat-
ment sludge generated by more stringent
water and air quality standards and pending
regulations to control the disposal of hazard-
ous wastes have focused new emphasis on
possible solutions to the problem of sludge
disposal. Disposal alternatives currently ap-
plicable to the textile industry include lagoons
landfilling, land farming, incineration, and wet
air oxidation. Current indications are that land-
filling and land farming disposal will predomi-
nate in the future, subject to pending regula-
tions. Incineration and wet air oxidation may
be used on special problem wastes, suck as
residuals generated by reverse osmosis treat-
ment of wastewaters.
INTRODUCTION
Recent estimates indicate that approximate-
ly 39,000 metric tons (dry weight) (43,000 tons)
of sludge are currently generated annually by
the textile industry, and this is projected to in-
crease to 167,000 metric tons, (164,000 tons) an-
nually by 1983 (ref. 1). Thus, adequate sludge
disposal is becoming a major concern in the in-
dustry, especially in EPA Regions I and IV
where approximately 65 percent of all textile
industry sludge is generated.
The disposal of sludge, including that of the
textile industry, has received increasing eni-
phasis in recent years, primarily as a result of
increasing sludge generation from water reuse
and pollution control systems. With passage of
the Clean Water Act of 1977, EPA respon-
sibilities for industrial point source effluent
controls have been specifically defined to in-
clude toxic pollutants, conventional pollutants,
and nonconventional pollutants. Control of 65
classes of toxic compounds is incorporated in
*Project Manager, SCS Engineers, Reston, VA.
jVice President, SCS Engineers, Long Beach, CA.
•fAssociate, SCS Engineers, Columbia, SC.
Section 307(b) as best available technology
(BAT) effluent limitations for 21 "heavy" in-
dustry categories, including the textile in-
dustry. Similar revised authority governing
conventional and nonconventional pollutants is
also included. Thus, increased textile sludge
generation is anticipated, although the
magnitude of the increase will depend on the
specific limitations adopted and the extent to
which pretreatment requirements are imposed
on textile plants.
In addition, the Resource Conservation and
Recovery Act (RCRA) of 1976 has expanded
EPA authority governing sludge disposal.
Specifically, Section 3001 (criteria, identifica-
tion, and listing of hazardous wastes) gives
EPA authority to define hazardous wastes.
The "status draft" of Section 3001 regulations
includes a list of processes considered to
generate hazardous waste, including six textile
processes. Thus, EPA regulations authorized
by Subtitle C of RCRA will be applicable to a
significant portion of the sludge generated by
the textile industry.
In an effort to address future sludge disposal
concerns, this paper focuses on treatment and
disposal methods currently in very limited use
but which appear to have promising applica-
tions.
TEXTILE INDUSTRY SLUDGE
CHARACTERISTICS
Before appropriate treatment and disposal
technology can be selected, the sludge charac-
teristics must be determined. First, data on
physical characteristics is needed, such as
specific gravity, solids concentration, settling
properties and particle size and distribution.
Subsequently, chemical properties such as fuel
value, fertilizer value, food value, and elec-
trophoretic properties, as well as specific
chemical characteristics (especially toxins),
should be evaluated. Lastly, biological charac-
teristics such as carcinogenicity, mutagenicity,
and enzymes may need to be evaluated.
It should be noted that characterization of
sludge can be a complex problem with an ex-
201
-------
pensive solution. Thus, the identifying charac-
teristics of interest need to be specified rela-
tive to the methods of treatment or disposal
that initially appear most feasible. For exam-
ple, if a sludge is to be thickened in a gravita-
tional thickener, a characteristic of primary in-
terest would be settling; whereas if the sludge
is to be used as a fertilizer, nutrients and toxic
constituents would be of primary interest.
Textile industry sludge characteristics are
the subject of another paper scheduled for this
symposium, and are therefore discussed only
briefly here. Concentrations of selected consti-
tuents of general interest in dry sludge from
six categories of textile mills are presented in
Table 1. Heavy metal concentrations range
from 2.5 to 2,370 mg/kg on a dry weight basis,
and the total chlorinated organics content is
between approximately 0.1 and 65 mg/kg. Of
the five heavy metals shown, cadmium concen-
trations were consistently the lowest (always
< 17 mg/kg) and zinc concentrations the high-
est (>550 mg/kg) for five of the six categories
(ref. 1).
Sludge produced from treatment of textile
industry wastewater is either retained in a
treatment system, such as a lagoon or pond, or
wasted from the system, as in an activated
sludge treatment unit. The solids content of
wasted sludge ranges from 2 to 10 percent,
depending on the industry category and the
method of disposal used. The solids content of
retained sludge ranges from 1 to 10 percent
(ref. 1).
OVERVIEW OF SLUDGE MANAGEMENT
Industrial sludges are typically disposed in
sanitary and chemical landfills, in lagoons or
ponds, in deep wells, or by incineration (ref. 2).
Some sludges are recycled or stockpiled and
some sludges are disposed by ocean dumping.
In addition, sludges are sometimes disposed by
open dumping. Incidences of air, ground, and
surface water pollution from improper land dis-
posal of industrial wastes have been reported
(ref. 3). Although adequate technology for in-
dustrial sludge management reportedly exists,
this technology is costly (approximately 5 to 20
times as expensive as current unacceptable
practices such as uncontrolled land dumping or
ocean disposal) (ref. 4).
In the textile industry, sludge treatment
may involve dewatering, wet air oxidation, in-
cineration, stabilization, or fixation, but no
treatment prior to disposal is currently typical
of the industry. Sludge disposal in lagoons, and
to a lesser extent landfills, is also typical, al-
Constituent
TABLE 1. CHEMICAL CHARACTERISTICS OF SELECTED
TEXTILE INDUSTRY SLUDGES
Categories/ing/kg of dry sludge §
Cadmium
Chromium
Boron
Zinc
Copper
Total Chlorinated
Organics
A
1.2
19
26
106
16
13
B
.7
267
170
1,130
117
0.11
D
4.4
1,196
36
2,370
652
15.2
E
4.5
33
52
550
410
64.7
F
10
112
110
1,790
211
26.2
G
2.5
31
25
1,505
264
40.1
§ Values presented are averages for each of the following categories:
A - wool scouring
B - wool fabric dyeing and finishing
D - woven fabric
E - knit fabric
F - tufted carpet dyeing and finishing
G - yarn and stock dyeing and finishing
Source: Reference 1.
202
-------
though dumps and land spreading are some-
times used.
Sludge dewatering is currently practiced at
relatively few textile plants, but is anticipated
to be used more as sludge volumes increase
and disposal requirements become more strin-
gent. Wet air oxidation has been used in at
least one textile plant for wastewater sludge
treatment, but has proven uneconomical to
operate due to the small volumes of sludge gen-
erated. Incineration is used to incinerate pro-
cess wastes at a few textile plants, but not
sludge, presumably due to the relatively high
cost (ref. 1).
In the future, the use of incineration and/or
wet air oxidation may increase for the treat-
ment of wastewater treatment residuals. In
particular, residuals produced by reverse
osmosis (RO) treatment of textile wastewaters
present a difficult disposal problem that may
only be adequately solved with these methods
of treatment. Similarly, stabilization and fixa-
tion will likely be used increasingly to enable
sludge management to comply with applicable
regulations.
Textile plant wastewater treatment systems
frequently include aerated lagoons. In many in-
stances, the sludge accumulation rate is slow
enough that sludge wasting is required very in-
frequently. In those plants that routinely
waste sludge, sludge disposal is frequently
handled by the textile plant, either onsite or
offsite. Thus, lagoons and landfills are the most
common disposal sites, with land spreading
and dumping used occasionally.
The methods of sludge disposal used in the
future will likely depend in large part on pend-
ing Federal hazardous waste regulations and
solid waste disposal criteria. It is anticipated
that use of offsite disposal facilities, such as
lined lagoons and landfills with leachate collec-
tion and treatment (if necessary) and land
spreading will become increasingly common.
SLUDGE TREATMENT
Selection of a sludge treatment method ob-
viously depends on the method of disposal.
Currently, lagoons are the most common
disposal method. Thus, sludge treatment in the
textile industry is atypical. However, it is anti-
cipated that landfill disposal will be used in-
creasingly. Stabilization and chemical fixation
are two methods of treatment used in other in-
dustries to facilitate landfill disposal, and
which appear to have promising applications in
the textile industry. The following discussion
is limited to these two techniques.
Stabilization
Raw sludges can be stabilized prior to
ultimate disposal to increase the number of ac-
ceptable disposal options. By far the most com-
mon sludge stabilization technique is anaerobic
digestion. Many variations of this technique
have been used from septic tanks through com-
pletely mixed heated digesters; however,
sludge toxicity can be a problem regardless of
digestor configuration. This problem is often
encountered in industrial waste treatment and
would likely be a problem for a significant
number of textile waste sources.
Aerobic sludge digestion has drifted in and
out of favor over the past 20 years. The process
is generally less affected by toxins than anaer-
obic digestion. However, the energy costs are
high. Because of increased emphasis on energy
conservation and high costs, it is unlikely that
use of aerobic sludge stabilization will become
as common for textile sludges.
Lime addition is another technique which
has long been used for sludge stabilization,
since it can increase sludge pH until most bio-
logical life is destroyed. Quicklime (CaO), also
has the ability to dewater sludge through a
hydration reaction. This has the advantage of
avoiding oxidation-reduction reactors which
can potentially produce hazardous byproducts.
In the South, where many textile operations
are located, an inexpensive lime source, marl,
can often be mined onsite for use in sludge
dewatering.
The primary disadvantage of lime stabiliza-
tion is that the effects are short-lived. It has
been shown by Paulsrud (ref. 5) that lime
sludges are not chemically stable and that it is
impossible to maintain a high pH even with
very high lime dosages. Chemical and/or bio-
logical action eventually seems to bring the pH
down and with that odors return.
Oxidants such as chlorine have also been us-
ed for sludge stabilization; however, the pro-
blems of recent concern regarding chloramines
203
-------
will probably discourage use of chlorine for
future applications.
Lagooning of sludges by industry is a long-
practiced stabilization technique, arid is cer-
tainly the most common method currently used
in the textile industry. This is not to say that
the primary purpose of most sludge lagoons in
the textile industry is stabilization, since most
are designed either as temporary holding
basins or as a means for ultimate sludge dis-
posal. However, some lagoons are designed as
digesters. The requirement for thorough diges-
tion in a lagoon is generally about 3 years
detention time, 1 of the years being required
for resting without sludge addition. It should
be apparent that such lagoons are practical on-
ly in low-density areas where inexpensive land
is available.
Fixation
An increasing number of industries are
favoring dry disposal of their waste sludges. In
addition to the advantages in disposal as a
solid, transportation and handling are also
enhanced.
Many sludges are not readily dewatered
through conventional mechanical means, and
as such may not possess adequate strength for
use as fill material. The physical or chemical
fixation of these materials can provide the
added strength and reduced permeability
desirable for landfill disposal. Physical fixa-
tion, which may only involve the addition of
dry materials to increase the solids content, is
sometimes adequate if the disposal site is rea-
sonably secure and land reclamation is the pri-
mary concern. On the other hand, chemical
fixation gives the sludge both a predictably
high load-bearing strength and a low
permeability. In no case do these sludge treat-
ment processes attempt to remove contami-
nants from the material; they simply serve
either to improve physical properties or
reduce contaminant mobility.
The predominant use of fixation to date has
been in the utility industry. Because textile
plants do not necessarily have the large quan-
tities of fly ash and bottom ash available to mix
with waste sludges, chemical fixation may be
the best available alternative for dry disposal.
A recent study of FGD sludge fixation technol-
ogy identified a total of 16 companies which of-
fer sludge fixation processes or services. These
companies are displayed according to their ex-
perience in Table 2 (ref. 6).
Fixation companies typically offer not only a
fixation process itself, but may also offer
engineering and operating services to their
client. Of the 16 companies listed in the table,
two companies have actively and successfully
marketed their services to utilities for FGD
sludge management. These companies are IU
Conversion Systems, Inc. (IUCS) and the Dravo
Corporation. Both of the processes are pro-
prietary in nature, involving the addition of at
least one material typically foreign to power
plant operations.
IUCS markets a system of FGD sludge fixa-
tion using lime and fly ash. In the IUCS pro-
cess, vacuum-filtered sludge is mixed with fly
ash and lime in a pug mill. Dry fly ash is added
in proportions varying from 50 to 100 percent
of the dry sludge weight. Lime is added at a
rate of 3 to 4 percent. IUCS calls the finished
material Poz-0-Tec.
Disposal of the sludge is accomplished by
landfilling. The mixture is controlled to opti-
mize compaction properties. Pozzolanic reac-
tions between the lime and fly ash result in a
significant increase in load-bearing strength
after the material has cured. Compared to un-
treated sludge, Poz-0-Tec has lower moisture
content (15 to 25 percent), much reduced
permeability (105 to 106 cm/sec) and com-
pressibility, as well as increased bulk density.
As an alternative to landfill disposal, Poz-0-Tec
can be used as a structural construction
material. Possible uses include road bases;
pond, landfill, or reservoir lining; concrete or
asphalt aggregate; and artificial reefs.
Mechanical compaction of Poz-0-Tec is re-
quired to obtain the optimum strength and
minimum permeability. If the compacted
material is disturbed it will undergo a decrease
in density. Since the in situ Poz-0-Tec material
is unsaturated, disturbances will not reslurry
the material.
The Dravo fixation process (patented) in-
volves the addition of Calcilox, a proprietary
material manufactured from blast furnace slag,
to the sludge. Calcilox is a basic, pozzolanic ad-
ditive which improves sludge handling charac-
teristics and disposal properties.
Three treatment/disposal schemes are possi-
ble using Calcilox as a fixation agent: perma-
204
-------
TABLE 2. COMPANIES PROVIDING COMMERCIAL FIXATION OR DISPOSAL SERVICE
Companies
Aerojet Liquid Rocket
Amax Resource Recovery Systems,
Inc.
American Admixtures Company
(Formerly Chicago Flyash)
Anefco Company
Chemfix
Chem-Nuclear Systems, Inc.
Dravo Corporation
Environmental Technology Corp.
IUCS
Marston Associates
Ontario Liquid Waste
Diposal Limited
TUK, Inc.
Sludge Fixation Technology, Inc.
United Nuclear Industries
Wenran Engineering Corporation
Werner Pfleiderer Corporation
Commercial
experience
with other
residues
Extensive
Extensive
No
Extensive
Extensive
Extensive
Limited
Limited
Extensive
No
Extensive
Extensive
None
Extensive
Not known
Extensive
System
specific
for FGD
sludge
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Full-scale
experience
with FGD
sludge
None
None
Will County
(140 Mw)
None
None
None
Bruce
Mansfield
(1,650 Mw)
None
Phillips
(400 Mw)
Elrama
(500 Mw)
Petersburg
(515Mw)
Conesville
(403 Mw)
None
None
None
None
None
None
None
Bench scale
with FGD
sludge
Yes
Yes
Yes
None
Yes
None
Extensive
s
Yes
Extensive
Yes
Yes
None
Yes
Yes
Yes
None
rield tests
with FGD
sludge
None
None
Will County
None
Shawnee, Will County
None
Shawnee, Mohave, Phillips
None
Shawnee, Phillips, Mohave,
Elrama, Four Corners
St. Clair
Mississauga
None
None
None
Non
None
Source: References.
nent ponding; temporary ponding; and direct ,
landfilling. Permanent ponding involves ad-
ding Calcilox and hydrated lime to the sludge
before pumping the mixture to a holding pond.
The hydrated lime is required to adjust the pH
above 10.5. Calicox is added at the rate of 5 to
10 percent of sludge solids. Curing is com-
pleted in approximately 30 days. Supernatant
is then returned to the FGD system.
Where large tracts of land are not available,
temporary ponding may be preferable. After
the 30-day curing period, the stabilized sludge
is excavated from small holding ponds and
trucked to the landfill. More exact control must
be exercised in order to ensure that the sludge
cures properly before excavation.
In cases where a temporary storage site is
not available, direct landfilling can. be ac-
complished immediately following the addition
of dry fly ash, lime, and Calcilox to filtered
sludge solids. Sludge treated in this manner
ca:i be efficiently transported to the landfill
1 jcation, where only 5 to 6 days curing is need-
ed before spreading.
Costs -
Limited operating experience with sludge
fixation has similarly limited the disposal cost
data base. Several major studies have involved
detailed cost analyses of FGD sludge manage-
ment alternatives. Table 3 displays selected
cost estimates for the various processes des-
205
-------
cribed previously (ref. 7). In general, these
figures show that the dry forms of fixation are
typically only 10 to 20 percent more expensive
than full dewatering and landfilling. This simi-
larity is due to the common dewatering, trans-
port, and handling requirements for each proc-
ess. There is a greater differential in cost be-
tween wet disposal of untreated sludge and
fixed sludge, due primarily to the lack of treat-
ment, handling, and disposal requirements for
the untreated material. As expected, the dis-
posal of untreated (thickened only) FGD sludge
is the most economical form of disposal, but
may also be the most environmentally unac-
ceptable (ref. 8). As the state-of-the-art in
sludge fixation technology and full-service con-
tractuals evolves, the cost differentials will
become more distinct. The figures shown in
Table 3 represent the cost ranking for disposal
alternatives at a model power plant, and do not
represent an absolute ranking for all situa-
tions.
The cost of other fixation methods not dis-
cussed here is typically higher, due in part to
the use of additives that are less readily
available. The selection of a fixation process
therefore will be based upon the availability of
ash or other dry waste material and the vol-
ume of waste to be treated, due to certain
economies of scale.
DISPOSAL
Lagoon and landfill disposal of sludge are
common and well documented elsewhere.
Another disposal option that is less well
known, and is therefore the focus of the follow-
ing discussion, is land spreading. Land spread-
TABLE 3. COST ESTIMATES FOR SELECTED FGD SLUDGE DISPOSAL ALTERNATIVES
Revenue requirements
$/Dry ton $/Wet ton
Process
Disposal option
Range
Base
Range
Base
Limestone
Scrubbing
Lime
Scrubbing
Dravo Fixation'
IUCS Fixation*
Untreated11
Dravo Fixation'
IUCS Fixation*
Untreated11
20.41-10.87
20.68-8.23
12.12-5.55
17.77-9.47
23.77-9.46
13.93-6.38
15.32
12.55
8.08
13.34
13.88
8.87
10.21-5.44
12.59-5.02
6.06-2.77
8.89-4.74
14.47-5.77
6.97-3.18
7.66
7.63
4.04
6.67
8.46
4.44
' Base, Dravo
New 500-MW plant (30-yr life); midwest plant location; coal analysis (by wt): 3.5 percent sulfur (dry
basis), 16 percent ash, Flyash removed with S02- Limestone process with 1.5 stoichiometry based on S02
removed. Pond characteristics: clay-lined, 1 mi from scrubber facilities, 50 percent solids settled density,
21.38 ft total depth, 414 acres. Calcilox added at 7 percent of dry solids.
*Base, IUCS
New 500-MW plant (30-yr life); midwest plant locations; coal analysis (by wt): 3.5 percent sulfur (dry
basis), 16 percent ash. Flyash removed with S02- Limestone process with 1.5 stoichiometry based on SO2
removed. Disposal in landfill 1 mi from scrubber facilities (by truck) 60 percent solids cakes. Lime added
at 4 percent of dry solids.
"Base, Untreated
New 500-MW plant (30-yr life); midwest plant location; coal analysis (by wt): 3.5 percent sulfur (dry
basis), 16 percent ash. Flyash removed with S02. Limestone process with 1.5 stoichiometry based on S02
removed. Pond characteristics: clay-lined, 1 mi from scrubber facilities, 50 percent solids settled density,
21.21 ft total depth, 407 acres.
Source: Reference 7.
206
-------
ing has the potential advantages of permitting
utilization of the waste and minimizing the
pretreatment required. The extent to which
these potential advantages can be realized in
the textile industry will depend in large part
on future regulations being developed under
RCRA.
Land spreading is a disposal technique by
which organic wastewater or sludge is mixed
with the surface soil to achieve volume reduc-
tion and degradation (ref. 9). This same dis-
posal method is also referred to as land cul-
tivation, land farming, sludge farming, soil in-
corporation, and other names. The practice in-
volves several steps: (1) application of waste
onto the surface soil; (2) mixing the waste with
the soil to aerate the mass and expose the
waste to soil microorganisms; (3) possibly
adding nutrients or other soil amendments
(i.e., limestone) during site preparation; and
(4) remixing the soil/waste mass periodically to
promote biodegradation. In practice, sludges
are hauled to the disposal site either directly
from the wastewater treatment plant or from
an interim storage lagoon. The sludges are ap-
plied to the land by spraying, spreading, or
subsurface injection. The field is then disced or
plowed by conventional farm cultivation equip-
ment.
Important processes that contribute to
waste volume reduction include microbial deg-
radation; microbiological degradation, includ-
ing chemical and photochemical degradation;
and evaporation and volatilization. There are
ordinarily large numbers of diverse microorga-
nisms in soils, consisting of several groups that
are predominantly aerobic in well-drained
soils. These microorganisms normally adhere
to the surfaces of the soil colloids. The majority
of groups are heterotrophic, deriving energy
from the breakdown of organic substances, and
these are the dominant microorganisms re-
sponsible for waste decomposition. The prin-
cipal groups of organisms (in surface soils) in-
clude bacteria, actinomycetes, fungi, algae, and
protozoa (ref. 10). The bacteria are the most
numerous, the most biochemically active, and
generally the most significant in terms of
waste degradation.
The dependence of microorganism activity
on environmental conditions such as soil pH,
air temperature, soil water content, and nutri-
ents has been well documented (refs. 11,12,13,
14). In general, pH between 5 and 8, warm air
temperatures, moderately well-drained soils,
and adequate nutrient availability favor soil
microorganism activity and waste degradation.
For any set of environmental conditions, how-
ever, waste degradation depends significantly
on waste characteristics.
In recent years, considerable emphasis has
been placed on the biodegradability of organic
substances that are considered potential envi-
ronmental toxins. Studies on the biodegrada-
tion of pesticides in soils illustrate the impor-
tance of waste characteristics. The chemical
structure, nature, and position of substituting
groups of the pesticides affect the extent and
rate of microbial degradation, as demonstrated
by the persistence of various pesticides in soils
(ref. 11). Although some of the toxic organic
substances do persist in soils, microbial degra-
dation will eventually proceed if the sub-
stances are adsorbed in the soil surface for a
sufficient period of time. There is, however,
limited information on environmental factors
affecting degradation of these substances as
well as the microorganisms and enzymatic
reactions involved.
Nonbiological degradation processes also
play an important role in the dissipation of
many organic substances in soils, including
pesticides. In particular, considerable evidence
has been reported on the importance of chemi-
cal hydrolysis and photochemical degradation
(ref. 15). Other reactions, including oxidation-
reduction, are important for certain com-
pounds.
Chemical degradation of waste materials in
soil is a widespread and complex phenomenon.
Various mechanisms of chemical degradation
or transformation have been found or postu-
lated, including oxidation, reduction, hydroly-
sis, isomerization, and polymerization. The
degradation rate depends on whether the reac-
tion occurs in solution or on an absorbent sur-
face, and is usually a function of pH; redox
potential (Eh); surface acidity; and the nature,
concentration, and availability of catalytic
sites (ref. 9). Hydrolysis reactions are impor-
tant steps in the degradation of many organic
compounds, including chloro-triazine herbi-
cides and organophosphate insecticides. These
reactions are catalyzed by clay present in the
soil (ref. 16).
Chemical reactions induced by electromag-
207
-------
netic radiation represent a potentially impor-
tant pathway for alteration and/or degradation
of many organic substances applied to soil.
However, photochemical degradation should
play a minor or negligible role in most land
spreading applications since the waste is incor-
porated into the soil, thus reducing the amount
of light received by the waste. Similarly, photo-
chemical degradation assumes minor impor-
tance when mobile wastes are leached from the
zone of photolytic influence.
Evaporation is a major mechanism of volume
reduction. The rate of evaporation from a wet
soil is controlled by the environmental condi-
tions that control the rate of evaporation from
bulk water and not by the properties of surface
soil. However, frequent mixing of the waste
with soil will increase evaporation loss and
decrease the probability of developing anaero-
bic conditions.
The waste volatilization process is depend-
ent on the vapor pressure of the compound and
its rate of movement away from the evaporat-
ing surface. The magnitude of volatilization
loss depends on the soil moisture content,
chemical and physical properties of the waste
and soil, atmospheric conditions (temperature,
wind velocity, relative humidity, etc.), and ap-
plication method (ref. 17). In land spreading,
mixing the waste with soil should significantly
reduce volatilization loss due to increased ad-
sorption of the chemical by soil organic matter
and clay, and the decreased vapor pressure of
the waste material.
Municipal wastewater treatment sludges
are widely applied to land for treatment and
disposal, although new regulations may limit
this practice for sludges which contain signifi-
cant quantities of environmental toxins.
Although to a lesser extent, land spreading is
also used as a sluge disposal method in several
industries, including food processing, wood
preserving, pulp and paper, Pharmaceuticals,
soap and detergents, organic chemicals, petro-
leum refining, leather tanning, and textiles
(ref. 18).
The extent to which the estimated 39,000
metric tons of sludge generated annually by
the textile industry (1977) is amenable to land
spreading disposal will depend on specific
waste characteristics, applicable regulations to
be promulgated under RCRA, and site-specific
conditions. Thus, estimates of the quantity of
textile sludge destined for land spreading dis-
posal are not currently available.
Textile-Specific Considerations
To date, there has been limited experience
with land spreading of textile sludge. The
physical characteristics of textile industry
sludge are reasonably similar to municipal
sludge, so that equipment used for municipal
sludge should be appropriate for textile
sludge. However, other design and operational
considerations such as waste loading rate, site
selection, site management, and their asso-
ciated environmental impacts may be signifi-
cantly different than those associated with
land spreading of municipal sludge due to dif-
ferent sludge characteristics. In addition, even
sludges that appear to be similar based on ana-
lytical measurements may behave differently
in response to land spreading if they have
significantly different origins. Thus, additional
investigation of textile sludge land spreading
is required to determine the extent of its ap-
plicability.
Some limited work has been done on land
spreading of sludge from textile and related in-
dustries which indicates that the subject is
worth pursuing. For example, both prelimi-
nary greenhouse experiments and 3 years of
full-scale field experience with land spreading
of wastewater treatment sludge from an organ-
ic chemical plant which produces primarily
dyes, dye intermediates and carriers, and
pigments for use in the textile industry in-
dicates that the practice has some beneficial
and no apparent detrimental effects on turf
grass production. It should be noted that the
heavy metal content of the sludge is maintain-
ed at low levels (generally less than 5 ppm wet
weight) through pretreatment of individual
process waste streams before discharge to the
treatment plant. Thus, future use of the dis-
posal site for production of food chain crops
should be unlimited (ref. 18).
In another instance, wastewater treatment
sludge from an integrated denim mill is land
spread on land used for sorghum and corn pro-
duction. Detailed monitoring data from 2 years
of full-scale field operation is scheduled for
release in the spring. Preliminary results in-
dicate that relatively low concentrations of
heavy metals are present in the sludge and
208
-------
that plant uptake of heavy metals is at low
enough levels to protect public health (ref. 19).
Costs
Costs, as well as technical feasibility, are
always an important concern with any disposal
method. Cost data on land spreading of textile
sludge were not available for preparation of
this paper. The unit costs for land spreading of
sludge generated by five other industries are
shown in Figure 1 (ref. 18). Also shown is the
expected range of costs for a hypothetical com-
mercial land cultivation site, estimated on the
basis of a conceptual design. All of the unit
costs presented include expenditures for land,
equipment, labor, interim onsite storage of
waste, and environmental monitoring. Costs
for waste transport to the site are Dt included.
Capital costs for the conceptual Jesign range
from $262,000 to $746,000, for waste delivery
rates from 1,000 to 4,000 dry metric tons (5,350
to 21,370 x 106 gal) per year of sludge at 5 per-
cent solids. Costs for disposal of uncontained
industrial wastes by sanitary landfill methods
generally range from $5.30 to $22.00 per m3
($0.02 to $0.08 per gal), depending on location,
site volume, and other factors. For industrial
waste disposed in drums or other containers,
costs can be substantially higher, ranging up to
$194 per m3 ($0.72 per gal). Thus, land cultiva-
tion cost is comparable with or lower than con-
ventional disposal'methods.
20
j
18
© DETERGENT
1
16
•• \
14
T 0
^ i
CO ]
E '
> in
£ 10 ( "
-J • 1
_j
o ORGANIC CHE^
5 ®
i
A ! 4-
1ICAL
*^ —
2 j ^L.
•
M
®,
IXED WASTE
h- •
"*" +—
!
CONCEPTUAL
DESIGN COSTS
/
© OIL
i 4
1
i
PUL
i
1
i
i
i
,
|
i
i
CP
.p
1
10 20 30 40 50 60 70 80 90 9b
WASTE RECEIVED - (1,000 m /yr)
Figure 1. Land spreading costs.
209
-------
REFERENCES
1. Versar, Inc., Assessment of Industrial
Hazardous Waste Practices — Textiles
Industry, U.S. Environmental Protec-
tion Agency Contract No. 68-01-3178,
Springfield, VA, June 1976.
2. P. W. Powers, How to Dispose of Toxic
Substances and Industrial Wastes,
Noyes Data Corporation, Park Ridge,
NJ, no date.
3. E. C. Lazar, "Summary of Damage In-
cidents From Improper Land Disposal,"
Proceedings of the National Conference
on Management and Disposal of
Residues From the Treatment of In-
dustrial Wastewaters, Washington, DC,
pp. 253-257, February 1975.
4. M. P. Lehman, "Industrial Waste
Disposal Overview," Proceedings of the
National Conference on Management
and Disposal of Residue From the Treat-
ment of Industrial Wastewaters,
Washington, DC, pp. 7-12, February
1S.75.
5. B. Paulsrud, unpublished data, Norwe-
gian Institute for Water Research, Oslo,
Norway.
6. Michael Baker, Jr., Inc., "State-of-the-
Art of FGD Sludge Fixation," prepared
for Electric Power Research Institute,
FP-671, Research Project 786-1, Final
Report, January 1978.
7. J. W. Barrier, H. J. Faucett, and L. J.
Henson, "Comparative Economics of
FGD Sludge Disposal," for presentation
at the 71st Annual Meeting of the Air
Pollution Control Association, Houston,
Texas, June 25-30, 1978.
8. Dallas E. Weaver, Curtis J. Schmidt, and
John P. Woodyard, "Data Base for Stan-
dards/Regulations Development for
Land Disposal of Flue Gas Cleaning
Sludges," Municipal Environmental
Research Laboratory, Office of Research
and Development, U.S. Environmental
Protection Agency, EPA-600/7-77-118,
December 1977.
9. C. R. Phillips and J. Nathwani, "Soil-
Waste Interaction: A State-of-the-Art
Review," Solid Waste Management
Report EPS 3-EC-76-14, Environment
Canada, October 1976.
10. F. E. Broadbent, "Organics," Pro-
ceedings of the Joint Conference on
Recycling Municipal Sludges and Ef-
fluents on Land, Champaign, Illinois, pp.
97-101, July 9-13,1973.
11. M. Alexander, Introduction to Soil
Microbiology, John Wiley and Sons,
New York, NY, 1961.
12. M. Alexander, "Biodegradation: Pro-
blems of Molecularrecalcitrance and
Microbial Fallibility," Adv. Applied
Microbiol, Vol. 7, pp. 35-80, 1965.
13. J. P. Martin and D. D. Focht, "Biological
Properties of Soils," Soils for Manage-
ment of Organic Wastes and Waste
Waters, American Society of Agronomy,
Madison, Wisconsin, pp. 115-169,1977.
14. R. H. Miller, "The Soil as a Biological
Filter," Proceedings on Conference on
Recycling Treated Municipal Waste-
water through Forest and Cropland,
EPA 660/2-74003, U.S. Environmental
Protection Agency, 1974.
15. D. E. Armstrong and J. G. Konrad, "Non-
biological Degradation of Pesticides,"
Pesticides in Soil and Water, W. D.
Guzeni, ed., Soil Science Society of
America, Inc., Madison, Wisconsin, pp.
123-131, 1974.
16. J. G. Konrad et al., "Soil Degradation of
Diazinon, a Phosphorothioate Insecti-
cide," Agron, Vol. 59, pp. 591-594,1967.
17. W. F. Spencer and M. J. Cliath,
"Vaporization of Chemicals," Environ-
mental Dynamics of Pesticides, R. Haque
and V. H. Freed, eds., Plenum Publishing
Co., New York, NY, pp. 61-78,1975.
18. SCS Engineers, "Land Cultivation of In-
dustrial Wastes and Municipal Solid
Waste: State-of-the-Art Study," EPA-
600/2-78-140a & b, U.S. Environmental
Protection Agency, Cincinnati, Ohio,
April 1978.
19. Dr. Larry King, Dept. of Soil Science,
North Carolina State University,
Raleigh, North Carolina, personal com-
munication to David Bauer, SCS
Engineers, October 5,1978.
210
-------
Session V: ASSESSMENT METHODOLOGY
Dale A. Denny, Session Chairman
211
-------
LEVEL 1 MEASUREMENT PROCEDURES FOR
EFFLUENT CHARACTERIZATION
R. G. Merrill, Jr., L. D. Johnson, J. A. Dorsey*
Abstract
The EPA/IERL-RTP phased approach and
its applications to environmental assessment
of textile wastewaters are discussed in this
presentation. In the first phase, Level 1,
methods have been developed and applied to a
variety of industrial and energy sources. The
procedures are multimedia and accordingly
have application to liquid effluent streams as
well as air and solid waste streams. The Level 1
procedures have been applied to the effluent
streams from 15 textile plants. The results of
the chemical and biological analyses are dis-
cussed in light of their successful use in rank-
ing the plants for further biological and chem-
ical analysis and for evaluation of best avail-
able control technology (BAT). The use of se-
lected Level 1 biological tests for control tech-
nology evaluation is also outlined.
INTRODUCTION
New Environmental Protection Agency
(EPA) regulations and guidelines, as mandated
by Congress and the Clean Air Act, place in-
creasingly greater demands on EPA's research
programs. Not only are EPA's research labora-
tories given the mission of evaluating pollution
control devices, but they must also develop ef-
fective methods of providing the Agency and
the public with a thorough understanding of
the nature and magnitude of current and fu-
ture environmental assessment problems. The
comprehensive assessment of environmental
insult posed by stationary sources is one of the
major missions of the Industrial Environmen-
tal Research Laboratory (IERL) at Research
Triangle Park, NC. An outgrowth of this mis-
sion is the IERL-RTP Environmental Assess-
ment Program.
*U.S. Environmental Protection Agency, Industrial En-
vironmental Research Laboratory, Research Triangle
The Environmental Assessment (EA) Pro-
gram is designed to pursue four major goals in
the environmental assessment of a stationary
source. They are:
1. A systematic evaluation of the physical,
chemical, and biological characteristics of
all streams associated with a process.
2. Predictions of the probable effects of
those streams on the environment.
3. Prioritization ofthose streams relative to
their individual hazard potential.
4. Identification of any necessary pollution
control technology programs.
Outputs of the EA program are primarily in-
tended to define the need for and design of con-
trol technology, and to provide a data base suf-
ficient to meet the planning needs of the regu-
latory and standard-setting parts of the Agen-
cy.
In order to focus available resources (both
manpower and dollars) on emissions and fulfill
the goals of the EA program, a 3-phased ap-
proach was developed after consideration of
and comparison to more traditional approa "hes
(refs. 1,2).
The phased approach (refs. 3,4,5,6,7) utilizes
three closely linked levels of sampling and
analysis. Level 1 is a screening or survey
phase, Level 2 is directed toward confirmation
of Level 1 information based on more detailed
analysis, and Level 3 involves monitoring of
selected pollutants determined by the results
of Levels 1 and 2. Level 3 may also include the
study of chronic sublethal effects constituting
partial or complete assessment of risk.
Each level includes consideration of the com-
m" ;surate sampling, chemical, and biological
n.ethods in order to meet the goals of the as-
sessment. Level 1 methods have been devel-
oped and applied to a variety of industrial and
energy sources including textile wastewaters.
This experience with Level 1 procedures has
provided firsthand knowledge of the limita-
tions of the techniques and aided early revision
of areas found to be deficient. The Level 2
sampling and chemical analysis approach is
213
-------
nearly complete in a broad conceptual sense.
Close monitoring of field and laboratory im-
plementation is now required to verify that
Level 2 meets the required goals.
The conceptual approach to Level 2 biolog-
ical testing is well underway and planning for
Level 3 approaches to sampling and chemical
analysis is also currently underway.
LEVEL 1 OVERVIEW
The multimedia Level 1 sampling, sample
preparation, and analytical procedures (shown
in Figure 1) have been developed to address
liquid, gaseous, and solid waste streams (refs.
8,9,10). Briefly, the procedures provide screen-
ing results within broad general limits for the
chemical composition and rate of emission of in-
organic species (as elements) and organic spe-
cies (as compound classes). The Level 1 proce-
dures also provide an indication of direct bio-
logical response in the health, aquatic ecolog-
ical, and terrestrial ecological areas. The
overall goal for accuracy at Level 1 is to report
the results within a factor of 3. In simple terms,
this means that a reported value between 30
and 300 is acceptable when the true measure-
ment is 100. The methods currently in use are
capable of much greater accuracy and preci-
sion individually, but the overall relaxation of
PHYSICAL
SOLIDS MORPHOLOGY
FIELD
SAMPLES
INORGANIC
ELEMENTAL ANALYSIS
(SPARK SOURCE MASS
AND ATOMIC ABSORPTION
SPECTROMETRY)
ORGANIC
LIQUID CHROMATOGRAPHY
INFRARED AND
LOW RESOLUTION
MASS SPECTROMETRY
REPORT
INPUT TO
IMPACT
ANALYSIS
BIOASSAY
in vitro CYTOTOXICITY
BACTERIAL MUTAGENICITY
ECOLOGICAL TESTING
in vivo TOXICITY
Figure 1. Flow chart of Level 1 scheme.
214
-------
traditional analytical limits allows a significant
cost savings. These savings are typified by the
use of grab sampling instead of integrated sam-
pling, where possible, and by the use of gravi-
metric results, where they are applicable,
rather than the use of higher priced analytical
techniques. When necessary, more costly sam-
pling or analytical methods are used (ref. 9) as
in the case of sampling particulate-laden stack
gas streams in order to collect representative
samples.
For aqueous fugitive effluents such as water
and leachate runoff, samples may be acquired
by means of simple plug collectors; however,
the sampling plan may be complex and may in-
volve careful consideration of procedures for
the interpretation of the results.
Liquids and slurries contained in ditches,
pipes, or holding areas (such as textile waste-
waters) are sampled by the dipper or bucket
grab procedure. If mechanical sampling de-
vices are in place and available, these may be
used to obtain Level 1 samples; however, they
are not specified for Level 1 screening proce-
dures. In fact, the Level 1 procedures leave
considerable flexibility in sample acquisition
design in order to allow grab or integrated
sampling to be tailored to the source (ref. 8).
For a reasonably consistent source, such as a
holding pond effluent, a single point grab sam-
ple is usually acceptable. However, if serious
changes occur or are suspected in the process
stream over a period of time, then a 24-h com-
posite should be collected from hourly subsam-
ples.
Samples tor laboratory chemical analysis are
preserved by standard techniques; for biolog-
ical testing they are shipped in ice without
preservation. Additional details on sample ac-
quisition and sample size requirements are
given in the appropriate reference documents
(refs. 8,9,10,11,12).
LEVEL 1 ANALYSIS METHODS APPLIED
TO TEXTILE WASTEWATERS
An overview of sample handling for organic,
inorganic, and biological analyses of aqueous
samples is shown in Figure 2. Field procedures
are indicated by solid lines and laboratory
analyses by broken lines. Field analyses such
as dissolved oxygen and pH are performed
electrochemically with the appropriate elec-
trode and are run on the neat unfiltered sam-
ple. Extraction for organic analysis is accom-
plished in the field at acidic and basic pH's. Fil-
tration for total suspended solids is also done
in the field and in accordance with standard
EPA methods.
Tests performed in the laboratory for inor-
ganic solids, filtrate, organic, and biological
analyses are shown in Figures 3, 4, 5, and 6,
respectively. Most of the inorganic analysis
methods may be readily arranged into three
major categories: spark source mass spectrom-
etry, atomic absorption spectrometry, and
standard water tests.
Spark source mass spectrometry is an excep-
tionally powerful survey tool, which yields
results for approximately 70 elements with one
analysis. This multielement capability gives
the technique overwhelming cost or informa-
tion advantages over competing techniques of
atomic absorption, neutron activation, optical
spectroscopy, and inductively coupled argon
plasma emission spectroscopy. The primary
disadvantage is lower accuracy; however, the
results usually fall within the requirements for
Level 1. Special problems with arsenic, anti-
mony, and mercury analysis require applica-
tion of atomic absorption techniques. The
standard water tests are applied by use of a
field test kit. These kits have accuracy suffi-
cient for Level 1, and represent time and cost
savings, as well as the option of field or
laboratory application. These kits, as well as
the standard methods from which they are
derived, are widely known and will not be
discussed in detail (refs. 13,14).
The one test that does not fit into the three
categories for inorganic analysis is the leach-
able anions test. This test is only performed if
impounding of the removed solids is evident or
likely. Storage of large quantities of the solid
materials could lead to drainage or runoff prob-
lems if toxic leachables are present. The cur-
rent Level 1 specification for this test calls for
test kit anion analysis of a distilled water
leachate and an HC1 leachate. New leachate
procedures are being specified by the Hazard-
ous Waste Management Division of EPA's Of-
fice of Solid Waste Management Programs.
These procedures are under review by IERL-
RTP, and will be adopted for Level 1 use if
appropriate. The use of ion chromatography
for analysis of leachable anions is also under
215
-------
WASTEWATER SAMPLE
DISSOLVED O2
pH
rBIOTEST~}-
1 blUI tb
EXTRACTION
!~ORGANIC
[ANALYSIS^
ACIDIFY
pH<2
-[BOD, COD]
FILTER
FILTRATE
DIVIDE
UNTREATED
INORGANIC ANALYSIS
SOLIDS
ANALYSIS
BASIFY
pH 8-12
Figure 2. Sample handling summary.
216
-------
SOLIDS
Hg, Sb, As
BYAA
DRY, WEIGHT
FOR SOLIDS
ELEMENTS
BY SSMS
LEACHABLE
AN IONS TEST
Figure 3. Inorganic analysis of solids.
FILTRATE
UNTREATED
ACIDITY
ALKALINITY
CONDUCTIVITY
HARDNESS
BY . EST KIT
BASIFIED
SELECTED
ANION
HCN
H2S
BY TEST KIT
Figure 4. Inorganic analysis of filtrate.
217
-------
CH2 Cl2 EXTRACT PREPARED
AS USUAL: 100-200 ml
TOTAL VOLUME
1
CONCENTRATE AS
NECESSARY
(RQTAVAPOR, K-D, ETC.)
YES
1
ALIQUOT
FORIR
'" ~*
ALIQUOT
FORLC
TCO+GRAV
> 100 mg/ml
NO
1
CONCENTRATE TO NOT
GREATER THAN 100 mg/ml
BUT NOT LESS THAN 2 ml
I
1 ml ALIQUOT
SOLVENT EXCHANGE
ALIQUOT FORLC
TCO + GRAV
> 15 mg/ml
1 ml HEXANE PLUS
SILICA GEL
LC
1 1 1
1 2 3 *
\
I 5
1 1
6 7
EACH FRACTION:
IR ON GRAV SAMPLE
LRMS
Figure 5. Modified Level 1 organic analysis procedure.
218
-------
BIOTEST SAMPLE
CYTO-
TOXICITY
FATHEAD
MINNOW
TOXIC ITY
AMES
TEST,
MUTA-
GENICITY
ALGAE,
TOXICITY
STIMU-
LATION
DAPHNIA,
TOXICITY
SOIL
MICROCOSM,
ECOLOGY
RAT
TOXICITY
Figure 6. Biotest analysis.
review and will be adopted where applicable
for use in Level 1.
The organic analysis scheme requires more
description for understanding of its nature.
Specific compound identification should not, in
general, be expected at costs commensurate
with the Level 1 philosophy. Therefore, the
scheme presented is relatively simple and inex-
pensive, yet it produces information that can
be utilized to decide whether more sophisti-
cated and expensive methods are justified. The
Level 1 organic analysis produces data on chro-
matographic classes of compounds, character-
istics of infrared absorption bands, and further
class information from low resolution mass
spectrometry (LRMS).
The Level 1 organic analysis strategy con-
tains four analytical operations that are cen-
tral to the scheme: liquid chromatographic
separation, total organics content determina-
tion, infrared absorption spectrophotometry,
and LRMS.
The first organic analysis operation is the
determination of total organics content. This
operation allows initial quantitation of the
organics in the unfractionated sample for ali-
quot size selection for optimum column opera-
tion and subsequent quantitation of material
from each chromatographic fraction. The orig-
inal Level 1 scheme (ref. 1), as well as the first
revision, depended entirely upon reduction to
dryness and weighing for total organics deter-
mination. More recent data show that many
materials in the boiling range below 275° C
may be partially lost by that approach (ref. 15).
Accordingly, a gas chromatography procedure
for volatile organics has been adopted as a part
of the Level 1 strategy (ref. 5). Total organic
content is obtained by addition of the gravime-
tric results and the total chromatographable
organics (TCO).
Liquid chromatographic separation, the sec-
ond organic analysis operation, is the heart of
the whole approach. It is an analytical step (in
that behavior of a given class of compounds is
predictable) as well as a separation step (since
the fractions may be further analyzed much
more readily than the original mixture). Distri-
bution of a few selected compounds is shown in
Table 1.
The third organic analysis operation is in-
frared absorption spectrophotometry. This
classical technique is often overlooked in to-
day's mass-spectrometry-dominated labora-
tory, but still remains a powerful tool provid-
ing considerable information at a moderate
cost. Infrared spectra of the seven chromato-
graphic fractions may be used to confirm the
absence or presence of particular compound
classes or functional groups as indicated by the
chromatographic data. It is occasionally possi-
ble to obtain specific compound identification
219
-------
TABLE 1. PERCENT DISTRIBUTION IN LC FRACTIONS
(FROM REFERENCE 9)
Compound
1
456
Hexadecane 85
Cumene
Dichlorobiphenyl 25
Acenapthene
Tetrachloroethane
o-Nitrotoluehe
Benzaldehyde
Dihexyl ether
N-methyl aniline
Quinoline
Diethyl phthalate
2-ethyl hexanol
Phenol
15
82 17
69 5
69 31
81 19
30 70
22 75 3
18 77 4
3
94
100
100
99
100
2
0.7
from the infrared spectra. But, as previously
mentioned, the complexity of .most environ-
mental samples makes this the exception
rather than the rule.
The fourth organic analytical operation of
the Level 1 organic scheme is LRMS. The orig-
inal Level 1 scheme did not contain LRMS (ref.
10), but it was included in the modified strat-
egy (ref. 8) to prevent potential triggering of
Level 2 efforts based on large amounts of sus-
picious, but innocuous, organics. LRMS can be
a very powerful tool, especially when combined
with the other Level 1 components. In many
cases, compound identification and quantifica-
tion are possible when the entire scheme is ap-
plied.
There are advantages to employing LRMS
rather than the more powerful high resolution
mass spectrometry (HRMS) or the more popu-
lar gas chromatography/mass spectrometry
(GCMS). HRMS is roughly 4 times as expensive
as LRMS. The detailed information and com-
pound specificity available from this technique
are far beyond the original goal of Level 1, and
HRMS is not readily available for the quantity
of samples envisioned. GCMS is also more ex-
pensive than LRMS and it has the added disad-
vantage of detecting only chromatographable
materials. Both HRMS and GCMS are consid-
ered excellent Level 2 techniques.
The Level 1 biological tests (ref. 12) typically
applied to aqueous samples, such as textile
wastewaters, are shown in Figure 6. In vitro
analysis for mutagenicity is performed using
four salmonella typhimurium strains in the
Ames assay. Cellular toxicity of aqueous ef-
fluents is assayed using WI-38 human lung fi-
broblasts. Acute in vivo testing in rodents is
performed as a quantal (all or none) assay. If
gross toxic responses occur, a second quanti-
tative assay is performed. The aquatic ecolog-
ical effects tests include a vertebrate (fish), an
invertebrate (daphnia or shrimp), and an algae
of either fresh or marine origin depending on
the impact of the effluent to fresh or marine en-
vironments. The terrestrial ecological effects
tests include assays with higher plants and in-
tact soil cores.
The results of the biological tests at Level 1
not only provide a measure of direct biological
response in health, aquatic ecological, and ter-
restrial ecological areas, but also aid in overall
ranking of the streams from one or a series of
stationary sources.
RESULTS OF LEVEL 1 SAMPLING AND
ANALYSIS
A few summary examples of the data and
their use are given in Tables 2 and 3. These
tables reflect the manner in which the chemical
data can be used to rank 15 textile effluents
using either a human health model or an aquat-
ic ecological model. Table 4 details the outcome
of using only Level 1 biological results to rank
the same streams. Together, the chemical and
biological rankings provide complementary in-
formation leading to a minimun possibility of
220
-------
TABLE 2. HEALTH EFFECTS: SUMMARY OF CHEMICAL
AND BIOLOGICAL TEST RESULTS
OF TEXTILE EFFLUENTS
Plant
U
S
L
W
T
N
V
F
B
K
A
C
G
X
E
la
DOH
8.86 x 1Q2
8.35 x 102
8.36 x 1fl3
4.99 x Ifl3
3.35 x 1fl3
3.28 x 1fl3
5.19X1Q2
4.52 x Ifl2
3.30 x 1fl2
1.43 x 102
1.14 X1Q2
9.90 x Ifll
9.40 x IQl
3.60 x IQl
3.30 X 10'
1b
DOH
2.9474
2.9217
3.9224
3.6984
3.5245
3.5161
2.7152
2.6551
2.5185
2.1553
2.0569
1.9956
1.9731
1.5563
1.5185
* = data not available
M = n
loderate toxii
city rating
2
3
4
5
AMES RAM WI-38 RAT
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
2 =
3 =
N
*
M
L
M
M
N
L
N
N
N
L
N
L
N
AMES =
RAM =
*
*
L
*
*
L
*
*
*
*
*
*
*
*
*
~- mutagenicity
rabbit alveolar
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
test
L = low toxicity rating
N = no detectable toxicity
la = DOH = degree of hazard
1b = DOH = degree of hazard
Oog10)
macrophage
4 = WI-38 = human embryonic
lung cells
5 = RAT = rodent acute
toxicity
underestimating a potential environmental in-
sult.
RELEVANT ONGOING STUDIES
Although considerable effort has gone into
development of the phased approach to envi-
ronmental assessment, there are still many
areas requiring further investigation and im-
provement. The results of data gathered dur-
ing the application of the phased approach to
environmental assessments have provided
guidance to IERL-RTP in planning supportive
investigations for Levels 1, 2, and 3. Listed
below are relevant studies that have resulted
from the experience gained from assessments
conducted to date.
• Use of porous polymer resins for sampling
of organics from water.
• Sampling and analysis of highly water-sol-
uble organics.
• Interpretation and synthesis of Level 1
chemical, biological, and engineering data.
• Increased accuracy of spark source mass
spectrometry (SSMS) analysis, and exten-
sion to arsenic and antimony.
• Use of ion chromatography for anion analy-
sis.
• Evaluation and upgrading of the terres-
trial biological tests.
The objective of Level 1 sampling and
analysis is to provide cost-effective screening
information for source assessment. This objec-
tive requires a comprehensive approach rather
than a specific one. Specificity is left for Lev-
el 2.
Evaluation of control technology for its ef-
fectiveness and ability to perform over typical
variations in stream composition is an integral
part of Level 3. The use of selected Level 1
analyses specifically from the biological testing
scheme is a distinct possibility at Level 3. This
221
-------
TABLE 3. ECOLOGICAL EFFECTS: SUMMARY OF CHEMICAL AND
BIOLOGICAL TEST RESULTS OF TEXTILE EFFLUENTS
Plant
N
L
C
B
E
V
F
T
U
A
G
K
S
W
X
6a
DON
1.51 xtfl5
6.07 x 10*
3.62X1Q4
3.05 X 1fl4
2.91X104
2.S4X104
2.23 X104
1.56x104"
1.46X1Q4
1.36X104
1.15X104
1.06 X1Q4
9.48X103
5.62 x 1fl3
3.43 x 102
6b
DON
5.1806
4.7830
4.5589
4.4836
4.4638
4.4051
4.3482
4.1939
4.1646
4.1345
4.061 8 '
4.0249
3.9770
3.7494
2.5353
7
FW
ALGAL
M
N
N
L
M
N
N
N
,N
N
N
N
N
N
N
8
FW
FISH
L
L
L
N
N
L
N
L
N
L
L
N
N
L
N
9
FW
DAPHNIA
H
L
L
N
M
M
L
N
L
'M
L
N
*
M
N
10
sw
ALGAL
M
M
L
N
L
L
L
L
N
*
L
L
N
L
N
11
SW
FISH
L
N
L
N
N
*
N
L
N
L
N
N
N
L
N
12
SW GRASS
SHRIMP
L
. N
L
N
N
*
N
L
N
L
N
N
N
L
N
* = data not available
H = high toxicity rating
M = moderate toxicity rating
L = low toxicity rating
N = no detectable toxicity
6a = DOH = degree of hazard
6b = DOH = degree of hazard
7 = FW ALGAL = freshwater algal
8 = FW FISH = fathead minnow
9 = FW DAPHNIA = freshwater invertebrate
10 = SWALGAL = marine algal
11 = SW FISH = marine sheepshead minnow
12 = SW G R ASS SH R IMP = saltwater shrimp
TABLE 4. PRIORITIZATION OF TEXTILE PLANTS BY
TOXICITY OF SECONDARY EFFLUENT
Toxicity ranking
Most toxic
Least Toxic
Nontoxic
Plant
N.A
W
C,T
V,L
S
B, E, F, G, K,
U,X
approach has been employed to determine the
reduction in toxicity before and after tertiary
treatment at plants selected as high priority
from the 15 sites screened by Level 1.
Many methods and their uses have been
touched upon in a fairly cursory manner in an
attempt to present an overview in a reasonable
amount of time and space. More details can be
found in the references cited.
REFERENCES
1. J. W. Hammersma and S. L. Reynolds, "Field
Test Sampling/Analytical Strategies and
Implementation Cost Estimates: Coal Gas-
ification and Flue Gas Desulfurization,"
EPA-600/2-76-093b, NTIS No. PB 254166,
April 1976.
2. J. Vlahakis and H. Abelson, "Environmen-
tal Assessment Sampling and Analytical
Strategy Program," EPA-600/2-76-093a,
NTIS No. PB 261259, May 1976.
3. L. D. Johnson, "Proceedings for Environ-
222
-------
mental Assessment of Industrial Waste-
water," Symposium Proceedings: Third
Annual Conference on Treatment and Dis-
posal of Industrial Wastewaters and Resi-
dues, Houston, TX, April 1978.
4. J. A. Dorsey, L. D. Johnson, R. M. Statnick,
and C. H. Lochmuller, "Environmental As-
sessment Sampling and Analysis: Phased
Approach and Techniques for Level 1,"
EPA-600/2-77-115, NTIS No. PB 268563,
June 1977.
5. L. D. Johnson and R. G. Merrill, "Organic
Analysis for Environmental Assessment,"
Symposium Proceedings: Environmental
Aspects of Fuel Conversion Technology,
HI (September 1977, Hollywood, FL),
EPA-600/7-78-063, NTIS No. PB 282429,
April 1978.
6. R. M. Statnick and L. D. Johnson,
"Measurements Programs for Environ-
mental Assessment," Symposium Proceed-
ings: Environmental Aspects of Fuel Con-
version Technology, II (December 1975,
Hollywood, FL), EPA-600/2-76-149, NTIS
No. PB 257182, June 1976.
7. L. D. Johnson and J. A. Dorsey, "Environ-
mental Protection Agency Level 1 Chem-
istry: Status and Experiences," Sympo-
sium Proceedings: Symposium on Poten-
tial Health and Environmental Effects of
Synthetic Fossil Fuel Technologies, Gatlin-
burg, TN, September 1978.
8. J. W. Hammersma, S. L. Reynolds, and
R. F. Maddalone, "IERL-RTP Procedures
Manual: Level 1 Environmental Assess-
ment," EPA-600/2-76-160a, (NTIS No. PB
257850, June 1976.
9. D. B. Harris, W. B. Kuykendal, and L. D.
Johnson, "Development of a Source As-
sessment Sampling System," presented at
the Fourth National Conference on Energy
and the Environment, Cincinnati, OH, Oc-
tober 1976.
10. P. W. Jones, A. P. Graffeo, R. Detrick, P. A.
Clarke, and R. J. Jakobsen, Technical
Manual for Analysis of Organic Materials
in Process Streams, EPA-600/2-76-072,
NTIS No. PB 259299, March 1976.
11. G. T. Brookman, J. J. Binder, and W. A.
Wade, "Sampling and Modeling of Non-
Point Sources at a Coal Fired Utility,"
EPA-600/2-77-199, NTIS No. PB 274369,
September 1977.
12. K. M. Duke, M. E. Davis, and A. J. Dennis,
IERL-RTP Procedures Manual: Level 1
Environmental Assessment Biological
Tests for Pilot Studies, EPA-600/7-77-043,
NTIS No. PB 268484, April 1977.
13. American Public Health Association,
Standard Methods for the Examination of
Water and Wastewater, 14th Edition,
APHA, Washington, DC, 1975.
14. U.S. Environmental Protection Agency,
Manual of Methods for Chemical Analysis
of Water and Wastes, EPA-625/6-74-003,
1974.
15. Arthur D. Little, Inc., Monthly Progress
Report on EPA Contract 68-02-2150, April
1977, to be published in a forthcoming EPA
report.
223
-------
SHORT-TERM BIOTESTING OF ENVIRONMENTAL EFFLUENTS
Shahbeg S. Sandhu*
Abstract
The U.S. Environmental Protection Agency
has proposed a phased approach for evaluating
the health and ecological impact of industrial ef-
fluents. Biotesting is considered an integral part
of this approach. The bioassays included in the
first phase, Level 1, are relatively simple, inex-
pensive, and rapid. The data generated by these
bioassays are considered to be of preliminary na-
ture and are used only for prioritizing waste
streams for further attention regarding imple-
mentation of control technology. Included in
Level 1 are bioassays for the evaluation of effects
on health, terrestrial ecology, freshwater ecology,
and marine ecology. With the exception of one
bioassay (Ames test), the tests included in Level
1 provide only the acute toxic effects. More ex-
tensive testing is required to verify the results
obtained from the Level 1 test matrix.
INTRODUCTION
Environmental pollution created by the in-
dustrial nations has significantly increased dur-
ing the last three decades. The revolution in the
chemical industry after World War II and the
increased demand for chemical products and by-
products have been responsible for increasing
the quantities and distribution of hazardous
substances in the biosphere. The need for devel-
oping new sources of energy involving synthetic
fuels and coal conversion technologies has
added new dimensions to the existing en-
vironmental problems.
Recent episodes arising from the lack of in-
stalling proper process control technology as in
the case of Seveso in Italy, and mishandling of
industrial waste as in the cases of the Love
Canal in the United States and B. T. Kemi in
Sweden have aroused public awareness to the
hazards associated with industrial effluents
and wastes.
Environmental assessment is an attempt to
'Health Effects Research Laboratory, U.S. Environmen-
tal Protection Agency, Research Triangle Park, NC.
evaluate the overall health and ecological im-
pact of industrial emissions, effluents, and
wastes. The approach to environmental assess-
ment is complicated by the awesome number of
technologies that must be evaluated. It is not
practical to provide a toxicological and ecolog-
ical assessment of the emissions and effluents
from these industrial sources by conventional
means. Therefore, the development of inexpen-
sive, short-term, and simple bioassays is a crit-
ical step in evaluating the health and ecological
hazard from industrial technologies.
It is clear that chemical and biological analy-
ses have a dual role in the process of environ-
mental assessment. Chemical analysis, while
indicative, cannot provide sufficient data for
complete evaluation of potential pollutant ef-
fects because the biological activity of complex
samples cannot be consistently predicted. Al-
though bioassays can indicate the biological ac-
tivity of a given sample, they cannot specify
which components of a crude sample are re-
sponsible for the observed toxicity.
In considering the application of analytical
chemical methodology, it is important to real-
ize that the presence or absence of a ' nown
toxic component within a complex cample
neither indicates nor precludes a relationship
between that component and the biological ac-
tivity of the sample. To avoid the possibility of
overlooking unanticipated biologically active
components, chemical analysis should not be
restricted to determination of preselected
known toxic compounds or suspected hazard-
ous components of a complex sample. However,
it may be technically overwhelming to analyze
for all of the known or suspected hazardous
components in a large number of complex sam-
p: ;s. For this reason, and for cost effective-
ness, it can be argued that a phased approach
involving stepwise application of biological and
chemical methodology is appropriate in eval-
uating the potential health and ecological haz-
ard of complex mixtures.
The Industrial Environmental Research
Laboratory of the Environmental Protection
Agency, Research Triangle Park, NC (IERL-
RTP), has delineated a 3-phase approach for
225
-------
performing an environmental source assess-
ment in order to determine the need for control
technology (refs. 1,2). The scheme, shown in
Table 1, involves the evaluation of process
streams beginning with the least expensive
and most rapid assays, mostly qualitative in
nature, followed by more expensive and time-
consuming assays mostly quantitative in na-
ture. This approach provides an integrated ap-
plication of physical, chemical, and biological
tests at each level.
The first level (Level 1) involves rapid
screening of a large number of industrial waste
materials for their toxic and mutagenic effects.
These results are used to prioritize streams for
further evaluation in Level 2. In Level 2, the
results obtained in Level 1 are validated and
confirmed through more stringent and compre-
hensive toxicological techniques and the chem-
ical and/or physical properties of the waste
materials are determined in greater detail.
Level 3 source assessment examines the com-
position and long-term variations of the more
toxic waste streams and focuses on the long-
term effects of these waste streams on biolog-
ical systems.
The Level 1 bioassays include health and
ecology related bioassays that, with the excep-
tion of one biotest, measure acute toxic effects.
The matrix of Level 1 bioassays is shown in
Table 2.
Because of the diverse physical forms of the
industrial discharges (i.e., liquid, solid, and
gaseous), all the Level 1 bioassays listed in
Table 2 may not be applicable to all the sam-
ples. A suggested minimal test matrix is shown
in Table 3. This matrix represents a minimum
test battery required to indicate potentially
adverse health and and ecological effects
caused by an industrial waste stream.
DATA MANAGEMENT
The number and diversity of proposed
bioassays employed in environmental assess-
ment require the development of a uniform
data reporting format that permits the ex-
amination, interpretation, and correlation of
the results. The data from each bioassay are ar-
bitrarily reduced to levels of toxicity—high,
moderate, weak, or nondetectable.
The criteria for determining the level of tox-
icity (also shown in Table 4) are listed below.
Nondetectable—no significant response as
determined by a preestablished set of crite-
ria including statistical analysis where ap-
propriate. The nondetectable classification
would take into account certain toxicants
that could be present in a sample below the
level of detectability for the assay. This
classification would also take into considera-
tion methods for concentration and/or ex-
traction.
Low—a significant response or LDso (ECso)
ranging from the maximum dose applicable
to the assay system (defined as MAD and es-
tablished in advance) to one-tenth of that
value. For example, if it is established that
10 g/kg will define the upper limit of sample
application to rats in the acute rodent toxic-
ity test, then any sample with an LD^ value
< 10 g/kg down to a concentration of 1 g/kg
TABLE 1. THE IERL-RTP PHASED APPROACH TO ENVIRONMENTAL ASSESSMENT
Level
Objective
Bioassays
Use
Screening, detection
Verification, identification
of toxic components
Long-term effects
Simple, inexpensive, rapid
Mostly qualitative data
Mostly quantitative data,
well-defined end points
Expensive and comprehensive
Animal and ecosystem studies,
chronic exposure
Prioritization for
further testing
Development of
control tech-
nology
Health and eco-
logical hazard
assessment
226
-------
TABLE Z. IERL-RTP LEVEL 1 BIOASSAYS
Test system
Health-related in vitro tests
Ames salmonella/microsome
mutagenicity
RAM-cytotoxicity
WI-38-cytotoxicity
CHO-cytotoxicity
Study time
(weeks)
3
3
3
3
Sample
1 g/50 ml
0.5 g/50 ml
0.5 g/50 ml
0 g/50 ml
Approximate
cost($)
600
600
340
340
Health-related in vivo tests
Rodent range-finding test
Freshwater toxicity ecology tests
Algal
Daphnia-static
Fish
Marine ecology tests
Fish
Shrimp
Terrestial ecology related bioassays
(Under development)
3
1
1
100g/liter
200 liters
200 liters
200 liters
50 liters
50 liters
640
1,570
1,000
1,570
1,570
TABLE 3. LEVEL 1 MINIMAL TEST MATRIX
Sample
Health effects tests
Ecological effects tests
Liquids
Solids
(aqueous extract)
Particulates
Sorbent (extract)
Gases
Ames, CHO
Ames, RAM, whole animal
Ames, RAM, whole animal
Ames, CHO
Ames
Aquatic, seed germination, nitrogen and
seedling growth fixation
Aquatic, seed germination, nitrogen and
seedling growth fixation
Seed germination and seedling growth
Aquatic, seed germination, nitrogen and
seedling growth fixation
Plant stress ethylene and foliar injury
insect toxicity
227
-------
TABLE 4. OPERATING CHARACTERISTICS OF WASTE TREATMENT SYSTEMS FOR TEXTILES
Range of concentration defining
Assay
Ames
Rodent toxicity
Fresh water and marine
aquatic assays §
RAM
CHO/WI-38 toxicity
MAD
5 mg/plate or 500 /Jl
10mg/literor100%
100 mg/literor 100%
1,000 /jg/ml orBOO/jl/ml
1,OOOM9/ml or 600 Ml/ml
High
10g/kgor
equivalent.
LDsfj at > 100 mg/liter
or 100%.
ECso at >1, 000 Mg/ml or
600 I/ml.
ECBfjat >1,OOOMg/mlor
600 Ml/ml.
Criteria for positive or negative interpretation of the Ames test are given in Appendix C.
s MAO for aquatic studies is set at 100 mg/liter or 100% if the total sample concentration is less than 100 mg/ml. Compound solubility can be used to set an upper
limit in this assay.
Notes: MAO = Maximum Applied Oose
MTD = Maximum Tolerated Dose
LD5Q = Lethal Dose 50%
EC5Q = Effective concentration (50% may be substituted).
will be defined as having low toxicity. The
degree of any concentration procedure will
be factored into the calculation procedure
and will reduce the level of toxicity accord-
ingly.
Moderate—a. significant response or an LDso
(ECso) value ranging from a concentration
less than 1/10 of the MAD (the lower limit of
the low toxicity) to 1/100 of the MAD. Again
using the same illustration, an LDso (ECso)
obtained at a concentration ranging from 0.1
g/kg to 1.0 g/kg in the rodent toxicity assay
will be considered as moderate. The same
considerations of sample concentration will
apply.
High—a significant response or an LDso
(ECso) value less than 1/100 of the MAD for
the particular assay. Thus, any sample with
an LDso less than 0.1 g/kg in the rodent assay
will be considered to have high toxicity.
THE PILOT STUDIES
Most short-term bioassays employed in the
Level 1 environmental assessment program
were developed for testing relatively "pure"
compounds. Their reliability for testing com-
plex mixtures was not known. It was decided,
therefore, to test the feasibility of biotesting of
industrial samples in pilot studies.
Samples from three industries —textile
wastewater, fluidized-bed combustor, and coal
gasification — were evaluated in these pilot
studies. The contract laboratories and EPA
technical advisors involved in conducting these
pilot studies are listed in Figure 1. The details
of bioassay protocol, data interpretation, and
data evaluation are presented elsewhere in
this document. Only the overall summary and
conclusions from the application of the short-
term Level 1 bioassay matrix are presented
here.
Biological activity, especially genetic ac-
tivity, may be masked by toxicity and may
require chemical fractionation in order to
demonstrate activity in the sample.
For liquid affluents, tests for potential eco-
logical effects may be more sensitive than
tests for potential health effects due to
concentration dilution problems.
The prediction of relative toxicity on the
basis of chemical analysis alone is subject
to error.
Results of biological and chemical tests
are complementary—considered together
the two types of tests provide useful infor-
mation not obtainable when considered
separately.
228
-------
EPA TASK OFFICER
MRC PROJECT LEADER
SAMPLE COLLECTION
MRC
_L
BIOASSAY AND
TESTING LABORATORY
MUTAGENICITY
SRI AND MRC
CYTOTOXICITY
NORTHROP AND MRC
FATHEAD MINNOW
AND DAPHNIA
EPA
FRESHWATER ALGAE
EPA
SHEEPSHEAD MINNOW
AND GRASS SHRIMP
BIONOMICS
MARINE ALGAE
EPA
14 - DAY RAT
ACUTE TOXICITY TES1
LITTON BIONETICS
SOIL MICROCOSM
EPA
J
EPA TECHNICAL
ADVISOR
HERL-RTP
M. WATERS
ERL-NEWTOWN
W. HORNING
ERL- GULF BREEZE
J. WALSH
HERL-CINC
J. STARA
EPA - CORVALLIS
B. LIGHTHART
Figure 1. Laboratories and EPA technical advisors involved in biotesting of effluent samples
in pilot studies.
229
-------
RELIABILITY OF LEVEL 1 BIOASSAYS
The utility of short-term bioassays included
in the Level 1 environmental assessment pro-
gram is for the identification of hazardous
streams and to prioritize the streams for fur-
ther testing in higher order bioassays, based
on positive results across Level 1 test systems.
The data from Level 1 testing are not used for
making decisions on the extent of hazard or
safety of the streams. The bioassays included
in Level 1 are in various stages of develop-
ment. Some of them, such as the Ames Salmon-
e//a/microsome assay, have been reasonably
well validated for testing a variety of com-
pounds; whereas others, such as assays to
evaluate terrestrial effects, are not as fully
developed. The results from the pilot study on
textile wastewater suggest that it may be nec-
essary to concentrate certain liquid samples
before these can be subjected to biotesting by
in vitro techniques. The nature of in vitro pro-
cedures, dictates the use of very small aliquots
of aqueous samples. In a complex mixture the
bioactive components may be present in very
small amounts. Then detection, therefore,
becomes difficult because of the limitations of
the bioassay. To obtain reliable results from in
vitro tests, the liquid samples frequently will
have to be concentrated before testing. Also,
especially for the in vivo range-finding rat test,
the possible need for bioconcentration should
be considered. The test animals should receive
a daily dose of the test material rather than a
single quantal dose as suggested in the IERL-
RTP Level 1 Bioassay Procedures Manual (ref.
1).
BIOASSAY-DIRECTED FRACTIONATION
It may be difficult to detect and/or delineate
the biological effects of a complex mixture
without separating it into subfractions because
the active compounds may be present in very
low concentrations. It is therefore desirable to
separate the mixtures into similar components
for the purpose of concentrating and eventual-
ly identifying the active compounds. A limited
fractionation approach has been successfully
applied to the study of synfuel products and
emissions from diesel combustion. In this
scheme, the test sample is fractionated into
four fractions; i.e., acidic, basic, neutral, and
residual. Each fraction is further tested for
genotoxic activity in microbial tests. The
overall conclusion from the synfuel studies is
that the results from the unfractionated com-
plex mixture sample may be misleading. In the
majority of cases, cumulative activity of the in-
dividual fractions is several fold higher than
the unfractionated sample (ref. 3). It is antic-
ipated that a simple fractionated scheme will
be adopted for biotesting of industrial efflu-
ents and streams.
REFERENCES
1. K. M. Duke, M. E. Davis, and A. J. Dennis,
IERL-RTP Procedures Manual Level 1
Environmental Assessment Biological
Tests for Pilot Studies, EPA-600/7-77-043,
1977.
2. James A. Dorsey, Larry D. Johnson, Rob-
ert M. Statnick, and Charles H. Lock-
miiller, Environmental Assessment Sam-
pling and Analysis: Phased Approach and
Techniques for Level 1, EPA-600/2-77-115,
June 1977.
3. M. D. Waters and J. L. Epler, "Status of
Bioscreening of Emissions and Effluents
from Energy Technologies," presented at
Third National Conference on the Inter-
agency Energy/Environment R&D Pro-
gram, June 1-2,1978, Washington, DC.
230
-------
THE TESTING OF ENVIRONMENTAL SAMPLES FOR MUTAGENICITY
AND CARCINOGENICITY USING MICROBIAL ASSAY SYSTEMS
Larry D. Claxton, Joellen Huisingh, Michael Waters*
Abstract
This paper will present a brief description of
several microbial assay systems for mutagenic-
ity and DNA damage. Using the Salmonella
mutagenicity assay, an attempt will be made
to identify and discuss the technical difficul-
ties encountered in the screening of complex
environmental samples. A second objective is
to present available procedural modifications
together with the utility of these modifica-
tions. The reproducibility and interpretation of
the data from short-term bioassays will be dis-
cussed.
INTRODUCTION
It has been reported that over 4 million
distinct chemicals are now known and that
63,000 chemicals are thought to be in common
use (ref. 1). Trying to determine the public
health effects of such a large number of
chemicals is a massive undertaking (ref. 2).
Microbial mutagenicity assays are now being
used to identify agents showing genetic activi-
ty and as a carcinogen prescreen. A variety of
short-term tests are available; however, the
Salmonella mutagenicity assay has been the
most extensively used. This paper will explore
the basic protocol, precautions, and modifica-
tions (Ames) of the Salmonella typhimurium
mammalian microsome assay. The application
of this assay to complex mixtures such as in-
dustrial emissions and effluents will be dis-
cussed and methods of concentration, extrac-
tion, and fractionation will be considered. Since
interpretation of results may sometimes be dif-
ficult, some discussion will concern the presen-
tation and interpretation of data. Although the
discussion will be mainly limited to the Salmon-
ella assay, many of the considerations can be
*Biochemistry Branch, U.S. Environmental Protection
Agency, Environmental Toxicology Division, Health El'
lects Research Laboratory. Research Triangle Park,
NC.
applied to other in vitro tests for mutagenicity.
THE BASIC SALMONELLA PLATE
INCORPORATION TEST
The reference protocol is described by Ames
et al. (ref. 3) and will not be repeated here. The
outline of the basic procedural protocol is a
very simple one (Figure 1). Basically, a specific
amount of the chemical to be tested and a mix-
ture of the indicator bacterial strain are added
to melted soft agar, mixed, and poured into a
minimal media plate that will select for bac-
teria that have undergone mutation. The plate
is incubated for 48 hours at 37° C and then
scored for the number of colonies per plate.
This procedure is repeated for each dose of the
chemical; for each bacterial strain; for each
routine modification (e.g., with and without
microsomal activation); for each needed con-
trol; and for each duplication. Therefore, a
chemical tested at 5 doses, with 3 replicate
plates at each dose, with 5 bacterial strains,
with appropriate controls, and done both with
and without microsomal activation, will cause
this basic procedure to be repeated a minimum
of 210 times, thus generating 210 "data points"
for each chemical to be analyzed.
The procedure scores for gene mutations
that occur in the histidine locus in strains of
Salmonella typhimurium created by Dr. Bruce
Ames. Each of these strains has undergone a
specific mutation within the histidine locus and
thus is unable to grow unless histidine is sup-
plied. These strains are said to have a histi-
dine-negative genome (his -). When these
strains mutate to a histidine-positive genome
they can be detected on a nutrient medium
that lacks histidine. Since these strains are
"mutating back" to a wild type condition, a
reverse mutation is said to occur; however, the
biochemical mechanisms are the same as occur
in forward mutation. Systems that detect for-
ward mutation, though, are generally not lim-
ited to a specific type of genetic event (e.g., a
frameshift or a point mutation), whereas a re-
verse mutation system does detect a specific
231
-------
PLATE INCORPORATION TEST:
S-9 MICROSOMES
BACTERIA-
-CHEMICAL
SOFT AGAR OVERLAY
INCUBATE
_». 48 hrs., 37° C
PLATE WITH OVERLAY
RESULTS:
CONTROL
INCREASING DOSAGE
(MUTAGENIC WITH NO KILLING)
Figure 1. Outline of basic procedural protocol.
type of mutation. Since screening systems
need to score for a variety of genetic muta-
tional events, several different Salmonella
strains have been developed. Each strain de-
tects a specific type of mutation and incor-
porates other changes needed to increase its
sensitivity. There are now five strains in com-
mon use (Table 1). Three of these strains
(TA1537, TA1538, TA98) detect frameshift mu-
tations. TA1537 detects compounds such as
9-aminoacridine. TA1538 and TA98 both detect
compounds such as 2-nitrosofluorenone, but
differ in the fact that TA98 contains an R fac-
tor plasmid. Chemicals that produce base-pair
substitutions, especially alkylating agents, are
detected in TA1535. Strain TA100 was devel-
oped from TA1535 by the addition of the R fac-
tor plasmid. TA100 has a rather high sponta-
neous mutation rate and may detect chemicals
that normally yield frameshift mutations as
well as base-pair substitutions. **
In addition to the histidine locus alteration,
these strains carry an rfa mutation that cre-
ates a deficiency in the bacterial cell wall
lipopolysaccharides. This deficiency increases
the cell's permeability to large molecules. The
uyr B mutation was added to these strains to
increase their susceptibility to several classes
of mutagens by decreasing genetic repair. The
R factor plasmid found in strains TA98 and
TA100 increases further the sensitivity of the
tester strains by participating in a type of
error-prone repair. The plasmid is detected by
ampicillan resistance. These strains also differ
in the number of spontaneous revertants per
plate generally found. Table 1 summarizes the
characteristics of the five strains.
232
-------
TABLE 1. CHARACTERISITICS OF FIVE SALMONELLA TYPHIMURIUM
TESTER STRAINS
Strain
TA1535
TA1537
TA1538
TA98
TA100
his-mutation
his G46
his C3076
hjs_D3052
his D3052
his G46
rfa
Yes
Yes
Yes
Yes
Yes
uvr B
Yes
Yes
Yes
Yes
Yes
R factor
No
No
No
Yes
Yes
Spontaneous
revertants/plate
5-50
2-25
540
15-75
80-200
MAMMALIAN METABOLIC ACTIVATION
Some compounds that are carcinogenic for
mammals are not mutagenic for bacteria;
therefore, a mammalian metabolic activation
system has been added to the assay. This addi-
tion allows for some of the metabolites as well
as the agent itself to be tested for mutagenic-
ity. The activation system described by Ames
consists of a 9,000 x g supernatant of a rat
liver homogenate. Aroclor induction of the rats
is done in order to induce the hepatic enzymes
referred to collectively as mixed function oxy-
genases (MFO). An induced MFO system gives
the potential for increased levels of metab-
olites and a corresponding increase in revert-
ant colonies due to mutagenic metabolites.
PRECAUTIONS AND PROCEDURAL
MODIFICATIONS
Although the reference protocol has been
described in great detail, separate laboratories
may make certain modifications to meet spe-
cial needs. The authors will attempt to
describe procedures likely to be neglected or
altered and the potential effects this is liable to
have on the reliability of the generated data.
This should be helpful to both the laboratory
investigator using the test and the reviewer
who examines only the results. The points con-
sidered will follow as closely as possible the se-
quence encountered within a testing protocol.
Checking of Tester Strains
Upon the receipt of tester strains and at
regular intervals, the tester strains should be
checked for UV sensitivity, ampicillan resist-
ance, and crystal violet killing. A failure to
check these parameters could allow for mixed
cultures and decreased sensitivity or false neg-
atives and positives. If these parameters are
not checked, the test should be considered un-
acceptable. Although spontaneous and positive
controls are conducted with each test, the in-
validating of some tests can be prevented by
checking spontaneous and induced mutant lev-
els when creating master plates.
Proper Overnight Culture Growths
Variation between bacterial strains, media
used, and testing conditions can cause consid-
erable variation in the bacterial culture used
for testing. Three conditions should alwpys be
met for the routine screening of chemicals.
First, a single colony isolate should be used to
initiate the culture. This helps to insure
genetic uniformity of the broth culture. Sec-
ond, the test culture should have a concentra-
tion of 1 to 2 x 109 viable cells per milliliter.
Note that it is viable cells per milliliter. This
can best be insured by using the culture in the
early portion of the stationary growth phase.
Next, if cells cannot be used immediately upon
harvesting, they should be stored in an ice-
wa.tr f bath.
mammalian Activation System
For general screening the 9,000 x g super-
natant (S9) from rat liver induced with Aroclor
1254 is generally recommended. This is an area
of the protocol that can be modified in a variety
of ways. One may modify the organism (rat,
mouse, guinea pig, plant, bacterial flora, etc.);
the tissue source (skin, liver, kidney, em-
233
-------
bryoes, etc.); the biochemical fraction (S9,
purified microsomes, etc.); induction chemicals
(Aroclor, phenobarbitol, ethanol, etc.); and the
induction schedule. One may also use in vivo
activation by examining the urine or bile of
treated animals.
These modifications are not generally ap-
plied. However, there are typically two indica-
tions for using one or more of these modifica-
tions. First, if the route of exposure or the con-
centration of the substance is associated with a
specific organ, that tissue could be used in ad-
dition to liver. Second, when compounds of like
chemical structure are shown to be positive
only when one of these modifications is used,
that modification should be used. As an
example, since human exposure to flame
retardants occurs via the skin, Rosenkranz
tested the activation of tris (2,3-dibromogropyl)
phosphate ("Tris") using skin enzymes (ref. 7).
He found that the skin of newborn animals was
capable of transforming "Tris" to a mutagenic
intermediate and that this activity was en-
hanced by inducers such as Aroclor.
The Substance to be Tested
The test sample brings the greatest var-
iability into the test system; therefore, the ef-
fects of the substance tested should be closely
defined. First, an adequate number of concen-
trations and range of concentrations should be
tested. With the initial screening of a sub-
stance, at least five doses at half-log intervals
should be tested if possible. Compounds de-
monstrating little or no toxicity should
preferably be tested in the 5- to 10-mg per
plate range before being reported as negative.
The solubility of the test substance should
be noted when placed both into the solvent and
into agar overlay. The precipitating of sub-
stances may not allow for adequate testing.
Highly volatile substances may evaporate so
quickly that they are not adequately tested.
These volatile substances should be tested in
sealed bags or dessicators.
Some indication of the toxicity should be
recorded. For example, if clearing of the test
plates is seen at all doses, lower doses should
be tested in order to assure that the effect seen
is not due to toxicity. Also, a negative mutation
response when no toxicity is seen is a question-
able result. When toxicity does result, some of
the bacteria may lyse and contribute histidine
to media; therefore, cross feeding may result
and colonies that are not revertants may be
found. Usually these colonies are small pin-
point colonies, yet that is not always the case.
Colonies from representative plates should be
transferred to minimal media plates without
histidine to insure that they are revertant col-
onies. Replica plating is an excellent means of
accomplishing this check.
Other substances, especially complex mix-
tures and technical-grade chemicals, may
either contain histidine or histidine-like-acting
components. Strain TA100 is especially sensi-
tive to small increases in histidine. One should
determine if the test substance is related to a
known compound that has been previously
tested. Some chemical classes (such as the low-
molecular-weight nitrosamines) are known to
give negative results in the typical plate
incorporation test. Either a modification of the
Salmonella test or another bacterial species
should be used under these circumstances.
Within-Experiment Controls
Every test should contain appropriate
positive, negative, and sterility controls. Table
2 records some of the acceptable positive con-
trols for the five commonly used tester strains.
The negative control should consist of the en-
tire system without the compound to check
spontaneous mutation rate. The activation mix
and test substance should be checked for ste-
rility.
Minimal Design Criteria for Assigning
a Negative Response to the Plate Test
Since most positive results are easily
distinguished, the experimental design should
be done to give some definition to a negative
result. In the authors' opinion the following are
the minimal testing requirements needed to
operationally designate a test negative. (More
will be said about defining a negative in a later
section.)
• Minimal of five doses at half-log intervals,
with the highest dose being highly toxic or
presented at a minimum of 5 mg per plate;
• Spontaneous and positive controls done at
least in duplicate;
234
-------
Strain
TABLE 2. CONTROL MUTAGENS
Without activation
With activation
TA1535 MNNG, or sodium azide, or 2-Anthramine
methylmethane sulfonate
TA1537 9-Aminoacridine 2-Anthramine
TA1538 2-Nitrofluorene, or 2-Anthramine
4-Nitro-o-phenylenediamine 2-Anthramine
TA 98 2-Nitrofluorene, or 2-Anthramine
hycanthone methanesulfonate
TA100 MNNG, or sodium azide, or 2-Anthramine
MMS, or nitrofurantoin
• Positive controls (in duplicate) for any bac-
terial/microsomal activation combination
used;
• At least two replicates per dose;
• At least one replicate experiment, if pos-
sible, done with a narrower dose range;
• Tested in TA1535, TA1537, TA1538,
TA98, TA100 (this requirement may
change with the development of new
strains and/or data);
• Proper sterility controls; and
• Checking of colonies for genotype.
Procedural Modifications Possible
Spot Test-
This is a very qualitative form of the plate
test where bacteria and the activation system
(if used) are overlayed onto a minimal media
plate and the chemical is spotted onto the plate
or onto a paper disc on the plate. This test is
used for rapid screening of many compounds or
for substances limited by availability. The dif-
fusibility of the substance strongly influences
the results (ref. 3).
Well Test-
The well test is a modification of the spot
test, which is to be used where the supply of
the sample is very limited. After the bacteria
are overlayed, wells are cut in a specified pat-
tern. Into each well soft agar with either the
activation system, the positive control, the con-
trol solvent, or a test substance is placed. The
advantage of the well test is that one or two
substances can be placed, along with controls,
on a single plate for mutagenicity both with
and without activation. Diffusibility again is
the main limiting factor and the results are
only qualitative (ref. 5).
Liquid Suspension Assay—
This is a more quanitative test than the plate
test; however, it is much more time consuming
and costly. Incubated together in a buffered
suspension are the bacteria, any activation
system used, and the compound used. After a
specified time of exposure, aliquots are trans-
ferred to appropriate plating media to score
for both survival and revertant colonies. This
procedure gives an approximation of actual
mutation frequency (ref. 6).
Preincubation Test—
Performed otherwise in the same manner as
the plate test, the preincubation test also re-
quires that the bacteria and chemical be pre-
incubated together at room temperature, 37° C,
or 45° C before soft agar is added and plating is
done. This modification allows for the detec-
tion of mutagenicity in some compounds that
are not detected in the plate test, and it is a
less complex test than the liquid suspension
assay (ref. 7).
Fluctuation Test—
Bacteria are incubated in suspension with
the test substance without histidine, and the
235
-------
number of tubes demonstrating growth are
scored. This is a very sensitive system for de-
tecting some mutagens and is amenable to
automation (ref. 8).
Anaerobic Testing—
The plate test is performed and the plates in-
cubated in Gas Pak jars in the dark at 37° C for
14 hours and then incubated aerobically. Cer-
tain chemicals (e.g., aga thioprine) can be de-
monstrated as mutagenic only in this way (ref.
4).
Host-Mediated Assay—
With this protocol, the bacterial indicator
organism and the substance to be tested are
administered to a mammal via different routes.
After a specified period of time, the bacteria
are recovered and handled as in a liquid sus-
pension test. Although the substance is dis-
tributed and metabolized throughout the
whole body, a variety of technical problems
complicate the test. These include: bacterial
recovery, bacterial-metabolite proximity, im-
mune response to bacteria, etc. (ref. 9).
Special Procedures for Complex Mixtures
Complex mixtures present a variety of
technical difficulties but are quite capable of
being tested. Most complex mixtures are di-
rect environmental pollutants and include a
diversity of sample types. Some of these
samples are very low in organics while others
are entirely organic.
Liquid secondary effluents from 22 textile
plants were initially tested without concentra-
tion or extraction with four Salmonella strains
and in a battery of other microbial bioassays
(ref. 10). No positive mutagenic response was
detected for any of these samples. Quantitative
chemical analytical data on selected organics
showed the presence of suspected carcinogens
and mutagens; the concentrations of these
chemicals in the effluent, however, were below
the concentration expected to be detectable in
the Salmonella assay (ref. 10).
The mutagenic activity of a complex mixture
may not be detectable to a variety of problems.
Examples of these would be: (1) dilution (as in
the textile wastewater example); (2) presence
of a sample matrix which prevents release of
the mutagens into the bacterial media (as in
the case with certain carbonaceous particulate
samples), or (3) the presence of highly toxic
compounds in the mixture — which results in
sufficient bacterial killing to prevent detection
of bacterial mutants. To overcome these prob-
lems, complex mixtures must often be concen-
trated, extracted, or fractionated prior to
bioassay. Model applications of these method
are described below.
Concentration—
There are a variety of methods by which
samples can be concentrated. If the sample is
particulate or precipitates out of the solvent
used, filtration will concentrate the sample.
The type of filtration will depend upon the
sample. For example, air particulates are
usually filtered by the use of polflex or glass
fiber filters. Some water and solvent samples
can be filtered for particulates with Millipore
and Nucleopore filters and other similar filters.
Air particulates also may be concentrated by
impaction devices and electrostatic precipita-
tion. Water and other fluid samples have been
concentrated by chromatography methods.
The XAD columns appear to be especially
suited to water and urine samples (refs.
11,12,13). The researcher needs to realize that
concentration may enhance chemical reactions
between the constitutents of a complex mix-
ture and thereby influence bioassay results.
Extraction—
Active organic compounds may be attached
to a variety of solid materials. Since bacteria
cannot ingest most of these solid materials (as
some mammalian cells are able to), the organics
must be extracted from the solid matter. A
variety of solvent systems have been used in-
cluding benzene, dichloromethane, cyclohex-
ane, acetone, methanol, methanol-benzene, and
dimethyl sulfoxide. Sonication along with a sol-
vent system has also been used. Usually the
mixture must be solvent exchanged before
testing in the Ames test. These approaches to
evaluating the organics associated with air, in-
dustrial, and mobile source particulates were
recently reviewed (ref. 14).
Extraction has not been as extensively ap-
plied to liquid samples. However, Kulik (ref. 15)
has applied this approach to several selected
secondary effluents from the textile plants
referred to above. When any of these methods
236
-------
are used, several questions must be addressed.
What types of compounds are extracted and
are not extracted by the procedure? Does the
solvent interact with any of the compounds to
produce artifactual components? When the sol-
vent itself is mutagenic, has it been effectively
removed before testing? What special controls
should be used with testing?
Fractionation—
Epler (ref. 16) has shown that the mutagenic
activity of some complex mixtures cannot be
detected until the mixture is fractionated. This
appears to happen because toxic components of
the complex mixture are removed by fractiona-
tion.
Safety Precautions
The following is a listing of some safety pro-
cedures established within the author's labora-
tory.
1. All samples for testing and positive con-
trol chemicals are housed in a special
weighing and storage area. These
samples and chemicals are recorded and
logged in and out.
2. All weighings and measurings of stock
chemicals are handled in a special weigh-
ing room within glove boxes that are
vented to the outside through charcoal
filters located on the roof.
3. Incubators, storage cabinets, and
refrigerators are vented. Upon the open-
ing of any of these units the air within the
unit is exhausted.
4. All work is performed in laminar-flow ex-
haust hoods or total-exhaust biological
hoods.
5. All work is done in designated areas by
employees properly clothed with dispos-
able clothing.
6. No mouth pipetting is allowed for any pro-
cedure.
7. All waste is double bagged and then
placed in drums which are sealed and sent
by a contractor to special landfill areas for
disposal.
Data Interpretation
Since actual mutational frequencies cannot
be calculated with the plate test, disagreement
exists as to the best statistical methods to
employ. The following, however, -hould be re-
ported and evaluated for ideal j dging of the
adequacy of a test:
• proper controls - positive, negative, ste-
rility;
• solubility and stability of test chemical;
• replication of plates at each dose and rep-
lication of each test;
• the number and range of doses used;
• the toxicity of the sample;
• contamination of the sample;
• relatedness of sample to other known
compounds;
• proportionate,increase of individual doses
above spontaneous levels; and
• presence or absence of dose response
curve.
In order to say that a definite positive has
been seen in the author's laboratory, the fol-
lowing is expected:
• at least a 2.5-fold increase at one or more
doses;
• a dose response curve with some type of
"regular curve"—i.e., a single high re-
sponse would not be accepted;
• expected responses from all controls; and
• apparent absence of histidine cross feed-
ing and/or contamination.
A dose response increase seen wit' proper
controls but never reaching a 2.5-fold increase
is considered questionable.
One may present summarized data in a
variety of ways. It has been suggested, how-
ever, that summarized data not be given with-
out its supporting data (ref. 17). Some of the
methods for data presentation follow.
1. Fold increase or mutagenic index —
average plate counts divided by average
spontaneous count for the same bacterial
strain;
2. Net increase in revertants — average
counts or individual counts minus average
spontaneous count (this is generally
considered unacceptable);
3. Specific activity — slope of the linear por-
tion of the dose response curve at a speci-
fied dose (e.g., 1 mg);
4. Specific activity fold increases —the spe-
cific activity divided by the average spon-
taneous count;
237
-------
5. Specific response activity — calculated
dose needed to give a specific number of
revertants per plate (e.g., 200 revert-
ants/plate) for a specific strain;
6. Maximum response—maximum number
of revertants per plate seen.
These methods are used to compare the re-
sults of a series of tests but do not necessarily
reveal the adequacy or positiveness of the
results. A generally accepted statistical test
for examining the results of the Salmonella
plate test has not been published and until this
occurs, people will be forced to examine the
data by these less exacting means.
REFERENCES
1. Thomas H. Maugh, "Chemicals: How Many
Are There?" Science, Vol. 199, No. 4325,
p. 162, 1978.
2. Larry D. Claxton and Patricia Z. Barry,
"Chemical Mutagenesis: An Emerging Is-
rue for Public Health," American Journal
of Public Health, Vol. 67, No. 11,
pp. 1037-1042, 1977.
3. Bruce N. Ames, et al., "Methods for
Detecting Carcinogens and Mutagens
With the Salmonella/Mammalian-Micro-
some Mutagenicity Test," Mutation Res.,
Vol. 31, pp. 347-364, 1975.
4. Herbert S. Rosenkranz, et al., "The Use of
Microbial Assay Systems in the Detection
of Environmental Mutagens in Complex
Mixtures," Application of Short Term
Bioassays in the Fractionation and Anal-
ysis of Complex Environmental Mixtures,
M. Waters et al., eds., U.S. EPA, 1978, in
press.
5. Thomas J. Hughes, et al., "Agar Diffusion
Well Test for Mutagehic Screening of Air
Particulates," 1978, in review.
6. C. N. Frantz and H. V. Mailing, "The
Quantitative Microsomal Mutagenesis
Assay Method," Mutation Res., Vol. 31,
No. 6, pp. 365-380, 1975.
7. T. Yahagi, M. Nagao, Y. Seino, T. Mat-
sushima, T. Sugimura, and M. Okada,
"Mutagenicity of N-nitrosamines on Sal-
monella," Mutation Res., Vol. 48,
pp. 121-130,1977.
8. M. H. L. Green, A. M. Rogers, W. J.
Muriel, A. C. Ward, and D. R. McCalla,
"Use of a Simplified Fluctuation Test to
Detect and Characterize Mutagenesis by
Nitrofurons," Mutation Res., Vol. 44,
pp. 139-143, 1977.
9. H. V. Mailing, "The Host-Mediated
Assay," Proceedings of the Fourth Inter-
national Congress, Motulsky and Lentz,
eds., pp. 207-211, 1973.
10. D. C. Poole, and V. F. Simon, final report
of "In Vitro Microbiological Studies of
Twenty-two Wastewater Effluent Sam-
ples Reported in Source Assessment:
Textile Plant Wastewater Toxics Study
Phase I," EPA Publication 600/2-78-004h,
March 1978.
11. E. Yamasaki and B. N. Ames, "Concentra-
tion of Mutagens From Urine by Absorp-
tion With the Nonpolar Resin XAD-2: Cig-
arette Smokers Have Mutagenic Urine,"
Proc. Natl. Acad. Sci. USA, Vol. 71, No. 8,
pp. 3555-3559, 1977.
12. C. C. Smith, "Strategy for Collection of
Drinking Water Concentrates: Applica-
tion of Short-term Bioassays in the Frac-
tionation and Analysis of Complex Envi-
ronmental Mixtures," EPA Publication
600/9-78-027, pp. 227-246, September 1978.
13. C. D. Chriswell, B. A. Glatz, J. S. Fritz,
and H. S. Svec, "Mutagenic Analysis of
Drinking Water," EPA Publication 600/
9-78-027, pp. 447-494, September 1978.
14. M. D. Waters, S. Nesnow, J. L. Huisingh,
S. S. Sandhu, and L. Claxton, eds., "Appli-
cation of Short-term Bioassays in the
Fractionation and Analysis of Complex
Environmental Mixtures," EPA Publica-
tion 600/9-78-027, September 1978.
15. F. A. Kulik, "Concentration of Potential
Mutagenic Compounds in Textile Plant
Effluents for Application to the Salmon-
ella Mutagenicity Test," EPA Publication
600/9-78-027, p. 584, September 1978.
16. J. L. Epler, B. R. Clark, C-h. Ho, M. R.
Guerin, T. K. Rao, "Short-term Bioassay
of Complex Organic Mixtures: Part II,
Mutagenicity Testing," EPA Publication
600/9-78-027, pp. 269-290.
17. NIEHS-sponsored conference, "Salmon-
ella/Microsome Assay —Workshop on
Data Production and Analysis," Meeting
Report, 1978, in preparation.
238
-------
CELLULAR TOXICITY OF LIQUID EFFLUENTS FROM TEXTILE MILLS
James A. Campbell,* Neil E. Garrett,*
Joellen L. Huisingh,* Michael D. Waters*
Abstract
The rabbit alveolar macrophage (RAM} short-
term bioassay was used to determine the cellu-
lar toxicity of 23 water samples containing tex-
tile mill effluents. WI-38 human lung fibro-
blasts and Chinese hamster ovary (CHO) cells
were used for comparative cytotoxicity stud-
ies. Parameters examined in estimating cellu-
lar toxicity included: viability by trypan blue
dye exclusion; cell numbers by optical enumer-
ation; cellular ATP by luminescence assay; and
clonal growth in CHO cells. Less than 10 per-
cent of the textile mill effluent samples studied
measurably affected macrophage viability. Ap-
proximately 30 percent significantly affected
total ATP of the macrophage. RAM and WI-38
cells responded similarly to a selected sample
of four effluents. Cellular ATP content in the
RAM and WI-38 was a sensitive index of toxic-
ity. In general, CHO clonal growth was a more
sensitive index of toxicity than ATP content of
either the RAM or WI-38.
INTRODUCTION
In vitro testing of environmental effluents
from manufacturing processes can be per-
formed in a wide variety of biological systems.
The parameters evaluated in determining cel-
lular toxicity frequently include morphological
changes detectable by light and electron micro-
scopy, alteration in cell growth and division,
biochemical alterations, and cytochemical
changes. In many cases environmental toxi-
cants exert their effects through complex and
indirect mechanisms. Primary cell cultures,
such as the macrophage, exhibit many of the
metabolic and functional attributes of the orig-
*Northrop Services, Int., Cytotoxicity and Biochemistry
Section, Environmental Sciences Group, Research
.Triangle Park, NC.
'U.S. Environmental Protection Agency, Biochemistry
Branch, Environmental Toxicology Division, Health Ef-
fects Research Laboratory, Research Triangle Park,
NC.
inal'in vivo state. The rabbit alveolar «iacfo-
phage has been used to evaluate the toxicity of
a variety of environmental agents including
gases (refs. 1,2), salts of soluble metals (refs.
3,4,5), metal oxides (refs. 5,6), and environmen-
tal particulates (ref. 7). Chemical analysis has
revealed a complex mixture of organic and me-
tallic compounds in liquid effluents from textile
mills. The objective of this study was to exam-
ine the potential applicability of the RAM
system to determine the relative cellular tox-
icity of 23 water samples containing textUe mill
effluents. In order to compare the cytotoxic
responses in different cell types, four of the
textile mill effluents were also evaluated using
the diploid WI-38 human lung fibroblast and
the aneuploid CHO cell. Toxieity of the ef-
fluents was evaluated in cultures of macrt>-
phages or WI-38 cells after exposure for a 20-h
period. After incubation, cell number, viability,
and cellular adenosine triphosphate (ATP)
were measured in control and treated cultures.
Toxicity of the effluent samples to the CHO
cell was evaluated by counting the number of
colonies formed after a 6-day exposure period.
METHODS
Sample Handling
Liquid samples were stored at 4°C and
tested at concentrations of 1 to 600 fil/ml in
culture medium. Dilutions were made with sup-
plemented culture media, sample, and/or ster-
ile distilled water to maintain proper isotonic-
ity. Samples were filtered through a 0.45 /*
Millipore filter prior to testing. Experiments
were performed which indicated that toxicity
observed from filtered and unfiltered samples
was approximately equivalent.
Before cytotoxicity testing of unfiltered
samples, an antibiotic sensitivity test was con-
ducted to ascertain if antibiotics were needed
to suppress growth of any bacteria in the
sample. Nutrient agar plates were streaked
with 0.5 ml of sample and antibiotic sensitivity
disks were added for a 24-h incubation period.
239
-------
The antibiotics present in the culture medium
were capable of inhibiting bacterial growth.
RAM Testing Procedure
Procurement of Macrophages and
Treatment of Cultures—
New Zealand white rabbits of both sexes
were used. Clinically healthy animals were
sacrificed by means of 150-mg injections of
sodium pentobarbital into the marginal ear
vein. Tracheostomies were performed using
sterile operating procedures. Lung lavage was
performed in situ (ref. 1). Thirty ml of 0.85 per-
cent saline (23° C) was introduced into the
lungs and allowed to remain for 15 minutes.
After removal of this initial lavage fluid, five
subsequent infusions were introduced and re-
moved immediately. Lavage fluid found to con-
tain blood or mucous was discarded. The cells
were washed once in normal saline by centrif-
ugation at 450 x g for 15 minutes at 0° to 4°C.
After a second centrifugation, the cells were
resuspended in tissue culture Medium 199 with
Hanks Salts (0° to 4°C). All media and sera
were obtained from Grand Island Biological
Company, Grand Island, NY. Media was sup-
plemented with 10 percent heat-inactivated
fetal calf serum, 100 units/ml penicillin, and 100
/ig/ml streptomycin and kanamycin. Cells were
counted and viability determined by means of a
hemocytometer or Biophysics Cytograf. The
cell suspension was then diluted to 2.0 x 106
cells per milliliter. If cell viability was less than
90 percent the cells were not used. A differen-
tial cell count was performed with a minimum
of 200 cells being scored.
Liquid samples in media (total volume of 1.5
ml) were then added to each of three cluster
dish wells, followed by 0.5 ml of supplemented
Medium 199 (10 percent fetal calf serum and
antibiotics) containing 2 x 106 cells per milli-
liter. No pH adjustment was made for the ini-
tial testing. If a sample, toxic at initial testing,
caused a drop in pH below 6.8 or an increase
above 7.6, the sample was retested under both
pH-adjusted and unadjusted conditions. The
adjustments were made using ultrapure HC1
and NaOH to pH 7.2. After gentle mixing, the
dishes were incubated on a rocker platform for
a 20-h period at 37° C in a humidified atmos-
phere containing 5 percent CC-2 in air.
Harvest of Macrophages, Cell
Enumeration, and Biochemical Analyses—
At the end of the incubation, media contain-
ing unattached cells was removed from each
dish and maintained separately in an ice bath.
The attached cells were removed with 1.0
ml/well of 0.25 percent trypsin and combined
with the decanted media. One-tenth ml of the
combined cell suspension was removed for
ATP determinations, and 0.5 ml for viability
and cell number measurement.
Viability was determined by trypan blue dye
exclusion. For hemocytometer counts, 0.1 ml of
0.4 percent trypan blue was added to 0.5 ml of
cell suspension and counted after 5 minutes.
Simultaneous determinations of cell viability
and cell numbers were made.
Adenosine triphosphate (ATP) was deter-
mined using a Du Pont Model 760 Lumines-
cence Biometer according to a procedure sup-
plied with the instrument. Briefly, dimethyl-
sulfoxide (0.4 ml) was used to extract ATP from
a 0.1 ml aliquot of trypsinized cell suspension.
After 2 minutes at room temperature, 2.5 ml of
cold 0.01 molar morpholinopropane sulfonic
acid at pH 7.4 was added to buffer the ex-
tracted sample. The extract was then placed in
an ice bath. Aliquots of 10 /il of each extract
were injected into the luminescence biometer's
react'on cuvette containing 0.7 millimolar
luciferin, 100 units crystalline luciferase, 0.01
molar magnesium sulfate, and filtered deion-
ized H2 (total volume of 100 /d). Light emitted
from the reaction cuvette was measured photo-
metrically in the luminescence meter and was
proportional to the ATP concentration of the
sample.
Human Lung Fibroblast (WI-38) Assay
Human lung fibroblasts were obtained from
the American Type Culture Collection, Rock-
ville, Maryland. Cells were subcultivated twice
weekly by use of 0.25 percent trypsin in GIBCO
solution A at a 1:2 split ratio. Cultures were
not employed beyond the 35th subcultivation.
Cells were maintained in Basal Medium Eagle
with Earle's Salts plus 10 percent fetal calf
serum (virus screened) and 2 /tmol/ml-gluta-
mine. Cells maintained under these conditions
showed a period of rapid growth from 24 to 72
hours after subcultivation, during which time
240
-------
the experiments were performed. Antibiotics
were routinely removed from the maiuteaance
cultures to -determine thejpresetffie of contain
toting mieirooTf anisms. IFoi testing eel! viability, and ATP determinations
•p«rf0?oredS as described iprevtau&ly tot the
RAM assay.
Chinese Hainster Ovacy (CHO-KI)
Clonal Cytotoxicity Test
the Chinese Hamster Ovary ? t§ minutes. The flasks were the.0
washed with dekmized H2 and allowed to dry.
response experiments were
used to tfetewnioe an estimate of the sample
«conc&ntrations needetJ to reduce cell viability
ind tell ATP content by 20 percent (£620) or 50
percent (BCgok ECaj anil EC5.o values were ob-
tained from the dose response f elation by in-
verse prediction from the simple regression
line. All dose response data were mathemat-
ically transformed and then fitted to a straight
One by the method of least squares. Concentra-
tions weft expressed as the natural logarithm
of fche dose in ,/sJ/ml. Viability was transformed
to the ate sine of the s^ua?e WSJt of the fraction
of viable cells. ATP data expfgs&ed as a per-
cent of control were fitted t& a straight line
•with the logarithm of the dose. Confidence
limits on the estimated £620 and ECso values
were obtained using standard statistical proce-
dures tref. 8).
RESULTS AND DISCUSSION
f fte maerop-hage assay was used to evaluate
the toxicity of 23' textife mill effluents. The
response of RAM cells to a toxic effluent from
textile mill waste water (sample N) is shown in
Figure 1. After a 20-h exposure period the ef-
fect on macrophage viability, viability index*
tref. 3), ATP, and cell number was determined.
Cell viability and viability index were reduced
from 94 percent in control cultures to 58 per-
cent and 62 percent, respectively, in cultures
treated with 600 /tl/ml of sample. Sample N was
not appreciably cytolytic as evidenced by the
fact that total cell number was not affected.
Cellular ATP was the most sensitive param-
eter measured. In cultures treated with 600
pi/ml, ATP was reduced to 7 percent of the con-
trol value.
Dose respxrase eurwes as illustrated in Fig-
ure I wer® obtained and the data were used to
toSex is (fefiaed as:
•stability ittftex =» viability 4%)
asm of sells ifl test culture
MB-
is sontrol culture
2*1
-------
100
50
100
50
i —D
0 20 60
200 400
Concentration jul/ml
600
Figure 1. Response of rabbit alveolar macrophage to textile mill effluent sample N after a
20-hour exposure (• — », viability index; O — O viability; D — D cells/ml; and
• — • ATP [fg/cell]). Cells/ml and ATP are expressed as the percent of the
control value.
determine an estimate of the sample concen-
trations needed to decrease, cell viability and
cell ATP content by 20 percent (EC2o) or 50
percent (EC50). Textile mill effluents for which
either a 20 percent or 50 percent response
could be determined are shown in Table 1. The
EC20 and ECso values are accompanied by the
95 percent confidence intervals. Only two
samples, N and C, caused a significant reduc-
tion in cell viability. Duplicate experiments
with sample N were performed. Two experi-
ments were also performed with sample C.
This sample increased the pH of the cell cul-
ture medium to 8.2 at 200 /d/ml and pH 9.1 at
600 /il/ml. In one of the experiments sample C
was evaluated with the pH adjusted to 7.2 and
no toxicity was detected.
As shown in Table 1, seven samples signifi-
cantly reduced cell ATP content. This result in-
dicated that ATP, as measured by lumines-
cence assay, is a more sensitive index of
cellular toxicity than is viability as determined
by trypan blue dye exclusion. Three samples,
L, N, and C, caused a 50 percent reduction in
cell ATP. No detectable toxicity was observed
with sample C after adjustment of the pH of
the cell culture media, further indicating that
the toxicity of this sample was due to its
alkaline nature.
Textile mill samples not listed in Table 1
242
-------
SAMPLE
TABLE 1. EFFECT OF TEXTILE MILL EFFLUENTS ON RABBIT ALVEOLAR
MACROPHAGE ATP AND VIABILITY
ATP»
ECSO (Ml/ml)
EC20 (Ml/ml)
Viability*
EC60 (Ml/ml)
T
N
N
L
X
ct
F
W
102 (60-197)
128(90-188)
351 (198-849)
—
335(222-616)
—
— —
25 (.02-26)
29 (14-50)
38 (22-56)
40 (23-63)
48 (27-80)
61 (38-89)
94 (33-228)
137(89-192)
133(61458)
221 (129481)
^_
168(106-301)
—
•Estimated concentration (Ml/ml) which should result in 50 percent and 20 percent response after 20-h exposure. The 95 percent
confidence intervals are represented by numbers in parentheses.
tpH not adjusted before testing.
yielded such high estimates of sample concen-
tration necessary to reduce cell viability or
ATP that they could not be considered valid
estimates of cytotoxicity.
Less than 10 percent of the 23 textile mill ef-
fluent samples studied measurably affected
cellular viability. However, approximately 30
percent measurably affected cellular ATP in
the macrophage cultures. These results indi-
cate that measurement of cellular ATP may
form a sensitive assay for evaluation of the
cytotoxicity of complex environmental mix-
tures.
In order to further examine the cellular tox-
icity of the textile mill effluents, four samples
evaluated in the RAM assay, including the
more toxic samples (L and N), and the less toxic
samples (M and R), were selected for additional
studies. Figures 2 and 3 show the effect of the
four samples on cellular ATP content in both
the macrophage (Figure 2) and in the WI-38
human lung fibroblast (Figure 3). ATP content
of macrophages and fibroblasts was depressed
by samples L and N. Samples M and R were not
appreciably toxic in either cell system. The
similarity of response of RAM and WI-38 cells
in culture has previously been reported for
soluble metallic compounds (ref. 6). Such data
are interesting in view of the species differ-
ence and the fact that unlike WI-38 fibroblasts,
macrophages do not divide in culture. These
data suggest that a variety of cell types can be
used to provide an effective ranking of en-
vironmental effluents. In previous studies of
the cytotoxicity of metallic salts in WI-38 and
RAM cells (refs. 3,6), cellular ATP was found to
be a more sensitive index of toxicity than cell
viability. ATP was also a more sensitive index
of toxicity of textile effluents in RAM and
WI-38 cells than was cell viability. It is note-
worthy that the liquid effluents from textile
mills contain a variety of metallic compounds.
The measurement of cellular growth and di-
vision is one of the most widely used criteria of
cellular toxicity. The CHO cell system exhibits
the useful property of forming discrete col-
onies from single cells and, thus, can readily be
used to measure cellular growth. The CHO cell
system has been reported to be sensitive to a
variety of environmental agents (ref. 9) includ-
ing particulate materials (ref. 7). Toxicity of the
textile mill effluent samples was evaluated by
examination of clonal growth after a 6-day incu-
bation period. A comparison of the dose re-
sponse curves from the CHO clonal assay and
ATP content in RAM and WI-38 cells after a
20-h exposure is shown in Figure 4 for cultures
treated with sample N. As can be seen, clonal
growth of cells was a more sensitive method
for determining toxicity of sample N than
assay of ATP. It was reported previously that
measurement of clonal growth in CHO cells
was equivalent in sensitivity to assay of ATP
in the macrophage for evaluation of the toxici-
243
-------
0
100 200 300 400
Concentration (jui/mt)
500
600
Figure 2. The effect of textile mill effluent samples M (n - n), R (o - 0), L (O - O), and
N (A — A) on ATP content of the rabbit alveolar macrophage after a 20-h
exposure.
244
-------
0
0
100 200 300 400
Concentration (jul/ml)
500
600
Figure 3: The effect of textile mill effluent san ^ies M (a - a), R (0 - 0), L (O - O), and
N (A - A) on ATP content of the WI-38 human lung fibroblast after a 20-h
exposure.
245
-------
50 100
Concentration (jLil/ml)
150
200
Figure 4. The effect of textile mill effluent sample N on ATP content of the rabbit
alveolar macrophage (• - •) and WI-38 human lung fibroblast (A - A),
and clonal growth in Chinese hamster ovary cells (D — o).
246
-------
160 180 200
Concentration (jut/ml)
Figure 5. Colony survival after treatment of Chinese hamster ovary cells with textile
mill effluent sample M (D - o), L (O - O), N (A - A), and R (0 - 0).
Toxicity of the samples was evaluated by counting the number of colonies
formed after a 6-day exposure period.
247
-------
ty of some particulate materials (ref. 7).
Figure 5 shows dose response curves for
samples L, M, N, and R, which were evaluated
in the CHO clonal assay. In this system, R was
the most toxic of the four samples, reducing
colony formation to less than 2 percent of the
control value at 100 /ul/ml of effluent. Samples L
and N were about equally toxic, and sample M
was the least toxic in the CHO test. There was
no immediate explanation for the toxicity of
sample R in the CHO system since this sample
did not appreciably affect cell viability, ATP,
or cell number when tested in RAM and WI-38
cultures.
Clonal assays may offer a different measure
of relative toxicity for some environmental ef-
fluents than proliferating cultures such as the
WI-38 or nondividing cultures such as the ma-
crophage. However, in experiments aimed at
the evaluation of toxicity of environmental par-
ticulates, good agreement was seen between
the RAM and CHO systems when biochemical
or cytochemical endpoints were utilized (ref. 7).
Furthermore, the agreement seen for a num-
ber of endpoints in the WI-38 and RAM sys-
tems suggests that similar mechanisms of eyto-
toxicity may be operative in these systems.
Questions involving comparative cytotoxicity
may be resolved by additonal studies using a
variety of cell systems and toxicants.
REFERENCES
1. D. L. Coffin, D. E. Gardner, R. S. Holzman,
and F. J. Wolock, "Influence of Ozone on
Pulmonary Cells," Arch Environ Health,
Vol. 16, pp. 633-636, 1968.
2. D. E. Gardner, R. S. Holzman, and D. L.
Coffin, "Effects of Nitrogen Dioxide on Pul-
monary Cell Population," J Bacterial, Vol.
98, pp. 1041-1043, 1969.
3. M. D. Waters, D. E. Gardner, C. Aranyi,
and D. L. Coffin, "Metal Toxicity for Rabbit
Alveolar Macrophages in vitro," Environ
Res, Vol. 9, pp. 32-47, 1975.
4. M. D. Waters, T. 0. Vaughan, D. J. Aber-
nathy, H. R. Garland, C. C. Cox, and D. L.
Coffin, "Toxicity of Platinum (IV) Salts for
Cells of Pulmonary Origin," Environ
Health -Perspectives, Vol. 12, pp. 45-56,
1975.
5. M. D. Waters, D. E. Gardner, and D. L. Cof-
fin, "Cytotoxic Effects of Vanadium on
Rabbit Alveolar Macrophages in vitro,"
Toxicol Appl Pharmacol, Vol. 28, pp.
253-263, 1974.
6. M. D. Waters, J. L. Huisingh, and N. E.
Garrett, "The Cellular Toxicity of Complex
Environmental Mixtures," Proceedings of
the Symposium on the Application of
Short-Term Bioassays in the Fractionation
and Analysis of Complex Environmental
Mixtures, Williamsburg, VA, 1978, in
press.
7. N. E. Garrett, J. A. Campbell, J. L. Hui-
singh, and M. D. Waters, "The Use of
Short-term Bioassay Systems in the
Evaluation of Environmental Particu-
lates," Proceedings of the Symposium on
the Transfer and Utilization of Particulate
Control Technology, Denver, CO, July 24,
1978, in press.
8. A. Hald, "Statistical Theory with Engi-
neering Applications," New York, John
Wiley and Sons, Inc., 1952.
9. M. T. Wininger, F. A. Kulik, and W. D.
Ross, "In Vitro Clonal Cytotoxicity Assay
Using Chinese Hamster Ovary Cells (CHO-
Kl) for Testing Environmental Chemicals,"
in vitro, Vol. 14, p. 381, 1978.
248
-------
REPEATED-EXPOSURE TOXICOLOGY STUDIES
Robert J. Weir*
Abstract
In the conduct of toxicology evaluation there
are acute and repeated studies on chemicals.
Acute studies are useful in providing informa-
tion on single-dose exposures. The results of
these studies are used to compare the acute
toxicity of chemicals in terms of their LDso's,
to provide labeling information, and to provide
the industrial clinician with information to pro-
tect or treat workers in cases of accidental ex-
posure.
Repeated exposure studies are used to pre-
dict cumulative effects in man in terms of
chemical exposure of a continuous or repeated
nature. The data from these studies are used to
explore safety and risk in intended repeated ex-
posure (i.e., food additives, drugs) or in acci-
dental exposure (pesticides, environmental con-
taminants).
Repeated exposure studies are usually car-
ried out on the basis of two durations: sub-
chronic (usually 90 days duration) or chronic
(usually lifetime exposure —defined as 24
months in rats and mice). Dosages for chronic
studies are established on the basis of results
from subchronic studies.
The actual designs of repeated exposure
studies vary from product to product. This
results from the requirements on toxicologists
in the different regulatory agencies and their
own individual preferences, rather than for
any rational reason. Our test animals cannot
differentiate among drugs, pesticides, food ad-
ditives, color additives, or environmental
chemicals. Efforts are now being made to unify
study designs.
Disregarding these interagency differences,
an adequate 90-day subchronic study would
have a minimum of four groups of test animals:
one control group and three groups of test
animals. The test group dosages would vary
one from another on a logarithmic basis. The
population of animals in each group would be
no smaller than 15 males and 15 females. The
test group doses would be selected so that an
effect is characterized in at least one group.
Every effort should be made to establish a
maximum tolerated dose for chronic studies;
i.e., that dosage which produces an effect that
would not be so severe as to interfere with the
course of the chro'nic study.
Parameters measured should include: clini-
cal observations, body weight, food consump-
tion, hematology, blood chemistry, urine analy-
sis, gross pathology, microscopic pathology,
and survival rate.
An adequate chronic study would have a
minimum of four groups of test animals, one of
which is a control group. The high-dosage
group would receive a maximum-tolerated
dose. The middle-dose group would receive half
the maximum-tolerated dose. The low dose
would have a relationship to the level expected
in the environment. Each control and test
group should contain a minimum of 50 male and
50 female animals, if rats or mice are jsed as
test subjects.
The parameters measured should be the
same as those in the subchronic study. Since
the chronic study can also be an oncogenesis
(cancer) study, emphasis should be placed on
the tissue masses during observation of signs
of toxicity and at gross necropsy. Microscopic
examination should be made on approximately
40 major tissues from each animal. The tissues
should be read from all animals in the control
and high-level dose groups, showing reduced
survival, and/or the highest group, showing no
• ;duced survival. Target organs should be ex-
amined in all low-level group animals.
Let us turn our attention to textile chem-
icals, biological assessment, and the Toxic Sub-
stances Control Act (TOSCA). If textile chem-
icals are found to have biological activity in
*Vici> President, Litton Bionetics, Inc., Kensington, MD.
249
-------
EPA's Level 1 Environmental Assessment or
when they are studied under TOSCA, they will
be examined according to the above or similar
protocols.
It is extremely important that the material
of commerce associated with activity in Level 1
Environmental Assessment and studies in tox-
icological experiments be characterized chem-
ically. It is possible that the material manu-
factured may not be identical to that found in
the environment. Special chemical synthesis
and toxicological evaluation may be required
under these circumstances.
If one eliminated the question of carcino-
genesis then, almost without exception, the
results of chronic studies could be predicted
from the results of the subchronic or 90-day
studies (ref. 1). The question of carcinogenesis
could be predicted on the basis of the Ames
Test, mouse lymphoma, or malignant transfor-
mation studies with a correlation factor above
90 percent. The textile manufacturer, but not
the regulatory agency, can be well assured on
the basis of these two studies that the material
is or is not a problem without conducting the
chronic study. With further research it may
eventually be possible to improve the correla-
tion of predictive short-term studies so that
regulatory agencies will be able to accept this
philosophy and reduce time and costs as well.
REFERENCES
1. M. A. Mehlman, R. Shapiro, and H. Blumen-
thal, Advances in Modern Toxicology - New
Concepts in Safety Evaluation, Vol. 1, Part
1, Chapter 4, Hemisphere Publ. Corp.,
Washington, DC, 1976.
250
-------
ASSESSMENT OF TEXTILE WASTE TOXICITY WITH
MARINE BIO ASS AYS*
Gerald E. Walsh, Lowell H. Bahnert
Abstract
Marine bioassays were conducted with tex-
tile wastes on an alga (Skeletonema costatum),
the grass shrimp (Palaemonetes pugio), and the
sheepshead minnow (Cyprinodon variegatus).
Effects measured were inhibition or stimula-
tion of algal growth and death of animals after
96-h exposure.
All wastes affected growth of S. costatum,
whereas only five were acutely toxic to the
animals. Grass shrimp were more sensitive
than juvenile sheepshead minnows: the range
of the LC50 was 12.8 to 34.5 percent waste for
the shrimp compared to 37.5 to 69.5percent for
the minnow. Four wastes stimulated growth of
algae (SC20 = 0.50 to 2.25 percent); four in-
hibited growth (EC50 = 1.5 to 84 percent); six
were both stimulatory at low concentrations
(SC20 = 1.5 to 21.75percent) and inhibitory at
high concentrations (EC50 = 50 to 93 percent).
The SC20, a term introduced here, is the calcu-
lated concentration of waste that would stim-
ulate growth by 20percent above the control.
Algae must be considered very useful for
assessment of possible effects of textile wastes
on aquatic systems because they responded to
all wastes tested.
It is suggested that bioassays be used to
assess the potential impact of whole waste on
aquatic systems; and after chemical analysis of
the waste, bioassays be performed on compo-
nents that may be bioactive. If the whole waste
or any of its components has an effect, then
that waste must be considered for application
of treatment technology.
INTRODUCTION
Population growth and the rate of industrial
development along coastlines in the United
'Contribution No. 376 from the Environmental Research
Laboratory, Gulf Breeze, FL.
TU.S. Environmental Protection Agency, Environment-
al Research Laboratory, Gulf Breeze, FL.
States have been accelerating within recent
years, and acceleration is expected to continue
in the forseeable future. In 1972 the U.S. De-
partment of Commerce (ref. 1) predicted that
the total human population density around 39
cities on the Atlantic, Pacific, and Gulf Coasts,
and in Hawaii would double between the years
1975 and 2020 (Figure 1). Concomitantly, total
manufacturing around and in those cities was
predicted to rise approximately 5.5 times and
textile products, approximately 4.5 times.
The possible ecological impact of such
growth upon estuarine and marine ecosystems
is great because industrial plants often emit
large volumes of effluents that contain bioac-
tive organic and inorganic substances. Also,
because the population of coastal areas will in-
crease, the volume of sewage and municipal
wastes discharged will increase, thus placing a
double pollution burden upon estuarine and
marine ecosystems.
Since estuaries are important commercial,
recreational, and aesthetic resources, it is im-
portant to know effects of wastes upon them
and to apply the proper treatment systems.
This paper discusses use of bioassays to evalu-
ate potential impacts of textile industry wastes
on estuarine biota. The tests were designed to
screen many wastes in a short period of time
and to rank their toxicities according to effects
on algae, shrimp, and fish. More sophisticated
bioassays are required under stringently con-
trolled conditions and with statistical analyses
to define toxicity more precisely. Such tests
will be done in IERL-RTP Level 2 bioassays.
The simplest, most direct, and inexpensive
method for testing effects of complex indus-
trial wastes upon estuarine organisms is
through exposure to waste in static tests. In
the work reported here, three species repre-
sentative of three major classes of estuarine
organisms were used:
1. Skeletonema costatum, a chain-forming
diatom that occurs in inshore and open-
ocean waters, is an algal species that has
been sensitive to a variety of pollutants
in our laboratory;
251
-------
2.00 r
1.00
1975
1985
2000 2010
YEAR
2020
Figure 1. Projected population growth in 39 coastal cities,
with 1975 as the base year.
Source: U.S. Department of Commerce (1972).
2. Palaemonetes pugio, the grass shrimp, is
an arthropod that occupies a broad geo-
graphical range along the Atlantic and
Gulf Coasts and is easy to maintain in
the laboratory; and
3. Juveniles of Cyprinodon variegatus, the
sheepshead minnow, were used because
juvenile fish are often more sensitive to
pollutants than adults and because C.
variegatus is common on the Atlantic
and Gulf coasts.
Since industrial wastes often contain mate-
rials that are toxic to algae and animals, and
nutrients that stimulate algal growth, labora-
tory results must be interpreted in relation to
death of animals, inhibition of algal growth,
and stimulation of algal growth with the possi-
ble establishment of eutrophic conditions. A
waste that has any of these effects in labora-
tory tests may affect an estuarine ecosystem
by changing relative numbers and kinds of
species, creating an assemblage of organisms
that reduces its resource value.
In this paper, we identify and rank those tex-
tile plant wastes that would be likely to cause
undesirable changes in estuarine ecosystems.
METHODS
Grab samples of wastes were distributed
under ice in glass containers by an IERL-RTP
contractor who also assigned code letters.
Each species was tested with 14 wastes. Skele-
tonema costatum was tested with Wastes B, C,
F, G, K, L, N, P, S, T, U, V, W, and X; P. pugio
and C. variegatus were tested with A, B, C, E,
F, G, K, L, N, S, T, U, W, and X. Details of
methods used are described in the U.S. Envi-
ronmental Protection Agency publication,
IERL-RTP Procedures Manual: Level 1 (ref.
2). A general overview of the methods is given
here.
Algae
Stocks of S. costatum were maintained in an
252
-------
TABLE 1. COMPOSITION OF MIXES ADDED TO
ALGAL GROWTH MEDIA
Nutrient mixes
* Add 15.0 ml/I to test solution.
§ Add 0.50 ml/1 to test solution.
1 Add 1.0 ml/1 to test solution.
Amount/liter
Metal mix*
FeCI2x6H20
MnCI2x4H20
ZnS04x7H20
CuS04x5H20
CoCl2x6H20
H3B03
Deionized water
Vitamin mix§
Thiamin hydrochloride
Biotin
B12
Deionized water
Minor salt mix'
K3P04
NaN03
Na2Si03x9H20
Deionized water
0.480 g
0.144 g
0.045 g
0.1 57 nig
0.404 mg
0.140 g
1 liter
50 mg
0.01 mg
O.IOmg
100ml
3.0 g
50.0 g
20.0 g
1 liter
artificial seawater medium (Rila® Salts*) of 30
parts per th
-------
burgh, PA). The absorbanee values of uninocu-
lated blank tubes were subtracted from those
of the inoculated tubes to give the absorbanee
due to algae. Optical density is closely related
to biomass (Walsh, unpublished), and this
method was used for both range-finding and
definitive tests.
Stimulation or inhibition of algal growth was
calculated by plotting the nonlinear regression
of absorbanee on waste concentration by an
equation developed for these studies:
1 + B (X-C) 1 + D (X-E)
where
A =
B =
C =
D =
E =
X =
Y =
mean maximum biomass
increasing slope
effluent concentration (increasing
slope) where Y = 0.5 x A
decreasing slope
effluent concentration (decreasing
slope) where Y = 0.5 x A
range of effluent concentrations tested
estimated population density over
range of X.
Calculations were made on a Digital Equip-
ment Corporation PDF 11/45 computer. The
concentration of waste that stimulated growth
by 20 percent (SC20) or inhibited growth by 50
percent (EC50) in relation to control cultures
was calculated from the regression curve.
Shrimp and Fish
Palaemonetes pugio and C. variegatus
juveniles were collected from the field and held
in flowing water for at least 2 days while being
acclimated to a salinity of 10 ppt. Each test was
performed in 3 liters of test solution in 4-liter
glass jars. Five animals were tested per jar
and all exposures were done in duplicate in
definitive tests. The jars were not aerated, but
oxygen content was measured each day by the
Winkler method. Temperature was 20 ± 1.0°C.
Salinity of the effluents was increased by 10
ppt by addition of Rila® Salts. Control and dilu-
tion water were prepared from Rila® Salts and
glass-distilled water.
Each effluent was subjected to a 48-h range-
finding test. If no deaths occurred in undiluted
waste, the exposure was continued to 96 hours.
If mortality was equal to or less than 50 per-
cent after 96 hours, no further tests were done.
If mortality was greater than 50 percent, 96-h
definitive tests were done.
Concentrations of wastes that would kill 50
percent of the animals in 96 hours (LC50) were
calculated by straight-line graphical interpola-
tion. Points representing the number of deaths
in concentrations lethal to more than one-half
and less than one*half of the animals exposed
were plotted on semilogarithmic coordinate
paper. The concentration at which a straight
line drawn between two points crossed the 50
percent mortality line was the estimated LC50.
RESULTS
Algae
Skeletonema costatum was a valuable
organism for detection of possible impacts of
textile industry wastes on natural waters. All
of the 14 wastes tested affected growth; 4 were
inhibitory only, 4 were stimulatory only, but 6
were stimulatory at low concentrations and in-
hibitory at high concentrations (Table 2).
The patterns of growth response in relation
to waste concentrations are given in Figures 2
to 16. Wastes B, S, U, and X were stimulatory
only; Wastes F, L, N, and P were inhibitory on-
ly; and Wastes C, G, K, T, V, and W were both
stimulatory and inhibitory. When inhibition
and stimulation occurred, the SC20 was much
lower than the EC50.
Shrimp
Only 5 of the 14 textile effluents caused
greater than 50 percent mortality of P. pugio
in concentrations up to 100 percent (Table 2).
The most toxic was Waste C (LC50 = 12.8 per-
cent), followed by Wastes W (19.6 percent), A
(21.2 percent), N (26.3 percent), and T (34.5 per-
cent). Three effluents caused mortality of less
than 50 percent and no LC50 could be calcu-
lated. The six remaining wastes (B, F, K, L, U,
and X) caused no deaths when grass shrimp
were exposed to 100 percent effluent for 96
hours.
The percentage of oxygen saturation of the
test solutions was never lower than 40. There-
fore, deaths were due directly to the wastes
and not to oxygen deficiency. There was no
mortality in the controls.
254
-------
TABLE 2. COMPARISON OF RESPONSE OF
SKELETONEMA COSTATUM. PALAEMONETESPUGIO. AND
, CYPRINODON VARIEGATUS TO TEXTILE WASTES
S. costatum
Waste
C
N
T
W
A
B
F
G
K
L
S
U
V
X
E
P
SC20
1.50
NE
2.00
1.50
—
0.50
NE
2.75
1.00
NE
2.25
1.50
21.75
0.50
—
NE
EC50
76
2.0
66
50
—
NE
84
59
75
1.5
NE
NE
93
NE
-
9.0
P. pugio
LC50
12.8
26.3
34.5
19.6
21.2
NE
NE
*
NE
NE
#
NE
NE
NE
t
—
C. variegatus
LC50
69.5
47.5
68.0
37.5
62.0
NE
NE
t
NE
*
*
NE
NE
NE
—
*20 percent mortality in 100 percent waste
"Uo percent motality in 100 percent waste
*50 percent mortality in 100 percent waste
§ 10 percent mortality in 100 percent waste
Notes: SC20 = percentage waste at which growth was stimulated by
20 percent;
EC50 = percentage waste at which growth was inhibited by
50 percent;
LC50 = percentage waste lethal to 50 percent of the animals;
NE = no effect.
Fish
Acute toxicities of the 14 textile effluents to
juvenile sheepshead minnows are shown in
Table 2. Only five wastes were toxic enough to
allow calculation of the LC50 (Waste W, 37.5
percent; N, 47.5 percent; A, 62.0 percent; T,
68.0 percent; and C, 69.5 percent). Mortalities
in Wastes E, G, L, and S were not great enough
to calculate LC50 values, and Wastes E, F, K,
U, and X had no effect on survival in 96 hours.
As in the shrimp tests, percentage satura-
tion of oxygen in test solutions was never
lower than 40, so deaths were not due to lack of
oxygen. There was no mortality in the controls.
DISCUSSION
The alga S. costatum was the most sensitive
organism in these tests: all wastes affected
growth of the alga, whereas only five were
acutely toxic to P. pugio and C. variegatus.
Table 2 contrasts responses of the three orga-
nisms to the wastes. These results were unex-
pected, since animals are usually more suscep-
tible to single pollutants than algae, and it
seems that algae are better indicators of ef-
fects of some complex wastes than are animals.
Also, the stimulatory effects of some wastes on
algae indicate that industrial wastes may have
a greater effect in bringing about eutrophica-
tion than previously realized.
Palaemonetes pugio was more sensitive to
the wastes than C. variegatus, but neither was
of value for ranking of all wastes because tox-
icity was detected in only five effluents. As
shown in Table 3 the two species also respond-
ed differently.
255
-------
8
7
6
5
4
u 3
i2
10
r-
o>
hi
3 9
> 6
Q
H 7
I 6
^ 5
4
3
2
RW>ER AND ALUEO PRODUCTS
1985 2OOO
YEAR
TOTAL MANUFACTURING
2010 2020
1975
1985
2000 2010 2020
9
8
7
6
si
4
3
2
I
9-
8
7
6
5
4
3
2
TEXTILE MLL PRODUCTS
1975 1985 2OOO 2010
YEAR
CHEMICAL AND ALLIED
PRODUCTS
2020
YEAR
1975 1985 2000 2OJO
YEAR
2020
Figure 2. Projected industrial growth in 39 coastal cities, with 1975 as the base year.
Source: U.S. Department of Commerce (1972).
256
-------
20 40 60
PER CENT EFFLUENT
80
100
Figure 3. Effects of Textile B effluent on growth of Skeletonema costatum.
257
-------
.300
.250
PER CENT EFFLUENT
Figure 4. Effects of Textile C effluent on growth of Skeletonema costatum.
258
-------
.300
.250
z .200
UJ
O
.I50
.100
.050
20 40 60
PER CENT EFFLUENT
80
100
Figure 5. Effects of Textile F effluent on growth of Skeletonema costatum.
259
-------
20 40 60
PER CENT EFFLUENT
80
160-
Figure 6. Effects of Textile G effluent on growth of Skeletonema costatum.
260
-------
20 40 60
PER CENT EFFLUENT
80
Figure 7. Effects of Textile K effluent on growth of Skeletonema costatum.
261
-------
PER CENT EFFLUENT
8
10
Figure 8. Effects of Textile L effluent on growth of Skeletonema costatum.
262
-------
.300
.250
>
£.200
2 4 6
PER CENT EFFLUENT
8
10
Figure 9. Effects of Textile N effluent on growth of Skeietonema costatum.
263
-------
246
PER CENT EFFLUENT
8
10
Figure 10. Effects of Textile P effluent on growth of Skeletonema costatum.
264
-------
20 40 60
PER CENT EFFLUENT
80
100
Figure 11. Effects of Textile S effluent on growth of Skeletonema costatum.
265
-------
PER CENT EFFLUENT
Figure 12. Effects of Textile T effluent on growth of Sketonema costatum.
266 _
-------
.300
20 40 60
PER CENT EFFLUENT
80
100
Figure 13. Effects of Textile U effluent on growth of Skeletonema costatum.
267
-------
PER CENT EFFLUENT
Figure 14. Effects of Textile V effluent on growth of Skeletonema costatum.
268
-------
20 40 60
PER CENT EFFLUENT
80
100
Figure 15. Effects of Textile W effluent on growth of Skeletonema costatum.
269
-------
.300
20 40 60
PER CENT EFFLUENT
Figure 16. Effects of Textile X effluent on growth of Sketetonema costatum.
270
-------
TABLE 3. RELATIVE TOXICITY OF FIVE TEXTILE
WASTES TO P. PUGIO AND C. VARIEGATUS
Degree of
toxicity
Most toxic
Least toxic
P. pugio
C
W
A
N
T
Waste
C. variegatus
W
N
A
T
C
As the table reflects, Waste T was least tox-
ic to P. pugio, with the LC50 of 34.5 percent.
Waste W was most toxic to C. variegatus, but
its LC50 was 37.5 percent (Table 2). Therefore,
for animals, ranking of wastes is according to
toxicity to P. pugio.
In contrast to the animal data, all wastes can
be ranked according to degree of effect in the
algal test~ although the ranking by stimula-
tion is di Cerent from that by inhibition (see
Table 4).
TABLE 4. RELATIVE RESPONSES OF
S. COSTATUM TO 14 TEXTILE WASTES
Degree of
effect
Greatest effect
Least effect
Waste
Stimulation
B,X
K
C.U.W
T
S
G
V
F,L,N,P
Inhibition
L
N
P
W
G
T
K
C
F
V
B,S,X,U
As shown in Table 4, there is a rough inverse
correlation between stimulation and inhibition,
although Waste W ranks high on both lists. In
general, wastes that were highly stimulatory
were not highly inhibitory, and those that were
highly inhibito* .' were not highly stimulatory.
It may be inierred from the data that the
response of algae to wastes that caused stimu-
lation at low concentrations and inhibition at
high concentrations is related to relative
amounts of nutrients and toxicants. If true,
high concentrations of either may mask effects
of the other in some wastes and the potential
impact of those wastes on aquatic systems
could not be evaluated by simple bioassays.
Also, responses of organisms in our bioassays
could be the result of synergistic, additive, and
antagonistic interactions between components
of the complex wastes.
Results of bioassays were analyzed in rela-
tion to chemical composition of each waste
reported by EPA (1978) (ref. 3). Neither stimu-
lation nor toxicity could be related to color,
BOD, COD, or to concentrations of sulfide,
phenol, suspended solids, cyanide, organic car-
bon, total organics extracted with methylene
chloride, metals, ammonia, nitrite, nitrate,
Kjeldahl nitrogen, ortho-phosphate, or total
phosphorus.
It is not surprising that chemical analyses
did not indicate toxicity of wastes. For exam-
ple, we have found (unpublished data) that the
EG50 of nickel and aluminum to several algae
is approximately 2 parts per million (ppm) in
the absence of a chelator. In the presence of
each other, or of an organic material such as
the pesticide 2,4-D, toxicities of the metals
were reduced. Waste X contained 7 ppm nickel
and 640 ppm aluminum in suspended solids and
0.006 ppm nickel and 8.2 ppm aluminum in fil-
tered effluent—enough to inhibit growth com-
pletely. It also contained 45 other metals and
diethyl phthalate; bis (2-ethylhexyl) phthalate;
hexachlorobenzene; 1,1,1-trichloroethane;
toluene; ethylbenzene; 1,1,2,2-tetrachloroethy-
lene; and phenol. In spite of this large number
of metals and organic compounds, Waste X did
not inhibit algal growth or kill animals. It did,
however, stimulate algal growth.
Our toxic waste data suggest that effects of
a complex effluent are the result of interac-
tions among toxicants and between toxicants
and growth stimulators. Potential impact of a
complex waste on aquatic systems may or may
not be predicted from either chemical or bio-
logical tests alone. It must be assumed that,
after a waste is dumped into a receiving body,
the individual components separate spatially
and the total impact may be greater than that
indicated by laboratory bioassay.
271
-------
We suggest that the potential impact of com-
plex industrial wastes be assessed in two ways:
(1) bioassays should be used to assess the po-
tential impact of the whole waste on aquatic
systems; and (2) after chemical analysis of each
waste, bioassays should be performed on com-
ponents that may be bioactive. If the whole
waste or any of its components has an effect,
then that waste must be considered for applica-
tion of treatment technology.
Wastes reported here that stimulated algae
at low concentrations (B, C, G, K, S, T, U, V, W,
and X), were toxic to algae at low concentra-
tions (L, P, and N) or were toxic to animals
would be ranked high on the list of those that
require treatment. Wastes C and W were stim-
ulatory to algae and toxic to animals. The com-
bination of algal, shrimp, and fish bioassays in-
dicate that all wastes, with the possible excep-
tion of E, require treatment. However, even
Waste E caused 40 percent of the shrimp to die
and is suspect.
CONCLUSIONS
1. Algae were more sensitive indicators of po-
tential environmental effects of textile
wastes than were shrimp or fish.
2. Algal response to each waste was either (a)
inhibition of growth, (b) stimulation of
growth, or (c) stimulation at low concentra-
tions and inhibition at high concentrations.
Fish and shrimp were killed by some
wastes.
3. Potential impacts of the wastes on aquatic
systems were death of animals, inhibition
of algal growth, and stimulation of algal
growth, which can lead to establishment of
eutrophic conditions.
4. It is suggested that, after complex wastes
enter receiving waters, their constituents
may be separated spatially, so that effects
of individual substances may occur at sites
removed from the source.
5. All wastes tested were bioactive toward
algae. Since bioactivity of the complex mix-
tures could not be related to chemical com-
position, we suggest: (a) bioassays should
be used to assess the potential impact of
the whole waste on aquatic systems, and (b)
after chemical analysis, bioassays should
be performed on components that may be
bioactive to aid in application of treatment
technology.
ACKNOWLEDGMENTS
We thank Sharon Edmisten and Shelley
Alexander for their technical assistance. Bio-
assays on P. pugio and C. variegatus were done
at the EG&G Bionomics Marine Laboratory,
Pensacola, Florida, under EPA Contract
68-02-1874.
REFERENCES
1. U.S. Department of Commerce, Population
and Economic Activity in the United
States and Standard Metropolitan Statisti-
cal Areas, Historical and Projected, 1950-
2020, Bureau of Economic Analysis,
Washington, DC., PB-216607, distributed
by National Technical Information Service,
Springfield, VA, 544 pp., 19"fc.
2. U.S. Environmental Protect n Agency,
IERL-RTP Procedures Manual Level 1,
"Environmental Assessment Biological
Tests for Pilot Studies," Industrial En-
vironmental Research Laboratory,
Research Triangle Park, NC, EPA-600/7-
77-043,106 pp., 1977.
3. U.S. Environmental Protection Agency,
"Source Assessment: Textile Plant Waste-
water Toxics Study," Phase I, Industrial
Environmental Research Laboratory,
Research Triangle Park, NC, 1978.
272
-------
FRESHWATER TOXICITY TESTS RELATED TO
THE TEXTILE INDUSTRY
William B. Horning, IP, Timothy W. Neiheiselt
Abstract
Freshwater toxicity tests were conducted on
samples of secondary treated wastewaters
from 23 textile mills as part of a joint U.S. En-
vironmental Protection Agency (EPA) and
American Textile Manufacturers Institute
(ATMI) study of the textile industry. The pri-
mary tests were (1) 96-A static acute toxicity
tests with fathead minnows, (2) 48-h static
acute toxicity tests with Daphnia pulex and (3)
the algal assay test. These tests were among a
series of ecological screening tests for potential
aquatic toxicity. The sublethal gill-purge rate
(GPR) test with bluegill sunfish was conducted
as a supplemental test on four of the textile-
mill effluent samples.
Thirteen of the textile samples indicated
acute toxicity. Seven samples were acutely tox-
ic to both fathead minnows and Daphnia pulex.
One was acutely toxic to fathead minnows on-
ly, and five were acutely toxic only to Daphnia
pulex. Ninety-six-hour LCgQ values ranged
from 19.0 to 64.7 percent effluent for fathead
minnows. Forty-eight-hour EC$o values ranged
from 0percent survival in all dilutions to 81.7
percent effluent for Daphnia pulex. In all cases,
most of the toxicity was exerted within the
first 24 hours. Extending the length of test
periods beyond 24 hours did not appreciably
alter LC50 or EC^ values.
The textile effluents were given two separate
relative rankings for each of the two species
tested. One was based on toxicity and consi-
dered only the fathead minnow LCgQ and the
Daphnia pulex 48-h ECgo values. The other in-
cluded both toxicity and the waste discharge
volume, which were used to calculate a daily
toxicity discharge expressed as million lethal
units/day (MLU/d). The order in the two rank-
*Chief, Newtown Fish Toxicology Station, U.S. EPA,
Cincinatti, OH.
tResearch Aquatic Biologist, Newtown Fish Toxicology
Station, U.S. EPA, Cincinatti, OH.
ing schemes was different. Several samples, in-
dicating relatively high acute toxicity (low
I/C50 values), are shown on a million lethal
units/day basis to be of less potential envi-
ronmental hazard than some other samples ex-
hibiting relatively low acute toxicity.
In the algal assay test, textile effluent
samples that were acutely toxic to fathead min-
nows, Daphnia pulex, or both, either stim-
ulated or inhibited algal cultures in 20 percent
dilution or less. Six were inhibitory and seven
were stimulatory. Of the four samples eval-
uated with the GPR test, only one elicited a
response.
INTRODUCTION
In 1971, the U.S. Environmental Protection
Agency published a state-of-the-art document
for waste treatment in the textile industry (ref.
1). Also a development document, the Develop-
ment Document for Effluent Limitations Guide-
lines and New Source Performance Standards
for the Textile Mills was published by the U.S.
Environmental Protection Agency in 1974 (ref.
2). Neither document addresses the issue of
textile waste toxicity. Both discuss various
waste treatment systems and their appli-
cability to the textile industry. The develop-
ment document speaks in terms of effluent
quality that can be attained by applying best
practical control technology currently avail-
able (BPCTCA). The guidelines were based on
this premise, and treatment to meet them was
to be achieved by July 1, 1977. Treatment to
achieve best available technology economically
achievable (BATEA) by July 1, 1983, was
recommended. However, all of the technologies
considered are concerned primarily with re-
ducing the BOD5, COD, total suspended solids,
oil and grease, color, chromium, sulfide,
phenols, fetil coliforms, pH, acidity, and
alkalinity being discharged into navigable
waters. No attempt was made to develop
guidelines to include limitations based on the
toxicity of textile wastes.
. 273
-------
The U.S. Environmental Protection
Agency's Industrial Environmental Research
Laboratory at Research Triangle Park, North
Carolina, (IERL-RTP) became involved in a
joint study with the American Textile Manu-
facturers Institute (ATMI) (EPA Grant No.
804329) to determine BATEA for textile plant
wastewaters. This study focused on only a
limited number of pollutants, i.e., BOD5, COD,
color, sulfides, total suspended solids, phenol,
and pH. Twenty-three textile mills having well-
operated secondary wastewater treatment
facilities and representing eight textile-
processing categories were selected for the
study.
A consent decree (resulting from National
Resources Defense Council et al. versus Train)
was issued by the U.S. District Court of Wash-
ington, D.C., on June 7, 1976. The decree re-
quired EPA to enhance development of ef-
fluent standards and focused Federal water
pollution control efforts on potentially toxic
and hazardous pollutants. As a response EPA
developed a list of 129 specific compounds
(priority pollutants) to be considered during
the standards-setting process. EPA decided to
conduct a parallel study to the ATMI/EPA
grant study, in which the objective was to
determine both the removal efficiencies for the
129 consent decree priority pollutants and the
reduction in toxicity by six tertiary treatment
technologies being investigated under the
original grant study.
Phase I of the overall textile wastewater
study was designed to establish baseline data
for toxicity and the level of priority pollutants
present in raw wastewater and secondary ef-
fluents at the 23 textile plants (ret. 3). These
data were used to screen the 23 plants and to
select those plants with toxic effluents for fur-
ther study. It is important to note that the tox-
icity tests were designed only to evaluate the
reduction in wastewater toxicity by control
technologies, and not to evaluate potential en-
vironmental impact of these wastes on receiv-
ing waters.
This paper presents a discussion of results
from static acute toxicity tests with fathead
minnows and Daphnia pulex and the algal
assay bottle tests. These tests were among a
battery of ecological and health effects screen-
ing tests that were used to evaluate toxieity of
the textile wastewaters. The sublethal gill
purge rate (GPR) test with bluegill sunfish was
also performed on several samples at the New-
town Fish Toxicology Station as a supple-
mentary test and part of our own research pro-
gram on complex effluents.
In order to evaluate the potential environ-
mental hazard of textile plant wastewater
discharges, the acute toxicity data were com-
bined with the effluent discharge volume to
provide a daily toxicity discharge rate labelled
million lethal units per day. This rate allows a
comparison of the relative importance of both
toxicity and volume of discharge.
METHODS
Eight individual grab samples of secondary
effluent were collected at equally spaced time
intervals during the normal working day at
each textile mill. Average hydraulic retention
time in the various wastewater treatment
plants ranged from 1 to 30 days, with an aver-
age of 5 days. Equal portions of each of the
eight grab samples were combined to make one
composite sample for chemical and biological
analyses. Chemical analyses included the crite-
ria pollutants and the 129 priority pollutants.
Aliquots of the composite samples were iced
and shipped air freight to the Newtown Fish
Toxicology Station, Cincinnati, Ohio, and to the
EPA Environmental Research Laboratory,
Corvallis, Oregon.
Ninety-six- and forty-eight-h static acute
toxicity tests were conducted with fathead
minnows and Daphnia pulex, respectively, at
the Newtown Fish Toxicology Station. These
tests were performed according to the pro-
tocols described by Duke (ref. 4). The Algal
Assay Bottle Test, using Selenastrum capricor-
nutum Prints as the test organism, was per-
formed at the Environmental Research Labo-
ratory, Corvallis, Oregon according to proce-
dures essentially described by Duke.
Seven effluent concentrations, ranging from
100 to 7.8 percent, and a control were used in
the static acute toxicity tests. Probit or
moving-average-angle statistical procedures
were used to calculate estimated LCso and
ECso values from the static acute toxicity
tests. For comparative purposes in this presen-
tation, the acute 96-h LCg, and 48-h ECso
values were quantified to a daily toxicity
discharge expressed as million lethal units per
274
-------
day. Quantification can be achieved by defining
lethal units per gallon (LU/g) as follows:
LU/g
100%
LC50 or ECso (in percent)
then,
LU/g x effluent flow (mgd) = MLU/d.
Thus, the toxicity data are quantified into
units comparable to pounds of BOD5. The
analogy between BOD5 and toxicity control is
essentially identical because in each case an
organism's response is used to measure the
combined effects from unknown materials in
unknown quantities.
The algal growth response was measured as
net biomass produced. The response was ex-
pressed as percent stimulation or inhibition at
the 20 percent wastewater concentration, as
compared to growth in a control sample.
The gill purge rate test was done with only
four of the textile-mill effluents used in the
static acut'- toxicity tests. The gill purge rate
is determined by placing individual fish in a
chamber having a stainless steel electrode at
each end. The bioelectric potential associated
with breathing (ventilation) movement is elec-
tronically recorded, and strip chart recordings
are visually counted for gill purges. The gill
purges show up as a higher than usual spike in
the normal breathing record or as a spike with
a double peak.
The equipment and procedures used in the
present study were similar to those described
by Carlson and Drummond (ref. 5) with the
following exceptions. The test concentrations
were 75, 25, 8.3, 2.7, and 0.9 percent effluent
plus a control. Four fish were exposed in 4
liters of solution at each concentration and con-
trol water. Gill purge rates of each fish were
determined for five 10-min periods over a 12- to
24-h preexposure period. Gill purge rates were
determined in the same manner between 12
and 24 hours after initial exposure and again
between 36 and 48 hours. Dunnett's 1-sided
t-test was used to determine a significant
change in gill purge rate at the 0.05 probability
level (ref. 6).
RESULTS AND DISCUSSION
The acute effects of the 23 textile plants
TABLE 1. TEXTILE PLANT WASTEWATERS: STATIC ACUTE TESTS INDICATING
TOXICITY AND THE ALGAL ASSAY BOTTLE TEST
Fathead minnow
Effluent
sample
A
C
E
F*
G
L*
M
N
T*
U
V
W
z
96-hr LC50
% effluent
19.0
46.5
NAT
NAT
64.7
23.5
NAT
48.8
46.5
NAT
36.0
55.2
NAT
MLU/d
1.1
2.2
_
_
3.2
3.2
2.1
1.3
_
2.2
0.5
-
Daphnia
48-h EC50
% effluent
9.0
41.0
7.8
81.7
62.4
28.0
60.0
<7.8t
NAT
12.1
9.4
6.3
42.6
pulex
MLU/d
2.2
2.4
25.6
2.4
3.2
2.7
15.3
-
—
2.5
8.5
4.0
4.6
Algal assay bottle test
Inhibition/
stimulation
I
S
I
S
S
1
S
1
S
S
S
i
1
% effluent
20
20
20
20
20
10
20
20
on
20
Aft
20
nn
20
1 n
10
on
20
* gill purge test samples.
t100 percent dead in all test dilutions.
Note: NAT = no at ite toxicity; I = inhibition; S = stimulation.
275
-------
wastewaters are summarized and discussed in
the EPA publication, Source Assessment: Tex-
tile Plant Wastewater Toxics Study, Phase I
(ref. 3). Briefly, the data indicate that 8 textile
effluents were acutely toxic to the fathead min-
now and 12 were acutely toxic to Daphnia
pulex (Table 1). Effluent from one plant (T) was
acutely toxic to the fathead minnow and not to
Daphnia pulex. Effluent from five textile
plants was acutely toxic to Daphnia pulex and
not to the fathead minnow. In five of the seven
cases where the wastes were toxic to both
species, the sensitivity of the two tests was
essentially the same.
Direct comparison of the algal assay bottle-
test data with the animal-test data is not possi-
ble because the highest dilution tested with
algae was only 20 percent. However, as shown
in Table 1, the 13 samples that were acutely
toxic to the fathead minnow, Daphnia pulex, or
both, either inhibited or stimulated algal
growth at 20 percent dilution or less. Six of
these wastes were inhibitory, and seven were
stimulatory. Additionally, algal growth was
stimulated by the nine samples that did not
acutely affect fathead minnows or Daphnia
pulex. Thus, a battery of tests, using different
organisms, better represents the effects of a
waste on aquatic organisms than a single
species test.
Closer examination of the data suggests
several significant aspects frequently over-
looked when interpreting results from acute
toxicity tests. In all cases where the textile ef-
fluents were acutely toxic, most of the toxicity
was exerted within the first 24 hours (Table 2).
Extending the length of the test periods
beyond 24 hours did not significantly affect the
LC5Q or ECso values. Eleven previous onsite
studies, representing eight different industrial
categories, have shown similar results with
well-designed and -operated industrial waste-
water treatment systems (Table 3). Thus, there
are good indications that 24-h acute toxicity
tests are sufficient for initial screening pur-
poses.
The textile effluent samples can be ranked
on a relative basis from most toxic to least tox-
ic based on fathead minnow LC50 and/or
Daphnia pulex EC50 values (Table 4). However,
this ranking system does not necessarily* in-
dicate the relative potential hazard of the in-
dividual effluents to the aquatic environment.
When the effluents are ranked, based on mil-
TABLE 2. TEXTILE PLANT WASTEWATERS: ACUTE TOXICITY TIME COMPARISON
Effluent
sample
A
C
E
F
G
L
M
N
T
U
V
w
z
24-h LC5Q
% effluent
24.6
46.5
NAT
NAT
68.2
37.5
NAT
55.2
63.1
NAT
39.1
68.2
NAT
Fathead minnow
48-h LC50
% effluent
19.0
46.5
NAT
NAT
68.2
32.5
NAT
48.5
46.5
NAT
37.5
66.4
NAT
Daphnia pulex
96-h LC50
% effluent
19.0
46.5
NAT
NAT
64.7
23.5
NAT
48.8
46.5
NAT
36.0
55.2
NAT
24-h EC50
% effluent
15.0
41.0
12.6
91.8
64.4
36.0
60
<7.8*
NAT
20.5
11.5
could not
calculate
63.4
48-h EC50
% effluent
9.0
41.0
7.8
81.7
62.4
28.0
60
<7.8*
NAT
12.1
9.4
6.3
42.6
*100 percent dead in all test dilutions.
Note: NAT = no acute toxicity.
276
-------
TABLE 3. INDUSTRIAL PLANT WASTEWATER: INCREASES IN TOXICITY
TO FISH BETWEEN 24 AND 96 HOURS IN STATIC ACUTE TESTS
OF TREATED EFFLUENT FROM 11 INDUSTRIAL PLANTS
Effluent source
!!• II •lll»..» • ^.l.» •!! !•!•• •!-
Rubber-rainwear
plant
Oil refinery
Metal-plating
company A
Metal-plating
company B
Paper company A
Paper company B
Glass works
Aluminum company B
Carpet mill
Organic chemical
plant A
Organic chemical
plant B
Number of
samples tested
— • • — • i
, 4
10
18
9
2
2
2
4
6
3
5
Number of samples
showing toxicity
24-h 96-h
" ii
4
1
3
6
2
0
0
0
2
3
0
4
3
3
6
2
2
0
0
4
3
2
Number of
samples showing
increased toxicity
* " - - !•
4
3
2
0
0
2
0
0
3
3
2
Number of samples
statistically showing
a significant increase
-" - • i in
0
2
1
0
0
0
0
0
0
2
0
lion lethal units/day (total amount of toxicity
discharged), the relative order is significantly
altered (Table 4). For example, effluent A is
shown to be one of the most acutely toxic
(LCso = 19.0 percent) to fathead minnows,
based on LC^ values, but is less of a potential
threat to the environment on a million lethal
units/day basis. The opposite situation occurs
with effluent G (LC50 = 64.7 percent). (The dif-
ferences between the two ranking systems are
more striking with Daphnia pulex.)
The lethal-unit concept considers the units of
toxicity as being analagous to BOD50. This pro-
vides a means for evaluating the capabilities of
a wastewater treatment system to remove
total acute toxicity as is done in determining
BOD50 reductions. All that is required is deter-
mining toxicity before and after treatment,
quantifying the data into lethal units, and mak-
ing a comparison. Additionally, million lethal
units/day appears to be a realistic means of ex-
pressing the potential environmental impact
for a specific discharge on the receiving
stream. To assess the acutal impact of a waste
in a given situation, receiving-stream flow
must be considered.
The gill purge is considered to be part of the
normal gill cleaning reflex of fish, in which
water is forcefully and rapidly flushed across
the gills in the same direction as in normal
breathing or in the reverse direction. The gill
purge rate has been shown to increase when
fish are exposed to suspended solids, some
pure chemicals, and complex effluents. The
response can be observed visually as well as by
recording cl anges in bioelectric potential
associated with breathing.
Three of the four textile effluent samples (F,
T, K) did not elicit any significant increase in
the bluegill gill purge rate. In addition, sample
277
-------
TABLE 4. TEXTILE PLANT EFFLUENTS: RELATIVE RANKINGS BASED ON
96-HOUR FATHEAD MINNOW LC50 VALUES, 48-HOUR DAPHNIA PULEX
EC5Q VALUES, AND MILLION LETHAL UNITS PER DAY
Fathead minnow
Most
toxic
Least
toxic
"50
A
L
V
C,T
N
W
G
MLU/d
G,L
C,V
N
T
A
W
Daphnia pulex
"50
W
E
A
V
U
L
C
Z
M
G
F
MLU/d
E
M
V
Z
W
G
L
U
C,F
A
K was not acutely toxic to either fathead min-
nows or Daphnia pulex. In all cases, the fish
survived, including those exposed to 75 per-
cent effluent. However, effluent sample L did
cause a significant response in 75 and 25 per-
cent dilutions. This sample, as indicated
previously, was acutely toxic to both fathead
minnows (96-h LC5o = 23.5 percent) and
Daphnia pulex (49-h EC50 = 28.0 percent).
Sample F indicated low acute toxicity (48-h
ECso = 81.7 percent) to Daphnia pulex, and
sample T was moderately toxic to fathead min-
nows (96-h LCso = 46.5 percent). As indicated
previously, the bluegills showed no response to
these effluents. The gill purge rate tests are
generally believed to be more sensitive and to
indicate toxicity at lower concentrations than
acute lethal toxicity tests. This suggests that
the bluegills were less sensitive to the toxic
materials present in samples F, L, and T than
were the fathead minnows and Daphnia pulex.
Thus, the usefulness of a battery of tests for
biologically evaluating the toxicity of complex
wastes is further demonstrated. It is apparent
that not all organisms or all tests are equally
sensitive to all chemicals or wastes.
REFERENCES
1. U.S. Environmental Protection Agency,
State-of-the-Art of Textile Waste Treat-
ment, U.S. Environmental Protection
Agency Report 12090 ECS 02/71, February
1971.
2. U.S. Environmental Protection Agency,
Development Document for Effluent
Limitations Guidelines and New Source
Performance Standards for the Textile
Mills, Point Source Category, U.S. En-
vironmental Protection Agency Report
440/1-74-022-a, June 1974.
3. G. D. Rawlings, Source Assessment: Tex-
tile Plant Wastewater Toxics Study, Phase
I, U.S. Environmental Protection Agency
Report 600/2-78-004h, March 1978.
4. K. M. Duke, M. E. Davis, and A. J. Dennis,
IERL-RTP Procedures Manual Level 1
Environmental Assessment Biological
Tests for Pilot Studies, U.S. Environment-
al Protection Agency Report 600/7-77-043,
April 1977.
5. R. W. Carlson, and R. A. Drummond, "Fish
Cough Response—A Method for Evaluat-
278
-------
ing Quality of Treated Complex Effluents,"
Water Res., No. 12, pp. 1-6,1978.
6. Allen L. Edwards, Experimental Design in
Psychological Research, Holt, Reinhart
and Winston, Inc., New York, 488 pp., 1972.
279
-------
EFFLUENT GUIDELINES DIVISION PROCEDURES FOR
MEASURING PRIORITY POLLUTANTS
Roger D. Holm*
Abstract
Priority pollutant Compounds are arranged
into several groups, each requiring a different
mode of approach* Volatile compounds that
can be sparged from water at room tempera-
ture are concentrated from 5 ml of water onto a
collection tube and then transferred onto a
chromatogrdphic column of analysis. Less
volatile compounds are extracted from waste-
water using methylene chloride in the case of
basic, neutral, and acid compounds, and
hexane-methylene chloride for pesticides. Ex-
tracts are concentrated to 1 ml to achieve
1,000- or 2,000-fold concentration enhancement.
Except for the pesticides, all organic analyses
are carried out using gas chromatography/mass
spectrometry (GCMS) techniques. A brief intro-
duction to the technique will be given. Ex-
amples of real samples will be provided. Pesti-
cides are characterized using gas chromatog-
raphy (GC) with electron capture detection.
Metals are measured using atomic absorption
arid inductively coupled argon plasma (ICAP)
spectrometry.
In this paper I would like to tell you about
the steps we, -at Monsanto, take in analyzing
priority pollutants. I shall introduce examples
of the classes of compounds currently being
measured, and then provide a brief introduc-
tion to the type of data supplied by the GCMS
instrumentation, our most powerful tool in
resolving and identifying compounds. Fol-
lowing this discussion, the analytical steps for
the volatile compounds that can be sparged
from water, the base-neutral and acid com-
pounds obtainable by methylene chloride ex-
traction, the pesticides obtainable by largely
hexane extraction, and the metals shall be con-
sidered.
*Group Leader, Monsanto Research Corporation,
Dayton, OH.
Let me begin by illustrating (Figure 1)
typical compounds or classes encountered in
priority pollutant analyses. Each category re-
quires a somewhat different analysis approach.
The volatile compounds are those that can be
quantitatively sparged from water and re-
covered under certain conditions. Examples
are carbon tetrachloride and chloroform. The
next group contains less volatile materials that
are extractable into methylene chloride. At a
pH above 11, the base-neutral fraction will con-
tain neutral compounds such as chlorobenzene
and phthalate esters, which seem to be every-
where, and amines such as dichlorobenzidine
or nitrosoamines. Phenols are extractable from
acid solution. Pesticides such as DDT are ex-
tracted in a separate scheme. Lastly, metals
such as Pb, Hg, and Cd are also included among
the priority pollutants.
To aid us later on in understanding the illus-
trations, I want to consider the type of data ob-
tained from a gas chromatographic mass spec-
trometer.
A small sample is injected by syringe into
the chromatograph. Various solute compounds
exit from the instrument at different times.
The mass spectrometer measures and can dis-
play the fragmentation pattern behavior of
these compounds as they issue from the
chromatograph. Figure 2 shows five mass spec-
tra obtained at 8-s intervals for a sample. Mass
number is displayed along the abscissa. Note
that a new compound having two strong
masses has appeared in the third scan, and
reached a maximum in the fourth scan.
Figure 3 shows a similar series of successive
mass spectra in a 3-dimensional display. Time
and accumulated scan number proceed along
the x-axis; the mass number from 60 to 180
atomic mass units is shown along the y-axis;
and signal amplitude is shown on the z-axis.
The display is that of dioctyl phthalate, which
produces two strong masses at 149 and 167 as
it elutes. In this example, I want to develop the
concept of a changing panorama of mass frag-
281
-------
VOLATILE COMPOUNDS
CARBON TETRACHLORIDE
HALOMETHANES, -ETHANES, -PROFANES
BASE-NEUTRAL COMPOUNDS
CHLOROBENZENES
POLYNUCLEAR AROMATICS
PHTHALATES
NITROSOAMINES
ACID COMPOUNDS
CHLORO-AND NITROPHENOLS
PESTICIDES
CHLORINATED COMPOUNDS
POLYCHLORINATED BIPHENYLS
METALS
HEAVY METALS
f
CYANIDE
Figure 1. Representative classes and compounds among the priority pollutants.
282
-------
MASS SPECTRA OF 5 SUCCESSIVE SCANS
-1 1
Ml H J.
.i..
• 171
n.c
• 178
33.7
» 173
S3.9
• 174
S3.9
1ST
Figure 2. Successive mass spectra scans as compounds elute from the gas chromatograph.
ments as chemicals elute from the chroma-
tograph. The computer software can connect
the peaks of the respective masses of each scan
to form a reconstructed chromatogram for an
individual mass, which is shown in Figure 4.
In this figure the capability is illustrated for
the mass range 140 to 180 amu. We have here a
chromatogram in which the data system is dis-
playing the appearance of mass 149. It can be
seen that here dioctylphthalate eluted around
scan number 196, about 27 minutes from the
start. In general, the compound identity is
determined by characteristic mass fragments,
its appearance at the appropriate time in the
chromatogram, and relative mass intensities
standing in the same ratios as those of stan-
dard compounds. We have considered the pro-
duction of these ion mass chromatograms
because we shall use them frequently during
the remainder of this talk.
Let us cons'ier now the analysis of the
volatile components of the priority pollutants.
Small vials of wastewater are collected with
care to avoid air bubbles. In the lab the water
is spiked with an internal standard, and it is
sparged for 12 minutes in a helium stream flow-
ing at 40 ml/minute.
The compounds are swept from the solution
onto a collection tube and are adsorbed on
Tenax-GC porous polymer. At the end of 12
minutes, the collection tube is heated to
transfer the compounds onto the head of a eold
chromatographic column. After 4 minutes of
heated desorption, the chromatograph run is
started and the column temperature is raised
slowly to 170° C.
Compounds elute at different times. In Fig-
ure 5 three volatile compounds are shown. The
chloroform is eluted at about 15 minutes. We
can tell it is chloroform by its characteristic
retention time and by its two significant
masses at 83 and 85 amu. It can be seen that
1,1,2-trichloroethane elutes later, also produc-
ing significant fragments with masses of 83
283
-------
MASS SPECTRA FROfl 15 SUCCESSIVE SCANS
FRN • SI440, SAMPLE SPECTRUM
NAME< OUEENV 301 9/H ON 3 HI
1ISC PATAt 1-9-78,.J.. SCHAAR
PERIODIC, QC RUN - 40 MIN. 1ST MASS - 36
*IASS RANGEi 50- 180
SCANSI 190- 803
THRESHOLD • 10.0
V EXPANSION • 100.0*1
SCAN
Figure 3. Three-dimensional view of successive mass spectral scans of bis(2-ethylhexyl)-phthalate
centered at scan 196.
CHROHATOGRAF1 OF SPECIFIC MASSES
FMN • 21440, SAMPLE SPCCTMUn
QUEENV 301 >XN ON B HI
C OATAt 1-9-78,. J.. SCHAAft
PERIODIC, QC RUN • 40 HIM. 1ST MASS - 35
nASS RANGEi 140- 180
SCANS' 130- 830
THRESHOLD - 50.0
V EXPANSION - 100.0k
130140 150 160 170 180 190
140
230 SCAN
Figure 4. Three-dimensional view of ion mass chromatogram for bis(2-ethylhexyl)-phthalate
centered at scan 196.
284
-------
** SPECTRUM DISPLAYXEDIT **
PECIAL CHOLOfclNE STDS, C2X'S>,GN 8HI
- 10-78, 1545, J
21661
1ST SCxPGi
*• 1.0O V-
85.0
TI
M_^%A_ J YA » A -. .
83.0 -. _J^_n_^LTl
55.0.A,
IB 11 iff 13 14 IB te 17 ift 10 ae ai ak ^i aU ak afe a? P'B 3
Figure 5. Display of ion mass chromatograms for three volatile compounds.
and 85. The ratio of intensity, however, is dif-
ferent. For confirmation, a third mass, 97 in the
case of 1,1,2-trichloroethane, must also be ob-
served, and the spectrum must match that of
the standard.
Finally Figure 6 illustrates a real sample
having 13 volatile priority pollutants. In this
case the total ion current of the detector is
plotted vs. time to provide a chromatogram
display.
Let us turn to the extractable compounds
not volatile enough to be quantitatively
sparged. The steps are summarized in Figure
7.
Our procedure is to extract a 2-liter sample,
adjusted to pH 11, with methylene chloride.
Neutral compounds and basic amines will con-
centrate in the methylene chloride, and the
phenols and acids will remain in the water as
salts. The organic layer is removed, dried, and
concentrated ultimately to 1 milliliter. The
aqueous solution containing the phenols is then
acidified to pH 2 and extracted again. The ex-
tracts are concentrated, also to 1 milliliter. In
this way a 2,000-fold concentration enhance-
ment is achieved.
Figure 8 shows a 2-liter extraction in proc-
ess. Figure 9 shows the extract in the. process
of evaporation using a Kuderna-Danish evap-
orator.
Upon analysis the base-neutral concentrate
may appear relatively complex, as is shown in
Figure 10. Hidden among many of these peaks
are quite a number of polynuclear aromatics.
This sample was from an aluminum plant.
Figure 11 illustrates the total ion trace of a
wastewater extract which had a great number
of compounds merging together as they eluted
285
-------
*S SPECTRUM DXSPLAV'EDZT
O UOL 36O R«U UATER QMBHX
ia-9-77 J.HILLER
**
FRM 81349
1ST SCXPCl 68
x* .se v- i.«e
TI
13
1
2
3
4
5
6
7
methylene chloride
bromochloromethane
chloroform
1,2-dichloroethane
1,1,1-trlchloroethane
carbon tetrachloride
trichloroethylene
8 benzene
9 1,1,2,2-tetrachloroethene
10 toluene
11 1,4-dichlorobutane
12 chlorobenzene
13 ethylbenzene
Figure 6. Total ion chromatogram of a volatile sample.
from the chromatograph. Despite all the com-
pounds, only two priority pollutants were
observed, however, in this sample.
In Figure 12 the lower plot is the total ion
chromatogram we saw before. The peak with
mass 149 at 27 minutes is bis (2-ethylhexyl)
phthalate. The small peak at 19 minutes is
d10-anthracene, which was added as an internal
standard for quantifications.
A significant effort goes into data interpre-
tation, and an example will illustrate this.
Figure 13 shows the mass spectrum of the
material eluting at this time. The hydrocarbon
background shown in this figure can be re-
moved in the interactive computer-based data
interpretation process to give the resulting
mass spectrum of the material alone. Thus,
Figure 14 shows the three characteristic
masses of d10-anthracene at 188, 94, and 106.
Figure 15 shows phenols observed in an acid
sample. The extraction from acid solution and
workup are identical to those of the base-
neutral fraction. In this case five phenols were
observed, some hidden beneath the envelope.
286
-------
pH > 11
W/SODIUM HYDROXIDE
METHYLENE CHLORIDE
EXTRACTION
BASES & NEUTRALS
ACIDS. UNEXTRACTABLES. PHENOLS
BOTTOM LAYER
DRIED ON
ANHY. SODIUM SULFATE
CONCENTRATED
INK-D EVAP.
FROM 10-3 m3 TO 10-6 m'
GC/MS
IDENTIFICATION &
QUANTITATION
TOP LAYER
CHANGE pH < 2
W/HYDROCHLORICACID
METHYIENE CHLORIDE
EXTRACTION
Figure 7. Extraction and workup scheme for base-neutral and acid compounds.
Figure 8. Extraction of priority pollutants with methylene chloride.
287
-------
Figure 9. Kuderna-Danish evaporator to reduce methylene chloride
extract volume.
These were phenol, 2-nitrophenol, 2,6-dimethyl-
phenol, 4-nitrophenol, and pentachlorophenol.
A variety of steps are taken to maintain
quality control in these analyses. Organic free
water is periodically run through the entire ex-
traction and cleanup process to ensure that
nothing is picked up in the laboratory opera-
tions. Duplicate runs are periodically made, in-
cluding samples divided in the field and sent to
the lab as blind duplicates. An internal stand-
ard is always added during the evaporation
stage. R-values, relative response values used
in quantification, are checked with standards
periodically to maintain accuracy and to ensure
proper instrumentation performance. Our ex-
perience with extractable compounds has been
that for a measurement performed in tripli-
cate, the reported value should lie within 34
percent of the true value 90 percent of the
time. Bear in mind that extraction efficiencies
in many waste solutions are not known, but
evidence suggests recoveries ranging from 10
percent to 100 percent, with many in the 30
percent to 80 percent range.
Pesticides are extracted using the method in
the Federal Register. Generally cleanup is re-
quired to remove from solution nonpesticides
that cause significant detector response. Fig-
ure 16 shows a typical pesticide chromatogram
using the electron capture detector before the
second-stage cleanup. The peaks indicated
were removed by an acetonitrile back ex-
288
-------
** SPECTRUM DISPLAVxCDIT **—
F Nj^tANT C-CSPtFft»M*v UASTC UMTa.aUL.OM 1M*
Figure 10. Typical total ion chromatogram of a base-neutral fraction.
*X SPECTRUM DISPLAY/EDIT *X
3O06 S UL INJ GN 6 LO
EI-GC
FRN 81867
1ST SCXPGI i
x- .s« v- i.eo
Figur ? 11. Total ion plot of a base-neutral sample illustrating overlapping peaks.
289
-------
149.(
«X SPECTRUM DISPLAVxEDIT X*
B'N 3006 8 UL INJ GN 6 LO
EI-QC 50x4xsx260xa0x30xspaas0 11x^x77 AU
FRN 21867
1ST SCXPG* 1
x- .so v- i.oo
94.0
188.
TI
A* 6169.
Monsanto
Figure 12. Illustration of individual compounds observed within the same sample.
FRN 51367 SPECTRUM 181
LARGST 4: 188.0,100.0 55.0. 53.4
LAST 4t 360.1, 1.6 B61.0., 1.4
RETENTION TIME 19.0
69.0. 44.9 95.0, 43.8
274.0. .6 276.0, .9
PAGE IV- l.«
Monsanto
Figure 13. Raw mass spectrum showing dig-anthracene plus hydrocarbon background.
290
-------
UORK AREA SPECTRUM FRN 81567
LARGST 4t 188.0,100.0 79.9, 17.9
PAGE i y 1
94.0, IS. 2 186.9, 14.3
80 100 180 140 160
Monsanto
Figure 14. Net mass spectrum of d-jQ-anthracene after removal of background.
*X SPECTRUM DISPLAV'EDIT X*
FRN
1ST SCXPGl 1
X- .35 V- 1.00
U 334319-A GN3LO
10-31-77 J.niLLER
d • 4-NITROPHENOL
e • PENTACHLORO PHENOL
f • d- ANTHRACENE
a « PHENOL
b = 2-NITROPHENOL
c • 2,6-DIMETHYLPHENOL
Monsanto
Figure 15. Total ion plot of typical chromatogram of a phenol sample.
291
-------
9.41
11.27
12.50
>-p 533Hfl
FlREfl 5S
RT
0.39
0.54
0.30
1.39
1.47
1.93
2.29
3.01
5.24
5.47
7.01
7.99
9.41
12.50
flREfl
28530000
3242000
3354000
1352000
1336000
347600
1167000
54970000
59320
1705000
51310
7361000
374400
1280006
flREfl '/.
27.786
3.157
3.753
1.304
1.350
0.339
1.137
53.535
0.053
1.661
B. 350
3.273
0.852
1.247
XFs 1.0000 E+ 0
Figure 17. Same pesticide chromatogram after back extraction with acetonitrile.
292
-------
Figure 16. Typical pesticide chromatogram before cleanup.
293
-------
traction of the hexane extract to yield the
cleaner chromatogram in Figure 17.
Sometimes further cleanup is required, in-
volving adsorption on silica gel, plus further
separation of components in different solvents
having a range of polarities. Note the peak that
appeared in the middle chromatogram of the 15
percent ether-85 percent petroleum ether frac-
tion in Figure 18.
Figure 19 lists two types of metal analysis
approaches. Metals are measured by acidifying
the solution and performing an atomic absorp-
tion measurement on the indicated four
volatile metals: Hg, Se, As, and Tl.
The remaining metals of concern are mea-
sured using an automated emission spectro-
graph having as a source an inductively cou-
pled argon plasma. This allows simple solution
aspiration excitation. Quality control is main-
tained using blind duplicate samples and
spiked samples.
We have looked briefly at several phases of
priority pollutant analyses, discussed typical
compounds, and looked at the way mass spec-
trometry data can be used to identify com-
pounds and how it can be selective. Also we
have discussed analyses of volatile compounds,
the base-neutral and acidic compounds, the
pesticides, and metals. There has been little
time to expand upon these techniques, but per-
haps you now have a feel for the efforts that
must be expended for these analyses.
*
V
*»
>Sfl%Jj'«ir. ether
if
Figure 18. Portions of typical chromatograms of solutions having increasing polarities
after Florisil column fractionation.
294
-------
METALS ANALYSES
ATOMIC ABSORPTION - (As, Hg, Se, Tl)
INDUCTIVELY COUPLED ARGON PLASMA (ICAP)
QUALITY CONTROL
• DUPLICATE SAMPLES
• SPIKED SAMPLES
Figure 19. Methods utilized for analysis of priority pollutant metals.
295
-------
ORGANIC SAMPLING AND ANALYSIS FOR
ENVIRONMENTAL ASSESSMENT
Philip L. Levins, Judith C. Harris*
Abstract
Sampling and analysis procedures for
organic species have been developed in support
of the phased approach to environmental
assessment developed by the Process Measure-
ments Branch of EPA, IERL-RTP. These
methodology studies focus on process and
emission streams.
Level 1 analysis procedures have been
evaluated and refined with the objective of pro-
viding a methodology suitable for use in any
experienced laboratory for the generation of
reliable data on most species likely to be en-
countered. A Level 1 reporting format has been
developed which translates the chemical data
into terms more readily used in assessing
relative hazardousness.
Level 2 sampling and analysis procedures are
being developed, based upon both a more com-
prehensive broad spectrum protocol and upon
specific procedures for particular analytes—
PAH's and phenols, for example.
INTRODUCTION
Organic sampling and analysis for en-
vironmental assessment calls for a phased ap-
proach including Level 1 (survey), Level 2 (con-
firmation), and Level 3 (monitoring) studies.
The Level 1 procedures developed by the Proc-
ess Measurements Branch have been in use for
about 2 years and procedures for Level 2 are
beginning to be both developed and applied.
The purpose of this paper is to discuss two
aspects of organic sampling and analysis
methodology which are particularly relevant
to environmental assessment of textile in-
dustry effluents. First, the status of Level 1
organic analysis procedures will be reviewed,
with emphasis on the methods of interpreting
and reporting the results. The second portion
of the paper will deal with the development of
alternative methods that are appropriate for
Level 2 sampling of polar, water-soluble
organics in aqueous effluent streams.
LEVEL 1 ORGANIC
ANALYSIS PROCEDURES
The philosophy of the phased approach to en-
vironmental assessment and the details of the
Level 1 sampling and analysis procedures are
described in other EPA publications (refs. 1,2).
The objectives and methodology of Level 1
organic analysis are therefore not discussed in
detail in this paper.
Level 1 sampling and analysis is designed to
provide identification of the major categories
of organic compounds and estimation of their
concentration in the gaseous, liquid, or solid ef-
fluent stream. Organic analyses are done
either directly on neat samples, such as gases
or organic liquid streams, or on methylene
chloride extracts of samples such as aqueous
streams, solids, or SASS train sample com-
ponents. Figure 1 illustrates the overall Level
1 analysis scheme for the organic components.
The procedure covers the quantitative analysis
of the major volatility classes by the pro-
cedures indicated below.
Volatility
Definition
Procedure
'Arthur D. Little, Inc., Cambridge, MA.
Gases Boiling point < 100° C Field GC
Volatile Boiling point 100-300° C TCO
Nonvolatile Boiling point > 300° C GRAV
The TCO procedure is a laboratory gas chroma-
tographic (GC) method for the volatile com-
ponents. The GRAV procedure is a gravimetric
method for the nonvolatile components.
Qualitative analysis to determine the major
chemical categories in each sample is per-
formed by first separating the sample accord-
ing to polarity on a silica gel liquid chromatog-
raphy (LC) column. Each of the fractions is
then identified by obtaining infrared (IR) and
low resolution mass spectral (LRMS) data.
297
-------
Aqu
sow, Solutions
Seven Fractions:
*
Analysis
t
Low Resolution
Mass Spectra
Analysis '
TCO1
Gravim
Anah
Figure 1. Level I organic analysis flow diagram.
REPORTING LEVEL 1
ANALYSIS RESULTS
At this time, sufficient Level 1 tests have
been run (refs. 3,4,5,6) to verify that the
specified organic analysis procedures are capa-
ble of identifying a broad range of organic com-
pound categories when present in environ-
mental samples. As the data have been ac-
cumulated, it has been possible to develop and
recommend with some confidence a format for
reporting and summarizing the organic
analysis results.
The first complete set of data to be
developed in the Level 1 organic analysis are
the LC results, which are reported in the form
shown in Table 1. The volatile (TCO) and non-
volatile (GRAY) components of each LC frac-
tion are reported after computing back to
equivalent quantities in the entire sample. The
total quantity of each fraction is then derived
from the sum of TCO + GRAY and the concen-
tration in the sampled stream is calculated
from the sample quantity and volume sampled.
Results of the IR analysis of each sample and
LC fraction are tabulated in terms of the fre-
quency of peak maxima, probable assignment,
and intensity (weak, medium, or strong). A
typical example is given in Table 2.
The LRMS results are reported in terms of
the compound categories found in the sample
and their estimated relative abundances. A
typical report using this format is shown in
Table 3. Most of the subsequent initial Level 1
interpretation of data is done using only the
major compound category data. Compound
298
-------
TABLE 1. EXAMPLE LC REPORT
Total Sample1'
Taken for LC*
Recovered*
Fraction
1
2
3
4
5
6
7
TCO
mg
106
23
17
TCO*
mg
1.4
15
55
0.5
0.9
3.2
0.5
GRAV
mg
386
84
74
GRAV§
mg
11.5
4.6
259.4
10.1
31.3
16.6
2.8
Total
mg
492
107
91
Total§
mg
13
20
314
11
32
20
3
Concentration*
mg/m^
82
18
15
Concentration*
mg/m^
2.1
3.3
52
1.8
5.3
3.3
0.5
*Total mg divided by total volume.
tQuantity in entire sample, determined before LC.
^Portion of whole sample used for LC, actual mg.
^Quantity recovered from LC column, actual mg.
§ Total mg computed back to total sample.
TABLE 2. EXAMPLE IR REPORT
Frequency
(cm-1)
3400
3050
2850,2920
1710
1680
1600
1450
1060
740
Intensity
M
W
M
S
M
M
M
S
M
Assignment
OH, NH
Unsaturated CH
Saturated CH
Acid, ketones
Monosub amide, ketones
Ring vibrations
CH2
Si-0, ether
Subst. pyridine, C-CI
Comments
Broad
Broad
categories for LRMS interpretation are chosen
primarily from a list of about 24 general com-
pound groupings.
For each sample or sample extract subjected
to the Level 1 organic analysis procedure,
there will be an LC report, eight IR reports,
and up to sev n LRMS reports. This is an un-
wieldy body of data from which to make a deci-
sion. To reduce these data to a workable form,
the detailed information from the combined
LC, IR, and LRMS data is combined into a sum-
mary report. The organic extract summary re-
port basically tabulates the identity and quan-
tity of all compound categories found in each of
the LC fractions and sums them, to yield an
estimate of source concentrations. The results
from the sample shown in Table 4, for instance,
indicated a high concentration of aromatics and
some polar compounds in the sampled stream.
The considerable overlap in chemical class
composition that is generally expected be-
tween adjacent fractions in the Level 1 LC
299
-------
TABLE 3. EXAMPLE LRMS REPORT
Major categories
Intensity
100
100
10
10
10
Category
Ketones
Heterocyclic nitrogen compounds
Esters
Carboxylic acids
Phenols
MW range
180-280
167-253
Sub-categories, specific compounds
Intensity
100
100
10
10
10
10
10
10
10
10
10
10
10
Other
Category
Acridine
Fluorenone
Phenol
Cresol
Benzole Acid
Carbazole
Methylacridine
Methylfluorenone
Anthraquinoline
Benzanthrone
Dibenzofluorenone
Dibutylphthalate
Methylbenzanthrone
m/e
179
160
94
108
122
167
193
194
229
230
280
278
244
Composition
C13H9N
Cl3H80
CeHgO
C7H80
C7H602
Ci2HgN
C14HnN
C14H100
CiyHiiN
C17H100
C21H120
C16H220
Ci8H120
scheme is illustrated by the data in Table 4. It
is worth emphasizing that this LC procedure is
not a high-resolution technique and that its
function in Level 1 methodology is to facilitate
qualitative analysis by IR and LRMS, rather
than to isolate various compound categories in
discrete fractions. The reporting format devel-
oped takes account of this fact by combining
the information obtained on various LC frac-
tions to provide an integrated description of
the chemical composition of the sampled
stream.
In summary, the procedures developed for
interpreting and integrating the Level 1
organic analysis data result in a list of organic
chemical classes and their estimated concen-
trations in the stream sampled. This overall
estimate of stream chemistry is appropriate
for comparison with decision criteria such as
the EPA's Multimedia Environmental Goals
for Environmental Assessment (ref. 7).
If one or more of the estimated concentra-
tions exceeds the levels of concern, it may then
be desirable to reexamine the organic analysis
data in detail to determine more specifically
the nature of the material responsible for the
positive "trigger."
LEVEL 2 SAMPLING
OF WATER-SOLUBLE ORGANICS
The results of the Level 1 organic analysis
provide a basis for proceeding with and/or di-
recting more detailed Level 2 studies. Level 2
inquiries are expected to be directed primarily
at the identification, quantification, and con-
300
-------
TABLE 4. ORGANIC EXTRACT SUMMARY TABLE
Sample: sorbent extract
LC1
LC2
LC3
LC4
LC5
LC6
LC7
Total orpnics, rag/m^
TCO.mg
GRAV.mg
Category
Sulfur
Aliphatic HC's
Aromatics-benzenes
FusedArom216
Fused Arom216
Heterocycl-c S
Heterclytic N
Carboxylic acids
Phenols
Esters
0.61
5.2
13
100/0.6
10/0.06
0.74
19
3.3
10/0.06
100/0.6
10/0.06
10/0.06
8.4
73
180
100/4
100/4
10/0.4
1.0
6.7
23
Int/mg/m^
100/0.5
100/0.5
10/0.05
10/0.05
0.33
3.7
7.3
/0.1*
/0.01*
/O.I*
/0.01»
/0.01*
1.5
5.3
41
-
100/0.7
10/0.07
100/0.7
10/0.07
10/0.07
0.50
0.1
15
10/0.02
100/0.2
10/0.02
10/0.02
13
110
280
0.6
0.06
0.06
5
5
0.5
1
0.3
1
0.1
0.08
* Estimated assuming same relative intensities as LC6, since IR spectra of LC5 and LC6 very similar.
firmation of specific compounds whose pres-
ence could be inferred on the basis of the cate-
gorical analysis of Level 1. Level 2 analyses
may be conducted either on samples freshly
collected for that purpose or on samples re-
tained from the Level 1 sampling effort. The
choice between collecting a new sample or ana-
lyzing a retained sample or fraction for the
Level 2 data will be highly specific to the exact
question to be answered and to the nature of
the chemical species to be determined.
Most of the sampling procedures chosen for
Level 1 were the best available and are still ap-
propriate for Level 2 studies. On the other
hand, there are several organic compound
categories for which alternative sampling pro-
cedures are required to achieve optimal collec-
tion and recovery. Examples that are particu-
larly relevant to textile industry environment-
al assessment are very polar, water-soluble or-
ganics in aqueous effluents. This section of the
paper will describe work in progress to charac-
terize sampling approaches for these types of
compounds.
Solvent extraction of pH-adjusted samples is
the most widely used approach to sampling of
organic polluta' cs in aqueous industrial ef-
fluents. This approach, which is specified for
Level 1 survey analysis, is capable of giving
quantitative recoveries of a wide range of non-
polar organic materials. However, quantitative
recovery of very polar, water-soluble organic
species is not always achievable by solvent ex-
traction. Estimates of the maximum achievable
extraction efficiency can be based on the value
of the solvent/water partition coefficient for a
particular species. For example, three succes-
sive extractions of water sample, using 100 ml
of solvent per liter of sample for each extrac-
tion, can be expected to provide >90 percent
extraction efficiency for individual organic
species that have solvent/water partition coef-
ficients >12. Extensive listings of partition
coefficients are available (refs. 8,9), and these
can be used to verify extraction efficiencies for
individual organic species of concern. Some il-
lustrative values, presented in Table 5, in-
dicate that carboxylic acids, sulfonic acids,
glycols, and phenols are among the categories
of compounds not efficiently extracted.
An attractive alternative approach for col-
lection of organics from large quantities of
water is the use of an in situ continuous extrac-
tive sampler containing a solid sorbent
material. One of the earlir - devices of this
type, once widely used in the field, was the
301
-------
TABLE 5. ILLUSTRATIVE VALUES OF SOLVENT:
WATER PARTITION COEFFICIENTS*
Compound
Efficiently extracted
Propyl ether
Hexyl amine
Butyl mercaptan
Benzene
Chlorobenzene
Indene
Acridine
Thiophene
PCHCl3:H20
580
490
1,200
460
6,200
78,000
20,000
301
'>heptane:H20
34
30
98
24
1,100
1,500
9,300
13
pether:H20
18
16
34
15
150
180
575
10
Inefficiently extracted
Acetic acid
Phenol
Benzene sulfonic acid
Ethylene glycol
0.03
2
0.0001
0.002
1.0
0.05
0.000006
0.00001
0.4
31
0.002
0.005
* Estimated using data bank and methods of reference 8.
EPA's carbon absorption module (CAM) (ref.
10), with generally good collection efficiency
but low recoveries for organics. A number of
previous workers (refs. 11,12,13,14) have done
laboratory studies on the use of macroporous
resins and have determined overall recoveries
of selected organic compounds from water,
under conditions designed to simulate sam-
pling applications. Other investigators (refs.
15,16) have explored the use of polyurethane
foam plugs for collection of organics, especially
polycyclic aromatic hydrocarbons. Anion ex-
change resins have been used successfully in
several instances for the concentration of polar
or ionizable organic compounds (refs. 17,18).
The Ambersorb carbonaceous adsorbents have
not yet been extensively researched as possi-
ble adsorbents for PAH or other compounds
from water.
The present work involves a systematic lab-
oratory investigation of the applicability of
macroreticular resins for general and
compound-specific sampling of organics, based
on small-scale, chromatographic experiments.
The basic approach used in this work is the
generation and analysis of the frontal
breakthrough curve produced by challenging a
sorbent cartridge with an aqueous solution of
organic analyte at constant flow rate. The
experimental method has been described in
some detail elsewhere (ref. 19). Figure 2 is a
schematic representation of a frontal break-
through curve, indicating the 50 percent and
100 percent breakthrough points. The frontal
breakthrough curve generated in each experi-
ment was analyzed to determine the character-
istic 50 percent breakthrough parameters and
the adsorption isotherm.
The specific retention volume corresponding
to 50 percent breakthrough, Vgo, is the number
of sorbent bed volumes (BV) of aqueous solu-
tion that have passed through the bed up to the
time when the exit concentration of organic
species equals one-half of the challenge concen-
tration. The Vgo value, which is independent of
challenge concentration as long as the latter is
sufficiently low, allows one to determine
whether an organic compound is retained by a
given volume of sorbent at a specified flow rate
and sampling time. The weight capacity of the
sorbent for organic species, in mg/cm3 of resin,
corresponding to 50-percent breakthrough is
given by V50 x challenge concentration. The
weight capacity is thus a function of challenge
concentration, subject to an upper limit cor-
responding to saturation of the bed. The
weight capacity, while useful for comparisons
between laboratory experiments, is less
302
-------
C100%.
Breakthrough or
Concentration C
(C) 50%
,V
50-
r100
• Frontal
Profile
Volume (Time)
Figure 2. Schematic of frontal breakthrough curve.
valuable than V50 as a sampling device design
parameter.
The adsorption isotherm, a plot of uptake of
solute in mg/cm3. of resin versus percent
breakthrough, was calculated by integrating
the area in front of the breakthrough curve at
5,25,50,75, and 100 percent breakthrough. Ad-
sorption isotherms contain valuable informa-
tion on the affinity of the resin for a sorbate,
and allow the saturation capacity of the sor-
bent to be ascertained. If the breakthrough
curve is generated under equilibrium sorption
conditions, the isptfferm is expected to show a
Langmuir shape—linear in the low uptake
region and reaching to a plateau corresponding
to saturation of the available adsorption sites.
Deviation of the isotherm from the expected
Langmuir shape is an indication of significant
mass transfer effects in the frontal chroma-
tography experiment.
The initial studies in this program were done
using XAD-2, the resin most extensively
studied for collection of organic species from
water. A series of experiments was performed
to determine the effect of aqueous sample flow
rate on the adsorption of a moderately polar
organic species, acetophenone, on XAD-2. As
shown in Figure 3, distinctly non-Langmuir iso-
therms were obtained at sample flow rates cor-
responding to linear velocities through the sor-
bent bed of 2 cm/min or higher. Note that the
uptake at 50-percent breakthrough is also very
much lower at the higher flow rates tested.
These flow rate studies clearly show that mass
transfer effects have an important impact on
adsorption of acetophenone from water onto
XAD-2, even at relatively low flow rates. This
may indicate that adsorption of species for
which XAD-2 has limited affinity, such as ace-
tophenone, is limited by diffusion into the
pores of the macroreticular resin particles,
while more readily sorbed species, such as
polynuclear aromatic hydrocarbons (PAH), are
efficiently captured on the exterior surface of
303
-------
7.0
6.0
5.0
Uptake, 4.0
mg/cc of
Resin
Bed
3.0
2.0
1.0
Flowrate:
Flowrate:
Flowrate:
Flowrate:
0.88 ml/min
1.12cm/min
1.6 ml/min
2.0 cm/min
3.1 ml/min
3.9 cm/min
6.1 ml/min
7.8 cm/min
50 75
% Breakthrough
95100
Figure 3. Isotherms for adsorption of acetophenone on XAD-2
from water as a function of flow rate.
the particles. It was necessary to reduce the
flow rate to about 1 cm/min linear velocity in
order to achieve the expected isotherm shape
and maximum uptake. This linear velocity was
maintained in subsequent breakthrough exper-
iments with other analytes and sorbents. The
flow rate dependence study illustrates the fact
that sampling approaches proven effective for
PAH in water are often not directly applicable
to other organics.
A variety of organic compounds in aqueous
solutions were examined to determine their
frontal chromatographic behavior on XAD-2.
Because of the interest in extrapolating to field
sampling of industrial effluents, pH adjust-
ments were not made in these tests. The 50
percent breakthrough parameters are pre-
sented in the first column of Table 6. The
results confirm the observations of other
workers that XAD-2 has a high affinity for non-
polar aromatic species; no measureable break-
through was obtained for these compounds.
The XAD-2 also seems to be an adequate sor-
bent, for sampling purposes, for species such as
304
-------
TABLE 6. COMPARISON OF UPTAKE ON VARIOUS RESINS AT 50 PERCENT BREAKTHROUGH
FOR 10 PPM CHALLENGE CONCENTRATION
Resin
Analyte
Phenol
Benzole acid
Benzyl alcohol
Acetophenone
Chlorobenzene
Methyl benzoate
Naphthalene
Carbazole
Acridine
XAD-2
Wt.
capacity
mg/cm3
0.43
0.54
0.53
3.2
(>3.8)*
11.4
(>3.8)*
Vol.
capacity
BV
74.7
102
98.9
527
1,750
XAD-7
Wt.
capacity
mg/cm3
0.60
0"
0.89
Vol.
capacity
BV
67.9
39.2
96.6
IRA-904
Wt.
capacity
mg/cm3
(>14.3)t
(>14.3)t
0.164
0.58*
0.32*
IRA-93 XE-347
Wt. Wt.
capacity capacity
mg/cm3 mg/cm3
* *
# *
*
(>712)t
(>0.72)t
(>1.42)t
*No breakthrough in 4 hours, therefore, volumetric capacity not determinable. (Number in parenthesis is
amount sorbed up to that point.)
breakthrough in 16 hours, therefore, volumetric capacity not determinable. (Number in parenthesis is
amount sorbed up to that point.)
'Jptake at 40 percent breakthrough.
ketones and esters. The low affinity of XAD-2
for very polar organics such as phenol and ben-
zoic acid was confirmed.
Several other macroreticular resins were
tested to explore their potential for use as
alternatives or supplements to XAD-2 for col-
lection of polar species. The breakthrough
parameters for these resins are summarized in
Table 6. XAD-7 was found to have higher affini-
ty for phenol, but lower affinity for benzoic
acid, than XAD-2.
The ion exch?*rge materials, IRA-904 and
IRA-93, showed >10 times more capacity for
phenol and benzoic acid in distilled water than
did XAD-2 or XAD-7. However, breakthrough
on these resins was virtually instantaneous if
the aqueous solution contained as much as 250
ppm of sodium chloride, an amount commonly
found in aqueous effluents. This is not entirely
unexpected in that the manufacturer of these
resins recommends use of salt solutions as the
most effective regeneration method.
Carbonaceous resin, XE-347, showed a very
high affinity for all of the compounds tested.
Except for the very polar analytes, no break-
through was obi ?rved over a 16-h period. In-
vestigation of ti Ambersorb resins, including
quantitative studies of recovery of sorbed
material, is continuing. This sorbent medium
presently appears to represent a promising ap-
proach for sampling of very polar water-
soluble organics in aqueous streams.
ACKNOWLEDGEMENTS
This work was performed under EPA Con-
tract No. 68-02-2150 from the Process Mea-
surements Branch, Industrial Environmental
Research Laboratory, Research Triangle Park,
North Carolina. The authors wish to thank Ms.
Zoe A. Grosser and Ms. Karen C. Weaver for
their contributions.
REFERENCES
1. U.S. Environmental Protection Agency,
"Environmental Assessment Sampling
and Analysis: Phased Approach and
Techniques for Level 1," EPA-600/2-77-115,
June 1977.
2. U.S. Environmental Protection Agency,
IERL-RTP Procedures Manual Level 1
Environmental Assessment, EPA-600/2-76-
160a, PB 257-850/AS, June 1976.
3. U.S. Environmental Protection Agency,
"Compilation of Level 1 Environmental
305
-------
Assessment Data," EPA-600/2-78-211, Oc-
tober 1978.
4. N. H. Gaskins and F. W. Sexton, "Addi-
tional Chemical Data: A Supplement to
Compilation of Level 1 Environmental
Assessment Data," report prepared on
EPA Contract No. 68-02-2156.
5. U.S. Environmental Protection Agency,
"Source Assessment: Textile Plant
Wastewater Toxics Study: Phase 1,"
EPA-600/2-78-004h, March 1978.
6. J. L. Rudolph, J. C. Harris, Z. A. Grosser,
and P. L. Levins, "Ferroalloy Process
Emissions Measurement," draft report
prepared on EPA Contract No. 68-02-2150,
August 1978.
7. J. G. Cleland and G. L. Kingsbury,
Multimedia Environmental Goals for En-
vironmental Assessment, Vol. I, EPA-
600/7-77-136a, November 1977.
8. A. Leo, C. Hansch, and D. Elkins, "Parti-
tion Coefficients and Their Uses,"
Chemical Reviews, Vol. 71, No. 525, 1971.
9. A. Seidell, Solubility of Organic Com-
pounds, Vol. II, 3rd ed., Van Nostrand,
Princeton, NJ, 1941.
10. M. J. Taras, A. E. Greenberg, R. D. Hoak,
and M. C. Rand, eds., "Organic Con-
taminants," Standard Methods for the Ex-
amination of Water and Wastewater, 13th
ed., pp. 259-270, American Public Health
Association, Washington, DC, 1971.
11. G. A. Junk, J. J. Richard, J. S. Fritz, and H.
J. Svec, "Resin Sorption Methods for
Monitoring Selected Contaminants in
Water," Identification and Analysis of
Organic Pollutants in Water, L. H. Keith,
ed., Ann Arbor Science, Ann Arbor, MI,
1976.
12. P. E. Strup, J. E. Wilkinson, and P. W.
Jones, "Trace Analysis of Polycyclic
Aromatic Hydrocarbons in Aqueous
Systems Using XAD-2 Resin and Capillary
Column Gas Chromatography Mass Spec-
trometry Analysis," 1'olynuclear Aromatic
Hydrocarbons: Second International Sym-
posium on Analysis, Chemistry and Biol-
ogy, P. W. Jones and R. E. Freudenthal,
eds., Raven Press, New York, pp. 131-138,
1978.
13. R. G. Webb, "Isolating Organic Water
Pollutants: XAD Resins, Urethane Foams,
Solvent Extraction," EPA-660/4-75-003,
1975.
14. P. Van Rossum and R. Webb, "Isolation of
Organic Water Pollutants by XAD Resins
and Carbon," Journal of Chromatography,
Vol. 150, pp. 381-392, 1978.
15. D. K. Basu and J. Saxena, "Monitoring of
Polynuclear Aromatic Compounds in
Water II. Extraction and Recovery of Six
Representative Compounds With Polyure-
thane Foams," Environmental Science and
Technology, Vol. 12, pp. 791-795, 1978.
16. J. Navratil, R. Sievers, and H. Walton,
"Open-Pore Polyurethane Columns for Col-
lection and Preconcentration of Poly-
nuclear Aromatic Hydrocarbons from Wa-
ter," Analytical Chemistry, Vol. 49,
No. 14, pp. 2260-2263, 1977.
17. M. Chasanov, R. Kunin, and F. McGarvey,
"Sorption of Phenols by Anion Exchange
Resins," Industrial and Engineering
Chemistry, Vol. 48, No. 2, pp. 305-309,1956.
18. B. Kim, V. Snoeyink, and F. Saunders, "Ad-
sorption of Organic Compounds by Syn-
thetic Re-sins," J. WPCF, Vol. 48, No. 1,
pp. 120-133, 1976.
19. Z. A. Grosser, J. C. Harris, and P. L.
Levins, "Quantitative Extraction of
Polycyclic Aromatic Hydrocarbons and
Other Hazardous Organic Species From
Process Streams Using Macroreticular
Resins," presented at the 3rd International
Symposium on Polynuclear Aromatic
Hydrocarbons, October 1978.
306
-------
MEASUREMENT OF COLOR IN TEXTILE DYEING WASTEWATERS
Linda W. Little*
Abstract
All presently accepted "standard methods"
for determining color of water and wastewaters
are based on spectrophotometric analysis. The
standard American Public Health Association
determination of color, initially designed for
application to natural waters, is based on com-
parison of true (soluble) color of the test sample
to knoum concentrations of yellow-brown color.
Since even slight amounts of turbidity in-
terfere with spectral analysis, turbid samples
require filtration prior to analysis. Application
of the methods to textile dyeing wastewater is
complicated due to the presence in many of
these wastewaters of insoluble dyes and dyes
of hues substantially different from the yellow-
brown color of natural waters. Recent ap-
proaches to modification of standard methods
to improve applicability to all hues /American
Dye Manufacturers^ Institute (AMDI)
Method/ and for improved and simplified
calculation of color values (CIELAB, CIELUV)
will be discussed, as well as possible approach-
es to obtaining realistic color values for
wastewaters containing dispersed dyes and
other dyes that are removed by filtration.
The presently accepted standard methods
for determination of color in water and
wastewater were developed for potable, sur-
face, and other natural waters in which the col-
or is generally soluble and yellow-brown in
hue. In the standard American Public Health
Association (APHA) procedure, "true color" is
defined as "the color of water from which the
turbidity has been removed," and "apparent
color" as the "color due to substances in solu-
tion, but also that are due to suspended
matter." (ref. 1) Apparent color is determined
on the sample without prior filtration or cen-
trifugation. In practice, the reported color
value is generally the true or soluble color.
"Chemistry and Life Sciences Division, Research
Triangle Institute, Research Triangle Park, NC.
Accurate color determinations of effluents
from textile and carpet dyeing are complicated
by two factors: (1) the colors may be drastically
different from the yellow-brown hue to which
the APHA value is related and (2) the color
components may be insoluble in aqueous
media.
To overcome the first problem, members of
the American Dye Manufacturers Institute
(ADMI) have developed a modified color
method that is an extension of the APHA
method and that is "(1) applicable to any hue,
(2) sensitive to small color differences, (3) re-
lated to APHA values, and (4) requires
relatively inexpensive instrumentation" (ref.
2). This method has now been adopted by EPA
(ref. 3) and is under consideration for the 15th
edition of Standard Methods for the Examina-
tion of Water and Wastewater.
THE ADMI COLOR METHOD (ref. 2)
As stated above, the ADMI measurement
was designed to be an extension of the APHA
color method. As such, the color is defined as
soluble color and the samples must be prefil-
tered through a celite-coated filter crucible or
celite-coated glass fiber filter paper. Such
filtration will obviously remove particulate col-
or and may also remove certain dyes through
adsorption. In brief, the ADMI method in-
volves:
• prefiltration to remove insoluble mate-
rials,
• determination of tristimulus values with
the aid of a spectrophotometer or a
tristimulus colorimeter,
• conversion of the tristimulus values to
corresponding Munsell values by use of
appropriate tables or by calculation,
• calculation of tve value DE, (the Adams-
Nickerson Color Difference related to a
definition of the color space designed as
ANLAB), and
• conversion of DE to the ADMI value by
calculation.
307
-------
A variety of spectrophotometric equipment
can be used, ranging from a relatively inexpen-
sive colorimeter especially designed for the
procedure to a "double beam recording spec-
trophotometer coupled to a digital computer
(which) will eliminate calculation time and
generate ADMI values within a matter of sec-
onds" (ref. 2). Without the aid of tables and of
programmable calculators or computers, the
calcultions are extremely tedious.
The ADMI procedure has been subjected to
interlaboratory testing and a sample having a
color value of 100 units has been shown to have
a 95 percent confidence limit of less than ± 5
units.
CONCEPT OF COLOR SPACE
As noted above, the ADMI procedure is bas-
ed on a color difference formula devised by
Adams and Nickerson and recommended by
the Society of Dyers and Colorists in 1970. This
formula is defined as the difference "between
any two specimens...as the distance between
their positions in the uniform colour space
known as ANLAB" (ref. 4). The Adams-
Nickerson formula is quite complex, involving
a fifth-degree polynomial equation to quantify
the relationship between one of the three
tristimulus values and the Munsell value.
At this point, it may be appropriate to brief-
ly discuss the notion of color space, a very com-
plex concept accompanied by very complex
mathematical representations. The human eye
responds differently to various hues and
brightnesses. In the ideal color space represen-
tation, one unit of color difference at any one
area of the color space (e.g., in the yellow area)
would be exactly equal to one unit in any other
area (e.g., in the blue area). Therefore, a blue
solution of 100 color units would have the same
color depth as a yellow solution of 100 color
units. The color space is 3-dimensional and
nonspherical. A given color can be accurately
defined, or located in the color space, by three
primary or tristimulus values, providing these
values are properly weighted in the calcula-
tion. In practice, the three tristimulus values
correspond to standard red, blue, and green
filters. A relatively simple explanation of the
nature of color, accompanied by full color
diagrams, may be found in reference 5. Suffice
it to say that a multitude of mathematical for-
mulae have been devised to define the color
space.
The international authority for all aspects of
color is the Commission Internationale de
1'Eclairage (CIE). This organization has valiant-
ly attempted unification of the variety of dif-
ferent evaluation practices in the hopes of com-
ing up with "a single uniform color-space and
color-difference formula...which will serve
most industry applications satisfactorily" (ref.
6). The two most promising practices at the
present time are: (1) a color difference formula
based on uniform color spaces as defined by
CIE in 1964 and designated as CIELUV, and (2)
a cube root version of the Adams-Nickerson
color difference formula designated as
CIELAB. The latter has found more favor with
the textile and dyestuff industries. The ADMI
procedure can be modified to incorporate the
simplified CIELAB formula, thus greatly sim-
plifying calculation of color values. Detailed
discussions of the CIELAB may be found in ar-
ticles by McLaren and associates (refs. 4,7).
Other detailed discussions of color measure-
ment in the textile industry are included in
proceedings of a recent AATCC workshop/
seminar (ref. 8).
APPLICATION OF COLOR
MEASUREMENT PROCEDURES
TO ACTUAL WASTEWATERS
We have, then, a variety of ways to obtain a
number representing "color" of waters and
wastewaters. But what is the significance of
the measurement we have obtained? How does
this measurement, however precise and repro-
ducible, relate to the basic question: How much
color is leaving the effluent from the dyeing
operation and entering into the receiving
stream? Given this concern, one is forced to
question the relationship between the effluent
and the sample that is actually subjected to
analysis, i.e., the portion of the effluent that is
capable of passing through a celite filter. Dur-
ing extensive research on the characterization,
bioassay, and treatability of dyes and dyeing
wastewaters (refs. 9 through 13), we observed
that with certain wastewaters, such as those
from disperse dyeings, measurement of either
true color or apparent color failed to adequate-
ly represent the color intensity obvious to the
308
-------
eye. It appeared that the discrepancy was due
to (1) interference by turbidity, especially non-
colored suspended solids, in the apparent color
determination; and (2) removal of most of the
color in the prefiltration step of the true color
determination. Preliminary research on sev-
eral synthetic dyeing wastewaters (i.e., those
prepared in the laboratory with disperse dyes
and dispersants) substantiated the removal of
most of the color by such processes as mem-
brane filtration, diatomite filtration, glass fiber
filtration, and centrifugation (ref. 14). Obvious
ly the true color bears little resemblance to the
real color of the effluent unless the effluent has
been subjected to pretreatment to give color
removals equivalent to these processes.
The study just described also revealed that
with disperse dyeing mixtures, addition of
organic solvents such as acetone prior to the
filtration step would facilitate measurement of
most of the color. Much additional testing with
a variety of solvents, dyes, and dyeing waste-
waters will be required to determine the op-
timum conditions for measurement of color.
The first criticism is apt to be that addition of
solvent will lead to the overestimation of color
in the effluent and that unrealistic require-
ments for color removal will thereby result.
SUMMARY
At the present time there is not a perfect or
even a reasonably good method for accurately
estimating the total color of complex waste-
waters containing undissolved dyes. The
challenge facing us is development of an ac-
curate method that can be readily performed
by wastewater operators or by monitoring per-
sonnel, preferably with moderately priced
equipment. Two approaches to such a method
are modification of the existing procedures for
color analysis (for exam-pie, by solvent addi^
tion) or development of new procedures that
measure color in either the soluble or insoluble
state. The ultimate goal of either approach will
be the development of a method that gives a
realistic estimate of color in actual effluents
from textile and dyeing operations.
REFERENCES
1. American Public Health Association,
American Water Works Association, and
Water Pollution Control Federation, Stan-
dard Methods for the Examination of
Water and Wastewater, 14th ed., Method
204, "Color," American Public Health
Association, Washington, DC, 1976.
2. W. Allen, et al., "Determination of Color of
Water and Wastewater by Means of ADMI
Color Values," Proceedings, 28th In-
dustrial Waste Conference at Purdue
University, Engineering Bulletin of Pur-
due University, Engineering Extension
Series No. 142, p. 661, 1973.
3. U.S. Environmental Protection Agency,
Methods for Chemical Analysis of Water
and Waste, addendum, "Color (ADMI
units)," Storet Nos. 0082 and 0083, En-
vironmental Protection Agency, Office of
Technology Transfer, Washington, DC, un-
dated.
4. K. McLaren and B. Rigg, "XII-The SDC
Recommended Colour-Difference Formula:
Change to CIELAB," Jour. Soc. Dyers and
Colorists, Vol. 92, p. 337, 1976.
5. M. M. Lih, "Color Technology," Chem.
Engng., August 12, 1968.
6. Commission Internationale de 1'Eclairage,
Official Recommendations of Uniform Col-
or Spaces, Color-Difference Equations,
Metric Color Term, Supplement No. 2 to
CIE Publication No. 15, Colorimetry
(E-l.3.1) 1971, 19 pp., 1976.
7. K. McLaren, "XIII-The Development of
the CIE 1976 (L*a*b*) Uniform Colour
Space and Colour-Difference Formula,"
Jour. Soc. Dyers and Colorists, Vol. 92,
p. 338, 1976.
8. AATCC, Color Measurement in the Textile
Industry, American Association of Textile
Chemists and Colorists, Research Triangle
Park, NC, 1977.
9. L. W. Little and J. C. Lamb III, "Acute
Toxicity of 46 Selected Dyes to the
Fathead Minnow, Pimephales promelas,"
Dyes and the Environment, American Dye
Manufacturers Institute, Inc., NY, 1973.
309
-------
10. L. W. Little, J. C. Lamb III, M. A. Chill-
ingworth, and W. B. Durkin, "Acute Toxici-
ty of Selected Commercial Dyes to the
Fathead Minnow and Evaluation of
Biological Treatment for Reduction of Tox-
icity," Proceedings of the 29th Industrial
Waste Conference, Engineering Bulletin
of Purdue University, Engineering Exten-
sion Series No. 145, p. 524,1974.
11. J. W. Ericson and L. W. Little, "Ap-
plicability of ATP Measurements for
Determination of Dye Toxicity in Short
Term Algal Assays," J. Elisha Mitchell
Scientific Soc., Vol. 92, No. 2, 1976.
12. L. W. Little, and J. W. Ericson, "Biological
Treatability of Wastewaters from Textile
and Carpet Dyeing Processes," Pro-
ceedings of the 8th Mid-Atlantic Industrial
Waste Conference, January 12, 1976,
Newark, DE, pp. 201-216.
13. T. Crowe, C. R. O'Melia, and L. Little, "The
Coagulation of Disperse Dyes," Pro-
ceedings of the 32nd Purdue Industrial
Waste Conference (May 1977), Ann Arbor
Science, Ann Arbor, MI, pp. 655-662, 1978.
14. R. Goodman, "Evaluation of the ADMI Pro-
cedure for Measuring Color in Textile Dye-
ing Wastewater, Master's report (under
the direction of L. W. Little), Department
of Environmental Sciences & Engineering,
University of North Carolina, Chapel Hill,
NC, 1977.
310
-------
USE OF BIOASSAY SCREENING TECHNIQUES FOR NPDES* PERMITS
W. H. Peltier, L. B. Tebo, Jr.t
Abstract
Effluents limits placed in the NPDES per-
mits do not take into account the presence of
byproducts, degradation compounds, unre-
ported compounds, the synergistic action of
these compounds, nor receiving water charac-
teristics in the control of toxicity to aquatic
organisms. In order to evaluate the toxicity of
complex effluents being discharged to fresh-
water and saltwater environs, EPA Region IV
is including, in selected NPDES permits, a
bioassay screening requirement. This require-
ment is in the form of periodic 24-h acute static
toxicity tests which result in a pass or fail
criteria. The criteria are based on 50 percent
survival of selected standard test organisms
exposed to a single percent effluent volume.
Failure to meet the criteria results in requiring
the permittee to conduct additional studies
such as: acute and/or chronic flow through tox-
icity testing, receiving water biological
monitoring, bioaccumulation studies, waste-
water persistence studies, flavor impairment,
and avoidance studies. The bioassay screening
requirement is not intended to replace the
traditionally required parameters in a permit
such as BOD, COD, and chemical-by-chemical
limitations, but is an additional compliance
parameter.
One of the goals of the 1977 Clean Water Act
states that, "it is a national policy that the
discharge of toxic pollutants in toxic amounts
be prohibited." Historically in Region IV, the
U.S. Environmental Protection Agency (EPA)
has been conducting bioassays at facilities
discharging complex wastewater to a receiving
water. These bioassays were conducted to
evaluate the lethality of complex wastewater
discharges to selected aquatic organisms.
Present EPA regulatory programs to con-
trol the discharge of toxic wastewaters em-
*National Pollutant Discharge Elimination System.
'U.S. EPA, Region IV, Surveillance and Analysis Divi-
sion, Athens, GA.
phasize specific limits on individual chemicals.
For chemicals of known toxicity and for which
economical methods of identification and quan-
tification are available, the specific chemical
limit is unquestionably a most efficient ap-
proach.
Unfortunately, the individual chemical ap-
proach provides no control of the tens-of-
thousands of chemicals for which toxicity data
are lacking or inadequate and chemical charac-
terization techniques are unavailable or pro-
hibitively expensive. The individual chemical
approach also provides little means for con-
trolling the toxicity of wastewater discharges
where toxicity is a function of the mixture of
chemicals in the wastewater. In this situation,
toxicity testing must be used to control the
discharge of toxicants.
A measure of the lethality problem of com-
plex wastewater discharges is indicated by
Region IV's experience with 64 onsite 96-h
flow-through toxicity tests where 49 of the
discharges were lethal to selected aquatic
organisms. Twenty-seven of fifty complex
wastewater discharges tested at the EPA
Athens, Georgia laboratory using 96-h static
toxicity tests were lethal.
The chemical analyses of many of the waste-
waters tested for toxicity disclosed the inade-
quacies of the chemical by chemical approach
in protecting the aquatic community. Up to 50
unpermitted chemicals have been detected in a
single wastewater source being discharged to
the receiving water. Of course, many of these
chemicals were byproducts on degradation
products of which very little is known regar-
ding their toxicity.
It is apparent that regulatory agencies must
have a means of controlling the discharge of
toxic complex wastewaters and the innumer-
able pollutants of unknown toxicity. The most
direct approach to this problem is estab-
lishment of limits based on the response of
aquatic organisms to potentially toxic
wastewater discharges. Therefore, EPA is
developing a program for the second round of
NPDES permitting whereby toxicity limits
311
-------
based on the lethality of the wastewater to
aquatic organisms is made a part of the
NPDES permit. The permittee would be re-
quired to conduct toxicity screening tests at
the wastewater concentration (percent waste-
water) specified as a toxicity limit in the per-
mit.
The screening test would be based on a
single wastewater concentration utilizing 24-h
acute static tests. A monthly testing frequency
would be required using a selected test
organism. The selected organisms would be
based on whether the wastewater is
freshwater or saltwater. If the wastewater is
fresh, then either the fathead minnow
(Pimephales promelas) or water flea (Daphnia
magna or pulex) will be used. On the other
hand, if the wastewater is saltwater, then
either the sheepshead minnow (Cyprinidon
variegatus) or mysid shrimp (Mysidopsis sp.)
would be used as the test organism. The pro-
tocol for the screening test is in the EPA
publication Methods for Measuring the Acute
Toxicity of Effluents to Aquatic Organisms
(ref. 1).
The toxicity limit calculation for the single
test concentration is based on the following
formula:
ill result in a safe concentra-
LC
LC
where:
AFi
LC
50
we
AF =
the concentration (expressed as per-
cent) at which the wastewater is
lethal to 50 percent of the test
organisms,
receiving water wastewater concen-
tration (expressed as percent) at the
7Q10 flow in a river, or for a lake or
estuary, the concentration at the
boundary of the mixing zone, and
numerical value which, when ap-
plied to acutely toxic concentration
+If the result of this calculation exceeds 100 percent for a
river, then the screening toxicity test must be conducted on
undiluted (100 percent) wastewater with a 90 percent or
greater survival rate required of the test organisms.
§If the permittee can demonstrate that the wastewater
toxicity is not persistent and/or bioaccumulative in the en-
vironment, then an application factor of 0.05 can be used,
otherwise an application factor of 0.01 is utilized in the for-
mula.
tion.
Example 1 Wastewater discharged to a
river via a diffuser system to promote rapid
mixing.
Data required by permit writer are as
follows:
Mean daily wastewater flow = 2 mgd
7Q10 = 270 mgd
Application factor = 0.01 for a per-
sistent wastewater.
Toxicity limit:
QW
> QW + QR
—
X 100
AF
LC50 ^
2 + 270
X 100
0.01
^ 74 percent.
Thus, in conducting the screening test, the
permittee would use 74 percent wastewater by
volume and 26 percent dilution water.
Example 2 - Wastewater discharged to a
lake or estuary via a diffuser system to pro-
mote rapid mixing.
Data required by permit writer is as follows:
•Permittee is required to display on a map
isopleths of wastewater concentrations
occurring in the receiving water (Figure 1).
•Numbers of hectares within each isopleth.
•Ecologically and economically important
resources in the receiving water are to be
identified for each isopleth.
•Application factor to be used for persistent
wastewater is 0.01.
Toxicity limit:
T _, . mixing zone boundary
L,Csn £ ———
0.25 ^
'50
LC50 ^
Percent-
Thus, in conducting the screening test the
permittee would use 25 percent wastewater by
volume and 75 percent dilution water.
Failure of a screening test must be reported
in writing within 5 working days after the test
312
-------
is completed. Within 90 days after any test
that caused a lethality of 50 percent or more,
the permittee must develop and submit to
EPA Region IV a plan and schedule for reduc-
ing the toxicity of the wastewater to a safe
level. Once the toxicity is reduced to the
NPDES permit toxicity limit, a return to the
monthly screening toxicity tests are to be in-
itiated.
In summary, the screening toxicity test
described in this paper is an additional ap-
proach being used by EPA Region IV in the is-
suance of second round NPDES permits. Both
specific limits on known toxic chemicals as well
as screening toxicity tests will be used to con-
trol toxicants discharged into our receiving
waters.
REFERENCE
1. U.S. Environmental Protection Agency,
Methods for Measuring the Acute Toxicity
of Effluents to Aquatic Organisms, U.S.
Environmental Protection Agency, En-
vironmental Monitoring and Support
Laboratory, EPA 600/4-78-012, p. 52,1978.
313
-------
Session VI: ENERGY AND MATERIALS CONSERVATION
John R. Rossmeissl, Session Chairman
315
-------
RECLAMATION OF WARP SIZES USING THERMAL PRECIPITATION
Warren S. Perkins, Robert P. Walker, Leo J. Hirth*
Abstract
A novel method for size reclamation based on
precipitation of the size material by heating
the desize washwater is reported. The method
uses hydroxypropylatedcellulose (HPC), which
is soluble in cool water but insoluble in warm
water. Separation of the thermally precipitated
hydroxypropylcellulose from water is dis-
cussed along with the applicability of the HPC
as a warp size and the economics of the proc-
ess.
INTRODUCTION
Size materials constitute a major portion of
the biochemical oxygen demand (BOD) and dis-
solved sol:ds in the wastewater from woven
fabrics finishing processes. The BOD contrib-
uted by desizing starch-sized fabrics may be 45
percent or more of the BOD load from the fin-
ishing plant. The BOD load from desizing may
be lowered by using synthetic sizing polymers
such as polyvinyl alcohol (PVA) or carboxy-
methylcellulose (CMC) either alone or in mix-
tures with starch. However, these synthetic
polymers do not easily biodegrade and remain
dissolved in the finishing plant wastewater.
A process to recover and recycle warp sizes
probably has the potential to lower the BOD,
COD, and dissolved solids in textile finishing
plant effluents more than any other single
change in textile processing. Such processes
have been the subject of much research and
development work in recent years. The follow-
ing are specific examples of approaches to size
reclamation:
• Reclamation of carboxymethylcellulose by
chemical precipitation from the desize
wastewater has been studied but not com-
mercialized.
• Recovery of polyvinyl alcohol by chemical
'Department of Cc-xtilc Engineering, Auburn Univer-
sity, Auburn, AL.
precipitation has also been subject to in-
vestigation.
• Studies of recovery of warp sizes using
solvent sizing and desizing have been
made, and a solvent sizing and desizing
system has been developed by two Italian
machinery companies.
• Recovery of PVA by ultrafiltration of the
desize wastewater is now being done com-
mercially.
The tremendous interest of the U.S. Textile
Industry in the reclamation of warp sizes is un-
derstandable since the economic outlook for
such a development is as favorable as are the
environmental considerations. The size mate-
rial adds no value to the fabric. It is applied to
the yarn in slashing to improve the efficiency
with which the yarn can be woven into fabric.
It is normally removed in desizing and dis-
posed of in a wastewater treatment facility.
Therefore, the size material itself adds to the
cost of fabric production. A process for recla-
mation of the size material would prove profit-
able to the textile mill if the cost of reclaimed
size is lower than that of virgin size.
PRINCIPLE OF SIZE RECLAMATION USING
THERMAL PRECIPITATION
The size reclamation process that is the sub-
ject of this paper uses a size selected from a
group of polymers having the unusual property
of being soluble in water at room temperature
but insoluble in water at some higher tempera-
ture. Hydroxypropylcellulose (HPC), on which
most of the research work at Auburn has been
based, is such a polymer. HPC, which is soluble
in water at room temperature, becomes almost
completely insoluble in water above about
45° C. Therefore, precipitation of HPC from
desize wastewater can be accomplished by
merely heating the desize wastewater to above
45° C. A schematic diagram of the desizing and
recovery system being studied is shown in Fig-
ure 1. Washing of the HPC from the fabric
takes place in water at a temperature below
40° C. The desize water containing HPC at a
317
-------
Fabric in
*^\
^_ Fabric out
Desize Washwater
(1.01 size)
Desize
Water at < 40
Reclaimed Water
Fresh Water
(approx. 101 of total)
Filter
for fiber
and insolu-
ble impuri-
ties.
Cooling
of
reclaimed
water
Heat to
• 45°C
Water/Size
Separator
reclaimed
water
Concentrated
<|r size to slasher
(8-101)
Figure 1. Schematic of desize and recovery system.
concentration of 0.5 to 1.5 percent is filtered to
remove insoluble impurities and heated to
above 45° C to precipitate the HPC. The precip-
itated HPC is then separated from most of the
water. The HPC redissolves when cooled to
yield a solution that can be reused in slashing.
The recovered water can be cooled to below
40° C and reused in the desize washer.
Hydroxypropylmethylcellulose (HPMC) also
exhibits the thermal precipitation phenomenon
and can be used in the process.
DESIGN OF WATER/HPC SEPARATOR
The heart of the reclamation system being
studied is the separation of the precipitated
particles of HPC from most of the water to ob-
tain a size solution of sufficient concentration
to reuse in slashing. Separation of the precip-
itated HPC by filtration and centrifugation
have each appeared feasible in the laboratory.
However, the most promising separation tech-
nique tested was settling by gravity.
Thermally precipitated HPC separates from
water as a colloidal dispersion. The HPC par-
ticles have a negative zeta potential and can be
made to coagulate and settle rapidly by adding
a cationic agent. A variety of these settling
assistants were tested for their benefit to the
HPC recovery process. Ions such as Na + ,
CA "*" 2, and Al+ 3 cause thermally precipitated
HPC particles to coagulate and settle rapidly.
The coagulating effect increases with higher
cationic charge and with increase in concentra-
tion of cations in the concentration ranges used
in this work. Cationic polyelectrolytes may
also be used to enhance the settling of HPC
particles in water. The most useful settling
assistant used in this work was aluminum sul-
fate. One-half to one percent aluminum sulfate
based on the weight of the HPC is sufficient to
cause rapid settling of thermally precipitated
HPC in water. This level of aluminum sulfate
does not affect the solubility of HPC in water.
When the water temperature is below 45° C
the HPC is readily soluble in water containing
318
-------
500
400
300
200
160
3 120 -I
•M
I
-P
•p
Q)
80
40 ,
2 4 6 8 10 12 14 16 18 20 22
Time '(mmites)
Figure 2. Settling rate of precipitated HPC at 50° C.
319
-------
1 percent aluminum sulfate on the weight of
the HPC. Analysis of the recovered HPC (con-
centrate) and the recovered water revealed
that the aluminum ions concentrated them-
selves in the recovered HPC. That is, the alum-
inum ions become a part of the reclaimed HPC
solution and addition of aluminum to the desize
washwater is necessary only when virgin HPC
is used. Several cycles of use and reclamation
of HPC can be run without any additions of
aluminum sulfate after the initial run. Studies
of physical properties of films of HPC revealed
that 1 percent by weight of aluminum sulfate in
the HPC caused no deterioration of properties
of the material.
Figure 2 illustrates the settling rate of
precipitated HPC in water. The volume of the
test solution was 450 ml and it contained 4.5 g
of HPC and 0.030 g of aluminum sulfate. In 15
to 20 minutes virtually all of the HPC had set-
tled into the bottom 45 ml of the container, a
500-ml graduated cylinder. Experiments of this
type have yielded solutions containing up to 30
percent by weight of reclaimed HPC at typical
recovery levels of 85 to 95 percent of the HPC
in the desize washwater.
UTILITY OF HPC FOR SIZING
The economic success of a textile greige mill
depends to a large degree on the efficiency
with which yarns may be woven into fabric.
Therefore, the performance of the size mate-
rial in protecting the yarn is of utmost impor-
tance. Since HPC has not previously been used
as a warp size, an important aspect of the work
that is being done at Auburn is the evaluation
of the performance of HPC as a warp size.
Three types of studies regarding the perform-
ance of HPC as a warp size are in progress.
They are:
1. Properties of films of HPC.
2. Properties of yarns sized with HPC.
3. Weaving performance.
The primary (and sometimes only) ingredient
in a size formulation is a film-forming agent.
Studies of size materials in the form of films
allows evaluation of the effect of various ad-
ditives and contaminants on material proper-
ties that may affect performance of the size
material. Tensile strength and elongation of
films of HPC, PVA, and CMC are compared in
Table 1. The strength of HPC is similar to that
of CMC, which is considerably lower than that
of PVA. However, the strength of HPC is ade-
quate for warp sizing. PVA is sometimes con-
sidered too strong for fine yarns and can cause
problems if it does not split easily enough at
the slasher lease rods. The elongation of films
of HPC is intermediate between that of CMC
and PVA, both of which are successfully used
as warp sizes. Other properties of HPC such as
adhesion to various fibers; resolubility in
water during desizing; and stability to me-
chanical action, chemical action, and moderate
heat in use appear to be adequate for the recla-
mation process.
Typical results of tests of properties of
yarns sized with HPC are shown in Table 2 and
compared to properties of yarns sized with
PVA. The effect of HPC on the breaking
strength and elongation of the polyester/cotton
yarns was about the same as the effect of PVA.
The elongation of the sized yarns was about 2
to 3 percent lower than that of unsized yarns.
The strength of the yarns increased by about
15 percent as a result of sizing with either PVA
TABLE 1. PROPERTIES OF HPC*, CMC* and PVA*
Property
Tensile strength (psi)
Elongation at break (%)
PVA
3,100
200
Materials
HPC
800-1,300
75
CMC
1,000-1,500
7
*Hercules, Inc., Klucel J.
* Hercules, Inc., Warp Size Grade CMC.
*Du Pont, Elvanol T-25.
320
-------
TABLE 2. PROPERTIES OF 36/1, SO/SO POLYESTER/COTTON
YARNS SIZED WITH PVA OR HPC
Sample description
Unsized control
PVA with lubricant
HPC
Size concentration
in formulation
—
10
10
Break factor
(oz. x counts)
432
500
501
Elongation
(%}
13.4
10.7
11.0
Abrasion
Resistance
(#cycles)
20
2,000+
1,200
or HPC. The abrasion resistance of the yarns
improved drastically as a result of sizing with
HPC or PVA. These yarns must withstand at
least 300 to 400 abrasion cycles to be
weaveable. The physical properties of yarns
sized with HPC are adequate for weaving
loosely constructed fabrics, but the abrasion
resistance is borderline for tightly constructed
fabrics such as sheeting.
ECONOMICS OF SIZE RECLAMATION
Estimates of the economics of reclaiming
HPC are shown in Table 3. These figures are
based on settling the HPC in large tanks. Also
shown are cost figures on an ultrafiltration
system for PVA recovery and some costs for
use of conventional desizing followed by waste-
water treatment.
The initial capital investment for an ultrafil-
tration reclamation system for 272,232 kg
(600,000 pounds) of fabric per week is esti-
mated to be $805,000 while that for a thermal
precipitation system is $262,600. Annual oper-
ating costs including labor, utilities, insurance,
taxes, maintenance materials, and depreciation
total $249,500 for ultrafiltration and $82,850
for thermal precipitation. The cost of energy
for running a thermal precipitation recovery
system will be about $54,000 per year less than
that required for a conventional desize process
or an ultrafiltration system because of lower
water temperature requirements. The cost of
size for a thermal precipitation system is
$507,500 annually, considerably higher than
the $460,000 for a conventional system and the
$252,000 for an ultrafiltration system using
PVA. Total annual operating costs for a ther-
TABLE 3. ECONOMICS OF SIZE RECLAMATION (BASIS IS 600,000 POUNDS
OF FABRIC PER WEEK)
Initial capital investment
Annual operating costs*
Energy for hot water
Size materials
Wastewater treatment §
Total annual operating costs
Ultrafiltration
$805,000
249,500
75,000
252,000f
20,000
$596,500
Thermal
percipitation
$262,600
82,850
21,000
507,500*
14,000
$625,350
Conventional
—
—
75,000
460,000
198,000
$733,000
•Excluding size material costs, cost of energy for hot water and cost of wastewater treatment.
t Based on 20 percent loss of PVA.
*Base' on 15 percent loss of HPC.
S Estimated for 1983.
321
-------
mal precipitation system are estimated to be
slightly higher than for ultrafiltration but less
than conventional desizing systems.
ACKNOWLEDGMENT
The work reported in this paper was per-
formed under the partial sponsorship of the
U.S. Environmental Protection Agency through
Grant No. R 805128-010. The work was per-
formed by the Auburn University Department
of Textile Engineering in cooperation with Ala-
bama Textile Education Foundation through
the Auburn University Water Resources Re-
search Institute with assistance from the Au-
burn University Engineering Experiment Sta-
tion. This paper covers work completed by
August 31, 1978. The project is scheduled for
completion by December 20, 1979.
322
-------
AN INDUSTRIAL WASTEWATER RECIRCULATION SYSTEM FOR
THE FIBROUS GLASS TEXTILE INDUSTRY
S. H. Thomas, D. R. Walch*
Abstract
In this presentation the concept, develop-
ment, and installation of an industrial waste-
water recirculation system for the fibrous glass
textile industry will be addressed.
INTRODUCTION
For those not familiar with fibrous glass
processes and products, I'll briefly describe
them. At Owens-Corning Fiberglas facilities,
the process begins with the mixing and feeding
of raw batch to a furnace. Furnaces for textile
processes are either gas- or oil-fired and are of
the regenerative or recuperative type. The
molten glass from any of the furnaces drains in-
to a forehearth and through a bushing. Bush-
ings are of a platinum material with a number
of orifices in various sizes to satisfy the end
product. The fibers have either binders or sizes
applied and are then cured. The final products
of our process include fibrous glass in forms for
further manufacture, fabrication, or in many
instances, for finished product.
Owens-Corning, product-wise, is essentially
divided into four categories: textile, insulation,
roofing, and special products. Owens-Corning
manufactures everything from a fiber one-
tenth of one-thousandth of an inch in diameter,
to bathtubs and huge underground gasoline
storage tanks, to the new 105-acre Fiberglas
fabric roof for the Haj Terminal at the Jeddah
International Airport in Saudia Arabia. How-
ever, this report will be limited to discussion of
textile products as manufactured at our Ander-
son, South Carolina, facility.
To start my story, I will have to take you
back 20 years to 1958 when I first started to
work with Owens-Corning Fiberglas and the in-
dustrial wastewater systems of that era. Like
most other industries of that time, we were
•Owens-Corning Fiberglas Corporation, Toledo, OH.
busy trying to treat industrial wastewater for
suitable discharge. My first industrial waste-
water problem was finding a suitable manner
to discharge approximately 4,000 gallons j er
minute of process water with a high BOD and
COD, and with suspended solids and phenol in
concentrations of up to 600 ppm from an insula-
tion facility.
The first step was to look for methods of
treatment available. At that time little was
known about the treatment of phenols and no
information was available from any fibrous
glass facilities either in the United States or
throughout the world. We started from scratch
using the famous old method of trial and error.
By 1960 I had developed an extended aeration
system for the degradation of the phenol as
well as the other parameters listed. Phenols
could be reduced from 1,000 ppm to about 2
ppm, which was a safe limit for fish as deter-
mined by bioassay tests. I recognized at this
time that this system would provide me with
about 10 years of breathing room before find-
ing alternate methods to further reduce the
discharge of contaminants into the waterways
of our nation. In that same year, 1960,1 also in-
herited the four textile operations in the com-
pany. It might be of interest at this point to
mention that in 1960 Owens-Corning had nine
operating facilities in seven States. Today we
have 90 facilities in 27 States.
Listening to the thoughts and ideas being
mentioned at that time in Washington, DC, and
State capitols convinced me that all our ini-
tiative should be directed toward total recir-
culation of all industrial wastewater within our
corporation. It was an idealistic goal, but it was
worth a desperate effort. The benefits at that
time seemed justifiable and, of course, today
are stupendous.
The textile industrial wastewater treatment
systems of that era consisted of primary and
secondary systems for the combined sanitary
and industrial wastewater. These systems ade-
quately reduced the contaminants for safe dis-
cha -ge to waterways. At this time, we were us-
323
-------
ing starch sizes for our textile products; this
was many years before we started using exotic
chemicals for sizes to meet product uses and
qualifications.
Being more familiar with the insulation prod-
ucts' wastewater from previous work, this ap-
peared to be the area of greatest concern in the
future. I directed my attention to improving
the insulation products' wastewater systems.
With this strong urge and my own headstrong-
ness, I quietly developed an initial concept for
a totally recycled system.
In the next 8 years I had to listen to the "No,
it won't work," etc., etc., but continually tried
to find answers and ways around the ever-
rising problems. By 1968, we turned the valves
and started our first totally recycled industrial
wastewater system at Barrington, New Jer-
sey. I might add at this point that during these
trying years top-level management was behind
me, helping wherever necessary.
In 1969, modifications were made to two
other insulation facilities in Texas and Georgia,
and these facilities converted to totally recir-
culated industrial wastewater systems (See
Figure 1). One new facility has been added
since then and it too is on total recycle, as will
all other new facilities of this type in the
future. Our record on insulation facilities as of
today is five plants on total recycle and two
more under construction, both to be on recycle
systems by 1980.
With the taste of success at our insulation
plants, we directed our attention to the textile
facilities. The first step for any process to be
adaptable to a totally recycled system is to con-
vince management, and particularly plant man-
agement, that you can develop the concept into
a workable system, with the necesssary
justification. Once this is accomplished, all
other steps are left completely up to the tech-
nical competence of the team working on the
City W = 92
Process Uses
S = 204
W = 204
Cooling Uses
S
146.3
Losses to
W
85.6
Atmosphere
S = Summer Flow, GPM
W = Winter flow, GPM
Reclaimed Water to Distribution S = 259.3
W = 198.6
Figure!. Overall water balance for recirculation.
324
-------
project. The second step is to reduce any and
all waste flows at the facility so that the de-
mand for water is greater than the discharge.
In the case of the Anderson textile facility, it
was necessary to devise piggyback systems, or
higher quality water usage discharged to lower
quality water usage systems, before final
discharge to the treatment facilities.
Once the demand and supply balance had
been established, it was necessary to analyze
the final effluent as well as determine the
parameters required for water usage in the
various systems. This is a very crucial phase
since complex facilities have discharges where
the parameters can vary frequently. Usually a
long-term sampling program is required to es-
tablish the basis for treatment as well as reuse.
Generally there are one or more systems that -
have more critical requirements for treatment
than all others. At Anderson one such system
existed.
At the Anderson facility it was found
necessary and desirable to develop the treat-
ment required to satisfy this system, then
design a small pilot trial including actual use of
the field conditions to provide a more reliable
system. At Anderson the critical water usage
system was the cooling of a heat exchanger
which in turn cooled the glass fibers as they
were formed (See Figure 2). The temperature
of the glass fibers at the tip of the heat ex-
changers is approximately 1,800° F. This must
be significantly reduced before the sizing can
be applied. The outside temperature of the
heat exchanger ranges from 150° to 190° F.
The cooling water flow through the heat ex-
changer is regulated to maintain an exit water
temperature of approximately 100° F. It is
Reclaimed Water
Makeup
iupply
Pump
Storage Reservoir
(4300 Gallon)
Gas Chlorlnatlon
i
Carbon Adsorption)
Sand
Existing
Treatment
Facilities
Cooling System
Cold Well
(1200 Gallon)
Five Manufacturing Machines
(Approximately 12 coollg loops per machine)
Figure 2. Schematic flow diagram - final pilot cooling system.
325
-------
under this hostile environment that the cooling
water and heat exchangers must function prop-
erly if the high integrity of the glass fiber is to
be maintained. This meant that parameters of
the reuse water for such a system had to be
such that there could not be physical separa-
tion that would cause buildup on the heat ex-
changers which in turn would greatly affect
the process—if not shut down the process com-
pletely. The hot water would then pass
through cooling ponds and again no physical
separation could take place. The preliminary
design of the treatment included the primary
and secondary systems that already existed
plus sand filtration, carbon adsorption for
organic removal, and finally, disinfection. To
make the whole treatment workable, it was
necessary to retrace some of our steps, and im-
prove and optimize the primary and secondary
phases. When this was accomplished, the pri-
mary and secondary phases were optimized in
the field, based on the lab bench studies. A
properly sized sand filter was designed and in-
stalled to handle 10 to 20 gallons per minute. A
carbon column of like size was also installed. Of
course, all of these steps were reviewed on a
constant basis with the plant personnel and
more particularly the production personnel.
This type of coordination makes the plant a
part of the whole process and the desire for a
total recycle system "rubs off on its key per-
sonnel.
The pilot system operated off the effluent
from the secondary or activated sludge sys-
tem. It was previously determined by the pro-
duction and management personnel that the
system would have to operate successfully for
a period of at least 120 days. As always, there
are bugs to work out in any new system. This
was so at Anderson. It was necessary to oper-
ate the pilot system for 210 days instead of 120
days.
The first plateau had been reached. The pilot
system proved practical and feasible — confirm-
ing our ability to construct a full-scale system
to satisfy the most critical situations as well as
all the variables at a complex manufacturing
facility.
The next step, of course, is to collect all the
data accumulated to date including param-
eters, preliminary design drawings, cost esti-
mates, approvals, comments, benefits, justi-
fication, etc., and present this to management
"for approval of funds for construction. It usual-
ly is necessary to obtain approvals from pollu-
tion control agencies.
I found it desirable to locate a consultant at
the beginning to work with the corporation
personnel throughout the project as required.
This keeps the consultant up to date so that
when work is required, no time is lost in
reviews. Owens-Corning Fiberglas had review-
ed consultants and their personnel during the
concept stage and selected Engineering
Science, Inc., of the Atlanta office, to work with
us frequently throughout the project. They
performed such duties as assisting on the
water balance, assisting on analysis of param-
eters, developing the pilot system from con-
cept, operating the pilot system, collecting
data, report writing, final design, equipment
specifications, and studying and evaluating as
required. All other work was accomplished by
corporate staff and plant personnel.
Once all the approvals and funds were
received, it was necessary to convert all the
data into a final design, develop the equipment
specifications, purchase materials and equip-
ment, let contracts, and begin the construction
work. This phase of the work is generally pro-
cedural; however, there are always questions,
changes, and modifications to be taken care of
so that the final system is workable.
The design work was completed in March
1977 and the actual fieldwork started at the
same time. There were more problems than
ever anticipated during this phase due to the
use of newly developed materials for the first
time ever in such a project. The major com-
ponents were completed by September 1977.
Those phases of the system were put into oper-
ation; however, recycling was never really ac-
complished until May 1978 and was on an inter-
mittent basis. There were many additional
problems encountered during this phase,
mainly because the primary and secondary sys-
tems' efficiency was greatly deteriorated by
the cold weather during the fall and winter of
1976 and 1977. At a time in the spring of 1977
when the systems should have snapped back,
nothing happened. It was necessary to go all
the way back to step one and concentrate on
the primary and secondary phases of the opera-
tion. It took from the spring of 1977 until
January 1978 to define the problems and make
the corrections required. During this time
326
-------
many investigative techniques were per-
formed including:
* review ol chemical usage dn production,
» Jar testing of coagulants,
* biological treatabiHty studies, and
» nutrient balance.
A review of production fecords and chemical
usage indicated that not only had the cold
weather affected the treatment operations, but
also some changes in chemical usage in produc-
tion (that were originally thought to be in-
significant) had a combined effect. However,
after a thorough review we still could not fully
account for the poor treatment efficiencies.
Therefore, jar tests were performed on the
primary clarifiers to reduce colloidal emulsion
carryover. Earlier studies, performed during
the pilot stage of the recycle system, indicated
that primary colloidal carryover would pass
through further treatment, remaining, in the
tertiary-treated effluent. As it turned out, the
jar tests were successful at developing a chem-
ical addition scheme which could remove the
suspended solids found in the raw wastewater.
Once this phase of the system was properly
corrected, the loading on the secondary system
was reduced.
Biological treatability studies were per-
formed next on the "improved" primary ef-
fluent. These studies indicated a nutrient defi-
ciency existed as a result of the shift in
chemical usage during production. Therefore,
supplemental nutrient addition was imple-
mented. The results were astounding—the
most significant to date. With each incremental
addition of nutrient, secondary effluent quality
improved. In addition, analyses were per-
formed on the secondary effluent to monitor
and control the nutrient addition. These
modifications enabled the primary and second-
ary systems to perform as they had during the
original pilot design period. Efforts were made
to further continue these improvements. These
Huded modifying the sludge wastage pro-
cedure, wasting from the aeration tanks rather
than the usual approach «f wasting from the
return sludge line. This allowed the plant
operators to more «asilf control the sludge
wastage.
The wasted sludge -was piped into the
aerated equalization tanks to further extract
any usefulness "f this biological sludge. The in-
tent was to utilize the sludge in equalization
tanks as a roughing system to reduce the im-
pact of the raw waste on the main body of the
secondary system. Our experience in defining
and solving problems in this system serves as a
resource upon which future projects can be ap-
proached with new insight and greater oppor-
tunity for achievement.
At this point in the presentation, it is best
for me to describe our total system (See Fig-
ures 3 and 4). The industrial wastewater is col-
lected from all of the operations at the facility
in a separate sewer system. Up until the time
work was started on this project, all industrial
wastewater was collected in a combined sewer
system with all the sanitary wastewater from
the facility (which employs approximately
2,000 personnel). The industrial wastewater
enters the treatment facility through a bar
screen and then flows into one of the two
equalization tanks. The flow to the treatment
plant averages 295 gallons per minute and flow
range is generally between 200 and 360 gallons
per minute. The two equalization tanks are of
capacities of 250,000 and 150,000 gallons, and
both are equipped to provide continued aera-
tion. The influent is collected in one equaliza-
tion tank and pumped to the influent weir,
which is adjusted daily, if necessary, to control
the flow through the system. Excess water is
collected in the second equalization tank and
continually blended. I should add that there
are three tanks with a total capacity of 55,000
gallons ahead of equalization that are utilized
to collect shock loads. This is accomplished by
cooperation and communication throughout
the entire product facility. When someone has
strong or other unusual wastes to discharge, a
phone call is made to the treatment facility
operators and the waste is tested for and, if
need be, collected in one of the storage tanks.
After tests are conducted, and experience
plays an important role here, the waste is
treated chemically and then metered into the
treatment system at a predetermined slow
rate to reduce the effect on the system.
Once equalization has been accomplished
and the flow regulated into the system, the
water enters a chemical mix tank where ferric
chloride at the average rate of 10 ppm is added.
Caustic soda is med to control pH, when re-
quired^
" In the flocculation tank, which is next in line,
327
-------
Chemical
Factory
"A", "B", & "D" Factories
•[Bar Screen ]^^^-
Mat Lines
o
Basket
Equalization
Flocculation
& Neutralization
Betsy
Creek
Figure 3. Process flow diagram — existing wastewater treatment facilities.
an anionic polymer is added to aid flocculation,
and nutrients (ammonium nitrate and phos-
phoric acid) are added if required. Wasted ac-
tivated sludge is also added at the entry of this
tank. It has been determined that the wasted
activated sludge aids in coagulation and floc-
culation and does reduce chemical usage.
The water is mixed and then distributed to
the five clarifiers. These five clarifiers were
part of the original construction and were
utilized as they existed, thus providing addi-
tional flexibility in the system. The total
capacity of the clarifiers is 44,500 gallons pro-
viding a retention rate of 2.5 hours at average
flow. The flow from the clarifiers is collected
and mixed with activated sludge which is re-
cycled at the rate of 67 percent before entering
the three activated sludge tanks. Swing arm
diffusers are used for aeration and approx-
imately 1,000 ft3/min of air is provided at
average flow. The effluents from these tanks
are collected and then equally distributed to
the three final clarifiers. These three tanks
have a total capacity of 37,400 gallons and thus
provide a 2.2-hour retention time at average
flow. Activated sludge is pumped off continual-
ly at a controlled flow and recycled to the head
of the aeration tanks, and the wasted sludge is
pumped back to the influent to the flocculation
tanks. The effluent from the final clarifiers is
measured in a Parshall Flume and flows by
gravity to the sand filter. The sand filter
design rate is 4 gpm/ft2.
The flow is collected in a clear well and
pumped into one of three carbon columns. Each
of these carbon columns is 10 feet in diameter
and 20 feet high with a capacity of 20,000
pounds of activated carbon. The system has
the flexibility of utilizing a lead column and a
lag column in series, and a third column is on
standby, ready for use. There is also a fourth
column in the grouping that is used for new or
spent carbon storage. The entire carbon ad-
sorption system operation is semiautomatic.
328
-------
Filter a Adiorber Backwash
Secondary
ClartNer Effluent
City H2o Distribution Tank
To Equalization
Basins
rank
Ott-
Speclflcatlon
Basin
Reclaimed
Wastewater
Storage
Basin
roP
ForR
Figure 4. Process flow diagram - advanced wastewater treatment facilities.
The personnel have immediate control of the
system; however, valves and other operating
controls operate automatically once the
changes are started.
The flow from the carbon system is chlori-
nated and measured prior to discharge to the
230,000-gallon distribution tank. This clear
well is covered to protect it from any fugitive
dust and an oversow line discharges any sur-
plus treated water to a 1,500,000-gallon holding
pond. At the present time, the water in the
holding pond is cycled through the sand filter
and carbon adsorption system before reuse;
however, a study is underway to determine if
this water could be used directly as makeup in-
to the distribution tanks. It may be necessary
to cover the entire pond or a portion of it for
protection.
The treated water is pumped from the distri-
bution tank back to all phases of product opera-
tions at the Anderson plant. This is accomplish-
ed by two pumps that maintain a minimum of
44 pounds of pr ;sure throughout the system
and at a flow rtquired by the product opera-
tions. A completely new distribution system
has been installed to each area of production
requiring water. The main water usage (about
60 percent) at the facility is equipment and
area washdown. The remainder of the usage is
in the many cooling systems throughout the
plant. Cooling water is used in furnace cooling,
fiber-forming heat exchangers, room and area
cooling, and controlling humidity. The
blowdown for the entire system is evaporation
losses and wind losses throughout all of the
operations. Any makeup water added at the
facility over and above those few areas that re-
quire potable water is added at the distribu-
tion tank. This is done to maintain control of
the water usage at the facility and continue a
zero discharge system.
It may be necessary to bypass the sand fil-
tration and/or the carbon adsorption system at
times; therefore, a second pond was installed to
collect this water as well as sand filter and car-
bon backwash water. This pond water can be
pumped to the head of the system for treat-
ment and reuse.
329
-------
There have been a number of problems dur-
ing the past 12 months; however, these are
slowly being reduced through modifications
and improvements. We have had re.circulation
since May 1978 with intermittent discharges to
the stream. These were more frequent at first
but reduced to lesser amounts until October 7,
1978, when total recirculation without any dis-
charge was achieved. The system has operated
that way ever since.
I have not made much mention of sludge in
my presentation. It is a subject in itself with
the new regulations being promulgated under
the Resource Conservation and Recovery Act.
We have been periodically making studies on
the sludge, but because of the complexity of
the system, it has been decided to utilize our
existing system until such time as we deter-
mine that the water treatment system has
been stabilized. Then we will make a compre-
hensive study of the total sludge problem and
develop, design, and install the facilities re-
quired to properly meet all regulations.
The present sludge system includes an auto-
matic drawoff system from each of the five
clarifiers. This sludge is pumped to an aerobic
digester with a capacity of 230,000 gallons.
Digested sludge is drawn from the digester
and deposited in a sludge lagoon. Supernatant
from both the digester and the lagoon is
pumped back to the equalization tanks. As
mentioned previously, wasted activated sludge
is also pumped to the hr rl of the treatment
system.
It is impossible to delve into all the details
for such a system as is operational at the An-
derson facility. To discuss any one of the spe-
cific phases is a full presentation in itself;
therefore, I have attempted to review the the
history of our company in the development of
zero discharge. Information on the pilot work
and all of the details are available in the report
submitted to Mr. Max Samfield, Grants Admin-
istration Division, Industrial Environmental
Research Laboratory, Research Triangle Park,
North Carolina, entitled "Preliminary Engi-
neering Report —Wastewater Reuse System
for Fiberglass Textile Industry," dated Feb-
ruary 1977.
The system at Anderson is presently oper-
ating with no discharge —but one week or one
month does not prove a state-of-the-art. We
will not be satisfied that we have developed a
zero discharge system until we get through the
winter and on to next summer without upsets,
but again, we have faith in the concept, the
equipment, and the personnel. And, we have
tasted success and are determined to make it a
reality, thus meeting the 1983 standards 6
years before required to do so by law.
330
-------
OPTIMIZATION OF HYPERFILTRATION SYSTEM
FOR RENOVATION OF TEXTILE WASTEWATERS*
S. M. Ko and P. G. Grodzkaf
Abstract
A computer program for design and simula-
tion of a multistage hyperfiltration system for
renovation of textile wastewater has been de-
veloped. The program is capable of practical
design, parametric simulation, and cost projec-
tion of the multistage hyperfiltration system
with tapered innerstages. The mathematical
model is formulated based on Sourirajan'spref-
erential sorption and solute diffusion theory.
Experimental rejection and flux data of a test
hyperfiltration module are required as input
parameters. Empirical correlations and test re-
sults available from recent EPA-sponsored pro-
grams are utilized to calculate membrane
transport parameters. Computed results for
sample cases using cellulose acetate and dy-
namic membranes are presented. Various de-
sign and operating parameters are considered
in the numerical computations to show effects
of these parameters on economics of the sys-
tem. This simulation program was developed
in a general manner and is readily adaptable
for evaluation of other reverse osmosis hyper-
filtration applications.
INTRODUCTION
Hyperfiltration (also termed reverse osmo-
sis) as a textile wastewater treatment and ren-
ovation process has been studied under EPA/
IERL-RTP* sponsorship to investigate the
technical feasibility and the economic practica-
bility of the separation process. A recent inves-
tigation (ref. 1) of a pilot-scale hyperfiltration
facility at La France Industries, a division of
Riegel Textile Corporation, successfully de-
*Study sponsored by the U.S. Environmental Protection
Agency under EPA Contract 68-02-2614.
'Lockheed Missiles & Space Company, Inc., Huntsville,
.All.
^Environmental Protection Agency, Industrial Environ-
mental Research Laboratory, Research Triangle Park,
NC.
monstrated the feasibility of hyperfiltration
membranes for the in-plant recycle and reuse
of composite textile dyeing and finishing
wastewater. The applicability of the concept to
a variety of composite textile wastewaters has
been confirmed in a more recent study (ref. 2)
at eight different textile mills encompassing
eight different subcategories of the textile
mills point source category (ref. 3). The scope of
these studies, however, was limited to testing
and evaluating a few commercial membranes
using plant composite wastewater and the
mixed dyehouse effluent. Furthermore, optimi-
zation of process parameters in regard to more
favorable economics, e.g. energy conservation,
byproduct recovery, and effluent control of the
process system, were not investigated in de-
tail.
Textile finishing wastewater contains a
variety of chemicals depending upon the par-
ticular dyeing and finishing operations. The
types of chemicals in the wastewater greatly
affect the performance of the membranes. The
effectiveness of membranes is also sensitive to
the temperature of the wastewater being
treated. In the interest of energy recovery,
high temperature operation of the process is
desirable. Thus, a parametric investigation of
the separation process directed toward effi-
cient and cost-effective design of the treatment
system is required to provide the optimum per-
formance and design information essential in
the development of a full-scale system. Since
experimental investigations of these param-
eters are not practical, such a parametric study
necessitates a theoretical analysis of funda-
mental mechanisms involved in the hyperfiltra-
tion process and the development of a comput-
er model of the process system. Unfortunately,
however, direct extension of general reverse
osmosis theory to mixed solute systems is com-
plicated due to nonavailability of pertinent fun-
damental physicochemical properties of both
membranes and solutes, and possible complex
interactions of chemicals in such a system.
Thus, the transport properties of solutes in the
composite wastewater must be calculated from
331
-------
experimental flux and rejection data. Where
such data are not available, these transport
parameters may be estimated using empirical
correlations (refs. 4,5).
MATHEMATICAL MODEL
The model described herein is based on
Sourirajan's preferential sorption and solute
diffusion theory (ref. 6), which applies simple
boundary film theory (refs. 7,8) to obtain the
concentration polarization of the solute (ref. 9).
The effect of pressure drop due to friction
losses and momentum changes on the perform-
ance of the system (refs. 10,11) is included in
the model.
Membrane Model
Consider a tubular hyperfiltration module as
shown in Figure 1. From the overall material
balance for the differential control volume, the
change in average axial velocity (u) can be writ-
ten as
du 2
= — y
dx r
w
(1)
where vw is the permeation velocity of product
water (permeate) through the membrane (ft/s).
Similarly, the solute material balance can be
written as
d(uCAl)
dx
2
r
(2)
where CAI and CAS are the molar concentra-
tions of solute in the bulk and the product, re-
spectively.
Now, an expression for the pressure drop in
a tube in terms of friction losses and momen-
tum changes may be obtained from an energy
balance (ref. 10)
u2f
- — = CM
dx
or substituting (1) yields
dP
dx
= CM-
U2f
u du
2vwu
gcr
(3)
where C is the molar density of solution (Ib-
mole/ft3), M is the average molecular weight of
solution, f is the Fanning friction factor, and gc
is the conversion factor.
By solving equations (1), (2), and (3) simulta-
neously, one can compute output variables u",
P, and CAI for the differential volume for a
given set of dependent variables, product rate
(vw), and solute transport (CAS)- These depend-
ent variables can be related to the feed rate (u),
the system operating pressure (P), and the con-
centration of feed stream (CAI) by transport
mechanisms of solute and permeate (product
water) through the membrane and associated
concentration polarization of solute in the high
pressure side of the membrane. To obtain the
expressions for vw and CAS, let us consider a
differential section of the tubular hyperfiltra-
tion module shown in Figure 2. The figure
Membrane
Product
Water
r
Feed
P'CA1.5
\ \
w
Figure 1. A tubular hyperfiltration module.
dx
t t
Reject
'A 2
Product
Water
332
-------
Preferentially Sorb-d-
Interfaclal Region
Concentrated Boundary
Solution
Bulk Feed Solution-
Under Operating
Pressure
to
o
ft
(9
«
,*
"A 2
'A3
Concentration
Product Solution at
Atmospheric Pressure
Spongy Porous Membrane
or Membrane Support
Dense Microporous
Membrane Surface
Figure 2. Concentration profile of the differential hyperfiltration module under steady state.
(C\2 and CA3 denote so'ute concentrations in the membrane phase in equilibrium
with respective concentrations in the liquid phases.)
shows a schematic of steady state concentra-
tion profile of the solute across the membrane.
Due to relatively high affinity of the membrane
to solvent (water) molecules, a concentration
gradient is established in the interface region.
Adsorbed water molecules are permeated
through the microporous structure of the mem-
brane at a rate cletermined by the character-
istic of the mer.brane and the pressure ex-
erted to overcome the osmotic pressure. Sol-
ute, on the other hand, diffuses to both direc-
tions from the interface. The preferential sorp-
tion of water molecules will develop high sol-
ute concentration right in the vicinity of the in-
terface and thus develop a maximum concen-
tration. This provides a driving force for the
solute diffusion to both directions from the in-
terface. •
The transport of product water through the
membrane (refs. 6,9) is proportional to the ef-
333
-------
fective operating pressure of the system. The
effective pressure is defined to be the differ-
ence between the operating pressure and the
osmotic pressure exerted by the concentrated
boundary layer. Assuming a linear dependency
of osmotic pressure on mole fraction, we obtain
the following expression. (Detailed derivation
of the expression is available in reference 12.)
NR
„ = — (4)
"w
AP
C
B
PC
(°A2
NB is the permeate water flux through the
membrane (Ib-mole/ft2-s), C is the molar concen-
tration of the bulk solution (lb-mole/ft3), A is
the pure water permeability of the membrane
(Ib-mole/ft2-s-psi), P is the operating pressure
(psi), and B is the proportionality constant for
osmotic pressure (psi).§
The expression for CAS can be obtained by
integrating the differential equation obtained
from a differential solute material balance
within the concentrated boundary layer. A de-
tailed derivation and solution of the boundary
equation are presented in reference 12. The re-
sulting expression is:
'A3
CA3B/CP+ CS/AP (5)
exp — <
'k/S (CA3B/CP + SC/AP)
S is the solute transport parameter* (ft/s) and k
is the mass transfer coefficient of the solute in
the solution phase (ft/s).
The set of equations (1) through (5) can be
numerically solved to obtain output conditions
§Note that the constant B is a very small number for or-
ganic solutes found in textile waste waters. For such
cases, the osmotic pressure term can be neglected and
the flux can be written simply as: vw = AP/C.
S is used to denote the solute transport parameter rep-
resented by (DAM/Ka) by Sourirajan (ref. 6). DAM is the
diffusivity of solute in the membrane phase (ft2/s). K is a
characteristic constant that represents preferential
sorbability of the solute on a given membrane surface. S
is uniquely determined by the characteristics of specific
membrane/solute combination.
from the differential section of the module for a
given input (or initial) conditions at the inlet of
the differential section. By repeating the pro-
cedure one can compute the overall perform-
ance of the module. The following input param-
eters are required for the performance predic-
tion:
1. Design specifications for the system, i.e.,
geometry of the module;
2. Initial process conditions at the inlet of
the system, that is, flow rate, feed concen-
trations, pressure, and temperature;
3. Physical properties of feed stream, that
is, viscosity and osmotic pressure as func-
tions of solute concentration and tempera-
ture; and
4. Transport parameters:
a. Permeability constant (A) as a function
of pressure and temperature,
b. Solute transport parameter (S) as a
function of pressure and temperature,
and
c. Mass transfer coefficient (k) as a func-
tion of dimensionless numbers!
Parameters 1 and 2 are defined by the sys-~
tern design specifications. The effect of these
parameters on performance/cost can be investi-
gated and an optimum set can be identified
through a series of numerical computations.
Parameter 3- can be obtained from the litera-
ture. In the event'these data are not available
from the literature, experimental measure-
ments may become .necessary." 'The relative
magnitude of the three transport coefficients
(parameter 4) will determine the shape of con-'
centration profile and the efficiency of the hy-
perfiltration process. The transport coefficient
can be obtained from a set of experimental
measurements of the product water rate, pure
water permeability of :the iriembrane (normally
given by the membrane manufacturer)!, and S01-
ute rejection efficiency/These parameters are
temperature and pressure dependent. The de-
pendencies can be .expressed by the following
equations (refs. 13;14): ,-
A at
S a P~ exp(-7/T).
a, /S, and 7 are constants, which can be
calculated from experimental results. Empiri-
cal correlations published in the literature can
334
-------
be used to calculate the solute mass transfer
coefficient (ref. 15).
Economic Model
The incremental cost for producing unit
quantity of permeate water is considered to
analyze the system economics. The cost ele-
ments contributing to the incremental product
cost are amortized capital cost, UCC ($/kgal),
and O&M costs ($/kgal). The amortized capital
cost is calculated based on the installed system
capital cost per unit membrane area ($/ft2 of
membrane surface). The O&M costs include
membrane replacement cost, UMRC ($/kgal),
pumping power cost, UPP ($/kgal), and other
O&M costs, UMOMC ($/kgal). The credits to
the product cost are credit from recovered
water, CRW ($/kgal), and credit from recovered
energy, CRE ($/kgal). The credit from recov-
ered chemicals is not included because recover-
ing chemicals from the concentrate usually re-
quires additional process modifications.
The unit incremental cost for product water,
UCPW ($/kgal), is written
UCPW = UCC + UMRC + UPP
+ UMOMC - CRW - CRE.
The unit costs can be obtained from vendors or
actual estimations.
DESCRIPTION OF SYSTEM AND METHOD
In practical applications of the reverse
osmosis process, a certain multistage design
concept is desirable to achieve desired levels of
product recovery and solute rejection. Three
such concepts with tapered inner stages are
presented in Figure 3.
The single-stage concept employs direct re-
cycle of a portion of the reject stream to con-
centrate the reject to the specified design
value. In this case, the system is operated at a
somewhat higher concentration than the con-
centration of the feed. In the 2-stage concept
shown in Figure 3b, the first stage is used as a
purification stage, and the second as a concen-
tration stage. Since the permeate from the sec-
ond stage has a higher concentration than the
design product concentration, it is recycled to
the first stage. In the 2-stage case, the concen-
tration of the combined feed to the first stage
is lower than the incoming feed concentration
(6)
to the hyperfiltration system. An additional
stage is employed in the 3-stage concept. The
purpose of this stage is to concentrate the
permeate from the second stage back to the
feed concentration. This scheme provides
higher system efficiency because of the re-
duced number of modules required for the sys-
tem.
For a given design rejection and product re-
covery factor, the best system efficiency can be
obtained when the recycle flow rate is the
minimum and the concentration of the recycle
stream is the same as the feed stream. Numer-
ical results indicated that the 3-stage concept
reduces the total number of modules by 30 per-
cent and 10 percent over the single and 2-stage
system, respectively. For this reason, the
3-stage concept is chosen as the system model.
The tapered innerstage design increases ef-
ficiency of the system since it reduces the con-
centration polarization in the boundary layer.
The feed rate at the inlet of each module is
maintained at the design value by reducing the
number of modules within an innerstage. The
number of modules for an innerstage is deter-
mined by the flow rate fed to the particular in-
nerstage. As the bulk flow rate decreases due
to permeation of water through the membrane,
the level of polarization in a module increases.
Minimizing the concentration polarization in-
creases the efficiency of solute rejection effi-
ciency as well as the product rate. The rejec-
tion increases due to reduced solute concentra-
tion in the boundary layer while higher effec-
tive operating pressure is responsible for the
improvement in the product rate.
For the reasons discussed above, the ta-
pered innerstage design is chosen for the
system design purposes. Note that the in-
nerstages are tapered with respect to the
number of modules rather than the physical
shape of the membrane support structure.
A schematic flow chart for the design simula-
tion of the 3-stage tapered hyperfiltration
system is shown in Figure 4. A computer pro-
gram developed to solve the system of equa-
tions presented in the previous section was
utilized to obtain design and economic results
for the staged system. The numerical proce-
dure for simulation of a single module employs
an iterative scheme based on the Newton-
Raphson method over a number of finite sub-
sections of the module. A similar iterative pro-
335
-------
a. Single Stage
Recycle
Feed
1
t
1=]
T
^
|51 _J=L
l^T* ""T^r
i
1
(Concen
trate^
t
Permeate
b. Two Stage
Feed
'Permeate
Reject
(Concen-
trate)
c. Three Stage
Feed
Reject
(Concen-
trate)
Permeate
Figure 3. Schematic diagrams of tapered multistage hyperfiltration systems.
336
-------
( Start J
Input Module Geometry,
Operating Conditions and
Test Module Data.
Calculate Membrane
Parameters
Assume a Value for
Stage 1 Permeate
Water Concentration.
*tf
Calculate Performance
of Stage 1 Inner stages.
Calculate Performance
of Stage 2 Innerstages.
Calculate Performance
of Stage 3 Innerstages.
Calculate Total
Number of Modules
and Costs.
C End J
Figure 4. Flow chart of the 3-stage hyperf iltration design program. (Subscripts P, R, and F
d ;note permeate, rejected concentrate, and feed, respectively.)
337
-------
cedure is utilized to match the concentrations
and flow rates to and from each stage.
RESULTS AND DISCUSSION
The numerical results of a single module per-
formance simulation are shown in Figure 5.
The case shown in the figure simulates a tubu-
lar Westinghouse module with a cellulose ace-
tate membrane on the inside tube wall. The op-
erating conditions and the geometry are indi-
cated in the figure. The pressure drop for the
100-ft module is approximately 1 percent. The
initial rejection is 96 percent and slightly
decreases as the concentration polarization in-
creases by 8 percent. Approximately 10 per-
cent of the feed is recovered as permeate
water and the resulting decline of the flow rate
is 10 percent. The numerical results are in good
0>
3 .1? 600
« »
f-l
Pu
595
O
I—I
fn
4.5
4.0
Rejectic
1.0
0.95
"
—
£ «t
-t-> t$
c ••?
0.1
O O
u nS
« f-t-,
1.1 ,4
2.3
2.2
L
_L
0
20
J
80
100
40 60
Tube Length (ft)
Figure 5. Performance simulation of a single tubular hyperfiltration module (feed rate = 10 gpm,
temperature = 70° F, tube diameter = 1 in., cellulose acetate membrane (by Westing-
house) with flux = 4.75 X 1(T5ft/s, feed concentration = 2.0 X 10~5 Ib-mole/ft3).
338
-------
agreement with the experimental results pub-
lished in references 1 and 2.
From the single module simulation results, it
is apparent that at least 8 stages in series are
required to obtain 80 percent product water re-
covery if the feed concentration is low enough.
For a normal textile wastewater, however, a
considerably higher number of stages are re-
quired due to additional purification necessary
to achieve the overall design rejection and
product rate. This will be evident in the follow-
ing discussions on the system design simula-
tion of two sample cases.
Table 1 summarizes the 3-stage design simu-
lation results for two sample 1-million-gallons-
per-day (MOD) hyperfiltration systems. One is
a Cellulose Acetate (CA) membrane system
(Westinghouse) operating 600 psi and 70° F; the
other is a Zr(IV)-PAA (polyacrylic acid) dynam-
ic membrane system (Selas) operating at 1,000
psi and 150° F. These two membranes are se-
lected because of readily available design and
TABLE 1. SUMMARY OF COMPUTER DESIGN SIMULATION RESULTS FOR
TWO SAMPLE HYPERFILTRATION SYSTEMS OF 1 M6D CAPACITY
Item
Design parameters
Test tube rejection
Test tube flux (ft/s)
Permeability (Ib-mole/ft^-s-psi)
Solute transport parameter (ft/s)
Design tube diameter (in.)
Design tube length (ft)
Design product recovery factor
Design rejection
Design feed concentration (Ib-mole/ft3)
Design temperature (°F)
Design pressure (psi)
Cellulose
acetate
membrane
0.96
4.75 X10"5
2.07 xlO"9
1.23 x10-G
1
100
0.8
0.95
2.0 x10'5
70
600
ZR(IV)-PAA
dynamic
membrane
0.96
11.3x10-5
2.88 x10"9
2.53 xlO'6
1
100
0.8
0.95
2.0x10-5
150
1,000
Design results
Number of innerstages, stage 1
stage 2
stage 3
Number of modules, stage 1
2
Total number of modules
Total membrane area (ft^)
Economic results*
Total installed capital cost ($)
Capital amortization cost (cents/kgal)
Membrane replacement cost (cents/kgal)
Pumping power cost (cents/kgal)
Other O&M costs (cents/kgal)
Credit for recovered water (cents/kgal)
Credit for recovered energy (cents/kgal)
Total unit cost (cents/kgal)
19
11
42
846
213
204
1,263
31,727
539,350
24
79
15
10
40
0
7
5
15
339
101
99
539
13,540
2,301,700
102
2
25
20
40
67
42
*The unit cost basis is obtained from a recent study estimation (ref. 2). More detailed
cost breakdown is described in the reference.
339
-------
cost information (refs. 1,2). The transport pa-
rameters are calculated from experimentally
measured rejection and flux data from the
references. The two systems are designed for a
product recovery factor and rejection of 0.8
and 0.95, respectively. The unit cost informa-
tion was obtained from vendors and recent
study estimations (refs. 1,2).
The computed total numbers of modules 1 in.
in diameter and 100 ft long for the CA and
dynamic membranes are 1,263 and 539, respec-
tively. The corresponding membrane areas re-
quired are 31,727 and 13,540 ft2. The higher
surface area requirement for the CA mem-
brane system is due to the smaller permeabil-
ity of water through the membrane. The unit
1.4
1.2
£0.8
N-t
O
Is
oo
\ 0.6
9)
O
.t; 0.4
c
0.2
— — — Cellulose Acetate Membrane (WestLnghouse)
Zr(IV)-PAA Dynamic Membrane (Selas)
DREJ = Design Rejection
DREJ =
0.95
0.93
_L
°-6 0.75 0.8 0.85
Design Product Rate Factor
0.9
FigureG. Effect of design product rate factor and rejection on unit cost (feed rate = 10gpm,
tube diameter = 1 in., temperature = 70° F for CA membrane and 150° F for
dynamic membrane, pressure = 600 psi for CA membrane and 1,000 psi for
dynamic membrane).
340
-------
cost for producing 1,000 gallons of permeate
water is 88 cents for the CA system and 42
cents for the dynamic membrane system. It is
noted that the major cost element for the CA
system is the membrane replacement cost,
while the capital amortization cost makes the
largest cost contribution, and the credit from
recowed energy is significant enough to
cause the dynamic membrane to be attractive.
It should be noted that the design and opera-
tion conditions used are for illustration pur-
poses and do not represent a typical case.
To demonstrate the range of applicability of
the developed model and to investigate the ef-
fects of various design parameters on the eco-
nomics of the system, a series of parametric
studies is performed. The results are shown in
Figures 6 and 7.
Figure 6 shows the effects of design product
rate factor and solute rejection on the unit
cost. The effect of design product rate factor on
the unit cost is relatively moderate. The design
3.5T
'3.0
«» 2.5
u
3
T3
O
o
i—i
00
Vi-
2.0
1.5
in
O
U
±J 1.0
c
D
0.5
•— — •— Cellulose Acetate Membrane (Westinghouse)
•• • Zr(IV)-PAA Dynamic Membrane (Solas)
P = Operating Pressure (psi)
\
1T6 80 100 UO 140
Operating Temperature (F)
341
-------
rejection factor has a more pronounced effect
on the cost of the dynamic membrane system.
Significantly lower unit cost can be obtained at
slightly lower design rejection for the dynamic
membrane. On the other hand, the effect of de-
sign rejection is less significant in the CA
membrane. Thus, a proper combination of
these two membrane systems may be desirable
when a higher design rejection is necessary
due to more stringent quality requirements for
reuse of the permeate water.
Figure 7 shows the effects of operating
temperature on the unit cost with operating
pressure as a parameter. Due to character-
istics of the membranes, the effects are shown
only for the respective applicable operating
temperature and pressure ranges. It is shown
that the unit cost is generally lower at higher
temperature and pressure and a strong func-
tion of temperature and pressure as compared
to those of product recovery factor and rejec-
tion shown in Figure 6.
The unit cost does not include credit from
possible reuse of the chemicals contained in the
concentrated reject stream. A net saving can
be realized if the chemical credit is included in
the computation of the unit cost. Although high
temperature and pressure operation is desir-
able in the interest of achieving lower unit cost
or even net savings, it is limited by excessive
maintenance costs at high operating tempera-
ture and pressure. It is noted that the mainte-
nance cost used in the computations does not
reflect this effect but rather a flat estimated
average is assumed.
In summary, the model developed is capable
of predicting system performance as well as
analyzing system economics to find an opti-
mum set of design parameters. The reliability
of computer results is largely dependent on the
availability of rigorous cost information oh the
system as well as the accuracy of the test mod-
ule data/membrane specifications.
ACKNOWLEDGMENT
The authors are especially grateful to Dr.
J. L. Gaddis of Clemson University and Dr.
C. A. Brandon of Carre, Inc. for their contribu-
tions to this study and for their comments on
the mathematical model. The EPA project offi-
cer for this project was Dr. Max Samfield of
IERL/RTP and his support and interest in this
study are also gratefully acknowledged.
REFERENCES
1. C. A. Brandon and J. J. Porter, "Hyper-
filtration for Renovation of Textile Fin-
ishing Plant Wastewater," EPA-600/2-76-
060, March 1976.
2. C. A. Brandon, J. J. Porter, and D. K.
Todd, "Hyperfiltration for Renovation of
Composite Wastewater at Eight Textile
Finishing Plants," EPA-600/2-78-047,
March 1978.
3. U.S. Environmental Protection Agency,
"Development Document for Effluent
Limitations Guidelines and New Source
Performance Standards for the Textile
Mills Point Source Category," EPA-440/
1-74-022-a, June 1974.
4. A. Shindler, unpublished results devel-
oped under EPA/RTP sponsorship.
5. H. G. Spencer and J. L. Gaddis, "Hyperfil-
tration of Nonelectrolytes: Dependence of
Rejection on Solubility Parameters,"
paper presented at EPA Symposium on
Textile Industry Technology, December
5-8, 1978.
6. S. Sourirajan, Reverse Osmosis, Aca-
demic Press, New York, pp. 176-184,1970.
7. S. Sourirajan, Reverse Osmosis, Aca-
demic Press, New York, pp. 185-188,1970.
8. P. L. T. Brian, "Mass Transport in Re-
verse Osmosis," Desalination by Reverse
Osmosis, Ulrich Merten, ed., The MIT
Press, Cambridge, 1971.
9. W. Pusch, "Concentration Polarization in
Hyperfiltration Systems," Reverse Osmo-
sis Membrane Research, H. K. Lonsdale
and H. E. Podall, eds., Plenum Press, New
York, pp. 43-57, 1972.
10. W. L. Griffith, R. M. Keller, and K. A.
Kraus, "Parametric Study of Hyperfiltra-
tion in Tubular Systems with High Per-
meability Membranes," Desalination, Vol.
4, pp. 203-308, 1968.
11. J. S. Johnson, Jr., L. Dresner, and K. A.
Kraus, "Hyperfiltration (Reverse Osmo-
sis), Principles of Desalination, K. S.
Spiegler, ed., Academic Press, New York,
pp. 345-439, 1966.
342
-------
12. S. M. Ko and P. G. Grodzka, "Study of Hy-
perfiltration Processes for Treatment and
Renovation of Textile Wastewater," to be
published.
13. S. Sourirajan, Reverse Osmosis, Aca-
demic Press, New York, pp. 191-202,1970.
14. J. L. Gaddis, private communication
dated May 24, 1978.
15. S. T. Hwang, and K. Kammermeyer,
Membranes in Separations, Wiley-Inter-
science, New York, pp. 351-359,1975.
NOTE: After the paper was given at the conference, objec-
tions were raised stating that the costs given for the
cellulose acetate membrane were too low. EPA, through its
contractor, solicited cost information from three cellulose
acetate membrane manufacturers. The manufacturers
refused to supply such information, and, therefore, the con-
tractor used the only information available as indicated in
the cited reference.
EPA would welcome additional cost information, which
should be sent directly to the Project Officer.
In reviewing this and the other papers on hyperfiltration
presented at the conference readers should be aware that,
in the projects discussed, one is dealing with dilute solu-
tions of largely nonelectrolytes which have low osmotic
pressures.
343
-------
SOLVENT RECOVERY FROM TEXTILE DRYERS
Nathan R. Shaw*
Abstract
Solvents are used throughout industry as
vehicles to carry coatings onto various prod-
ucts, diluents, cleaners, and intermediates in
the production of other products. In these
capacities, hydrocarbon solvents are used in
virtually all industries. The textile industry is
no exception. This paper will explore some sol-
vent uses.
Two frequently encountered uses of solvents
in the textile field have to do with the applica-
tion of coatings and, of course, in scouring
and/or cleaning of fabrics prior to dying or
other finishing operations. In both cases, the
final step is t/> remove the solvent that has re-
mained in the fabric by drying.
The most frequently utilized solvents for
these applications are trichloroethane, tri-
chloroethylene, or perchloroethylene. Eco-
nomic considerations make it desirable to
recover and reuse these solvents. Existing
and/or pending EPA regulations often make
their recovery mandatory. Fortunately, vapor-
phase carbon adsorption technology, which
satisfies both existing and pending EPA regu-
lations, is readily available for recovery of
these materials so solvents may be reused.
Most of us, by this time, are familiar with the
basic technology of vapor-phase carbon adsorp-
tion. For those of you not familiar with the
basic equipment of the process and as a re-
fresher for the rest, the basic equipment con-
figuration consists of a blower and motor
assembly to move the solvent-laden air
through the activated carbon bed, one or more
vessels containing activated carbon, a steam
distribution system used for regeneration of
the carbon beds, a condenser to convert the
solvent/steam vapor to a liquid, and a decanter
for separation of the organic and aqueous
phases.
•Eastern Sales Manner, VIC Manufacturing Company,
Minneapolis, MN.
During adsorption, solvent-laden air from
the process enters the system through the
blower and passes through the bed of activated
carbon where the organic material is adsorbed.
The clean air passes out the exhaust valve to
atmosphere.
When the carbon is saturated with the or-
ganic component, the bed is isolated from the
process and regenerated with low pressure
steam. During this period, the second bed is
placed on adsorption giving a continuous
operating cycle.
During desorption, low pressure steam flows
in a direction countercurrent to that direction
utilized during adsorption. The thermal energy
provided by the steam heats the carbon, revol-
atilizes the organic material, and then acts as a
carrier gas to move the organic vapor into the
condenser section where it is reduced to a liq-
uid. It then flows by gravity to the decanter
where, if the water is immiscible and is of a dif-
ferent density than the condensed steam, it
will separate forming an aqueous and organic
layer that can be removed from the decanter
by gravity flow.
Solvent recovery systems for coating and
scouring applications are usually small, skid-
mounted units. One such system consists of
two 36-in. diameter adsorption vessels and
related equipment, as already discussed. This
system is processing 700 SCFM of solvent-
laden air containing 24.9 kg (55 Ib) per hour of
1,1,1 trichloroethane. The solvent is used as a
vehicle in the application of an antishrink
chemical to a web of fabric. It literally locks the
fibers together to prevent shrinkage. This
process operates 24 hours per day, 6 days per
week. Solvent-laden air is picked up from the
entering and existing ports of the dryer.
Utilities required for operation of the
system are as follows:
• Electrical (3 hp) = 2.24 kWh
• Steam = 123 Ib/h at 15 psig
• Water = 14 gal/min
• Compressed air = negligible.
345
-------
Operating costs for the system can be
calculated to be less than $1,600 per year per
8-h shift using the following cost factors:
• Electrical power at $0.035/kWlj
• Low pressure steam at $3.25 per 453.6 kg
(1,000 Ib)
• Water at $0.20 per 3,785.4 liters (1,000
gal).
Value of recovered solvent at $0.21 per
pound is $92.40 per 8-h shift, or $27,700 per
year per shift. In other words, after subtract-
ing operating costs, the solvent recovery
system generates over $26,000 per year of
reusable solvent. Total system cost was less
than this amount as originally purchased. In-
stallation costs are easily recovered by the
second-shift operation, and the third shift is
profit. This system was fully amortized in 1
year.
Another solvent recovery system involves
the recovery of perchloroethylene from a wool
scouring operation. Three thousand SCFM of
solvent-laden air contaminated at a rate of 45.4
kg (100 Ib) per hour is being processed. Solvent-
laden air is from a bank of dryers and from the
deodorizing exhaust of two automatic scouring
machines. At a value of $0.20 per pound, this
system generates $40,000 per year of reusable
solvent. Regeneration of the beds occurs once
every 3 hours.
Utilities required for this system are:
• Electrical (20 hp) = 15 kWh
• Steam = 238.1 kg (525 Ib) per hour
• Water = 219.6 liters (58 gal) per minute
• Compressed air = negligible.
Using the same cost figures as in the
previous example, and remembering that the
utilities are used only approximately one-third
of the time, operating cost is calculated at ap-
proximately $3,000 per year. The value of
recovered solvent is more than $37,000 per
year after subtracting operating costs. The in-
itial cost of recovery equipment was less than
$20,000. Installation was less than $10,000.
This system was fully amortized in less than 1
year.
The expected life of the systems described
above is 7 to 10 years or longer with proper
maintenance. A most important additional
benefit is compliance with EPA regulations.
The examples used above describe textile
dryer applications well suited to vapor-phase
adsorption techniques. There certainly are
many more applications involving various
solvents which, if even with longer amortiza-
tion times, still offer a viable solution to energy
conservation and compliance with environ-
mental regulations.
346
-------
SOLVENT DYEING OF NYLON CARPET USING THE STX
SOLVENT DYEING PROCESS
Charles C. Wommack*. M. Pierre Favierf
Abstract
Clean technologies result from an efficient
process and the complete recycling of effluents.
The STX solvent dyeing process has been de-
veloped up to the pilot stage in the field of
nylon carpet piece dyeing. It utilizes, in place
of water, two solvent fluids—perchloroethylene
and methanol—which, although perfectly mis-
cible in each other, have different character-
istics. The French Textile Institute has shown
that a mixture of these two solvents penetrates
into the nylon fiber in a manner far superior to
the penetrating effect of water or of either of
the two solvents used separately. This effect is
maximized when the methanol is removed by
distillation. Good penetration of the fiber is re-
quired for best dye fastness, maximum dye ex-
haustion, and minimum energy consumption.
Rhone-Poulenc has developed the equipment
and has selected commercially available dyes
that are soluble in methanol and insoluble in
perchloroethylene. The dye exhaustion is com-
plete and no added chemicals are needed. Be-
sides the carpet and the dyes, the only prod-
ucts used are the two solvents, which are re-
cycled through a recovery system that removes
the maximum amount of residues from the car-
pet and the air. No process water is used in con-
tact with either the fiber or the dyes in the STX
solvent dyeing process. Rhone-Poulenc hopes
to support efforts by an American company to
conduct a demonstration program that will
provide a measurement of the benefits to the
carpet industry in terms of product quality,
supply and discharge of water, and reduction
of the energy requirements. An additional ob-
jective of the envisioned demonstration pro-
gram is to document the economic factors in-
volved in the full commercialization of the
process.
JWommack and Company, Atlanta, GA.
'Rhone-Poulenc-Chimie Fine, Paris, France.
INTRODUCTION
Rhone-Poulenc Industries is a large-scale
chemical manufacturing firm located in France.
As a result of its research in the area of dye
processes, a new and entirely clean technology
has been developed regarding the pilot-plant
stage in Europe. It is recognized that the car-
pet industry of the United States, due to its
concentration in small geographic areas and its
much greater reliance on the use of nylon fiber
in comparison to carpet manufacturers in Eu-
rope, would be able to realize the benefits of-
fered by the new dye process much more rapid-
ly than would be possible in Europe. Based on
these considerations, the STX solvent dyeing
process is being offered to American carpet
firms interested in seeking to achieve full com-
mercialization of the process.
Direct energy requirements of the U.S. tex-
tile industry during 1971 (the most recent year
for which comprehensive data are available)
were equal to approximately 67 million barrels
of oil. About 60 percent of the total energy
presently used by the textile industry is con-
sumed in wet processing. Approximately 125
billion gallons of water are consumed annually
in this processing. Much of this water, when
discharged, includes appreciable quantites of
both organic and inorganic chemicals which re-
quire extensive and expensive waste treat-
ment. The required treatment processes also
consume large volumes of energy. Thus, high
energy consumption associated with the dye-
ing process and with the treatment of waste
discharges are two primary concerns of the
textile industry.
The STX process was developed in France
as a joint venture of some of France's largest
industrial concerns and the government of
France. The STX solvent dyeing process, if
shown to be economically and technically fea-
sible on a large volume basis, could have
significant potential benefits that go well
beyond the immediate needs and concerns of
any single company. It is estimated that
energy savings through use of the STX process
347
-------
will be on the magnitude of 60 to 70 percent
when compared to present methods of dyeing.
Additionally, since the process eliminates the
need for process water, there are no effluent
discharges. Consequently, implementation
could result in significant industry-wide en-
vironmental benefits as well as significant sav-
ings in the energy required for industrial proc-
essing and treatment.
The objective of a full-scale demonstration
project would be to conclusively demonstrate
that the solvent dyeing process can result in
significant environmental benefits and produc-
tion cost savings (primarily through reduced
energy consumption), and also result in a
higher quality product or at least a product
comparable in quality to that produced with
conventional dyeing processes. The develop-
ment of an effluent-free technology is a dif-
ficult and long-range venture. When success
comes, the reward is a substantial savings of
energy and materials. These savings stem
from the fact that the first step in the new
development is a careful analysis of the condi-
tions to do a good job. It was an analysis of this
type in the field of textile dyeing which lead to
the STX dyeing systems.
CONDITIONS NEEDED TO PRODUCE
QUALITY TEXTILE DYEING
The process of imparting color to textiles in
all cases requires that the fibers of the textile
be brought into contact with a dye chemical. To
achieve good color fastness, the dye chemical
must penetrate deeply into the fiber or must
chemically react with the fiber, or both. Uni-
formity or evenness of coloration requires that
each fiber of the textile be afforded equal op-
portunity to contact the dye chemical. This uni-
formity is normally achieved either by agi-
tating or moving the textile, by pumping the
dye chemical through the textile, or both.
All dyeing machines are designed to produce
these conditions so that good quality results in
terms of color fastness and uniformity of col-
oration. Figure 1 represents the typical dyeing
machine and its various functions. In this dia-
gram, the dye chemical is a liquid in which the
textile is immersed. Penetration of the fiber is
enhanced by using the heater to raise the tem-
perature of the liquid and the textile. Uniform-
ity of coloration is achieved by pumping the liq-
uid.
In current practice, such a dyeing machine
would use water as the liquid for carrying the
dye chemical. Not only is water very inexpen-
sive, but its characteristics are familiar to
everyone and it performs the following func-
tions needed for dyeing fairly well.
• It acts as a wetting and penetrating agent.
• It carries the dye chemicals and spreads
them evenly.
• It is a good heat transfer agent for heating
the fiber and equalizing temperature.
• It ionizes chemicals.
There are, however, many compromises
which must be made chemically when using
water to constitute the dyebath liquid. These
compromises result in less than total transfer
of dye chemicals from the dyebath to the tex-
tile fibers. Even when an assortment of other
chemicals is used to promote the transfer of
dye from the dyebath to the fiber, a portion of
the dye chemicals remains in the water. Due to
these factors, dyeing systems which use water
as the carrier of the dye chemicals have the
following inherent characteristics:
• All of the dyes are not used to dye the
fiber.
• Some of the dyes remain in the bath,
which must be reclaimed.
• Some of the dyes are left on the surface of
the fiber when the water is removed. This
residue makes it necessary to rinse the
textile material, thus polluting the water
used for rinsing.
• The chemicals used to improve dye quality
are themselves a cause of pollution.
THE STX SOLVENT DYEING PROCESS
In the STX solvent dyeing process, no water
is used in contact with the textile fibers or the
dye chemicals. Instead of water, a mixture of
two liquids is used to carry the dye chemicals.
In the dyeing of nylon, the two liquids found to
perform best are perchloroethylene and meth-
anol. During the dye cycle in the STX process,
one of the two components of the dyebath is
progressively removed. Since the dye chem-
icals chosen are soluble in only the component
of the dyebath that is removed, complete trans-
348
-------
Figure 1. Basic dyeing principle.
fer from the dyebath to the fiber is assured
(see Figure 2).
In the case of nylon dyeing, the dye chem-
icals chosen are soluble in methanol and insol-
uble in perchloroethylene. During the dye cy-
cle, the methanol is removed by distillation.
The diagram shown in Figure 3 represents
the dyeing machine used in the STX solvent
dyeing pilot plant in Europe. As can be readily
seen, the principles involved are very similar
to those shown in Figure 1. The only difference
is the addition of a system for diverting a por-
tion of the liquid through a second heater,
which brings the temperature to a level high
enough to produce rapid distillation of meth-
anol. It should be noted that even this distilla-
tion temperatur ; is comparable to or lower
than the norm.! operating temperatures of
most conventional dyeing machines. In the
pilot plant, the maximum temperature reached
in the main circuit was 70° C (158° F) and in the
secondary circuit the temperature was brought
to 93° C (200° F). Since the mixture of per-
chloroethylene and methanol starts to boil at
65° C (149° F), it is evident that rapid and com-
plete distillation of methanol is achieved at the
temperature used.
FACTORS IN THE SELECTION
OF SOLVENTS
Before going into more detail about the dye-
ing cycle and the pollution problems encoun-
tered, an explanation of some of the guidelines
to the selection of the solvents for the STX
process is in order. The solvent mixture must
349
-------
A CHANGE OF COMPOSITION OF THE DYEBATH WILL MOVE
THE DYE TOWARDS THE FIBER
THIS CHANGE OF COMPOSITION WILL BE MADE :
IN CONVENTIONAL SYSTEMS
BY ADDING CHEMICALS TO THE WATER
IN THE STX SYSTEM
BY EXTRACTION OF ONE OF THE
SOLVENTS
Figure 2. Change of dyebath composition.
allow a good dissolution of the dyes and a good
penetration of the dyes into the fibers. The
dyes must then stay in the fibers. The only in-
dustrial machines built to date have used the
two solvents, perchloroethylene and methanol,
in the ratio of 90 parts of perchloroethylene to
10 parts of methanol. Other solvent combina-
tions have been used in the laboratory and may
prove to be better for fibers other than nylon.
For nylon dyeing, the perchloroethylene and
methanol combination was found to perform
best. Perchloroethylene is a widely used prod-
uct in the dry cleaning industry. Its boiling
point is fairly high 121° C (250° F), giving am-
ple freedom in the selection of the best dyeing
temperature. Methanol is also a readily avail-
able material, and its boiling point of 64° C
(148° F) is high enough to enable operation
above the fibers' transition point, yet low
enough to permit easy distillation.
As previously mentioned, the dyestuffs used
are soluble in methanol and insoluble in the
perchloroethylene. The result is that the mix-
ture of perchloroethylene-aleohol-dye pene-
trates the fiber, the distillation of the alcohol
deposits the dye into the fiber (this dissolution
is probably induced by the presence of free
NH2 radicals), and exceptionally high quality
dyeing is achieved.
The work done by the French Textile Insti-
tute following the development of the STX
process has determined that the role played by
methanol in the process is as a carrier enabling
effective dye penetration to the very heart of
the fibers.
Water will not swell polyamide until temper-
atures are above the transition (around 49° C
[120°]). Methanol, even below 21°C (70°F), will
wet the surface of the fiber and will quickly
penetrate to the center of the fiber even
though it is unable to open the finer pores.
Because of these factors, methanol is not re-
tained in the macromolecular material. This
result is shown by the reversible stretch under
tension of nylon fibers when they are im-
mersed into methanol (see Figure 4). The ten-
sion is low and does not affect the criu,p.
Perchloroethylene has a swelling effect
much lower than the swelling effect of water.
However, when perchloroethylene is mixed
with methanol, there is a carrying effect. The
mixture penetrates the fibers well below 60° C
(140° F). The amount of perchloroethylene ab-
sorbed depends on the amount of methanol in
the mixture as shown in Figure 5. A study of
the perchloroethylene content within a section
of fiber, which has been immersed for 1 hour at
21 °C (70° F) in a mixture of 90 percent per-
chloroethylene to 10 percent methanol, con-
firms the carrying effect of the methanol.
The penetration of the chlorinated solvent is
again shown in Figure 6. The fibers have been
impregnated for 1 hour at 21 °C (70° F) by the
mixture of solvents. They have then been
dryed at 38° C (100° F) for 24 hours, then at
175° F for 2 hours. The curves show the pene-
350
-------
condenser
heater H2
1. Main dyeing circuit
2. Flash circuit
flash
solvent
for reuse
Figure 3. Principle of flash (liquid phase PER + methanol).
351
-------
FIGURE 4
1st cycle
ond ii
3rd
Immersion cycles
in methanolat20°C
Figure 4. Immersion cycles in methanol at 20° C.
352
-------
"" — ~- ^^^^ CONCENTRATION
TEMPERATURE (»C) ~~" ^^__^
50
70
80
0
0,00
5
0,02
0,08
0,09
10
0,08
0,17
0,29
20
0,20
0,40
0,47
30
0,15
0,20
0,39
Figure 5. Quantity of perchloroethylene adsorbed by polyamide 6-6 in relation with
temperature and concentration of methanol (cm3 . G~1).
tration at different temperatures. It is espe-
cially high when perchloroethylene is replaced
by trichloroethane, yet is much more volatile.
Because o* this penetration, it is necessary,
industrially, to use a steam stripping of the
fibers in order to completely remove the chlori-
nated solvents from the fiber.
POLLUTION CONTROL
In the previous sections, we have described
a dyeing process in which:
• No chemicals are added to the dyebath.
• The dyes are entirely exhausted from the
dyebath.
• The dyes penetrate exceptionally deep
into the fiber.
We will now look at each phase of the dyeing
and drying cycle and discuss pollution hazards
involved in the STX solvent dyeing process.
Figure 3
The dyebath is a mixture of perchloro-
ethylene and methanol. The dyeing cycle has a
duration of 1.30 hours. During this time, the
only opening of the system to the outside at-
mosphere is through the methanol condenser.
The temperature in this location is kept well
below the boiling point of the dyebath (65° C
[149° F]).
We can thus v .isider the dyeing circuit as a
closed system. It should be noted that other ef-
forts made towards dyebath reuse with aque-
ous systems use batch-type closed circuits and
require fairly sophisticated pieces of equip-
ment.
At the end of the dyeing cycle, all of the
methanol has been extracted through the con-
denser. It can be reused for another cycle, even
with a different shade. The only liquid left in
the machine is perchloroethylene. This solvent
will have dissolved the spinning oils and other
greases contained in the carpet. Only a fraction
of this oil will be deposited back on the carpet
during the drying cycle. Our experience in the
pilot plant is that it is adequate to recycle 10
percent of the perchloroethylene used in the
machine, for each cycle, thus obtaining a per-
fectly clean carpet.
Figure 7
At the end of the dyeing cycle, the per-
chloroethylene is dumped into a storage tank.
Except for the 10 percent which will be recy-
cled, it will be immediately available for use in
the next dyeing cycle. The carpet, still in the
machine, is soaked with perchloroethylene.
The solvent will be mechanically extracted by
forcing air through the carpet at 71° C (160° F).
This could be done continuously while the car-
pet is being unrolled. In order to avoid air con-
tamination, the machine is equipped with a
closed drying circuit, quite similar to the dye-
ing circuit. The drying cycle lasts 2 hours; the
353
-------
TCE/Meth
PER/lsopr
100 110 120 130 140 150 160 170 180 190 200 210
Figure 6. Penetration of chlorinated solvent.
354
-------
condenser
condensed
solvent for reuse
Figure 7. Principle of drying.
355
-------
overall total dyeing and drying cycle is, there-
fore, approximately 3.5 hours.
Figure 8
At the end of the drying cycle, the air con-
tained in the closed circuit is still mixed with a
very small amount of solvent that cannot be
condensed at the temperature of the con-
denser. The opening of the autoclave at that
time would result in a slight air contamination.
The figure shows that the machine is venti-
lated with fresh air. This air is filtered with ac-
tivated carbon before release into the atmos-
phere. This could be done in a closed circuit,
but some carbon dust would deposit back on
the textile. When the activated carbon is satu-
rated with solvent, it is reclaimed by steaming.
Sizing Oils
The sizing oils are dissolved by the per-
chloroethylene. The separation and the re-
claiming of the perchloroethylene and the oils
is done in a boiler similar tc those used in com-
mercial dry cleaning machines. At the end of
the cycle, the last of the perchloroethylene is
stripped by steam. In conventional dyeing
cycles, the oils are lost in the effluents. The
same happens with the chemical agents used in
conventional dyeing. No elaboration is needed
on the savings resulting from the fact that no
chemical agents whatsoever are used in the
STX process.
Energy Consumption
Solvent dyeing would not be feasible if sol-
vent reclaiming required large amounts of
energy. Fortunately, specific and latent heat
figures are very low in the case of the solvents
used, as is shown in Figure 9.
The advantages as compared to conventional
water-based systems are shown in Figure 10.
The energy consumptions for solvent dyeing
are related to the use of a 1-metric-ton-capacity
machine. These consumptions would be re-
duced, as well as the bath ratio, for machines
having a larger capacity. Additional reductions
of energy consumption would be obtained by
operating under a slight vacuum, which would
help in the exhaustion of methanol.
Figure 11 shows that dyeing itself, perform-
ing as it is in fiber penetration, only represents
one-half of the task. The atmospheric contami-
nation hazard, due to the use of solvents, is
taken care of through closed circuit drying,
with a low energy consumption (14 percent).
Finally, it must be pointed out that solvent
reclaiming represents a large share (32 per-
cent) of the total energy consumption.
POTENTIAL BENEFITS AND CONCERNS
ASSOCIATED WITH WIDE-SCALE
IMPLEMENTATION OF THE STX PROCESS
Energy Conservation
The potential energy savings associated
with STX solvent dyeing are very significant
when compared with the conventional beck
dyeing process. The textile industry is heavily
dependent upon petroleum and natural gas (the
fuels currently in shortest supply) to meet a
major part of the process energy require-
ments. About 60 percent of the energy present-
ly utilized is consumed in wet (water) process-
ing, which includes the cleaning, dyeing, and
finishing processes. These processes are gen-
erally energy inefficient with efficiencies of
less than 20 percent typically encountered in
commercial practice.
Among the important energy saving fea-
tures of the STX solvent dyeing process, in
comparison to the conventional wet dyeing
processes, are:
• Energy consumption in the carpet dyeing
industry is mainly used in the heating and
boiling out of water. In the solvent dyeing
process, steam utilization is reduced by
over 50 percent.
• Shortages of natural gas have had a major
impact on the textile industry. The sol-
vent dyeing process does not utilize nat-
ural gas.
• The reclaiming of solvents requires only
about 10 percent of the energy needed to
evaporate water in the conventional sys-
tem.
• The use of chemicals in the dyeing process
is reduced by about 50 percent. An indi-
rect benefit is the reduction of energy con-
sumed in producing these chemicals.
• Carpet is dyed at lower temperatures
(about 70° C) in the solvent dyeing process
356
-------
to atmosphere
ac ivated carbon—
fresh air
Figure 8. Principle of the ventilation.
357
-------
PERCHLORETHYLENE
WATER
SPECIFIC HEAT
KCAL/KG/°C
0,216
LATENT HEAT
UNDER 76O MMHG
50
539
Figure 9. Specific and latent heat figures of solvents used.
ENERGY CONSUMED FOR
100 KG OF DYED CARPET
STEAM KILOGS
ELECTRICITY KWH
GAS M3
TOTAL K CALORIES
WA
WITHOUT EFFL
BECKS
935
53
20
803. 483
100 %
TER
UENT TREATMENT
CONTINUOUS
508
55
20
546. 248
66 %
SOLVENT
S T X
435
57
0
315. 032
38 %
Figure 10. Advantages over conventional water-based systems.
The costs of treating waste discharges from
the carpet industry are high. However, just as
important is that the advanced treatment proc-
esses required with wet process dyeing result
in the consumption of large amounts of energy.
Water Usage
The capacity of the solvent dyeing plant
developed by STX is approximately 6 tons of
carpet per day. To dye an equivalent amount of
carpet using the beck dyeing process would re-
quire 36,000 gallons of process water since ap-
proximately 3,000 gallons of water are used
with the conventional beck dye load of about
1,000 pounds.
Because of the concentration of the carpet in-
dustry in Dalton, Georgia, nearly 50 percent of
358
-------
BEAMING, UNBEAMING, STEAMING
DYEING
DRYING
SOLVENT RECLAIMING
KILOCALORIES
10 000 3
159 000
44 000
102 000
51
14
32
315 000 100
Figure 11. Breakdown of the energy consumption.
all raw water used in the carpet manufacturing
process is taken from local streams in north-
west Georgia. This restricts the use of these
waters for other purposes because of the need
to maintain flows sufficient to meet these proc-
ess demands and to assimilate wastes dis-
charged after completion of the carpet dyeing
processes.
Occupational Safety
The STX solvent dyeing process uses two
solvents: perchloroethylene and methanol. The
process used is self-contained and both per-
chloroethylene and methanol are completely
recycled and reclaimed.
The Occupational Safety and Health Admin-
istration (OSHA) has established an industrial
standard for exposure to perchloroethylene of
100 ppm for an extended time period. In the
plant built in Europe to test the solvent dyeing
system, probes located in the dye house never
showed a concentration above 30 to 40 ppm of
perchloroethylene in the atmosphere, and this
level was reached only during the limited time
that the autoclave was opened at the end of
each dyeing cycle.
The fact that the temperature utilized in the
solvent dyeing process is quite low and never
exceeds 70° C is another safety feature of the
STX system since it eliminates the need for
high temperature operations.
The solvent dyeing process never includes
the mixing of air and methanol. In the process,
perchloroethylene eliminates the air before
methanol is introduced. This is strictly moni-
tored and controlled.
The solvent dyeing plant in Europe was op-
erated for a period of 1 year. Documentation is
available that demonstrates that this plant has
met all applicable U.S. safety standards during
this period of operation.
Chemical Usage
The chemical usage in the STX process is
limited to the dyestuffs and the small quan-
tities of solvents lost in the process. This
chemical usage level compares very favorably
with the existing aqueous systems now in use,
which demand that a variety of chemicals be
used as dyeing assistants in order to achieve
good product characteristics. The elimination
of the need for these other chemicals can cut
the cost of dyeing chemicals by as much as 50
percent.
359
-------
Benefits to the Carpet Industry
The STX solvent dyeing system offers many
potential benefits to the carpet industry as
well as to other segments of the textile indus-
try. The most apparent benefits are:
• Since water is used only for cooling and
steam generation, the STX process elim-
inates all problems related to supply of
raw water and discharge of polluted
wastewater. Since carpet dyeing plants
are having to cope more and more with the
twofold problems of supply and pollution
of water, there is a great incentive to in-
vestigate any alternative process that
would solve this situation.
• The overall cost of dyeing carpet is re-
duced by the STX process since chemical
usage, water usage, and energy usage are
all reduced. There is also a cost savings in-
herent in the process as the process elim-
inates the 4 to 5 percent waste of carpet
fiber that is experienced in using aqueous
dyeing. These savings could potentially
yield an overall dyeing cost reduction ex-
ceeding 30 percent in the carpet industry
as compared to current aqueous dyeing
methods.
• The subtle components of product quality
in the carpet industry such as softness,
luster, color fastness, and texture are
either matched or improved by the STX
solvent dyeing process. In a highly com-
petitive market, such as the high quality
residential carpet market, these charac-
teristics can often be an overriding factor
in the buying decision of the consumer.
• A very important benefit to carpet
manufacturers offered by the STX proc-
ess is the potential for more accurate color
matching from one dye lot to another. If
sufficiently accurate controls can be devel-
oped, the STX process will enable a manu-
facturer to match one dye lot to another
close enough to permit installation of two
different dye lots side by side without any
noticeable difference in color. In conven-
tional beck dyeing, this is not now possible
even with the sophisticated measuring
equipment available. The capability for
color matching would offer many advan-
tages, both in marketing and manufactur-
ing.
• Production efficiency should be improved
with the STX solvent dyeing system since
dyeing cycle time is significantly reduced.
• The dyestuffs and other chemicals used in
the STX process are commonly used in the
carpet and textile industry, and, there-
fore, availability of these materials should
not pose any problems.
Patent Rights
Patent rights are held for all aspects of the
solvent dyeing process. Patents are held on the
dyeing system employed including its chemical
aspects and also on the equipment utilized.
Patents are also held on the use of this system
with different types of materials (nylon, wool,
etc.).
in comparison to the wet dyeing process,
which requires temperatures of about
100° C. This results in significant energy
savings.
• Elimination of the need for the processing
of raw water and the treating of process
wastewater results in significant cost sav-
ings.
It is estimated that the solvent dyeing proc-
ess would permit the savings of approximately
1 pound of fuel oil for every pound of carpet
dyed.
Water Pollution Control
It has been estimated that the discharges
from a dyeing process with a capacity of 6 tons
of carpet per day (the capacity of the plant
utilized in the solvent dyeing process) equal
the domestic wastes of 10,000 to 15,000 people.
This is indicative of the strength of the waste
discharges of the carpet industry, which in-
clude high quantities of both organic and in-
organic chemicals (dyes, pH control agents,
lubricants, etc.). These wastes require exten-
sive treatments prior to discharge to the
natural stream system. When the strengths of
these wastes are considered in the context of
the large volume of waste discharged (over 100
billion gallons of water per year), the potential
360
-------
impact of these discharges on the water quality
of receiving streams, and the costs of treating
the wastes prior to discharge, become very sig-
nificant.
The solvent dyeing process does not use
process water. Additionally, the solvents uti-
lized in the process are reclaimed, not dis-
cha^ged. Consequently, effluent problems and
water quality problems associated with carpet
dyeing (and potentially with all textile dyeing)
could be virtually eliminated through wide-
scale use of solvent dyeing.
Nylon fiber is used in nearly 90 percent of
the carpet now manufactured in the United
States. Since the solvent dyeing of nylon is
technically proven, the potential for immediate
water quality benefits could be significant in
the carpet industry.
361
-------
RECONSTITUTION AND REUSE OF DYEBATHS
Wayne C. Tincher,* Fred L. Cook,* L. Howard Olson,* W. W. Can-t
Abstract
A system for dyeing textiles has been
developed that permits reuse of dyebaths for
up to 15 times before the bath is discharged. At
the conclusion of the dyeing cycle the bath is
analyzed and reconstituted to the dye and
chemical concentrations required for the next
dyeing. The system has been used on a pilot
scale for dyeing nylon and polyester carpet,
nylon hosiery, and polyester yarn packages. A
plant demonstration of the reuse technology
for hosiery dyeing has been completed. Signifi-
cant reductions in dye (19 percent), chemical (35
percent), water (43 percent), and energy (57per-
cent) requirements were achieved with the
reuse system. The cost of materials and energy
for hosiery dyeing was reduced 37 percent by
reuse of the dyebath.
INTRODUCTION
The textile industry uses approximately
500 x 109 liters (125 billion gallons of water an-
nually (ref. 1). Much of this water is discharged
with appreciable quantities of organic and in-
organic chemicals (dyes, pH control agents,
lubricants, surfactants, auxiliaries) that re-
quire extensive treatment of the wastewater
and contribute to the pollution problems
associated with textile processing. In the past,
the pollution problem resulting from textile
processing has been attacked by construction
of waste treatment facilities. As requirements
became more stringent, the waste treatment
plant was expanded or additional treatment
processes were added to meet the standards.
This approach has consumed large sums of
nonincome-producing capital and has increased
the operating costs of many textile plants.
In addition to problems associated with
' i MI
•School of Textile Engineering, Georgia Institute of
Technology, Atlanta, GA.
'Engineering Experiment Station, Georgia Institute of
Technology, Atlanta, GA.
water supply and pollution control, the increas-
ing cost and decreasing supply of energy repre-
sent a major problem for the textile industry.
Energy requirements for the U.S. textile in-
dustry in 1971 (most recent comprehensive
data), were equal to approximately 67 million
barrels of oil (ref. 2). Industry is heavily de-
pendent on petroleum products and natural
gas (the fuel currently in shortest supply) for a
major part of process energy requirements. If
the industry is to adapt to the shortages or cur-
tailments of energy, new processes or process
modifications must be implemented on a broad,
industry-wide scale.
Two basic types of procedures are used for
textile coloration—pad-fix (continuous) proc-
esses and exhaust (batch) processes. Con-
tinuous dyeing is used primarily for long runs
of a given fabric style. Because of their con-
tinuous nature and the use of relatively small
volumes of dye liquor, these processes tend to
be more efficient than batch processes. Ex-
haust processes are generally very inefficient
in their use of chemicals, water, and energy
and generate large volumes of waste that must
be treated. Despite these disadvantages, the
versatility, ease of control, and short run
capability of exhaust processes make them
very attractive for coloration of many textile
products.
A number of different types of batch dyeing
processes (beck dyeing, jet dyeing, jig dyeing,
package dyeing, beam dyeing) are used in the
textile industry. Batch processes were used for
dyeing an estimated 5 billion pounds of textile
fibers and fabrics in 1973 (ref. 3). Batch proc-
esses readily lend themselves to process modi-
fications that can reduce energy and materials
requirements for dyeing. Such modifications
were a major objective of this research project.
DEVELOPMENT OF THE REUSE SYSTEM
In the conventional batch dyeing process the
dyeing machine is filled with water, the goods
to be dyed are entered, and the fabric moved
through the bath (or the bath moved through
363
-------
the goods) to saturate the fabric with water.
Chemical auxiliaries such as wetting agents,
pH control agents, leveling agents, chelating
agents, etc., are then added to the bath fol-
lowed by the dyes. Usually the beck is heated
from ambient to dyeing temperature at a rate
of approximately 2°C (4°F) per minute and
held at the dyeing temperature for the time re-
quired to complete the dyeing. The goods being
dyed are checked for proper shade and if on-
shade, the dyebath is discharged to drain. The
goods are then post-scoured and/or rinsed to
remove incompletely fixed dye. The goods are
removed from the machine and the machine
refilled with water for the next load.
If the dyebath is examined before and after
the dyeing cycle, two major changes have oc-
curred. First, most of the dye has been re-
moved from the bath by the yarn or fabric and
second, the bath is hot rather than cold. Most
of the auxiliary chemicals added to the bath
are still present in the same condition as they
were at the start of the dyeing cycle. When the
dyebath is discharged to the drain, large quan-
tities of energy, water, and useful chemicals
are lost. A more reasonable procedure would
be to analyze the spent dyebath for remaining
dye, to reconstitute the bath to the desired
strength, and to reuse it for subsequent dye-
ings. Reuse of dyebaths in this way should
significantly reduce the energy, water, and
chemical requirements in batch dyeing.
A number of technical problems required
solution before dyebath reuse could be broadly
applied in batch dyeing. First, an analytical
system had to be developed to simply, ac-
curately, and economically determine the con-
centration of dyes remaining in the bath. The
analytical technique has to be compatible with
existing dyehouse personnel, space, time, and
equipment constraints. Second, dyeings must
be started at elevated temperatures (~75° C).
The increased rate of dye adsorption from the
bath at these temperatures could lead to spot-
ting and poor levelness in the recycle dyeings.
Third, materials handling procedures had to be
worked out to give scouring, dyeing, and rins-
ing cycles compatible with current plant opera-
ting procedures. Fourth, evaluation pro-
cedures were required to insure that dyeings
in recycle baths were equivalent in quality to
conventionally dyed products.
Analysis System
The very strong absorption of dyes in the
visible region of the spectrum provides the
simplest and most precise method for deter-
mination of dye concentration. The absor-
bance, A, of a dye solution can be related to the
concentration by the modified Lambert-Beer
equation (ref. 4):
A = logI0II = Kc
where I0 is the intensity of the visible radiation
falling on the sample, I is the intensity of the
radiation transmitted by the sample, K is a con-
stant including the path length of radiation
through the sample and a constant related to
the absorptivity of the sample at a given
wavelength, and c is the concentration of the
absorbing species. In mixtures of absorbing
species, the absorbance at any wavelength is
the sum of the absorbance of each absorbing
species and is given by:
A =
K2c2 + K3c3 . . .
Kncn.
This characteristic of light absorption by dyes
is important in the analysis of dye mixtures of
the type found in spent dyebaths. For such dye
mixtures, the absorbance can be measured at a
number of wavelengths and the concentration
of the dyes determined by simultaneous solu-
tion of a set of linear equations of the type
shown above. The wavelengths selected for the
analysis are generally those for which one of
the dyes gave a maximum in absorbance.
A second procedure for obtaining dye con-
centrations from absorbance data was
developed for very complex dye mixtures. In
this procedure the absorbance of the spent
dyebath was measured at 16 equally spaced
wavelengths from 400 to 700 nm. The dye con-
centrations were determined from a least-
squares fit of the absorbance curve at these 16
wavelengths. A computer program in FOR-
TRAN IV was developed to determine the con-
centrations of up to six dyes in a mixture by
this technique.
The Lambert-Beer relationship is generally
invalid for absorbing species that are not in
solution. Most dye classes used in this work
(acid, basic, direct, reactive) are soluble in the
364
-------
dyebath and the speetrophotometric analysis
could be carried out directly on the dyebath.
Disperse dyes, however, are not water soluble
and required development of techniques to
give true dye solutions. In disperse dye anal-
ysis the spent dyebath sample was placed in a
separatory funnel, a measured quantity of an
organic solvent (benzene or toluene) was add-
ed, and the mixture shaken to extract the dye
into the organic layer. The speetrophotometric
analysis was then carried out on the dissolved
dye in the organic solvent. Standards for calcu
lation of the K values for disperse dyes were
treated in the same way to insure accuracy of
the procedure.
Correction for absorbance of species other
than dyes (background absorbance) was re-
quired also in some cases. For dyebaths con-
taining disperse dyes and using organic sol-
vent extraction, background absorbance pre-
sented no problem. The absorbing species
other than dyes remained in the water phase.
For dyes pleasured directly in the dyebath, a
correctio i procedure was necessary. Samples
of the textile product being dyed were treated
in a "blank" bath (a bath containing all chemi-
cals except the dyes) in exactly the same way
that the yarn or fabric would be dyed. Samples
of the blank dyebath were then used to obtain
standard absorption data for the blank dye-
bath and background K values were calculated
in the same way as described above for dyes.
The background absorbance could then be
treated like another dye. Thus, a dyebath with
three dyes would be treated as if it contained
four dyes with the fourth dye being the
background absorbance.
A further advantage of spectrophotometers
is the ready availability of a number of low cost
instruments with sufficient accuracy and re-
productivity for dyebath analysis. Much of the
work in the current study was carried out on a
single beam grating spectrophotometer
costing approximately $2,000. The computa-
tions necessary in the analysis can be conven-
iently carried out on low-cost desk calculators
or microprocessors. The calculations necessary
for a four-dye mixture (or three dyes plus back-
ground) can be handled on a system costing
less than $1,000. Even the least-squares fit of
16 points of the absorption spectrum can be
carried out on; $3,000 minicomputer. Develop-
ment of these low-cost instrument-
minicomputer systems is largely responsible
for consideration of dyebath reuse as a prac-
tical reality for textile dyeing.
Accurate analysis for dyebath components
other than dyes has not been required. These
dyebath additives generally control the dye-
bath environment and are not used up or
removed during the dyeing cycle. In general,
these components were added to the reuse
baths in direct proportion to the quantity of
makeup water added to baths between dye-
ings. For example, if 20 percent of the dyebath
was lost during the dyeing (primarily due to re-
moval with the wet goods) then 20 percent of
the original quantities of pH control agents,
wetting agents, etc., was added to the bath be-
ing reused. This procedure worked very well
for the series of 13 and 14 dyeings in the same
bath.
Evaluation Procedures
Three important characteristics of textile
products dyed by the reuse procedure were
measured to determine if the dyeings were of
acceptable quality—the uniformity, the repro-
ducibility, and the stability or "fastness."
The uniformity was assessed by selecting
representative samples from the dyed lot and
determining the color (tristimulus values) on a
standard colorimeter (DIANO^SCE Automate
System). The difference in color of each speci-
men from the average color of all specimens
was determined using the FMC II color dif-
ference formula (ref. 5). In this system one unit
of color difference is defined as the minimum
perceptible difference in color. Thus, spotting
or unlevel dyeings could be readily identified
by variations in color difference between speci-
mens from the same dyeing.
Reproducibility of the dyeings was deter-
mined by calculation of the color differences
between dyed lots and a standard dyeing of the
same shade. In this calculation the color dif-
ference was determined from average tristim-
ulus values of each dyed lot compared to a con-
ventional dyeing, again using the FMC II equa-
tion.
In addition to instrumental measurements,
samples dyed by the reuse procedure were ex-
amined' visually by expert dyers to further
assess the color uniformity and color repro-
ducibility.
365
-------
Dyed samples were also evaluated for
resistance to color change by rubbing (crock
fastness), exposure to light (light fastness), and
exposure to water (wet fastness). Standard
test methods recommended by the American
Association of Textile Chemists and Colorists
were used in the fastness evaluations.
APPLICATIONS OF THE REUSE SYSTEM
Carpet Dyeing
Initial studies of dyebath reuse were carried
out on carpet samples. Preliminary bench-scale
studies to evaluate the technical feasibility of
dyeing in spent baths and to develop the re-
quired analytical techniques were carried out
as part of a previously reported (refs. 6,7)
research program. The current work was
undertaken to scale-up the dyebath reuse
system for carpet dyeing. All dyeings were car-
ried out in a 1-m beck capable of dyeing 15-kg
(35-lb) carpet samples. This is the same equip-
ment normally used in a carpet dyehouse for
development work. Two of the carpet experi-
mental trials—dyeing of nylon carpet with
disperse dyes and dyeing of polyester carpet
with disperse dyes —are reported here.
Dyeing of Nylon Carpet With Diperse Dyes
The objective of the first trial run in the
pilot-scale experiments was to dye nylon
carpet to the same shade five times with reuse
of the dyebath. The carpet used in these
studies was tufted from Nylon 6 face yarn with
a polypropylene primary backing. Both regular
nylon and cationic dyeable yarns were used in
the carpet, but these types of yarns were dyed
to essentially the same color with disperse
dyes. The carpet surface contained both cut
pile and loops.
Approximately 15 kg (35 Ib) of prescoured
carpet were placed in the standard pilot-scale
1-m beck, cold water was added to give a liquor
ratio of 25:1, approximately 375 1 (100 gallons),
and the auxiliary chemicals (leveling agent, 1
percent owf; complexing agent, 0.5 percent
owf; pH control agent, 1 percent owf) were add-
ed. The carpet was circulated through the bath
for 5 minutes and the dyes required to give a
medium green shade (Disperse Blue 7,0.03 per-
cent owf; Disperse Yellow 3, 0.35 percent owf;
Disperse Red 55, 0.04 percent owf) were added
over a 15-min period. The bath was brought to
the boil at 2° C/minute with direct steam injec-
tion and held at the boil for 45 minutes. The
carpet was removed by hand at 95° C and
rinsed three times in a separate machine. The
water volume that was removed by the carpet
and by evaporation (85 liters) was added to the
remaining bath, and only auxiliary chemicals
calculated to bring the added water to the in-
itial concentration of auxiliary chemicals were
added. The residual dye concentrations were
determined using the analytical procedure
detailed above. A second batch of carpet was
placed in the beck after bringing the bath back
to 80° C. The dyestuffs were added to the beck
over a 15-min period at 80° C, the bath was
brought to the boil at 2° C/minute, and the cy-
cle continued as in the first run. The reuse se-
quence was repeated in three additional runs
to give a total of five dyed batches.
The pilot-scale dyeings of nylon (15-kg
samples) with disperse dyes were evaluated by
color measurement and color difference cal-
culations exactly as described above. Ten ran-
dom samples were cut from each of the five
nylon carpet sections (1m x 2m) that were dyed
medium green in the reuse sequence on the 1-m
beck. Color differences between the samples
from the four reuse runs and samples from the
initial (conventional) dyeing were calculated.
The tristimulus values X, Y, and Z were obtain-
ed on the Diano instrument by measuring each
sample three times against a white tile stan-
dard for a total of 30 color measurements for
each carpet length. The tristimulus values
were utilized with the FMC II equation to
determine the end-to-end shade variation of
the carpets by averaging the tristimulus
values from the 30 measurements and calcu-
lating color differences in MacAdam units bet-
ween the individual sample positions and the
average. The average of the tristimulus
values for all 30 measurements of each carpet
section was calculated, and the color differ-
ences between those averages and the average
tristimulus value for the conventionally dyed
standard were calculated in order to determine
overall lot-to-lot variation.
Color differences between each of ten in-
dividual measurements along the carpet sam-
ple and the average for the sample for each of
the pilot-scale dyeings are shown in Table 1.
366
-------
TABLE 1. ENO-TO-END COLOR DIFFERENCES IN MACADAM UNITS OF NYLON
CARPET DYED WITH DISPERSE DYES
Position
1
2
3
4
5
6
7
8
9
10
Run 1
1.68
0.67
1.04
1.99
0.35
1.60
0.72
3.39
0.98
1.10
Run II
0.76
1.33
3.90
1.82
3.57
2.38
3.21
0.84
2.53
5.78
Run III
2.61
4.02
0.97
1.00
3.75
3.15
1.41
4.17
1.96
1.23
Run IV
1.81
5.16
3.17
1.90
1.61
2.11
0.76
1.55
3.59
1.52
RunV
0.63
0.40
1.30
1.13
2.14
0.69
0.46
1.79
2.09
1.64
These data indicate that there is little dif-
ference in end-to-end color variation between
the sample dyed in the conventional manner
(Run 1) and samples dyed in recycled dyebaths
(Runs II to V).
Reproducibility of dyeings in the recycled
baths is shown in Table 2. The differences in
color between the first run (conventional dye-
ing) and dyeings in the recycled baths were
determined from the difference in the average
color of each run and the average for Run I. It
is clear from Table 2 that differences in color
for samples dyed in recycled baths are well
within the acceptable commercial range.
Dyeing of Polyester Carpet With Disperse
Dyes
The second trial run was conducted on
polyester carpet dyed with disperse dyes. In
this series three shades in a color line—a light
beige, a medium blue, and a rust —were
selected for dyeing. These colors were actual
large-volume shades currently being dyed by a
leading carpet firm. The objective of this run
was to dye polyester carpet seven times (3
beige, 2 blue, 2 rust) in the same dyebath utiliz-
ing the reuse technology. In addition, the run
was carried out using material handling pro-
cedures that were designed to be compatible
with in-plant dyeing. The following sequence
was used:
1. Dye a beige carpet in conventional
manner,
2. Pump dyebath to holding tank,
3. Rinse carpet 1 in beck,
4. Remove carpet 1 from beck,
5. Clean fiber from rinse water,
6. Enter carpet 2 for prescour,
7. Drop rinse-prescour water,
8. Pump dyebath to beck,
9. Dye carpet 2 beige, and
10. Repeat steps 2 through 9 for carpets 2
through 7.
TABLE 2 BETWEEN SAMPLE COLOR DIFFERENCES FOR
PILOT-SCALE DYEING OF NYLON CARPET WITH
DISPERSE DYES (IN MACADAM UNITS)
Run
t (control)
IV
V
Color difference
versus control
1.3
0.4
0.5
0.5
367
-------
The dyebeck was modified by the addition of a
pump and insulated holding tank for this series
of runs. It should also be noted that the final
rinse bath was used to give the subsequent
carpet a short prescour before dropping the
rinse bath.
In a typical dyeing in a previously used bath,
the beck was full at the beginning of the cycle
with the final rinse from the previous run (at
70° C). The carpet was entered and allowed to
run for 5 minutes to partially remove tufting
oils and fiber identification tints. The rinse-pre-
scour bath was then dropped and the dyebath
used for dyeing the previous carpet pumped to
the beck from the holding tanks. The dyebath
temperature was 75° C. The bath was made up
to volume with freshwater to give a 20:1 liquor
ratio. The quantities of all auxiliaries (except
carrier) were determined from the volume of
freshwater required and added to the beck.
The carpet was run for 5 minutes, the pH ad-
justed to between 5.0 and 5.2, and the required
quantities of dyes added to the bath over a
15-min period. The carrier (10 percent owf) was
then added slowly over a 15-min period. The
dyebath was raised to the boil at 3° C/minute
and run at the boil for 1 hour. At the end of the
dyeing the bath was returned to the holding
tank and analyzed for the residual dyes. The
carpet was post-scoured, rinsed, removed from
the beck, and air-dried.
Initial inspection of the dyed carpets sug-
gested that good color reproducibility and
uniformity had been achieved. Ten samples
were cut from each of the dyed samples and the
color measured on the Diano Colorimeter. Each
sample was measured three times to average
variations due to instrument changes and sam-
ple texture differences. Tristimulus values
(X,Y,Z) were calculated from the average of the
'three measurements.
The base color of each dyed sample was
determined by averaging the ten values ob-
tained from the ten different samples cut from
each carpet. Color uniformity was evaluated by
calculation of the color difference of each of the
ten samples from the average value for the car-
pet. These results for the seven dyed carpets
are shown in Table 3. The color difference
values were calculated using the FMC II color
difference equation.
Color differences between the beck-dyed
samples and standards prepared previously in
beaker dyeings by conventional dyeing pro-
cedures have also been determined. In these
calculations the average color for each carpet
was determined from ten measurements, and
in turn, the difference of this average from the
average of ten measurements on the standard
was determined. These calculations give a mea-
sure of the reproducibility of the dyeings in the
recycled baths and are shown in Table 4. In ad-
dition to the usual FMC II color differences,
the color differences in the new C.I.E. 1976
recommended color space, L*a*b*, are shown
for comparison. With the exception of the two
blue samples, the reproducibility was ex-
cellent. Reproducibility in the blue samples
TABLE 3. COLOR UNIFORMITY IN POLYESTER CARPET DYEINGS WITH DYEBATH REUSE
Ae in FMC II units
Sample
number
1
2
3
4
5
6
7
8
9
10
Beige
#1
1.14
2.97
3.30
0.99
0.89
4.28
0.66
3.03
4.59
4.49
Beige E
leige Blue
#2 #3 #1
1.93 1.97 1
.15
2.58 0.97 1.00
0.95 0.87 1
.27
3.40 0.99 0.64
1.89
1.44
0.85
0.93
1.81
.26 0
.96
.70 2.46
.27 2.56
.73 2.02
.33 0.76
1.64 0.74 2.33
Blue
#2
2.37
0.93
0.28
0.89
1.09
1.29
0.94
1.23
1.24
2.82
Rust
#1
1.16
1.51
1.00
0.76
1.42
0.53
2.77
1.39
0.89
2.92
Rust
#2
1.54
0.81
2.86
1.23
2.68
0.47
0.81
2.03
1.82
1.44
368
-------
TABLE 4. COLOR REPRODUCIBILITY IN POLYESTER CARPET
DYEINGS WITH DYEBATH REUSE
Sample
Beige #1
Beige #2
Beige #3
Blue #1
Blue #2
Rust#1
Rust #2
FMCII
Ae
0.44
0.76
0.27
4.85
3.7
1.70
1.80
CIE L*a*b»
Ae
1.22
0.43
0.09
3.64
2.50
1.17
1.25
was at the outer limit of what can be con-
sidered acceptable commercial carpet dyeing.
A comparison of the two blue samples with
each other showed that they were very close
together in color (Ae = 1.44), but different
from the standard. In both blue carpets the
yellow was higher than the standard, and the
samples were lighter than the standard. Addi-
tion of a s.nall quantity of blue dye would have
corrected this color difference between the
samples and the standard. Since we had
elected to do "no add" dyeing in this series,
this dye-add was not made. A slight modifica-
tion of the dye formula would have corrected
the small difference between the blue samples
and the standard.
In addition to instrumental color measure-
ments, the dyed carpets have been examined
visually by four experienced carpet dyehouse
supervisors. All four agreed that both the
nylon and polyester carpet dyeings were ac-
ceptable as first-quality in both color uniformi-
ty and color reproducibility.
Measurement of the fastness properties of
the dyeings to light, crocking, and water were
also carried out. These results are shown in
Tables 5,6, and 7. All dyeings appeared to have
commercially acceptable fastness properties.
Economic Evaluation of Carpet Dyeing
With Oyebath Reuse
The energy, materials, and cost savings that
are possible with reuse of the dyebath are
dependent on a number of factors. These in-
clude the type of yarns, the shade being dyed,
the number of times the bath is used, the type
of auxiliary chemicals used, and the tempera-
ture of incoming water. All of these will in-
fluence the possible reduction in materials,
energy, and cost. Calculations of these savings
have been carried out based on the two pilot-
scale dyeings of carpet discussed above. These
savings are shown in Table 8. The reductions in
dye, chemical, and water requirements with re-
cycle of the dyebath were determined directly
from the quantities of materials used in the
pilot-scale dyeings. Since the beck used was
TABLE 5 EVALUATION OF LIGHT FASTNESS FOR POLYESTER CARPET
DYED BY THE RECONSTITUTED DYE BATH METHOD
Color
Beige
Beige
Beige
Blue
Blue
Rus'
Ru>r
Dyeing
1
2
3
4
5
6
7
24 hours
5
5
5
5
5
5
5
48 hours
5
5
5
5
5
5
5
72 hours
5
5
5
5
5
5
5
96 hours
4-5
4-5
4-5
5
4-5
5
5
120 hours
4-5
4-5
4-5
4-5
—
369
-------
TABLE 6. EVALUATION OF FASTNESS TO DRY
CROCK FOR POLYESTER CARPETS DYED BY THE
RECONSTITUTED DYEBATH METHOD
Shade
Beige
Beige
Beige
Blue
Blue
Rust
Rust
Dyeing
number
1
2
3
4
5
6
7
Rating
Sample 1
5
5
5
5
5
5
4-5
Sample 2
5
5
5
5
5
5
4-5
TABLE 7. EVALUATION OF FASTNESS TO WATER FOR
POLYESTER CARPETS DYED BY THE RECONSTITUTED
DYEBATH METHOD
Shade
Dyeing no.
Rating
Beige
Beige
Beige
Blue
Blue
Rust
Rust
1
2
3
4
5
6
7
5
5
5
5
5
5
4-5
TABLE 8. REDUCTION IN MATERIALS, ENERGY, AND COST BY REUSE OF
DYEBATHS IN CARPET DYEING
Reduction in percent
Nylon, disperse dyes
5 cycles
Polyester, disperse dyes
7 cycles
Dyes
Chemicals
Water
Energy
Cost
3
65
34
27
27
7
7
39
27
12
Cost summaries for carpet dyeing
Nylon, disperse dyes
5 cycles
Dyes (data and cost)
Chemicals (data and cost)
Water (data and 45 cents/1,000)
Energy (calc. and $3 per 106 Btu)
Total cost
Cost per pound
Savings percent
Savings cents/pound
Conventional
2.18
1.35
0.45
3.99
$7.97
0.046
27.5
1.3
Reuse
2.11
0.47
0.29
2.91
5.78
0.033
Polyester, disperse dyes
7 cycles
Conventional
10.25
7.29
0.26
5.43
22.93
0.094
12.4
1.2
Reuse
9.70
6.57
0.07
3.74
20.08
0.082
not equipped with energy monitoring devices,
the energy reductions were calculated. An ex-
ample of the energy reduction calculations is
given in Table 9.
These results suggest that significant reduc-
tions in water use, chemical requirements,
energy requirements, and cost can be achieved
by dyeing carpet with reconstitution and recy-
cle of the dyebath.
The estimated cost for a medium size
(10-beck) carpet mill to convert 50 percent of
dyeing capacity to the dyebath reuse system is
shown in Table 10. The calculation assumes the
purchase of new equipment with a separate
370
-------
TABLE 9. ENERGY SAVINGS FOR PILOT-SCALE DYEINGS NYLON WITH DISPERSE DYES
(5 CYCLES)
Basis: 35 pounds Carpet samples, 875 pounds water (105 Gallons), 25:1 liquor ratio
Incoming water at 60° F
Makeup water required each cycle ~ 25 gallons
Dyeing temperature 212° F, rinse temperature 120° F
Efficiency of energy utilization at beck 90 percent
Reuse entry temperature 170° F
Steam required to hold at boll 1 pound/gal/h
Carpet retains 300 percent water between processes
Conventional dyeing
Heat bath 60° to 212° F (152° x 875 pounds x 1.11 x 1 Btu/pound) = 14.76 x 104 Btu
Heat carpet 60° to 212 ° F (152° x 35 pounds x 1.11 x 0.5 Btu/pound) = 0.30 x 104 Btu
Hold at boil for 45 min. (105 Gal x 1,000 Btu/lbx 0.75 h) = 7.88 xlO4 Btu
Rinse at 120° F (60° x 770 pounds x 1.11 - [140 poundsx 92° + 35 poundsx 92°x 0.5]) = 3.68x10* Btu
Total per cycle = 26.62 x 104 Btu
Total 5 cycles = 1.33 xlO6 Btu
Btu/pound carpet = 7,600
Reuse dyeing
Heat makeup water 60° to 170° F (110x100 poundsx 1.11x1 Btu) = 1.22 XlO4 Btu
Heat retained prescour water 120° to 170° F (50°x 100 pounds x 1.11 x 1 Btu) = 0.56 x 104 Btu
Heat bath 170° to 212° F (42° X 875 pounds x 1.11 x 1 Btu/pound) = 4.08 x 104 Btu
Heat carpet 120° to 212° F (92° x 35 pounds x 1.11 x 0.5 Btu/pound) = .18 x 104 Btu
Holdatboil45min.(105Galxl,OOOBtu/galx0.75h) =7.88 lO^Btu
Rinse at 120° F (60° x 770 pounds X 1.11 - [140 poundsX 92° + 35 pounds x 92° x 0.5]) - 3.68 IQ^Btu
Total per reuse cycle = 17.60 x 104 Btu
Total cycles (1 conventional + 4 reuse) = 0.97 xlO6 Btu
Btu/pound carpet dyed = 5,543
Energy savings - 27 percent
Energy savings - 2,057 Btu/pound carpet dyed
TABLE 10. PAYBACK ON CAPITAL COSTS
Basis: Plant with 10 becks-5 to be modified for reuse
Capacity 2,000 pounds per cycle, 6 cycles per day per beck
No additional personnel required
Swings using dyebath reuse, If per pound of carpet dyed
No productivity increases are assumed
Capital costs:
Returns:
Payout:
Hold tanks (new) 20,000 each X 5
Pumps 3,000 each X 5
Piping, valves 1,500 each x 5
Analysis equipment
Total cost
Production 5 becks x 2,000 pounds/beck x 6 cycles/day
Returns par day 60,000 pounds/day x 1* /pound
t129.SOO/$600
100,000
15,000
7,500
7,000
129,500
= 60,000 pounds/day
= $600/day
= 21 6 days
=====
371
-------
holding tank and pumping system for each of
the five becks. At an estimated savings of 1
cent per pound of carpet dyed (considerably
less than indicated in the previous cost analy-
sis) the payback time for the installation is less
than 1 year.
Batch Dyeing of Nylon Pantyhose
Hosiery is colored exclusively by batch dye-
ing processes. The two main types of equip-
ment utilized are the overhead paddle machine
and the rotating drum machine. In hosiery dye-
ing a very limited range of dyes and auxiliaries
is employed. Thus, hosiery dyeing presented
an excellent opportunity for application of the
reuse dyeing technology.
Nylon pantyhose dyeing was selected for ini-
tial pilot-scale studies. Both Nylon 6 and Nylon
66 are used in pantyhose and both types are
dyed to various shades of brown and black with
disperse dyes. The same three disperse dyes
are used to produce the majority of shades.
An 8-lb rotary drum machine was equipped
with a holding tank and pumping system for
the reuse studies. Recipes for three Nylon 66
shades (light, medium, and dark brown) and
two Nylon 6 shades (medium and dark brown)
were obtained from a leading hosiery manufac-
turer. Preliminary experiments indicated that
these shades could be dyed in any desired
order by reconstitution and reuse of the dye-
bath. The dyeing procedure developed for the
experimental trial is shown in Table 11. The
first cycle is a conventional dye cycle with each
reuse cycle after the first essentially identical.
Note that in the reuse runs the rinse water
from the previous cycle was used for a quick
prescour of the hosiery before dyeing. This
prescour was used to partially remove knitting
TABLE 11. PROCEDURE FOR DYEING PANTYHOSE WITH DYEBATH REUSE
I. Dyeing batch I — Conventional cycle
1. Load machine
2. Fill with cold water, raise to 90° F
3. Add auxiliary chemicals (1 percent dyeing assistant, 1 percent leveling agent,
0.5 percent wetting agent) run 10 minutes
4. Add dyes
5. Raise temperature 3° F/minute to 170° F
6. Run 20 minutes at 170° F
7. Sample, make adds if necessary
8. Pump dyebath to holding tank, sample and analyze
9. Refill with cold water
10. Heat rapidly to 110°F
11. Add softener and run 10 minutes at 110° F
12. Unload machine
Dyeing batch II - Reuse cycle
1. Load machine
Add prescour chemical (1 percent dyeing assistant) to finish bath in machine
Run 5 minutes
Discharge finish-prescour bath to drain
Pump dyebath from holding tank to dyeing machine
Add necessary auxiliaries to reconstitute the bath
Add dyes to reconstitute the bath slowly from add tank MO minutes)
Raise temperature 3° F/minute to 170° F
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
Run 20 minutes at 170 F
Sample, make adds if necessary
Pump dyebath to holding tank, sample and analyze
Finish as above (Steps 9-12)
111. Dyeing batches 111, IV, V, etc. - Same as batch 11
372
-------
oils and fiber finish that could build up in the
dyebath and affect the reuse dyeing.
The fibers and shade dyed in the pilot-scale
run are shown in Table 12. Approximately 8
pounds of pantyhose were dyed in each run. All
shades were dyed with the same red, yellow,
and blue disperse dyes. It should be noted that
both fiber types and light and dark shades
were randomly mixed during the 13 runs.
Five specimens were cut from the legs of
each of three pairs of pantyhose randomly
chosen from each of the 13 runs and the colors
of each specimen measured on the Diano LSCE
Automate. Color differences were calculated
between the samples from each run and plant
standards for each of the shades. These color
differences (calculated by the FMC II equation)
are shown in Table 12. Similar measurements
on color tolerance standards used by the plant
for two of the shades gave a color difference of
7.8 for the Medium II shade and 2.9 for the
Dark II shade. Thus, color differences for the
reuse dyeinps were well within commercial
standards for those shades. In fact, all dyeings
were well within normally accepted standards
for hosiery dyeing. The dyed hosiery were
returned to the participating company and
were shipped as first-quality goods.
PLANT DEMONSTRATION OF DYEBATH
REUSE TECHNOLOGY
Based on the successful pilot-scale work on
dyeing with reconstitution and reuse of
dyebaths, a plant-scale demonstration of the
reuse technology has been conducted. The
demonstration was carried out in the dyehouse
of the Ladies' Wear Division of the Adams-
Millis Hosiery Company.
A standard 45-kg (100-lb) rotary drum
machine was modified by addition of a holding
tank, pumping system, water meter, and steam
measuring equipment. The machine had been
equipped previously with a temperature con-
trol system and an add tank. Thus, the dyeing
machine could be operated in both the conven-
tional manner and in the dyebath reuse mode
with monitoring of all materials and energy in-
puts. In addition to machine modification, the
dyehouse laboratory purchased a Bausch and
Lomb Spectronic 100 spectrophotometer and a
Hewlett-Packard Model Number 9815A compu-
ter for use in dyebath analysis. Cost of the
analytical equipment and computer was ap-
proximately $7,000. An interface was con-
structed so that the computer could read ab-
sorbance data directly from the spectropho-
TABLE 12. DYEING OF NYLON PANTYHOSE WITH DYEBATH REUSE
Run
number
1
2
3
4
5
6
7
8
9
10
11
12
T3
Fiber
type
66
66
6
6
66
66
66
6
6
66
66
6
66
Shade*
Light 1
Medium 1
Medium II
Dark II
Dark!
Light 1
Medium 1
Medium II
Dark II
Dark!
Light 1
Medium II
Light 1
Color difference
versus standard
2.0
4.2
2.9
1.9
2.8
2.0
3.0
2.2
2.4
2.4
3.0
3.4
4.2
*A total of 5 shades were utilized. The numeral I refers to a Nylon 66 shade,
and the numeral II refers to a Nylon 6 shade.
373
-------
tometer. Programs were written to permit the
computer to calculate and print out the quan-
tities of dyes and auxiliary chemicals needed to
reconstitute the spent dyebath to the level
desired for the subsequent dyeing.
Before beginning the reuse study, 15 conven-
tional dye cycles were run to provide baseline
data for comparison with the reuse dyeings.
Energy, water, dye, and chemical require-
ments for pantyhose dyeings were determined
from this 15-cycle run.
Following the 15 conventional dye cycles, 3
series of dyeings were carried out with recon-
stitution and reuse of the dyebath. In the first
series the dyebath was used five times before
discharge to drain. This series served to check
out the reuse system and establish operating
procedures. The second series consisted of 17
dyeings in the same dyebath. In this series the
rinse water from each cycle was used to
prescour the next load to be dyed similar to t he
laboratory dyeings. The third series, con-
sisting of 14 dyeings in the same bath, was con-
ducted in the same manner as Series II except
that the hosiery was not scoured before dye-
ing. Materials and energy requirements for the
reuse dyeings were monitored in exactly the
same way as the conventional dyeings.
The sequence of dyeings in the Series II,
17-cycle run, is shown in Table 13. It should be
noted that both fiber types and light and dark
shades were indiscriminately mixed in the run.
TABLE 13. SERIES II SEVENTEEN REUSE CYCLES
21
22R
23R
24R
25R
26R
27 R
28 R
29 R
30R
SIR
32R
33R
34R
35 R
36R
37 R
Nylon 66
Nylon 66
Nylon 6
Nylon 6
Nylon 66
Nylon 66
Nylon 6
Nylon 66
Nylon 6
Nylon 6
Nylon 66
Nylon 66
Nylon 6
Nylon 6
Nylon 66
Nylon 66
Nylon 66
Light shade
Medium shade
Medium shade
Dark shade
Dark shade
Light shade
Medium shade
Light shade
Medium shade
Dark shade
Light shade
Medium shade
Medium shade
Dark shade
Medium shade
Light shade
Medium shade
No problems were encountered in dyeing a
light shade following a dark shade.
The materials and energy requirements for
the 15-cycle conventional dye series and for the
17- and 14-cycle reuse series combined are
shown in Table 14. It is apparent that the reuse
dyeings gave substantial reduction in energy,
water, chemical, and dye requirements. The
dye savings are related, of course, to the par-
ticular shades dyed in the various runs but
reduction in chemicals, water, and energy are
essentially independent of the shades being
dyed. Cost reductions achievable with reuse
dyeing are shown in Table 15. Dyeing costs
were calculated using current quoted prices
for chemicals, dyes, energy, water, and sewer
service.
Several important aspects of plant imple-
mentation of dyebath reuse were revealed by
the demonstration runs. First, the analysis
time was reduced during the course of the dem-
onstration from 30 minutes to approximately
10 to 15 minutes with no apparent loss in re-
quired accuracy. In this time the analysis could
be readily completed during the usual rinse cy-
cle and the dyes weighed and readied for the
next load.
Second, the third series of dyeings indicated
that prescouring of the hosiery was unneces-
sary. No differences in the analysis of the bath
or the quality of the dyeings were detected in
the third series. Removal of the prescour step
provided a dyeing cycle more compatible with
the normal plant operation. About the only dif-
ference in Series III runs and conventional
runs was that at the conclusion of dyeing; the
dyebath was pumped to a holding tank rather
than being discharged to the drain and the
beck was filled with used dyebath rather than
freshwater in the subsequent cycle.
Careful records were kept of the number of
adds required to produce the desired shade in
both conventional and reuse dyeings. Savings
from dyebath reuse could be rapidly lost if
more adds were necessary in reuse dyeings.
The 15 conventional cycles required a total of
11 adds for an add rate of 0.73 adds per cycle.
The 31 reuse cycles required 12 adds for a rate
of 0.39 adds per cycle. Thus, the number of
adds required in the reuse dyeings was actual-
ly 47 percent less.
The hosiery dyeing plant runs clearly dem-
onstrated that reuse dyeing can be incor-
374
-------
TABLE 14. MATERIALS AMD ENERGY REDUCTION REUSE DYEING
Energy
Water
Conventional
Reuse
Reduction
Conventional
Reuse
Reduction
Chemicals
Dyes
Conventional
Reuse
Reduction
Conventional
Reuse
Reduction
2.18 pounds steam/pound fiber dyed
0.94 pounds steam/pound fiber dyed
57 percent
4.08 gallons/pound fiber dyed
2.31 gallons/pound fiber dyed
43 percent
1.17 ounces/pound fiber dyed
0.76 ounces/pound fiber dyed
35 percent
0.0058 pounds/pound fiber dyed
0.0047 pounds/pound fiber dyed
19 percent
TABLE 15. ENERGY AND MATERIAL COST REDUCTION
IN REUSE DYEINGS
Cost per dye cycle:
Conventional $4.66
Reuse $2.98
Cost per pound of fiber dyed:
Conventional 5.25 if
Reuse 3.28 if
Cost reduction: 37 percent
porated in a commercial operation with no loss
in quality or productivity and with substantial
savings in materials and energy costs. Addi-
tional plant demonstrations of the reuse tech-
nology are planned for carpet dyeing and for
dyeing of polyester/cotton blend yarn
packages. Laboratory work is also being car-
ried out on jet dyeing of knit fabrics.
ACKNOWLEDGMENTS
The authors gratefully acknowledge the fun-
ding of this work by the U.S. Department of
Energy. Further details on reuse technology
will appear in reports issued by that depart-
ment.
The plant demonstration phase of this work
was made possible by the generous support
and assistance of the management and person-
nel of the Adams-Millis Hosiery Company.
The substantial contributions to the labora-
tory phase of the project by Mrs. Lynn
Averette and Mr. Jimmy Wadia, student
research assistants, are gratefully acknow-
ledged.
375
-------
REFERENCES
1. U.S. Environmental Protection Agency,
"In-Plant Control of Pollution: Upgrading
Textile Operations to Reduce Pollution,"
Environmental Protection Agency Tech-
nology Transfer Seminar Publication,
EPA-62513-74-004, October 1974.
2. J. F. Lowry, et al., "Energy Conservation
in the Textile Industry," Phase I Technical
Report, Department of Energy Project,
E(40-l)-5099, Engineering Experiment Sta-
tion, Georgia Institute of Technology,
Atlanta, GA, p. 36, April 1977.
3. J. F. Lowry, "Energy Conservation in the
Textile Industry," Phase I Technical
Report, Department of Energy Project,
E(40-l)-5099, Engineering Experiment Sta-
tion, Georgia Institute of Technology,
Atlanta, GA, p. 29, April 1977.
4. C. Giles, Notes for a Laboratory Course in
Dyeing, Society of Dyers and Colorists,
Yorkshire, England, p. 53,1966.
5. D. B. Judd and G. Wyszecki, Color in
Business, Science and Industry, Third Edi-
tion, John Wiley and Sons, New York,
p. 322,1957.
6. W. C. Tincher, "Conservation of Water,
Chemicals and Energy in Dyeing Nylon
Carpet," Final Technical Report
ERC-07-77, Office of Water Research and
Technology Project A-064-GA, School of
Textile Engineering, Georgia Institute of
Technology, Atlanta, GA, November 1977.
7. W. C. Tincher, American Dyestuff
Reporter, Vol. 66, No. 5, p. 36, 1977.
376
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before ct
EPA-600/2..79-TfU
3. RECIPIENT'S ACCESSION NO.
\ND SUBTITLE
-"^WVV^ktlllftH^
Symposium Proceedings: Textile Industry Technology
(December 1978, Williamsburg, VA)
5. REPORT DATE
May 1979
6. PERFORMING ORGANIZATION CODE
1OR1S)
Frank A. Ayers (Compiler)
8. PERFORMING ORGANIZATION REPORT NO.
'ERFORMING ORGANIZATION NAME AND ADDRESS ~
Research Triangle Institute
P.O. Box 12194
Research Triangle Park, North Carolina 27709
10. PROGRAM ELEMENT NO.
1AB604 and 1BB610
11. CONTRACT/GRANT NO.
68-02-2612, Task 37
12. SPONSORING AGENCY NAME AND ADDRESS
EPA, Office of Research and Development
Industrial Environmental Research Laboratory
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Proceedings; 11/77 - 3/79
14. SPONSORING AGENCY CODE
EPA/600/13
15. SUPPLEMENTARY NOTES ffiRL-RTP project officer is MaxSamfield, Mail Drop 62, 919/-
541-2547.
16. ABSTRACT The pj.^^^^ document most of the approximately 40 presentations at
the symposium, December 5-6, 1978, at Williamsburg, VA. The symposium, spon-
sored by EPA/IERL-RTP's Chemical Processes Branch, provided an exchange of
ideas between industry, government, and academic representatives in three related
areas: emissions control, energy conservation, and material recovery. There was
also a session on assessment methodology. This was the first symposium on textile
industry technology sponsored solely by the U.S. EPA. It represents EPA's most
extensive technology transfer activity dedicated to support the attainment of pollution
control and energy and material conservation.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS
COSATI Field/Group
Pollution
Textile Industry
Emission
Conservation
Energy
Materials Recovery
Assessments
Pollution Control
Stationary Sources
13B
HE
14B
07A,13H
18. DISTRIBUTION STATEMENT
Unlimited
19. SECURITY CLASS (ThisReport)
Unclassified
IO. OF PAGES
384
20. SECURITY CLASS (This page)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
377
------- |