4»EPA
United States
Environmental Protection
Agency
Office of Research and
Development
Washington, DC 20460
April 1991
Research and Development
Remedial Action,
Treatment, and Disposal
of Hazardous Waste
Proceedings of the
Seventeenth Annual
RREL Hazardous
Waste Research
Symposium
-------
-------
EPA/600/9-91/002
April 1991
REMEDIAL ACTION, TREATMENT," AND DISPOSAL
OF HAZARDOUS WASTE
Proceedings of the Seventeenth Annual Hazardous Waste Research Symposium
Cincinnati, OH, April 9-11, 1991
Sponsored by the U.S. EPA, Office of Research & Development
Risk Reduction Engineering Laboratory
Cincinnati, OH 45268
Coordinated by:
JACA Corp.
Fort Washington, PA 19034
under subcontract to:
Science Applications
International Corporation
Cincinnati, OH 45203
Project Officers:
Marta K. Richards
Gordon M. Evans
H. Paul Warner
RISK REDUCTION ENGINEERING LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OH 45268
Printed on Recycled Paper
-------
NOTICE
These Proceedings have been reviewed In accordance with the U.S.
Environmental Protection Agency's peer and administrative review policies
and approved for presentation and publication. Mention of trade names or
commercial products does not constitute endorsement or recommendation for
use.
-------
FOREWORD
Today's rapidly developing technologies and industrial practices
frequently carry with them the increased generation of materials, that if
improperly dealt with, can threaten both public health and the environment.
The U.S. Environmental Protection Agency is charged by Congress with
protecting the Nation's land, air, and water resources. Under a mandate of
national environmental laws, the Agency strives to formulate and implement
actions leading to a compatible balance between human activities and the
ability of natural systems to support and nurture life. These laws direct
the EPA to perform research to define our environmental problems, measure
the impacts, and search for solutions.
The Risk Reduction Engineering Laboratory is responsible for
planning, implementing, and managing research, development, and
demonstration programs to provide an authoritative, defensible
engineering basis in support of the policies, programs, and regulations of
the EPA with respect to drinking water, wastewater, pesticides, toxic
substances, solid and hazardous wastes, and Superfund-related activities.
This publication is one of the products of that research and provides a
vital communication link between researchers and users.
These Proceedings from the 1991 Symposium provide the results of
projects recently completed by RREL and current information on projects
presently underway. Those wishing additional information on these projects
are urged to contact the author or the EPA Project Officer.
RREL sponsors a symposium each year in order to assure that the
results of its research efforts are rapidly transmitted to the user
community.
E. Timothy Oppelt, Director
Risk Reduction Engineering Laboratory
-------
ABSTRACT
The Seventeenth Annual Research Symposium on Remedial Action,
Treatment, and Disposal of Hazardous Waste was held In Cincinnati, Ohio,
April 9-11, 1991. The purpose of this Symposium was to present the latest
significant research findings from ongoing and recently completed projects
funded by the Risk Reduction Engineering Laboratory (RREL).
These Proceedings are organized in three sections: Sessions A and B
consist of paper presentations. Session C contains the poster abstracts.
Subjects include remedial action treatment and control technologies for
waste disposal, landfill liner and cover systems, underground storage
tanks, and demonstration and development of Innovative/alternative
treatment technologies for hazardous waste. Alternative technology
subjects include thermal destruction of hazardous wastes, field
evaluations, existing treatment options, emerging treatment processes,
waste minimization, and biosystems for hazardous waste destruction.
IV
-------
CONTENTS
SESSION A
Page
Site Demonstration of Microfiltration Technology for Groundwater
Contaminated with Metals
John F. Martin, U.S. Environmental Protection Agency 1-4
Site Demonstration of the CF Systems Organic Extraction Process
Laurel J, Staley, U.S. Environmental Protection Agency 5-18
Summary Results of the SITE Demonstration for the CHEMFIX
Solidification/Stabilization Process
Edwin F. Barth, U.S. Environmental Protection Agency. 19-27
Soil Vapor Extraction: Air Permeability Testing and Estimation
Methods
Ghi-Yuan Fan, U.S. Environmental Protection Agency 28-42
Underground Storage Tanks Containing Hazardous Chemicals
Joseph Maresca, Vista Research, Inc 43-56
Subsurface Fate and Transport of Petroleum Hydrocarbons From
Leaking Underground Storage Tanks: A Basis for Evaluating
The Effectiveness of Corrective Actions
Warren J. Lyman, Camp, Dresser & McKee 57-65
The Incineration of Lead-Contaminated Soil Related to the
Comprehensive Environmental Response Compensation and Liability
Act (CERCLA) (Superfund)
Howard 0. Wall, U.S. Environmental Protection Agency 66-78
Full-Scale POHC Incinerability Ranking and Surrogate Testing
Andrew Trenholm, Midwest Research Institute 79-88
EPA's Mobile Volume Reduction Unit for Soil Washing
Hugh Masters, U.S. Environmental Protection Agency... 89-104
Guidance for Treatability Testing Under CERCLA: An Update
David Smith, U.S. Environmental Protection Agency 105-118
S.I.T.E. Demonstration of A Soil Washing System by Biotrol Inc.,
at a Wood Preserving Site in New Brighton, Minnesota
Mary K. Stinson, U.S. Environmental Protection Agency 119-130
Background Information on Clean Products Research and Implementation
Marjorie A. Franklin, Franklin Associates, Ltd 131-145
Industrial Pollution Prevention Strategy: Research Priorities
and Opportunities for the 1990's
Ivars J. Licis, U.S. Environmental Protection Agency. 146-161
-------
SESSION A (Continued)
Page
The Evaluation of an Advanced Reverse Osmosis System at the
Sunnyvale, California Hewlett-Packard Facility
Lisa M. Brown, U.S. Environmental Protection Agency.. 162-171
The Behavior of Trace Metals in Rotary Kiln Incineration; Results
of Incineration Research Facility Studies
D.J. Fournier, Jr., Acurex Corporation 172-189
Soil Heating Technologies for In Situ Treatment: A Review
Janet M. Houthoofd, U.S. Environmental Protection Agency....*... 190-203
The United States Environmental Protection Agency Municipal Waste
Combustion Residue Solidification/Stabilization Program
Carl ton C. Wiles, U.S. Environmental Protection Agency ..'. 204-224
Weathering of Selected Degradable Plastic Materials Under Outdoor
and Laboratory Exposure Conditions
Anthony L. Andrady, Research Triangle Institute... 225-237
Pollution Prevention Research Within the Federal Community
James S. Bridges, U.S.;Environmental Protection Agency.......... 238-257
Toxic Substance Reduction for Narrow-Web Flexographic Printers
Paul M. Randall, U.S. Environmental Protection Agency........... 258-286
VI
-------
SESSION B
The Treatability of Urban Stormwater Toxicants
Patricia F. Barren, Law Engineering 287-308
Report on Enhancing Effects of Low Frequency Vibration in Soil
Washing
Maivina Wilkens, U.S. Environmental Protection Agency , 309-328
Vacuum-Assisted Steam Stripping to Remove Pollutants from
Contaminated Soil: A Laboratory Study
Arthur E. Lord, Jr., Drexel University 329-352
Effectiveness of Commercial Microbial Products in Enhancing Oil
Degradation in Prince William Sound Field Plots
Albert D. Venosa, U.S. Environmental Protection Agency.......... 353-369
Biodegradation of Volatile Organic Compounds in Aerobic and
Anaerobic Biofliters
Rakesh Govind, University of Cincinnati......................... 370-387
Screening of Commercial Bioproducts for Enhancement of Oil
Biodegradation in Closed Microcosms
John R. Haines, U.S. Environmental Protection Agency..... 388-403
Alternating Current Electrocoagulation for Superfund Site
Remediation
Clifton W. Farrell, Electro-Pure Systems, Inc 404-415
High Energy Electron Beam Irradiation: An Emerging Technology for
the Removal of Hazardous Organic Chemicals From Water and Sludge -
An Introduction
William J. Copper, Florida International University..... 416-435
Soil Barrier Alternatives
Walter E. Grube, Jr., U.S. Environmental Protection Agency...... 436-444
Waste Minimization Assessment Centers - Cost Savings Recommended
and Implemented in Twelve Manufacturing Plants
F. William Kirsch, University City Science Center 445-459
Evaluating Final Covers for Hazardous Waste Landfills Using A
Rule-Based Knowledge System
James T. Decker, Computer Sciences Corporation 460-476
Evaluation of Asbestos Release from Building Demolition Following
the October 1989 California Earthquake
Bruce A. Hollett, U.S. Environmental Protection Agency 477-487
Assessing the Risk of Remedial Alternatives at Superfund Sites -
Implications for Technology Demonstrators
Patricia Lafornara, U.S. Environmental Protection Agency 488-495
vii
-------
SESSION B (Continued)
Page
Closure of a Dioxin-Contaminated Superfund Site
Joyce Perdek, U.S. Environmental Protection Agency 496-510
Thermal Desorption Attainable Remediation Levels
Paul R. dePercin, U.S. Environmental Protection Agency 511-520
Effectiveness of the Stabilization/Solidification Process in
Containing Metals from RCRA Electroplating Wastes
Ronald J. Turner, U.S. Environmental Protection Agency 521-531
Field Assessment of Air Emissions From Hazardous Waste
Stabilization Operations
Thomas C. Ponder, PEI Associates, Inc... , 532-542
Assessment of the Parameters Affecting the Measurement of
Hydraulic Conductivity for Solidified/Stabilized Wastes
D.J. Conrad, Alberta Environmental Centre 543-559
Ozone-Ultraviolet Light Treatment of Iron Cyanide Complexes
Sardar Q. Hassan, University of Cincinnati 560-573
Development of LDR Standards for Contaminated Soil and Debris
Carolyn K. Offutt, U.S. Environmental Protection Agency 574-593
Fate and Treatability of Dyes in Wastewater
Glenn M. Shaul, U.S. Environmental Protection Agency 594-609
Carbon-Assisted Anaerobic Treatment of Hazardous Leachates
A.T. Schroeder, University of Cincinnati 610-625
Anaerobic Pretreatment of An Industrial Waste Containing Several
VOC's
B. Narayanan, John Corollo Engineers.... 626-648
Development of Nonlinear Group Contribution Method for Prediction
of Biodegradation Kinetics from Respirometrically-Derived Kinetic
Data
Henry H. Tabak, U.S. Environmental Protection Agency 649-671
viii
-------
SESSION C
Page
The Effect of Chlorine on NOX Emissions From the Incineration of
Nitrogen-Containing Wastes
William P. Linak, U.S. Environmental Protection Agency 672
Pilot-Scale Evaluation of an Incinerability Ranking System for
Hazardous Organic Compounds
Gregory J. Carroll, U.S. Environmental Protection Agency 673
U.S. EPA Incineration Research Facility Update
O.W. Lee, Acurex Corporation 674
Industry Pollution Prevention Guides
Teresa M. Harten, U.S. Environmental Protection Agency.......... 675
Inorganic Recycling/Delco Region V Hazardous Waste Recycling
Determination
Brian A. Westfall, U.S. Environmental Protection Agency.... 676
New Jersey/EPA Waste Minimization Assessments
Mary Ann Curran, U.S. Environmental Protection Agency.... 677
Asbestos Control in Buildings
Thomas J. Powers, U.S. Environmental Protection Agency 678
Supercritical Water Oxidation Deep-Well Reactor Model Development
Earnest F. Gloyna, The University of Texas at Austin 679
Expansion of RREL Data Base to Include Soil, Debris and Sediment
Stephanie A. Hansen, Radian Corporation 680
Orismology: The Next Step in Quality Control
Guy F. Simes, U.S. Environmental Protection Agency 681
Research Opportunities at EPA's E-TEC Facility
Daniel Sullivan, U.S. Environmental Protection Agency 682
Accessing Leaking Underground Storage Tank Case Studies and
Publications Through the EPA's Computerized On-Line Information
System (COLIS)
Robert W. Hillger, U.S. Environmental Protection Agency......... 683
Metal Value Recovery from Electromachining Sludge Wastes
Larry G. Twidwell, Montana College of Mineral Science and
Techno! ogy 684
Bench-Scale Wet Air Oxidation of Dilute Organic Wastes at the
Environmental Protection Agency's Test and Evaluation Facility
Avi N. Patkar, IT Environmental Programs, Inc 685
ix
-------
SESSION C (Continued)
Page
Characteristics of Bevill Smelting Wastes
Henry Huppert, Science Applications International Corp.. 686
Test Program for Evaluation of Foam Scrubbing for Control of
Superfund Toxic Gas Releases
Pat Brown, Foster Wheeler Enviresponse 687
Vapor-Liquid Equilibrium Studies at the U.S. Environmental
Protection Agency's Test and Evaluation (T&E) Facility
Franklin Alvarez, U.S. Environmental Protection Agency 688
Sorption Isotherms for Azo Dyes Onto Activated Sludge Biomass
Richard J, Lieberman, U.S. Environmental Protection Agency 689
EPA's Synthetic Soil Matrix (SSM) Blending Facility
Raymond H. Frederick, U.S. Environmental Protection Agency 690
RPM/OSC Summary - Remedial/Removal Incineration Projects
Laurel J. Staley, U.S. Environmental Protection Agency 691
Treatability Study Results on Soil Contaminated with Heavy Metals,
Thiocyanates, Carbon Disulfate, Other Volatile and Semi volatile
Organic Compounds
Sarah Hokanson, PEI Associ ates 692
Evaluation of Temporal Changes in Soil Barrier Water Content
R.J. Luxmoore, Oak Ridge National Laboratory 693
Evaluation of Stress Cracking Resistance of Polyethylene Flexible
Membrane Liners
Yick Halse-Hsuan, Drexel University 694
Center Hill Solid and Hazardous Waste Research Facility
Gerard Roberto, University of Cincinnati 695
-------
SITE DEMONSTRATION OF MICROFILTRATION TECHNOLOGY FOR
GROUNDWATER CONTAMINATED WITH METALS
John F. Martin
Risk Reduction Engineering Laboratory
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
Kirankumar Topudurti
Stanley Labunski
PRC Environmental Management, Inc.
Chicago, Illinois 60601
ABSTRACT
The Superfund Innovative Technology Evaluation (SITE) Program has as its
major thrust the documentation of reliable performance and cost information
for innovative alternative technologies so that they are developed,
demonstrated, and made commercially available for the permanent cleanup of
Superfund sites. Demonstration projects identify limitations of the
technology, applicable wastes and waste media, potential operating problems,
and the approximate cost of applying the technology.
A demonstration project was conducted with E.I. DuPont de Nemours &
Company, Inc. and the Oberlin Filter Company to evaluate a microfiltration
technology for removal of suspended solids from wastewater. The
microfiltration system utilized DuPont's Tyvek® T-980 membrane filter media in
conjunction with the Oberlin automatic pressure filter. The project was
undertaken at the Palmerton Zinc Superfund site in April 1990 to evaluate the
ability of the technology to remove zinc from the site's shallow groundwater.
Pretreatment of the groundwater to precipitate dissolved zinc and other metals
was included as part of the demonstration program. The treated filtrate
indicated that the system removed precipitated zinc and other suspended solids
at greater than 99.9%, and the filter cake produced during the study passed
both the EP Toxicity test and the TCLP.
INTRODUCTION
Over the past few years, it has become increasingly evident that land
disposal of hazardous wastes is at least only a temporary solution for much of
the material present at Superfund sites. The need for more long-term,
permanent treatment solutions as alternatives to land disposal has been
stressed by recent legislation such as the Hazardous and Solid Waste
Amendments of the Resource Conservation and Recovery Act (RCRA) as well as the
Superfund Amendments and Reauthorization Act (SARA) of 1986. SARA directed
the U.S. Environmental Protection Agency to establish an "Alternative or
Innovative Treatment Technology Research and Demonstration Program," to
identify promising technologies, assist with their evaluation, and promote the
use of these technologies at Superfund sites. The Superfund Innovative
1
-------
Technology Evaluation (SITE) Program resulted from that mandate.
The SITE Program is now in its sixth year of demonstrating technologies
applicable to Superfund sites with 52 developers conducting 55 projects. The
Program offers several advantages to participants, in addition both the Agency
and technology developers benefit from the demonstrations. Primary benefits to
developers include; experience gained from operating a commercial, field-
scale process at a Superfund site; acquisition of valuable regulatory
background; increased public awareness of the technology and its capabilities;
and documentation of the applicability of the process to cleanup of hazardous
waste sites.
Under the Demonstration Program, the developer and EPA participate in a
joint venture to operate and evaluate a technology. In general, the developer
is required to operate the technology at the selected location while EPA is
primarily responsible for writing a demonstration plan, for all sampling and
analytical operations, and for reporting and technology transfer activities.
Demonstrations at Federal or State Superfund sites (remedial or removal
action sites), EPA test facilities, or at Federally owned sites are
encouraged; however, if such sites are not available or not applicable, a
developer's facility or a private site may be utilized. EPA is becoming
increasingly flexible in the designation of appropriate sites as the
Demonstration Program continues to evolve.
TECHNOLOGY DESCRIPTION
The demonstration project conducted by DuPont & Company, Inc., in
conjunction with the Oberlin Filter Company, features a microfiltration system
designed to remove solid particles from liquid wastes, forming a filter cake
typically ranging from 30 to 50 percent solids. The filtration unit can be
manufactured as an enclosed, trailer-mounted system, requiring little or no
attention during operation. The system utilizes Oberlin's automatic pressure
filter (APF) combined with DuPont's special Tyvek® T-980 filter media made of
spun-bonded olefin. The Tyvek® material is a thin, durable fabric with
openings of about one micron. During operation of the unit, it may be
possible to get filtration down to the half-micron range or less. A
microscopic view of standard Tyvek® material shows it to be only slightly
porous, whereas the newer T-980 material has increased porosity and sub-
micron filtration capability.
The APF, supplied by Oberlin, provides the support for pumping wastewater
through the Tyvek® where solids accumulate to form a filter cake.
Contaminated water is pumped across the filter fabric during the filtration
cycle until build-up of a filter cake causes the feed pressure to rise to
approximately 55 psig. At this point, the APF cuts the feed stream and
switches to the cake dewatering cycle where air is blown through the filter
cake to dry it prior to discharge. During the discharge cycle, the upper
portion of the filter is raised and the filter cake is conveyed out of the
filtration chamber on the used Tyvek® filter media as new material is drawn
from the clean media roll into the filter chamber. The upper half of the
filter then lowers, sealing the chamber above the Tyvek® for the next
filtration cycle. The unit cycles through the complete operation
automatically so that there is minimal worker exposure to hazardous materials.
-------
The Oberlin APF is available in a variety of sizes from 2.4 square feet up
to 36 square feet of filtering area. Similarly, Tyvek® filter media is
produced in bulk rolls in several standard widths. The demonstration unit
evaluated by the SITE Demonstration Program during the first three weeks of
April 1990 was a 2.4-square foot, skid-mounted filter. Treatability tests
conducted during July and October 1989, using groundwater from the
demonstration site and a synthetic wastewater designed to simulate the
groundwater, showed excellent performance by the filter in removing
precipitated zinc and other suspended solids. In the July 1989 treatability
test, on two separate groundwater runs with mean influent concentrations of
12,433 mg/1 and 6,640 mg/1 of TSS, the respective effluent concentrations of
TSS were 44 mg/1 and 23 mg/1.
DEMONSTRATION SITE ,
The microfiltration project was located at the Palmerton Zinc Superfund
site in the Lehigh Valley of Pennsylvania. Contamination at the site resulted
from smelting operations begun in 1889 by the New Jersey Zinc Company. In
1980 primary zinc smelting operations at the site were terminated, but
secondary metal refining and processing operations continued under the
ownership of the Zinc Corporation of America, a Division of Horsehead
Industries, Inc.
The solid process waste or slag from the smelting operations has been
disposed at the site since 1913, and by 1986 approximately 33 million tons of
slag had accumulated in a pile nearly 2.5 miles long. Because of elevated
levels of heavy metals in the surface water and groundwater of the Palmerton
area, the slag pile site was included on EPA's National Priorities List of
hazardous waste sites. Samples of the shallow groundwater at the site
indicate that zinc is present at the highest levels (300-500 nig/l), while
copper (0.02 mg/1), cadmium (1 mg/1), and selenium (0.05 mg/1) are present
down to trace levels.
TECHNOLOGY EVALUATION
The demonstration was proposed to evaluate the overall ability of the
DuPont/Oberlin treatment process to remove zinc from the groundwater at the
Palmerton site. In order to accomplish this objective, field studies were
designed to produce data relating to four primary aspects of the technology
application:
1. Precipitation of metals from the groundwater with emphasis on zinc.
2. Filtration and dewatering of the metals precipitate.
3. Production of filtrate and filter cake to meet applicable disposal
requirements.
4. Documentation of operating costs.
During Phase I of the demonstration program, operation of the APF remained
unchanged while lime doses for metals precipitation, and ProFix (a filter aid
material supplied by EnviroGuard, Inc.) doses for cake buildup and
-------
stabilization were varied. Nine separate runs (treatment batches) yielded
data to indicate that optimum chemical addition rates were: lime addition to
pH 9; ProFix addition at the rate of 12 grams per liter.
Optimum operating conditions for the APF were set during Phase II of the
project. During the rest of the evaluation the chemical addition parameters
were held constant. The two items varied during this phase were the pressure
at which air was blown through the filter cake to dry it, and the length of
time allowed for this function after all water was forced from the filter
chamber and air broke through the filter cake. Taking into account the
filtrate quality, the solids content of the cake, and the length of time
(relating to the cost of treatment) for the drying cycle, the optimum
operating conditions were set at a drying (blowdown) time of 0.5 minutes with
38 psig air.
Phase III provided two additional runs at the optimum operating conditions
so that reproducibility of the treatment process could be evaluated. The
influent zinc concentration in Phase II was reduced from 444 mg/1 to 0.22 mg/1
(99.95% removal), while in the two Phase III runs the zinc concentration was
reduced from 465 mg/1 to 0.24 and 0.28 mg/1 (99.94 and 99.95% removal). Total
suspended solids of 12,500 mg/1 in the Phase II influent were reduced to 10.9
mg/1 (99.91% removal), and TSS concentrations of 14,300 and 14,000 mg/1 in the
Phase III tests were lowered to 7.7 and 6.8 mg/1 (indicating 99.95% removal).
Phase IV was designed to test the reusability of Tyvek® in the filter
system. For this portion of the evaluation program, the filter media was
rolled back into the APF following cake discharge. The same area of Tyvek was
used for six filtration cycles with no apparent degradation or loss of
filtering capacity.
Economic evaluation of the system will be reported in the Applications
Analysis Report for this demonstration to be published in the Spring of 1991.
CONCLUSIONS
During optimum operating conditions the system removed zinc and TSS at a
rate of 99.95%. Filter cake solids varied from approximately 30 to 47% with
cake solids being 41% at optimum conditions for filtrate quality, chemical
.addition, and filter cycle time. The filtrate did meet applicable National
Pollutant Discharge Elimination System (NPDES) permit limits for discharge to
a local waterway for metals and TSS, (maximum daily discharge limits to
Aquashicola Creek for zinc and TSS are 2.4 and 30 mg/1) but pH limits were
consistently exceeded. The alkaline nature of the ProFix added to the feed
stream to increase filtration capability consistently raised the effluent pH
to 11.5, thereby violating the 6-9 limit. This condition is not critical,
however, and can be mitigated by adding a pH adjustment step as a
posttreatment option. The filter cake resulting from the process passed the
paint filter liquids test for free liquids at all operating conditions, and a
composite cake sample for the total demonstration successfully passed both the
EP Toxicity and TCLP tests. A large scale system operating over a longer time
might send the filter cake to a metals reclamation facility or a land disposal
site.
-------
SITE DEMONSTRATION
OF THE
CF SYSTEMS
ORGANIC EXTRACTION PROCESS
by
Laurel J. Staley
Risk Reduction Engineering Laboratory
U.S. Environmental Protection Agency
Cincinnati, Ohio 4S268
Richard Valentinetti
Air Pollution Control Division
Department of Environmental Conservation
Agency of Natural Resources
Waterbury, Vermont 05676
and
Jorge HcPherson
Science Applications International Corporation
8400 Westpark Drive
McLean, VA 22102
ABSTRACT
The CF Systems Organic Extraction Process was used to remove PCBs from
contaminated sediment dredged from the New Bedford Harbor. This work was done
as part of a field demonstration under EPA's Superfund Innovative Technology
Evaluation (SITE) program. The purpose of the SITE program is to provide an
independent and objective evaluation of innovative waste remediation processes.
The purpose of this paper is to present the results of the SITE demonstration
of this technology. Results of the demonstration tests show that the system,
which uses liquefied propane, successfully removed PCBs from contaminated
sediments in New Bedford Harbor. Ramoval efficiencies for all test runs exceeded
70%. Some operational problems occuYred during the demonstration which may have
affected the efficiency with which PCBs were removed from the dredged sediment.
Large amounts of residues were generated from this demonstration project. Costs
for using this process are estimated to be between $150/Ton and $450/Ton.
-------
INTRODUCTION
In 1986, the Superfund Amendments and Reauthorization Act (SARA) was passed
to renew the Superfund program. During the reauthorization process, Congress
recognized and acted to correct Superfund's heavy reliance on land disposal by
mandating the use of innovative or alternative technologies. Section 121 of
SARA, Cleanup Standards, requires EPA to favor remedial actions that employ
treatment technologies to permanently reduce the volume, toxicity, or mobility
of the hazardous substances present at the site. Further, Section 121 encourages
EPA to select remedial actions that utilize alternative treatment technologies
to the maximum extent possible. The Superfund Innovative Technology Evaluation
(SITE) program was established by EPA to identify alternative treatment
technologies that can comply with the cleanup standards set forth in Section 121
of SARA.
The SITE program has two main objectives. These are as follows.
1. To conduct demonstrations of promising technologies in order to establish
a catalog of reliable performance and cost information.
2. To disseminate this information to those involved with Superfund site
remediation.
Solvent extraction has been considered for use in decontaminating hazardous
wastes. The idea behind using solvent extraction and other separation
technologies is to reduce the amount of material that needs to be disposed of
as hazardous waste. This is done by separating waste contaminants from their
initial substrate and then concentrating them in a smaller volume. For example,
a sediment contaminated with a few hundred parts per million PCBs can be treated
as follows using separation. The PCBs can be separated from the sediment using
solvent extraction. However, since separation processes do not destroy the PCBs,
they must be captured after separation or they will be released into the
environment. In capturing the released PCBs, however, they are concentrated on
a substrate of smaller volume. As a result, a smaller volume of waste needs to
be treated and disposed of. This makes the overall cleanup process more
efficient. For example, using solvent extraction to decontaminate soil or
sediment containing only a few hundred parts per million PCBs is more efficient
than burning many tons of contaminated soil or sediment, especially if the
decontaminated soil or sediment can be redisposed on site.
This paper discusses the use of solvent extraction to decontaminate PCS
contaminated sediment from the New Bedford Harbor. This work was done under
the Superfund Innovative Technology Evaluation (SITE) Program. After describing
the CF Systems process and the contamination in New Bedford Harbor, this paper
will describe the tests conducted under the SITE demonstration, the results
obtained, and the anticipated costs of operation.
-------
PROCESS DESCRIPTION
The CF Systems Organic Extraction Process is a trailer-mounted, pilot-
scale unit designed to treat 20 barrels of contaminated sediments per day. A
schematic diagram of the process is presented in Figure 1. CF Systems refers
to this test unit as the PCU-20. Slurried sediments from the feed kettle (FK)
are pumped through a 1/8". mesh screen to remove oversized particles prior to
entering the first stage extractor (El). The El agitator (not shown) mixes the
sediments with liquefied hydrocarbon gasses (70% propane/ 30% n-butane, by
weight) from the second decanter (D2). Mixing increases the liquid-solid contact
surface area and enhances the solvent's ability to extract organics from the
sediment. The solyent-organics-sediment mixture flows (by pressure difference)
from El to the first decanter (Dl), where the mixture forms two immiscible
layers; solvent-organics on top and sediments at the bottom. The solvent-
organics overflow from Dl passes through a fine-mesh paper filter (Fl) in order
to remove entrained particulates and enters the solvent recovery column (SRC).
The sediment flows from Dl to the second stage extractor (E2) and is mixed with
freshly recycled solvent to further extract organic pollutants, The solvent-
organics-sediment mixture passes from E2 to D2 and separates as per Dl. The
treated sediments flow from D2 to the raffinate product tank (RPT). The RPT
holds treated sediment for subsequent discharge from the PCU-20 system. The
solvent-organics mixture passes from D2 to El,
The pressure in the SRC is low relative to the extraction subsystem. When
the solvent-organics stream enters the SRC, the solvent flashes into a vapor
state and rises to the top of the SRC. The organics, still in a liquid state,
collect at the bottom of the SRC in the column reboiler (CR) along with a small
amount of unflashed solvent. As the CR fills, the organics flow to the extract
product tank (EPT). the EPT holds the organics for subsequent discharge from
the PCU-20 system. The vaporized solvent leaves the SRC and combines with the
residual solvent vapors, scavenged from the RPT and EPT and compressed by the
secondary compressor (C2). The combined solvent vapor stream feeds the main
compressor (Cl) and is compressed to a liquid state. The hot, liquefied solvent
leaves Cl and passes through a shell and tube heat exchanger located at the
column reboiler. The heat exchanger transfers heat of compression from the
recycled solvent to the liquid organics to boil off any remaining unfl.ashed
solvent. The recycled solvent condenses, leaves the CR heat exchanger and
returns to the extraction subsystem via E2.
SITE DESCRIPTION AND TEST ACTIVITY
New Bedford Harbor has been contaminated for a number of years as a result
of industrial activity that has taken place near the harbor in New Bedford, Mass.
In 1982, the Harbor, was listed on the National Priority List (NPL) due to PCB
and toxic metal contamination. PCB contamination ranges from less than 50 ppm
-------
00
Feed Kettle
Sediments
Screen
E2
M
(I)
S-l
D2
El
Extractor
Dl
DecanterV/
Decanter ^^
Processed Sediments
Solvent I Solvent
Raftinate
Product
Tank
RPT
EPT
.1
Tank
i r Extracted
Processed Organ ics
Sediments
Compressor
C-2 C-l
Solvent
Fl
ract
iuci
k
tra<
rgan
:ted
ics /
V
SRC
CR
Solvent
Recover
Column
c
Column Reboiler
Propane Bullet
Solvent
•*>To system, as needed
Figure 1. CF Systems PCU-20 simplified process flow diagram
-------
to more than 30,000 ppm, with the majority of the site containing less than 50
ppm PCBs.(l) Roughly 20% of the volume of the contaminated sediment onsite,
however, contains between 50 and 500 ppm PCB contamination. Other organic
chemicals are present as well. Among these are naphthalene, acenaphthylene,
dibenzofuran, fluoranthene, and other polynuclear aromatic hydrocarbons. Metal
contamination is also present onsite, although not in high concentrations. Among
the metals present at greater than 1 ppm in the sediment are aluminum, calcium,
chromium, iron, manganese, potassium, sodium and zinc.(l) The untreated sediment
passes the EP Toxicity Test.
In an effort to look at new and better remediation alternatives, the Army
Corps of Engineers, who have been cleaning up the site for EPA Region I, decided
to conduct a pilot study of the CF Systems process. In a joint effort with the
EPA, the pilot study was conducted as a SITE demonstration.
Sediments was dredged from five New Bedford Harbor locations and stored
in 55-gallon drums. The raw sediments ranged in total PCB (Aroclors 1242 and
1254) concentration from approximately 160 to 26,000 milligrams per kilogram
(mg/kg) on a dry weight basis. In addition, the sediment contained between 30%
and 40% solids, was made up of 37% sand, 41% silt, and 22% clay, and had a
neutral pH.(l)
Raw sediment was screened to remove particles greater than 1/8" which
could damage system valves. Harbor water was then added to produce a pumpable
slurry which could be processed by the PCU-20. The prepared sediment was
blended to provide three drums of low PCB concentration feedstock (nominally
300 mg/kg) and one drum of high PCB concentration feedstock (nominally 5000
rag/kg).• These four drums of sediment were set aside for processing using the
CF Systems Pit Cleanup Unit (PCU) during the demonstration.
Using this material, a series of three tests was conducted using 2
different feed concentrations (of PCBs) and three different residence times,
where residence time is defined as the number of passes through the PCU. Table
1 summarizes the conditions for each run.
-------
TABLE 1 Demonstration Test Conditions
Test No.
1
2
4
5
Feed Cone.
360 ppm
350 ppm
288 ppm
No. Passes
3
10
2575 ppm 6
Toluene* 3
Purpose
Equipment Shakedown
Produce solid residue
containing < 10 ppm PCBs.
Reproduce the first three
passes of Test 2
Produce effluent containing
50-500 ppm PCBs
Decontamination: Produce
effluent containing < 10 ppm
PCBs.
* pure toluene was used to decontaminate the PCU-20
The feed and effluent streams were sampled after each pass. A feed-to-solvent
ratio of 1.5 was maintained throughout the test. These test conditions were
chosen on the basis of earlier bench scale tests conducted by CF Systems which
estimated that the PCU-20 could produce a residue containing 17 ppm PCBs from
a feedstream originally contaminated with 210 ppm PCBs using 10 passes and the
above mentioned feed-to-solvent ratio.
RESULTS
Extraction efficiencies were at least 72% for all 3 of the test runs. In
Test 4, the feedstream containing 2575 ppm PCBs was reduced by 61% to 1000 ppm
in 3 passes. Additional passes appeared to continue to reduce the PCB content.
After six passes the total reduction in PCB concentration was 92%. Table 2
summarizes the results of the 3 test runs.
TABLE 2 Demonstration Test Results (1)
Test No.
1
2
3
4
5
No.of Passes
N/A
10
3
6
3
PCB
Feed Cone.
360
350
288
2575
%Reduct1on
Shakedown
89%
72%
92%
Decontamination
10
-------
Table 3 shows the pass by pass reduction efficiency achieved.
TABLE 3 Pass by Pass PCB Reduction Efficiency (1)
Test No.
2
2
2
2
2
2
2
2
2
2
2
3
3
3
3
4
4
4
4
4
4
4
Pass No.
Feed
1
2
3
4
5
6
7
8
9
10
Feed
1
2
3
Feed
1
2
3
4
5
6
PCB Cone.
351 ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
77
52
20
66
59
41
36
29
8
40 ppm
288 ppm
47 ppm
72 ppm
82 ppm
2575ppm
lOOOppm
990 ppm
670 ppm
325 ppm
240 ppm
200 ppm
%Reduction bv Pass
N/A
78%
32%
62%
N/A
11%
31%
12%
19%
72%
N/A
N/A
N/A
N/A
N/A
N/A
61%
1%
32%
52%
26%
17%
Figure 2 plots extraction efficiency as a function of feed concentration.
Two observations can be made from the above data. First, it is difficult
to consistently achieve low effluent concentrations if the PCU-20 has not been
thoroughly decontaminated prior to the start of testing. For example, if the test
unit is decontaminated by flushing with toluene until the effluent concentration
is below 50 ppm, it may be difficult to show much extraction if the feed
concentration is at or below 50 ppm. This is because previously extracted
organic material .is trapped within the system and can result in cross
contamination from test to test.
During the demonstration it was noted that the internal system plumbing
was coated with oily residue. Since PCBs and other organic contaminants are
readily soluble in oily material, this could explain the cross contamination
that was observed. This cross contamination may have made it impossible to show
a pass-to-pass reduction for most of test 3 and some of tests 2 and 4.
11
-------
6 360
a,
a
~"I 300 -
e
o
"Z 260 -
<8
c 200 -
Test 2 PCB Reduction
o
O
m
o
a.
e
a
o
2
160 -
100 -
60 -
4 e
Extraction Pass No.
Test 3 PCB Reduction
10
E
a
a.
c
o
a
c
o
o
O
m
o
a.
e
to
o
2?
300 -
280 -
260 -
240 -
220 -
200 -
180 -
160 -
140 -
120 -
100 -
60 -
*"
4 ' '
A
* '" ',
i i i i i I i I i i
4 6
Extraction Pass No.
Test 4 PCB Reduction
10
e
Q.
0.
*"*
C
O
Z
2
c
u
c
o
O
O
O
D.
C
fS
»«-
2600 -
2400 -
2200 -
2000 -
1800 -
1600 -
1400 -
1200 -
1000 -
800 -
600 -
400 -
200 -
k.
A— -A..
A
A-~~A
~±
n i i i i t ! * i i i
4 6
Extraction Pass No.
10
Figure 2. Pass-by-pass extraction efficiency
12
-------
It is reasonable to assume that some PCBs washed from the extract adhered
to tank and pipe walls. PCBs are soluble in oils, and the amount of oil that
can adhere to internal hardware could be significant. For example, assume (1)
a wetted hardware surface area of 10 square meters, (2) a coating thickness of
1 millimeter, and (3) an oil density of 0.8 grams per cubic centimeter. This
is equivalent to 8000 grams of oil accumulation. In this limited throughput
demonstration, there were approximately 20,000 grams of oil contained in all four
feed drums combined. The demonstration should have included another test using
many drums of oil-rich sediments to determine how much extracted oil is required
to establish an equilibrated flow through the extract path.
Second, pass-to-pass reduction in PCB contamination is very difficult to
measure for two reasons. One, material does not pass through the system
uniformly. This is shown most dramatically in Test 4, At the end of pass 2,
the concentration of PCBs had only dropped by 10 ppm, from 1000 ppm to 990 ppm.
Either the material was not extracted by the system or excess PCBs, which had
been retained in the system from earlier test runs, were emitted during this pass
resulting in an apparently low extraction efficiency for this test. Material
balance calculations suggest .that the system irregularly retained and discharged
treated sediments For some passes, as much as 50% of the feed may have been
retained in the system, adhering to internal piping and tank surfaces. The
sporadic nature of solids accumulation and discharge was not anticipated. If
the PCU-20 had modular "add on" extraction stages and was operated in a pass-
through mode, solids retention and cross-contamination would be lesser concerns.
However, the PCU-20 has only two extraction stages, and recycling is the only
way to mimic the effect of a larger number of stages. In future tests, the unit
should be partially dismantled at the demonstration's conclusion and inspected
in order to reconcile the total solids balance. Larger versions of the PCU-20
have been built and operated by CF Systems. These units do not require that the
feed be recycled.
Two, variations in the PCB concentration in the feed contributed to
variations in the extraction efficiency. The normal variation in harbor sediment
PCB concentration, combined with the variability inherent in the analytical
methods used may result in wide variations in the reported extraction efficiency.
Despite these pass-to-pass variations, however, the overall PCB reductions
were, as mentioned earlier, high for each test run.
The metals present in the original feedstream remained in the solid
residuals after processing.(1) This was expected since this process does not
remove metallic contamination. As with the feedstream, this material passed
the EP Toxicity Test.
Two other problems occurred during the demonstration. First, dissolved
propane caused foaming in the treated sediment product tanks. This caused minor
operational problems and can be alleviated with further operational changes.
13
-------
Second, the volume of waste generated by this process exceeded the volume of
material originally treated. While only four drums of waste were treated during
the demonstration, 57 drums of process residue were produced. The contents of
the residues is summarized in Table 4.
TABLE 4 Residues Produced from the Demonstration (1)
Numberof Drums Drum Contents
6 Toluene (unit decontamination residue)
6 Toluene rinsewater
2 Naphtha-based fuel product and unit residue
15 Sediments
8 Sediments and rinsewater
20 Decontamination water
57 TOTAL
Thirty drums of used Tyvek suits and decontamination debris were also
generated as a result of the onsite activities of EPA's evaluation contractor.
Although this quantity of waste is unlikely to be generated during normal field
operations, it is important to minimize waste generation when using this process.
Thirty-two of the 57 drums of residue contained materials related to process
decontamination. Refined decontamination procedures that minimize the use of
toluene during this step in the process may reduce the volume of toluene and
toluene contaminated wastewater produced from this procedure.
Finally, 15 drums of sediment were produced. Under routine operation,
this sediment would be redisposed onsite and would not need further treatment.
Costs for this process, as estimated by CF Systems, ranged from $150/ton
to S450/ton depending upon whether a high volume, low contamination (base case)
waste was being treated ($150/ton) or a low volume high contamination (hot spot)
waste was being treated ($450/ton).(2) Cost analyses for these two cases were
prepared by CF Systems and are reproduced below in Tables 5 and 6. The unit
costs for treatment presented above do not include the cost of disposing of the
concentrated matrix into which the PCBs were extracted. Presumably, this
material will be relatively low volume and so will not add greatly to the cost
of remediating the site using this technology.
This cost analysis is presented to indicate which factors most affect cost
and is not intended to determine a cost for the use of this process at all sites.
As determined by CF Systems economic model, the costs associated with operating
this process were affected by several factors. These are as follows.
14
-------
1. The on-stream factor. Fluctuations in this variable significantly affected
costs. A decrease in on-stream factor from 85% to 70% increased costs by 20 %.
2. Waste Pretreatment. Elimination of the waste pretreatment step to decrease
the solids content can result in a 30% cost savings. Therefore, if the waste
is already a pumpable slurry to which no additional water need be added, using
this process will be less expensive. This savings occurs as a result of reduced
volumetric throughput, reduced equipment sizes and elimination of some pre-and
post-treatment steps. Eliminating the need to dilute the waste feed reduces cost
more than any other variable in the economic model.
3. Extraction Unit Costs. Costs specific to the extraction unit account for
53% to 68% of total remediation costs using this process.
4. Sediment Excavation andPre and Post-Treatment Costs. These costs account
for 28% to 41% of the total remediation costs.
CONCLUSIONS
Several conclusions can be drawn about the PCU-20 as a result of this
demonstration. These are as follows.
1. The PCU-20 was capable of removing at least 70% of the PCB contamination
in the New Bedford Harbor sediments treated during the demonstration.
2, System decontamination indicates that significant amounts of PCBs may
have coated the interior of the process. The resulting cross contamination
of the (recycled) sediments may have changed the effluent concentration.
3. The amount of water that must be added to the feed to produce a pumpable
slurry affects the cost of the process and the amount of residues
generated.
The Demonstration Report and the Applications Analysis Report for the
demonstration of this process will be available in June from the USEPA's Center
For Environmental Research Information in Cincinnati, Ohio. For further
information on this process, please contact:
Laurel J. Staley
Risk Reduction Engineering Laboratory
U.S. Environmental Protection Agency
26 W. Martin Luther King Dr.
Cincinnati, Ohio 45268
15
-------
TABLE 5 Base Case and Hot Spot Case Summary
Capacity Base Case HotSpot
Raw sludge (40% solids): cubic yards 695,000 50,000
tons 880,000 63,000
Processing Time: years 8.5 1.19
Operating Days 2,591 369
Raw Sludge Feed Rate (@ 40% solids)
tons/ operating day 339.5 171.5
Extractor Feed: % Solids 26.7 26.7
total tons processed 1,319,414 94,922
nominal system size (tons/day) 500 250
feed rate (tons/operating day) 509.2 257
Inlet PCB Concentration: ppm 580 10,000
Outlet PCB Concentration: ppm 50 10
PCB Reduction: % 91 99.9
Configuration * 1 2
Processing Fee (1989 $)
Facilities $5,170,676 $ 762,496
Extraction $62,109,781 $15,857,695
Pre/Post Treatment $46,172,028 $7,993,608
Contingency $11,345,248 $2,461,380
Project Management $5,672,624 $1,230,690
TOTAL $130,470,358 $28,305,869
Total Life Cycle Unit Cost ($/ton):
Extraction Only $ 71 $251
Total $148 $447
NOTES
Configuration: 1- Two extraction sections connected in parallel
feeding one solvent recovery section connected in
series.
16
-------
2- An extraction and solvent recovery section in
series connected in parallel with a second identical
extraction and solvent recovery section.
Base Case = High volume of waste /Low contamination level
Hot Spot = Low volume of waste/ High contamination level
TABLE 6 Estimated Cost (2)
1A IB 1C
ID
Case Base Case
Description (1)
Waste Volume (Tons) 880,000
PCB Reduction (%) 91
Solids Content (%) 27
On-Stream Factor (%) 85
Remediation Time (weeks) 434
Base Case
Reduced
On-Stream
Factor
880,000
91 .
27
70
527
Base Case
No Solids
Content
Reduction
880,000
91
27
85
280
Base Case Hot
Increased Spot
PCB Removal
Efficiency
880,000 63,000
98 99.9
27 27
85 85
347 64
Estimated Cost, $/Ton
Site Preparation
Extraction Unit
Pre/Post Treatment
Excavation
Equipment
Extraction .Unit
Pre/Post Treatment
Startup and Fixed Costs
Labor
Extraction Unit
Pre/Post Treatment
Supplies & Consummates
Extract. Unit Utilities
Pre/Post Trtrat. Utilit.
Analytical
TOTALS, $/Ton
3,02
1.95
21.44
48.39
23.86
6.76
10.72
10.80
17.06
2.29
1.98
148.27
2.96
1.95
26.03
58.52
28.98
8.21
13.02
13.11
19.51
2.78
2.41
177.48
3.02
1.95
13.97
31.53
15.50
4.39
6.97
7,01.
11.08
1.48
1.29
98.19
5.86
1.58
17.14
77.81
19.08
5.40
10.86
9.01
23.91
.1.83
1.59
174.09
47.53
23.57
43.94
173.57
48.91
13.73
33.68
24.09
29.20
4.68
4.07
446.97
17
-------
Notes: 1) These estimates are only intended for use in planning, scoping,
and the inviting of firm bids. The American Association
of Cost Engineers has established an accuracy goal of +50% to
-30% for preliminary estimates such as these.
3) The costs shown are based on a proprietary model developed
by CF Systems, Inc. Cost model outputs are presented in
Appendix B for the Base Case and the Hotspot Case.
REFERENCES
1. Technology Evaluation Report: CF Systems Organic Extraction System, New
Bedford, Massachusetts. EPA/540/5-90/002 January 1990 Cincinnati, Ohio
2. Applications Analysis Report: CF Systems Organic Extraction System, New
Bedford, Massachusetts. EPA 540/A5-90/002.
18
-------
Summary Results of the SITE Demonstration for the
CHEMPIX Solidification/Stabilization Process
by: Edwin F. Barth, P.E.
Center for Environmental Research Information
United States Environmental Protection Agency
Cincinnati, OH 45263
ABSTRACT
A demonstration of the CHEMFIX solidification/stabilization process was
conducted under the United States Environmental Protection Agency's (U.S. EPA)
Superfund Innovative Technology Evaluation (SITE) program. The demonstration
was conducted in March 1989, at the Portable Equipment Salvage Company (PESC)
uncontrolled hazardous waste site in Clackamas, Oregon. Wastes containing
lead, copper, and polychlorinated biphenyls (PCBs) from four different areas
of the site were treated. Results showed substantial reduction of leachable
lead and copper between the untreated waste and treated waste utilizing
the Toxicity Characteristics Leaching Procedure (TCLP) test. Long term TCLP
results were different than initial results. The effectiveness of this process
for immobilizing PCBs could not be determined since the raw waste did not
leach PCBs at high concentrations, utilizing the TCLP test. Data from other
leaching tests for lead and copper would need to be utilized as input into a
site specific ground water model to determine whether solidification/-
Stabilization would be an acceptable remedy for the site. Physical testing
results indicated durability in exposed conditions. Valuable lessons were
learned which have been useful in subsequent demonstrations.
Description of the CHEMFIX Process
The CHEMFIX process is a patented solidification/stabilization
(immobilization) process utilized for the treatment of liquids, sludges,
soils, and ashes containing heavy metals and organics. Soluble silicate
reagents are normally added to the waste of concern. Three classes of
reactions may
:
Soluble silicates react with cations in the waste matrix to form
immobile silicates?
- The silicious setting agents react with the remaining soluble
silicates to produce a gel structure;
- Hydrolysis, hydration, and neutralization reactions also occur
to further stabilize the waste.
t
The presence of certain organic constituents may necessitate the use of other
additives to the process.
19
-------
Contaminated soils (screened to minus one inch) were excavated with a
backhoe and delivered to an even feeder hopper. The soil was transported to
an electrically monitored weigh feeder via a conveyor. The waste was then
introduced into a series of mixing equipment apparatus (homogenizer followed
by a pug.mill) where liquid reagent and dry reagent were introduced. Reagent
addition was controlled electronically. Additional make-up water was not
always necessary, since the soils from the site were wet from the rainy
conditions encountered. Treated wastes are normally transported to an on-site
disposal area or transported off-site. CHEMFIX initially rated the capacity
of the system utilized at this site for the demonstration at 100 tons per hour
(tph).
Demonstration Objectives
A Data Quality Objectives (DQO) program was utilized to define the
objectives of the CHEMFIX process demonstration before site selection. Based
upon performance claims and previous treatment data submitted by CHEMFIX, the
following demonstration objectives were established:
* Evaluate the ability of the CHEMFIX process to meet or be below
land disposal banning levels for heavy metals waste (specifically lead)
established by the U.S. EPA;
* Determine the effectiveness of the CH1MFIX process (based upon percent
reduction) to reduce the mobility of heavy metals (specifically lead and
copper) and polychlorinated biphenyls (PCBs) after treatment utilizing the
TCLP test?
* Determine physical properties of waste treated by the CHEMFIX process
for reducing leaching potential and long-term durability indication.
Site Description
The PESC site was selected as a demonstration site because the wastes
present on the site were suitable for evaluating the demonstration objectives
described above. The PESC site operated as a transformer and metals salvage
facility from the early 1960*s to 1985. Operations at the site involved
scrapping and recycling power transformers containing PCBs in cooling oils.
Salvageable metals from internal wiring and transformer carcasses were
processed and recycled. Transformers and other recycled electrical equipment
were burned in a furnace to eliminate insulation and other non-economic
elements. Waste transformer oil was used to fire furnaces and metal smelters
at the site.
Pre-demonstration sampling activities were geared at isolating waste areas
that were sufficiently different in soil type or contaminant concentration to
operate the process over a range of characteristics. This approach was aimed
at defining the limits of the process for the site waste. Four selected areas
were identified as having different characteristics and are described in Table
1. The data showed ranges of contaminant concentrations and a large
percentage of debris, which are common on uncontrolled hazardous waste sites.
One of the areas contained an ash material as opposed to soil. It should be
noted that some lead concentration values encountered were extremely high (up
to 139,000 mg/kg).
20
-------
TABLE 1, Description of Waste Areas Evaluated (pre-demonstration sampling)
During the CHEMFIX Site Demonstration
Area Description
A Soil - High lead and copper concentrations (10,000 - 139,000 mg/kg)
Medium to high PCB concentrations (100 - 1,940 mg/kg)
C Soil - High lead and copper concentrations (up to 117,000 mg/kg)
High PCB concentrations (up to 1,350 mg/kg)
f
E Soil - High lead and copper concentrations (up to 110,000 mg/kg)
Low PCB concentrations (<100 mg/kg)
F Ash - High lead and copper concentrations (40,000 - 136,000 mg/kg)
Medium PCB concentrations (200 - 300 mg/kg)
Demonstration Program
Ten cubic yards of soil from each of areas A, C, and E and ten cubic yards
of ash from area F were individually processed for this demonstration. The
material was excavated, screened to minus one inch, and mixed on the ground by
a backhoe to better homogenize the waste before treatment. Only a total of
forty cubic yards of material were processed to minimize the amount of treated
material that would be generated on-site. Process mixing performance was
expected to be negatively impacted by processing such limited quantities of
waste in a high capacity system, because calibration time would be limited.
Area C soils were selected as the waste type where the majority of the
physical and chemical testing methods would be performed.
The screened material was visually estimated to be about 30% of the total
volume of material excavated. The quantities of all materials utilized were
documented for mass balance purposes.
Sampling and Analysis Program
Raw soil, samples were obtained the day of the demonstration and treated
soil samples were taken immediately after processing. Leaching test samples
were formed in plastic or cardboard molding tubes to eliminate destruction of
samples or perceived interferences of leaching results from coring operations.
Table 2 describes the leaching tests utilized in the demonstration.
21
-------
TABLE 2. Leaching Test Utilized in
the CHEMFIX SITE Demonstration
Name of Test
TCLP
Reference
40 CFR Part 268
Description
Ground material subject to 18 hour
extraction process with acetic acid
leachant to simulate co-disposal
environment with municipal waste
MEP
ANS 16.1
EPA method 1320
(SW 846)
American Nuclear
Society 16.1
Ground material subject to 24 hour
extraction with acetic acid leachant
followed by nine sequential extractions
with acidic rain simulated leachant
Monolithic material placed in distilled
water that is replaced over discrete
time intervals (diffusion model)
Baseline physical and chemical tests were performed on the raw waste for
comparison to treated waste samples. Table 3 lists the physical and chemical
teats utilized in the demonstration. An air monitoring system with
polyurethane foam (POF) was employed to determine if PCBs were volatilized
during the mixing or curing process.
TABLE 3. Physical and Chemical Test Utilized
in the CHEMFIX SITE Demonstration
Name of Test
Unconfined Compress-
ive Strength (UCS)
Hydraulic
Conductivity
Wet/Dry Resistance
Freeze/Thaw Resist-
ance
Ox idat ion/Reduction
Electrical
Conductivity
Reference
ASTM D1633
1PA draft protocol
ASTM D 4843
ASTM D 4842
EPA method 9045
(modified)
ASA 10. - 3.3
Description
Used to assess structural
integrity of monolith
Used to assess resistance
of material to water flow
Indication of durability in
wet/dry environment
Indication of durability in
freeze/thaw environment
Determine oxidation/reduction
state of waste matrix
Determine amount of ionic
materials present in solution
22
-------
Demonstration Results
Leaching Testing
Two values were established as the criterion for determining if the
treated waste met the U.S. EPA land ban criteria for lead. The first value
was .51 mg/L (TCLP leachate) based on the U.S. EPA standard for listed (F006
sludges) hazardous waste(2). A second, higher level of 5.0 mg/L was
arbitrarily chosen to recognize that soils may be more difficult to treat than
sludges. 65% of all samples tested from areas A, C, E, and F passed the 0.51
mg/L standard and 70% passed the higher standard of 5.0 mg/L.
Substantial reductions in the leachability of lead and copper in the raw
waste, as determined by the TCLP, were observed. Table 4 shows that lead
reductions ranged from 94% in area E to 99% in areas A, C, and F. Copper
reductions ranged from 95% in area C to 99% in areas A, E, and F. Data from
these TCLP results need to be kept in the limited perspective that this test
simulates a leaching environment involving co-disposal with municipal waste.
This scenario may not have applications for the PESC site.
TABLE 4. Mean Concentrations of Lead and Copper
from the CHEMFIX Demonstration
Area A
Lead
Copper
Area C
Lead
Copper
Area E
Lead
Copper
Area F
Lead
Copper
Untreated
Waste
(total)
21,000 mg/kg
18,000 mg/kg
140,000 mg/kg
18,000 mg/kg
92,000 mg/kg
74,000 mg/kg
11,000 mg/kg
33,000 mg/kg
TCLP From
Untreated
Waste
610 mg/L
45 mg/L
880 mg/L
12 mg/L
740 mg/L
120 mg/L
390 mg/L
120 mg/L
TCLP From
Treated
Waste
<.05 n»g/L
0.57 mg/L
2.5 mg/L
0.54 mg/L
47 mg/L
0.65 mg/L
0.10 mg/L
0.60 mg/L
Percent
Reduction
Of TCLP
Extractable
Metal
99
99
99
95
94
99
99
99
The reduction in mobility of PCBs, based on the TCLP, could not be
determined since the PCBs essentially did not leach in the raw waste. More
stringent leaching or extraction tests would be necessary to determine the
effectiveness of this process for stabilizing PCBs.
23
-------
Data from the MEP test were non-conclusive. In many instances, a
criterion of 5.0 mg/L for lead was exceeded during the first extraction, but
was not exceeded during the nine subsequent leaching extractions.
Data from the ANS 16.1 test for lead and copper are presented in Table 5.
Leaching data are calculated into a leachability index (LI) which is the
negative logarithm of the effective diffusivity coefficient. The treated
material successfully exceeded the Nuclear Regulatory Commission (NRC)
criterion for LI of six by several orders of magnitude^3'. However, the
acceptability of these results, based on protection of the public health and
environment, should be judged only after leaching values (fluxes) are
incorporated into a site specific ground water model.
TABLE 5. ANS 16.1 Leachability Index (LI) Data from the
CHEMFIX SITE Demonstration
Contaminant LI
Lead 13.2
Copper 15.2
Chemical Testing
Data from chemical testing between the raw and treated material can be
seen in Table 6. Oxidation/reduction potential of the treated waste was less
than the raw waste. Conductivity of the treated waste was considerably higher
than the raw waste, indicating that ions are leaching from the treated
material. However, loss of ions into solution may not be of concern,
depending upon the chemical nature of the ions. The pH of the raw waste, 6.6,
increased to 11.5 after reagent addition. The pH of the treated material was
not impacted to a large degree by the acetic acid leachant of the TCLP.
The air monitoring data suggested that there was no significant
volatilization of PCBs during the treatment operations at the site. It should
be noted that the wet, cool temperature environment encountered during the
demonstration would not promote volatilization.
Physical Testing
The results of the physical test performed on the treated wastes are shown
in Table 6. Weight loss during wet/dry and freeze/thaw cycle testing was less
than one percent. These data indicate durability in an exposed environment.
Unconfined compressive strength (UCS) values ranged from 27 pounds per square
inch (psi) to 307 psi. The U.S. EPA guidance value for solidified/stabilized
waste for UCS is 50 psi^. Hydraulic conductivity of the treated material
was in the range of 1 x 10~° to 1 x 10"7 cm/sec. Acceptable hydraulic
conductivity values should be compared to in-situ permeability measurements,
which were not obtained at the site. The volume increase in the waste
excavated material after treatment ranged from 20 to 50 percent.
24
-------
TABLE 6. Physical and Chemical Properties of
Treated and Untreated Wastes from
Area C of CHEMFIX Demonstration
Area C
Untreated
Wastes
Treated
Wastes
Sh (millivolts)
Conductivity (umhos/cm)
pH
28-day UCS (psi)
Wet/dry stress weight loss
290
130
6.6
N/A
N/A
24
3200
11.3-11.5
27-307
Freeze/thaw stress weight loss
Permeability (cm/sec)
N/A Not Applicable
N/\ A
10"4 to 10"6
< 7
10"6 to 10~7
Lone? Term Leaching Testing Results
Treated samples from area C were subjected to the TCLP test and the ANS
16.1 test after six months and one year. Leaching values for the TCLP test
progressively increase over time as can be seen in Table 7. The author feels
however, that surface carbonation reactions may occur on these samples over
time and possibly interfere with the true leaching results. If however, the
leaching results are true/ regulatory agencies may want to consider utilizing
longer cured samples in treatment evaluations. Leachability index results were
consistent from the 28 day sample through the one year sample, indicating no
change in physical or chemical behavior.
In an effort to reduce uncertainty of the long term leaching of stabilized
materials, the SITE program has established long term testing apparatus to
expose the stabilized material to extreme weathering conditions over time.
Laser holography, X ray diffraction and acoustic vibration are some of the
testing proceedures that weathered samples will encounter.
TABLE 7. LONG TERM LEACHING RESULTS FOR LEAD
FOR AREA C
SAMPLE
Raw Soil
28 day cured
6 month cured
1 year cured
TCLP MEAN CONC.
(mg/l)
880.0
2.5
14.0
24.0
LEACHABILITY INDEX
NA
13.2
13.7
13.8
N/A - Not Applicable
25
-------
Legspng,
Several potential problem areas for performing large scale demonstrations
of stabilization processes were observed during the demonstration planning and
implementation. The following is a brief summary of general problem areas s
* Debris such as metal wire or broken bricks were encountered and needed
to be physically separated from the contaminated soil to eliminate leaching
test interferences
* PCBs were not leachable in the raw waste utilizing the TCLP test,
making treatment performance efficiencies difficult to calculate
* The demonstration occurred at a lower processing rate than the
equipment was designed for, therefore accurate calibration of the mixing
process during the short processing time was not possible
* Chemical concentration results from preliminary sampling at the site
varied from the actual sampling results during the demonstration, negating
optimization of the binding mixture utilized
* Testing methods for determining mixing efficiency are not well
developed
Conclusions
The majority of the treated waste samples from the CHEMFIX
solidification/stabilization process met land ban standards criteria
established for the demonstration. Reductions of leachable lead, judged by
the TCLP test, ranged from 94 to 99%. Data from the ANS 16.1 test
successfully exceeded the NRG criterion of six by several orders of magnitude,
but these data need to be incorporated into a site specific ground water model
before the performance of the process can be judged for this site.
The CHEMFIX process generally produced treated material with acceptable
physical properties. The treated material had properties that indicated long-
term durability in exposed environments. Volume increase in the excavated
material after treatment ranged between 20 to 50%. Valuable lessons were
learned during the preparation and implementation of this demonstration that
have minimized problems on subsequent demonstrations.
References.
(1) CHSMFIX Technology Description Proposal, submitted to U.S. EPA for
SITE 002 Proposal Evaluation (1987). CHEMFIX Technologies Inc.
2424 Edenborn Ave. Metairie, Louisiana 70001
(2) Land Disposal Restrictions for First Third Scheduled Waste,
Federal Register, Volume 53 No. 159, August 17, 1988.
(3) Technical Position on Waste Form, Nuclear Regulatory Commission (1983).
(4) Prohibition on the Placement of Bulk Liquid Hazardous Waste in Landfills,
OSWER Policy Directive No. 9487.00-2A, EPA/530-SW-86-016 (1986).
26
-------
ftckowledqementB
The author would like to acknowledge the following people in implementing
this demonstrationJ Mark Evans, Nancy Willis, and Shin Ahn of PRC Inc., Danny
Jackson and Debra Bisson of Radian Corporation, Guy Simes and Dean Neptune of
EPA's Office of Research and Development, and Paul Lo and Phil Baldwin of
CHEMFIX Technologies Inc.
27
-------
SOIL VAPOR BJTRACTIOM; AIR PERMEABILITY
TESTING AND ESTIMATION METHODS
by: Katharine L. Sellers
Tom A. Pedersen
Camp Dresser & McKee Inc.
Cambridge, MA 02142
Chi-Yuan Fan, P.E.
U.S. EPA/RREL/RCB
Edison, NJ 08837
ABSTRACT
Soil-air permeability is a critical parameter used to assess the
feasibility of soil vapor extraction (SVE) technology for sites where
volatile organic compounds are present in the vadose zone. Field,
laboratory, and empirical correlation methods for estimating soil-air
permeability have been reviewed for their appropriateness in determining SVE
feasibility and for the development of SVE system design criteria.
Empirical methods are available to derive estimates of soil air
permeabilities from soil grain size distributions, hydraulic conductivity
measurements, or pump test drawdown data. Although these techniques provide
data which serve to determine if the use of SVE should be excluded from
further consideration, they do not provide adequate data for system design
criteria development. Laboratory soil-air permeability tests are -also
inappropriate for SVE system design because of the variability in soil field
permeability and the non-representative nature of soil cores collected in
the field. Field techniques employed for determining soil-air permeability
in surficial soils are likewise inappropriate for the evaluation of
contaminant releases that have migrated to depths of greater than one meter.
The in situ field borehole permeability techniques used by petroleum
engineers, and subsequently modified for use at relatively shallow soil
depths, hold the most promise for application to SVE design.
Although most SVE vendors and contractors use some type of soil
permeability estimation or measurement technique, description of the use of
these techniques in conjunction with SVE system design is often of a
proprietary nature. Additionally, no standard technique for uniform
measurement or reporting of soil-air permeability is available. Shell
Development Co. and Shell Oil Co. (Shell) have developed a practical
approach to the design, operation, monitoring of SVE, which has been
detailed in a paper by Johnson, Stanley, Kemblowski, Byers, and Colthart (1,
2). The determination of soil-air permeability is a critical component of
28
-------
this approach, and a procedure is presented for the calculation of soil-air
permeability. The Shell method also provides a basis for refining the
soil-air permeability measurement techniques used in the remediation
industry. However, successful implementation of the methodology as a
comparative site evaluation tool for SVE system design could benefit from
the development of standardized guidelines for the construction of vapor
extraction wells and monitoring veils.
This paper has been reviewed in accordance with the U.S. Environmental
Protection Agency's peer and administrative review policies and is approved
for presentation and publication.
INTRODUCTION
Soil vapor extraction (SVE) has become a commonly used technology for
the remediation of soils contaminated with volatile contaminants. Critical
to the application of SVE technology at a particular site is the ability to
achieve adequate vapor flow through the contaminated zone. Soil-air
permeability describes how easily air will flow through the soil and is
expressed in units of length squared (cm2). Vapor flow rates through porous
media, such as soil, are dependent upon soil characteristics including
porosity and permeability, as well as gas properties such as viscosity,
density, and pressure gradients. Gas is a fluid and as such its flow rate
through porous media is commonly characterized by Darcy's law. Darcy's lav
is valid for laminar, isothermal flow that is uniformly distributed across a
given cross sectional area. The general formulation of Darcy's law for
saturated fluid flow in one dimension is (2):
q » (kA/p)(dP/dm) (1)
where:
3
q = flow rate (cm /sec)
k = permeability (cm ) „
A » cross-sectional area (cm )
u = viscosity (g/cm-sec) „
dP/dm = pressure gradient ((g/cm-sec )/cm)
Since air permeability vill control the decision to implement SVE
technology at a contaminated site to a large degree, the importance of the
air permeability measurement or estimation technique is evident. To date,
no standard method has been advanced for determining soil-air permeabilities
for sites at which SVE techniques might be applied. In fact, many SVE
technology vendors utilize proprietary techniques and methods for estimating
cleanup times and establishing system design criteria. The conflicting
methods and claims pose problems to regulators required to make judgements
on the appropriateness of a proposed SVE system design. To provide some
guidance with regard to air permeability test methods, a review of available
techniques was undertaken. This paper provides an initial assessment of the
29
-------
applicability of various air permeability determination methods. Soil-air
permeability estimation and testing methods considered include: correlation
to empirical soil properties such as soil grain size distribution or soil
hydraulic conductivity; laboratory measurements; and in situ field
measurements. Table 1 lists some air permeability test methods and
summarizes the limitations with regards to site evaluation and SVE
feasibility determination.
CORRELATION METHODS
Soil air permeability may be estimated from known physical
characteristics of the soil sample, such as the grain size distribution or
saturated hydraulic conductivity,
Massman (4) discusses the use of a linear correlation between both soil
grain size distribution analyses or saturated hydraulic conductivity and
soil air permeability. These methods do not account for decreases in the
air permability due to increased moisture contents or visa versa, nor do
they account for in situ bulk density, soil structure, or heterogeneity of
the subsurface soils. The variability in permeabilities associated with
differing soil characteristics can be accounted for by sampling throughout
the area where SVE is to be applied and averaging the resulting air
permeabilities.
The use of saturated hydraulic conductivity values to estimate air
permeability is subject to several additional sources of er-ror including
soil moisture content, swelling soils, and gas slippage. First, both air
permeability and saturated hydraulic conductivity are influenced by the soil
water content. Hydraulic conductivity generally increases while air
permeability generally decreases as the water content of a soil increases
(5, 6). The second source of error results from the interaction of soil
particles with water. Theoretically, the intrinsic permeability for water
is only a function of the medium. However, in some cases, water alters the
structure of clay, due to expansion as the water content increases. This
dramatically reduces the intrinsic permeability. Therefore, the correlation
between saturated hydraulic conductivity and air permeability would hot be
valid for soi'ls with an appreciable content of expandable clay. The third
source of error in this method is due to the "slippage" of gas as they pass
along the soil pore wall. This factor commonly known as the "Klinkenberg
effect", accounts for the soil air permeability having a greater value than
the liquid permeability at low gas pressures in fine grain soils. Darcy's
law assumes that the flow velocity at the pore wall is zero. However, this
assumption becomes invalid, and the "Klinkenberg effect" becomes significant
in materials having an intrinsic permeability less than 10 millidarcies
(15).
Field data from aquifer slug tests or drawdown tests can be used to
estimate the soil air conductivity and soil gas storage coefficients using
graphical curve matching techniques analogous to the Theis, Hantush, or
30
-------
Jacobs methods developed for groundwater pump tests. The Theis curve
matching method (7) is commonly used in unsteady well hydraulics. For
unsteady flow applications, Darcy's law is expressed as a partial
differential equation that includes time as a variable. However since these
tests are performed in the saturated zone they do not provide a good
indication of unsaturated zone permeability.
Correlation methods provide a quick means of assessing the relative
permeability of soil but would not be appropriate for SVE system design
criteria development. However, both the grain size and saturated hydraulic
conductivity correlation methods only provide an order of magnitude estimate
LABORATORY METHODS
Laboratory methods used for the determination of air permeability were
pioneered in the petroleum industry (8, 9) and adapted for use in
agriculture (10, 11). Although the petroleum industry designed air
permeability measurement apparatus for relatively undisturbed porous rock
samples, and the agricultural industry designed apparatus for undisturbed
soil samples, the general features of the laboratory tests are similar.
Samples are placed in a pressure vessel (permeameter) and saturated with
water or another wetting fluid and air is injected to force the wetting
fluid out of the sample (desaturation) while air flow rates and air pressure
in the sample are measured. Measurements of porosity, air flow and pressure
differential are used in Darcy's law to estimate the air permeability as a
function of water content.
Laboratory air permeability measurement methods are subject to
significant error since air permeability is sensitive to the bulk density,
and the structure of the soil. These soil characteristics are altered when
a subsurface sample is exhumed and placed in the permeameter. In addition,
laboratory measurements can not account for heterogeneities encountered at a
site that affect the overall soil air permeability of a particular location.
Use of laboratory permeability measurements for SVE system design is not
considered appropriate because of limitations in the methods to account for
spatial variability associated with field conditions.
FIELD METHODS
There are a variety of in situ field methods that have been employed for
the determination of soil air permeability. All of the methods rely on
measuring the difference between the ambient atmospheric pressure and the
soil air pressure in the soil during subsurface vapor transport. The
methods described in this section include: air injection testing; subsurface
31
-------
barometric fluctuation measurements! oil field tests, such pressure build-up
tests, and drawdown tests; and a soil vacuum extraction method (1).
AIR INJECTION
Air injection tests are generally performed on surficial agricultural
soils, utilizing equipment consisting of a compressed air tank and a gas
flow/pressure regulator attached to a cylinder which is inserted into the
soil (12, 13, 14). The pressure differential is measured before and after a
known volume of air is injected through the cylinder into the soil over a
given time period. The soil air permeability is estimated using Darcy's law
with measured and known values for the pressure differential, air flow rate,
cross-sectional area of the cylinder, and viscosity of the air. Although
air injection testing has been applied principally to surficial soils it can
be applied to subsurface investigations.
The advantages of surfical air injection testing are relatively portable
measurement equipment, quick measurements, and low cost. However, air
injection testing may not be appropriate to determine the soil air
permeability needed for SVE implementation or developing SVE system design
criteria. When air is injected the soil particles may tend to disperse and
yield a higher permeability value. Therefore, the permeability measured by
injecting air into soil may not be the same as the permeability measured by
vacuum suction of air from soil resulting in lower values.
BAROMETRIC FLUCTUATIONS
Monitoring changes in barometric pressure can be used to determine the
vertical air permeability by finding an effective "pneumatic diffusivity" at
prevailing soil porosity and moisture content (15). The pneumatic
diffusivity is analogous to the hydraulic diffusivity terms used in
groundwater flow. Barometric pressure changes are measured using specially
designed piezometer wells nested within the unsaturated soil. Manifold
manometer systems are used to measure air pressure at each of the screened
levels. The permeability is estimated using an iterative process that
equates the air flux between adjoining subsurface layers where the
barometric pressure was measured. Weeks (15) provided a computer program
that solves for the vertical air permeability using a trial and error
numerical solution.
One major limitation of applying this method is that the manometer
system can only measure barometric pressure differences during the normal
diurnal barometric pressure change if the unsaturated zone is greater than
20 meters thick and has at least one layer with a permeability less than 2
to 3 darcies. However, permeabilities this low would probably limit the
effectiveness of SVE technology. For thinner, more permeable, unsaturated
soils, barometric pressure differences can only be obtained during the
passage of an atmospheric front in which the pressure changes a few
millibars in less than an hour. Since it is not feasible to predict the
occurrence of atmospheric conditions, and since this test principally
address vertical flow, the use of this technique for SVE is not appropriate.
32
-------
OIL FIELD TEST METHODS
The petroleum industry has used a number of in situ field tests to
determine air permeability. Host of the tests are performed using one or
more potential gas production wells. Production well natural gas differs
from SVE gases, in that natural gas in subsurface geological layers is under
higher pressures and temperatures than the organic vapors encountered in the
vadose zone. Thus, the modeling of natural gas flow and air permeability
estimates of natural gas production reservoirs usually incorporates
parameters for gas compressibility and temperature, not found in the vapor
transport equations used for SVE (1,2). Two commonly used methods of
determining air permeability of gas producing formations are the pressure
buildup test and the drawdown test.
Pressure build-up and drawdown tests can be performed using a single
well or by use of an extraction well and monitoring wells. The physical
relationship between the vapor extraction well and the monitoring well used
in drawdown tests is illustrated in Figure 1.
Pressure Buildup Test
The pressure buildup test is conducted by closing a gas well that has
been producing gas at a constant rate for a given period of time, and
monitoring the down hole pressure increase after the well is closed or
shut-in (16). Gas flow in the well is modeled by Darcy's law with radial
gas flow. The pressure buildup test is considered to be the result of two
superimposed effects? the pressure drawdown caused by the initial gas flow
from the well, and the increase in pressure that occurs when the well is
closed or shut-in. The pressure increase is modeled as a gas injection with
a flow rate equal in magnitude and opposite in sign (or direction) to the
gas flow rate from the well during pressure drawdown. There are a number
pressure time relationships presented in natural gas engineering texts (17,
18, 19).
The pressure build-up tests most likely are not applicable to soil air
permeability testing for SVE technology since confined deposits of vapor are
not usually associated with contaminant releases. The vapor pressures found
in contaminated soils may not create a significant pressure build-up when
the vapor extraction well is shut-in.
Pressure Drawdown Test
The pressure drawdown test is another common oil field method for
determining air permeability (17, 18, 19). In this test, gas is extracted
from a well at a constant flow rate while the pressure reduction in the well
is observed over time. Again, as in the case of the pressure build-up test,
the pressure versus time relationship is modeled using Darcy's law applied
to radial gas flow. Additional factors are included in the modeling
equation to account for the compressibility of the gas, and porosity
variations near the well screen cause by well installation (skin effect).
The pressure buildup test is sometimes preferred in oil field work over a
33
-------
pressure drawdown test, since it is difficult to achieve the constant gas
flow rate out of the well required for a drawdown test (16).
The mechanics and modeling of the drawdown test are similar to those
used to determine soil air permeability proposed by Johnson et al. (1, 2).
The additional gas compressibility and skin effect factors are probably not
significant in determining air permeability of low pressure vapors extracted
from shallow wells.
SVE AIR PERMEABILITY MEASUREMENT METHODS
SHELL'S APPROACH
The soil air permeability test developed by Shell (1, 2) for SVE
technology is similar to the oil field drawdown test. In the case of the
Shell method, the drawdown or vacuum pressure (P') is measured in a
monitoring point at a distance (r) from the vapor extraction well, while
removing vapors at a constant rate (Q). The equation that approximates
expected pressure changes over time is:
P' = Q/(4nm(k/p) [-0.5772 - ln[(r2ep)/(4kP . )] + ln(t)] (2)
3> 1.01
where:
P' = gauge pressure (g/cm-sec2) .,
Q - flow rate from vapor extraction well (cm /sec)
ra - stratum thickness (screen?length) (cm)
k = soil air permeability (cm )
p = dynamic viscosity of air (g/cm-sec)
it - 3.142
r = distance of sample probe well from vapor extraction well (cm)
P - ambient atmospheric pressure (g/cm-s )
e = vapor filled void fraction (0.0 - 1.0)
t • = time (sec)
As in the case of the oil field drawdown method, soil air permeability
(k) can be estimated graphically from field data. The first method assumes
that the volumetric vapor flow rate (Q) from the extraction well and the
thickness of the screened interval (m) from which vapors are being extracted
are known. The second method is applied when Q and m are unknown. Both
methods rely on calculating the slope of the regression line that relates
gauge pressure, P', measured at a sample probe well to the natural logarithm
to the time, In (t), from the initiation of vapor extraction. The
relationship of P' versus In (t) is based on equation (1), where the slope
(A) is:
34
-------
A - Q/4jim(k/p) (3)
and the y intercept (Pf axis) of the regression line is:
B . Q/4nm(k/p) [-0.5772 -ln(r2eu/4kPatm)] (4)
If the flow rate (Q) and the screened interval (m) are known, the soil
air permeability is calculated by solving equation (3) for k.
k m Qy/4Anm (5)
If flow rate (Q) and the screened interval (m) are not known then the
soil air permeability is calculated by substituting equation (3) into
equation (4) and solving for k.
k » r2cp/4P,..Bi [exp(B/A + 0.5772)1 (6)
3.TID
All of the parameters used to estimate permeability are measured in the
field with the exception of the dynamic viscosity of air which is estimated
as a function of air temperature.
Permeability values should be measured at a number of locations around
the vapor extraction well and then averaged to provide a reasonable estimate
of the areal variability soil air permeability.
TESTING METHOD APPLICATION CONSIDERATIONS
Air permeability testing as suggested by Shell (1, 2), assumes that as
the time from initiation of soil vapor extraction increases, the vacuum
pressure in the subsurface increases (ie: the absolute pressure becomes more
negative). However, this relationship may not always hold during the time
intervals over which SVE vacuum pressure data may be measured in the field
(days, weeks and months of continuous operation). As the time over which
SVE is implemented increases, soil moisture is removed from the subsurface,
increasing the effective porosity. As the effective porosity increases, the
subsurface vacuum pressure may decrease (become more positive). Therefore,
it is important to specify that the test be conducted for a short period.
The time interval over which air permeability measurements are made
should be long enough to extract at least one pore volume of air, yet short
enough not to be hampered by: variations in atmospheric pressure, and
effective porosity changes that occur after rainfall and when soil air
moisture condenses and evaporates during diurnal temperature changes.
. Because it is often difficult to maintain a_constant vapor extraction
rate during SVE operation, variations in the vapor extraction rate should be
recorded and used when evaluating data. The sensitivity of the air
35
-------
permeability measurement will be reduced as the variations in the vapor
extraction rate increase.
SVE is often operated concurrently with groundwater extraction and
treatment. If the groundwater extraction causes an increasing cone of
depression in the area where SVE is being implemented, then, over time, the
soil volume susceptable to vacuum pressures may increase, and the perceived
soil air permeability may vary.
The Shell soil-air permeability method provides a valuable tool for SVE
site assessment and design criteria development. More wide spread
application of these test methods might result if additional guidance or
standardization were provided for if standardized?
o extraction well diameter
o screened length
o monitoring point diameter
o radial spacing of monitoring points around extraction well
o extraction rates
o measurement intervals and period
In addition, to allow for correlation of data to other sites, critical
soil measurements should be obtained during extraction well installation
including soil density (standard penetration test), soil grain size
distribution (sand, silt and clay), and soil moisture content.
CONCLUSIONS
Soil air permeability data are critical to the assessment of SVE
feasibility and subsequent development of SVE design criteria at a
particular site. Use of correlation techniques to estimate air permeability
are appropriate for use in first step estimation of appropriateness of SVS.
Conversely, laboratory soil-air permeability determination and air injection
field tests are generally not appropriate for determining the feasibility of
SVE technology. The best estimates of soil-air permeability would probably
be obtained from in situ field drawdown tests modified for use in the
vadose zone.
For sites where SV1 is feasible, a field soil-air permeability test such
as that described by Johnson et al. (1990) should be undertaken to develop
SVE design criteria. The field program should identify the data collection
requirements and at a minimum the following guidelines should be followed.
o Field measurements of vacuum pressure from at least three monitoring
wells spaced around a vapor extraction well should be used.
o A constant extraction well flow rate should be maintained for the
duration of the test.
o Several air permeability measurements should be made over the site to
ensure that the lateral heterogeneity of the site is assessed.
36
-------
o Several air permeability measurements should be made over the site to
ensure that the lateral heterogeneity of the site is assessed.
The basic air permeability tests provide an order of magnitude estimate
for application of SVE. Depending on site conditions and contaminant
distribution the use of vapor flow modeling techniques may be required to
obtain an adequate understanding of the vapor flow regime necessary to
achieve remediation. Likewise, the use of laboratory treatability studies
may be required in order to develop accurate estimates of cleanup times,
REFERENCES
1. Johnson, P.C., Stanley, C.C., Kemblowski, M.W., Byers, D.L., and
Colthart, J.D. A practical approach to the design, operation, and
monitoring of in situ soil-venting systems. G¥MR Spring, 1990.
2. Johnson, P.C., Kemblowski, M.W., and Colthart, J.D. Quantitative
analysis for the cleanup of hydrocarbon-contaminated soils by in-situ
venting. Groundwater 28(3):413-429, 1990.
3. Bear, J. Hydraulics of Groundwater. HcGraw Hill, 1979.
4. Massman, J.V. Applying groundwater flow models in vapor extraction
system design. J. of Envir. Eng. 115(1):129-149» 1989.
5. Vyckoff, R.D. and Botset, E.G. The Flow of Gas-Liquid Mixtures through
Unconsolidated Sands. Physics 7s325-345, 1936.
6. Corey, A.T. Air Permeability. In: Methods of Soil Analysis, Part 1.
Physical and Minerological Methods. Soil Science of America Monogragh
no. 9, 1986.
7. Theis, C.V. The relationship between lowering of the piezometric
surface and the rate and duration of a well using ground-water storage.
Trans.Amer. Geophysical Union 16. 1935.
8. Hassler, G.L. U.S. Patent 2,345,935. 1944.
9. Osaba, J.S., Richardson, G.G., Kerver, J.K., Hafford J.A., and Blair,
P.M. Laboratory measurements of relative permeability. Trans. Am.
Inst. Min. Metall. Pet. Eng. 192-.47-56, 1951.
10. Brooks, R.H. and Corey, A.T. Hydraulic properties of porous media.
Hydrology Paper no. 3, Colorado State Univ., Fort Collins. 1964.
11. Corey, A.T. The interrelation between gas and oil in unsaturated soil.
Soil Sci. Soc. Am. Proc. 21:7-10, 1957.
37
-------
12. Evans, D.D. and Kirkland, D. Measurement of air permeability of soil in
situ. Soil Sci.Soc.Am. Proc. 14:65-73, 1949.
13. Grover, B.L. Simplified air permeameters for soil in place. Soil Sci.
Soc. Am. Proc. 19:414-418, 1955.
14. van Groenevoud, H. Methods and apparatus for measuring air permeability
of the soil. Soil Sci. 106(4):275-279, 1968.
15. Weeks, E.P. Field Determination of Vertical Permeability to Air in the
Unsaturated Zone. U.S.G.S. Prof. Pap. no. 1051. 41 pp, 1978.
16. Donohue, D.A.T., and Ertekin, T. Gaswell testing Theory, Practice and
Regulation. Internat. Human Res. Devel. Corp., Boston, MA. 214 pp,
1982.
17. Katz, D.L. and Tek, M.R. Underground Storage of Natural Gas an
Intensive Short Course. The Univ. of Mich. Eng. Summer Conferences. May
17-28, 1971, 1971.
18. Ikoku, C.U. Natural Gas Reservoir Engineering. J. ¥iley & Sons, N.Y.
503 pp, 1984.
19. Smith, R.V. Practical Natural Gas Engineering. PennWell Publ. Co.
Tulsa OK. 252 pp, 1983.
38
-------
TABLE 1
AIR PERMEABILITY TEST METHODS
CO
10
Method
Reference
Field or Lab
Parameters
Assumed
Parameters
Applicability to SVE Sites
Correlation with
Grain size
Distribution
Massman (1989)
Correlation with
Hydraulic
Conductivity
Massman (1989)
Laboratory
Determination
Corey (1986)
Air Injection
Van Groenwoud
(1968)
- grain size
distribution
vater saturated
hydraulic
conductivity
water flow rate
injected air pressure
soil porosity
cross-sectional area
of sample
injection of air
flow rate
differential injection
pressure
cross-sectional area
tested
- empirical - Can be used to determine feasibility to
constant SVE. Doesn't account for variable water
content, bulk density, soil structure,
or lateral heterogeneity of soils.
Multiple measurements needed.
- empirical - Can be used to determine feasibility of
constant SVE. Only valid for dry soils; doesn't
account for "slippage" of gas by pore
walls; or reduction of permeability by
expandable clays. Multiple measurements
needed.
- Not applicable to SVE. Soil structure
altered when placed in permeameter.
- Not applicable to SVE. Expands soil
particles, increasing measured air
permeability. Usually used to measure
air permeability in surficial soils.
-------
TABLE 1
AIR PERMEABILITY TEST METHODS
(Continued)
Method
Reference
Field or Lab
Parameters
Assumed
Parameters
Applicabilityto SVE Sites
o
Barometric
Pressure Change
Weeks (1978)
Pressure Buildup
Test (gas well
production test)
Donohue and
Ertekin (1982)
Pressure Drawdown
Test (gas well
production test)
Donohue and
Ertekin (1982)
atmospheric pressure
variation
vertical extent of
unsaturated zone
well pressure
static reservoir
pressure
gas flov rate
formation thickness
time at well shut-in
time interval during
well shut-in
viscosity of
air
- Not applicable to SVE. Can only
measure pressure differences if
unsaturated zone is greater that 20m
and has at least one layer with a
permeability of 2 to 3 darcies. Large
atmospheric fluctuations needed to
measure permeability are unpredictable.
viscosity of - Potentially applicable to determine
air SVE feasibility. Subsurface air
temperature pressure may not be high enough at SVE
of gas sites.
compressibility
factor
well pressure drawdown - gas compress-
static reservoir ibility
pressure - viscosity of
gas flow rate gas
formation thickness
time of experiment
soil porosity
distance of monitoring
well from extraction
well
temperature
of gas
- Can be used to determine SVE
feasibility and development of SVE
design criteria. However Gas
compressibility not important factor
at SVE sites. Multiple
measurements needed to account for
lateral soil heterogeneity.
-------
TABLE 1
AIR PERMEABILITY TEST METHODS
(Continued)
Method
Reference
Field or Lab
Parameters
Assumed
Parameters
Applicability to SVE Sites
Drawdown Curve
(applied to
drawdown test)
Theis (1935)
Shell Method
(developed for
SVE) Johnson
et al. (1990)
extraction well flow - viscosity of
rate air
thickness of subsurface
time of experiment
pressure drawdown
distance between
extraction well and
monitoring wells
extraction well flow
rate
monitoring well
pressure
porosity of soil
atmospheric pressure
distance between
extraction well and
monitoring wells
time of experiment
- viscosity of
air
- Can be used to determine feasibility
of SVE. Multiple monitoring well
test measurements needed to account
for lateral soil heterogeneity.
- Can be used to determine SVE
feasibility and development of SVE
design criteria. Multiple
monitoring well test measurements
needed to account for lateral
soil heterogeneity.
-------
P . VACUUM READING
Q- AIR FLOW RATE
r - RADIUS OF INFLUENCE
m
FIGURE 1 - AIR PERMEABILITY TEST COMPONENTS & MEASUREMENT PARAMETERS
42
-------
UNDERGROUND STORAGE TANKS
CONTAINING HAZARDOUS CHEMICALS
Richard F. Wise, James W. Starr, and Joseph W. Maresca, Jr.
Vista Research, Inc.
Mountain View, California 94042
Robert W. Hillger and Anthony N. Tafuri
Risk Reduction Engineering Laboratory
U. S. Environmental Protection Agency
. . Edison, New Jersey 08837
ABSTRACT
The regulations issued by the United States Environmental Protection Agency
(EPA) in 1988 require, with several exceptions, that underground storage tank
systems containing petroleum fuels and hazardous chemicals be routinely tested for
releases. This paper summarizes the release detection regulations for tank systems
containing chemicals and gives a preliminary assessment of the approaches to
release detection currently being used. To make this assessment, detailed
discussions were conducted with providers and manufacturers of leak detection
equipment'and testing services, owners or operators of different types of chemical
storage tank systems, and state and local regulators. While these discussions were
limited to a small percentage of each type of organization, certain observations are
sufficiently distinctive and important that they are reported for further investigation
and evaluation. To make it clearer why certain approaches are being used, this paper
also summarizes the types of chemicals being stored, the effectiveness of several
leak detection testing systems, and the number and characteristics of the tank
systems being used to store these products.
This paper has been reviewed in accordance with the U. S. Environmental
Protection Agency's peer and administrative review policies and approved for
presentation and publication.
INTRODUCTION
Federal underground storage tank regulations promulgated on 23 September 1988 establish
a broad range of minimum requirements for the design, installation, operation and testing of a
large fraction of tank systems in the United States. These regulations cover tank systems
containing petroleum fuels as well as those containing other hazardous chemicals [1,2]. They are
designed to help the underground storage tank community control and minimize the adverse
environmental impact caused by leakage of product from a tank or its associated piping. The
regulatory standards for leak detection in tank systems containing hazardous chemicals are more
stringent than for those containing petroleum motor fuels. This paper describes (1) the
43
-------
regulatory standards for leak detection in tank systems containing hazardous chemicals, (2) the
types of chemicals being stored, (3) the types of containers in which these chemicals are stored,
(4) the effectiveness of tank tightness tests and automatic tank gauging systems for detection of
leaks in tanks containing chemicals other than petroleum, and (5) the approaches to leak
detection being implemented by tank owners and operators. Because the first four items have
been described in detail elsewhere [3-5], this paper simply summarizes them. The main focus is
on the fifth item, specifically, the results of a preliminary survey of manufacturers of leak
detection equipment for chemical tank systems, owners and operators of these systems, and state
and federal regulators.
REGULATORY STANDARDS
The federal regulatory standards for release detection in underground storage tanks issued
by the EPA on 23 September 1988 [1] require that tank systems containing petroleum products
and hazardous chemicals be tested periodically for releases. (A hazardous chemical is any
substance defined by the Comprehensive Environmental Response, Compensation, and Liability
Act (CERCLA) [2].) The regulations for testing underground storage tanks containing
hazardous substances are similar to those for tank systems containing petroleum products.
During the first 10 years after the issuance of the regulations, all existing tank and pipeline
systems containing hazardous substances must meet the requirements specified for tank systems
containing petroleum products. After 10 years, all existing tank and pipeline systems must be
upgraded, if necessary, to meet a more stringent set of requirements. These requirements
emphasize the use of either double-wall tanks and piping or tanks and piping with secondary
containment, both with interstitial monitoring to detect a leak in the inner wall of the system.
These options are described in Section 280.42 (a) - (d) of the regulations [1]. If the tank system
is new or has been upgraded, single-wall tanks and piping are permitted provided that owners
and operators meet the following three criteria.
• Use any one of the release detection methods for tanks specified in Sections 280.43 (b)
- (h) of the regulation or demonstrate to the implementing agency that an alternative
method is at least as stringent. These include internal methods such as tank tightness
testing systems, automatic tank gauging systems, and manual tank gauging for tanks
7,600 L (2,000 gal) or less, as well as external methods such as groundwater- and
vapor-monitoring systems.
• Provide information to the implementing agency about health risks, effectiveness of
corrective action, properties of the stored substance and characteristics of the site. If
the health risks associated with the release of the chemical substance being stored are
no higher than those associated with the release of a petroleum product, and there exist
effective methods to clean up a release, then a single-wall tank system with release
detection would be appropriate.
44
-------
« Obtain approval from the implementing agency.
Although for some types of stored chemicals the single-wall tank system may be a highly
effective way to satisfy the regulations, this option is treated as a variance. The onus is on the
owner or operator to demonstrate to the implementing agency that the chemical substance will
not be any worse than petroleum if accidentally released.
During the 10-year period between 1988 and 1998, the EPA regulations allow tank
owners/operators to use either internal or external systems to test for releases. All systems
attached to or inserted into the tank, piping, or interstitial space of double-wall tanks or piping
are considered internal systems. Internal systems must meet a specific performance standard:
they must have a capability to detect a leak of specific size with a probability of detection (PD) of
95% and a probability of false alarm (PFA) of 5%. No performance standards are specified for
external systems, but specific requirements about conducting tests with such systems are given.
During this 10-year period, the regulation allows three general approaches to release
detection, any of which might be practically pursued. The first two approaches use internal
release detection systems and the third uses external monitoring systems. The first and most
popular approach is to conduct an annual tank or line tightness test to detect small releases and to
use more frequent monitoring by another method to detect large releases. All tank and line
tightness tests must be performed at least once per year and must be able to detect leaks of 0.38
L/h (0.1 gal/h). In all cases where annual tightness tests are used, the regulation requires an
additional form of leak detection in which tests on tanks are conducted at least monthly and those
on pressurized lines at least hourly; this ensures the detection of excessively large releases. For
tanks, daily inventory records must be reconciled monthly. For pressurized lines, leaks of 11.4
L/h (3 gal/h) must be reliably detected; this is usually accomplished by means of a mechanical
line leak detector. The second approach is to install an automatic tank gauge or automatic line
leak detector that is capable of detecting leaks of 0.76 L/h (0.2 gal/h); all monitoring tests must
be done at least once per month. As with the tank and line tightness testing approach, this option
also requires that there be a system for detecting large leaks. The tank gauge can be used to
satisfy the inventory control requirements, and most automatic line leak detectors are designed so
as to be able to satisfy the 11.4-L/h (3-gal/h) hourly test for pressurized piping. Interestingly, if
the tank gauge is used to satisfy the "Other" option in the EPA regulation rather than the
Automatic Tank Gauge option, inventory control is not required; however, owners or operators
who use this option do so because of the potential for better and more accurate control of
inventory. The third approach is to install an external monitoring system that can detect the
presence of the stored chemical in or on the groundwater or in the backfill and soil surrounding
45
-------
the tank system. Among other things, the success of external systems depends on the sensitivity
of the sensor, the ability of the sensor to distinguish the stored chemical from other chemicals
(i.e., its specificity), the ambient background noise level of the stored chemical, the migration
properties of the chemical, and the sampling network. In many instances both internal and
external methods are used in conjunction as a way to increase the probability of detection.
STORAGE OF HAZARDOUS CHEMICALS
Two surveys were conducted to estimate (1) the number of tanks storing hazardous
chemicals, (2) the types of stored chemicals by tank number and capacity and (3) the
characteristics of the tanks by capacity, construction material, and age. A detailed description of
these surveys can be found in [3,4].
The states participating in the program provided databases from thek underground storage
tank registration programs1 for compilation and analysis; a total of 16 state databases were used
in the analysis. The first survey, conducted in 1987, used data from the two largest states in
terms of population, California and New York [3]. In the second survey, conducted in 1990,
chemical tank data from New York and 13 other states were analyzed [4]. In selecting these
states, efforts were made to obtain representative national coverage while simultaneously
examining the more populous industrial states, which might be expected to have large numbers
of chemical tanks. The 14 states included in the 1990 survey were Delaware, Florida, Illinois,
Indiana, Maine, Massachusetts, Minnesota, New York, Missouri, Montana, Ohio, Texas,
Virginia, and Wisconsin. New York was included in the second survey so that changes in its
tank population since the earlier survey might be identified. Tables 1 through 5 summarize the
results of the survey.
TYPES OF CHEMICALS STORED
Solvents were found to comprise the single largest fraction of hazardous chemicals,
comprising over 85% of the total. Table i presents the distribution of the most commonly stored
chemicals by the number of tanks storing the chemical and by the total volume of product being
stored. The 1987 data from New York and California are based only on the population of tanks
containing hazardous chemicals, while the 1990 data from the 14 state databases are based on
the population of all chemical tanks; the 1990 tabulation includes tanks containing both
hazardous and nonhazardous chemicals. As illustrated in Table 1, acetone, toluene, xylene,
1 In 1984, as part of the amendments to the Resources Conservation and Recovery Act (RCRA), each state was
required to register all underground storage tanks.
46
-------
methanol and methyl ethyl ketone were found to be the most commonly stored chemical
substances. The 1987 survey indicated that these five substances accounted for as much as 60%
of all stored organic chemicals. After the fraction of tanks containing nonhazardous chemicals is
removed from the 1990 databases, it can be shown that the five most common organics comprise
49% of all tanks containing hazardous chemicals, a figure that is slightly less than the estimate
made from the survey of the two large states in 1987.
TABLE 1. SUMMARY OF THE MOST COMMONLY STORED ORGANIC CERCLA SOLVENTS
Chemical
Acetone
Toluene
Xylene
Methanol
Methyl
Ethyl
Ketone
TOTALS
1987
% by
Tank
Number
22.8
13.3
8.1
6.6
10.3
61.1
California Data
% by
Tank
Volume
18.0
14.2
6.3
5.5
9.6
53.6
1987
%by
Tank
Number
12.0
22.4
15.5
11.5
9.0
70.4
New York Data
%by
Tank
Volume
18.3
21.1
11.7
8.5
7.0
66.6
1990 Data
%by
Tank
Number
3.9
5.6
—
3.8
3.7
20.1
(14 States)
%by
Tank
Volume
4.2
9.2
2.5
3.3
2.9
22.1
CHARACTERISTICS OF THE TANKS STORING CHEMICALS
Tables 2 through 5 give information about the characteristics of the tanks used to store
chemicals. The characteristics tabulated are the number of tanks, the capacities of the tanks, the
construction materials, and the ages of the tanks. Table 2 presents the total number of tanks
compiled in the 1990 survey that contain hazardous substances. The 5,529 tanks containing
hazardous chemicals represent approximately 57% of the 9,656 registered tanks containing
products other than petroleum. The remaining statistics in the table (i.e., minimum, maximum,
mean, and standard deviation) are based on a tabulation of the number of hazardous-substance
tanks registered in each state. The mean number of tanks containing hazardous substances in
each state is 395. The large standard deviation, the large difference between the mean and the
median value, and the large spread between the states with the minimum and maximum number
of tanks indicate that the number of tanks per state is quite variable. In comparison to the
number of petroleum tanks, the number of tanks containing hazardous chemicals is only a very
small fraction of the total underground storage tank population. Based on these data, the number
of tanks containing hazardous materials throughout the United States should be between 1 to 2%
47
-------
of the total tank population, whether calculated by number or by tank volume. The tabulation
indicates that Illinois has more than twice the number of hazardous-substance tanks than any of
the other states surveyed [4].
TABLE 2, SUMMARY OF THE NUMBER OF TANKS CONTAINING HAZARDOUS CHEMICALS
COMPILED FROM 14 STATE DATABASES"
Statistics
Total for 14 States
Minimum
Maximum
Median
Mean per State
Standard Deviation
Number of Tanks
5,529
14
' 2,060
255
395
516
'The total number of registered tanks containing nonpetroleum chemicals was 9,656,
Table 3 summarizes the capacities of the storage tanks containing hazardous substances,
including the average volume of product. The percentage of tanks in each category (denoted in
the first row across in Table 3) is based on the entire population of tanks containing hazardous
chemicals. The statistics in the remaining rows, both percentages and average volumes, were
computed from the average percentages and average volumes reported for each state. It is
interesting to note that the states having the minimum, maximum, and median values vary
considerably with tank capacity. Roughly 60% of the tanks in the state databases had capacities
between 3,800 and 38,000 L (1,000 and 10,000 gal), with the average size of a tank (based on
data from all states) being 7,205 gallons. Over 27% of the tanks are larger than 38,000 L (10,000
gal).
TABLE 3. SUMMARY OFTANK SIZE DISTRffiUTIONS.COMPILED FROM THE 14 STATE
DATABASES AND EXPRESSED AS A PERCENTAGE OF THE NUMBER OF TANKS IN EACH STATE
Range of Tank Capacities (Gallons)
Statistical
Parameters
Total"
Minimum
Maximum
Median
Mean*
Standard Deviation
< 1,000
11.3
4.3
44.7
12.2
13.7
10.8
1,000-
<4,000
29.6
16.0
39.9
29.0
27.9
6.6
4,000-
<10,000
31.9
19.7
37.6
28.8
29.6
5.6
10,000-
<20,000
19.6
3.9
24.6
19.7
18.2
5.7
>20,000
7.6
0.5
29.6
7.0
9.2
8.6
Average
Volume
7,205"*
3,409
101,293
6,889
7,555"
2,460
Totals for New York, Indiana, and Montana are based on CERCLA chemicals only.
Does not include the Delaware data because one of the tanks in that state has a capacity of 430,000 gal, and
inclusion of these data would result in a misleading statistical estimate.
: 48
-------
Table 4 summarizes the types of materials from which chemical tanks are constructed; the
total is broken down according to the percentage of tanks constructed from steel,
fiberglass-reinforced plastic, and "other" material. As was the case for tank size (Table 3), the
percentage of tanks in each category (denoted in the the first row across in Table 4) is based on
the entire population of tanks. The data indicate that 86% of the tanks are fabricated from steel
and approximately 6% from fiberglass; about 4% are constructed of material(s) other than steel
or fiberglass, and for another 4%, the construction material is not known.
TABLE 4. SUMMARY OF TANK CONSTRUCTION MATERIALS COMPILED FROM LISTING OF
REGISTERED TANKS IN THE 14 STATE DATABASES AND EXPRESSED AS A PERCENTAGE OF
THE NUMBER OF TANKS IN EACH STATE'
Type of Construction Material
Statistical
Parameters
Total
Minimum
Maximum
Median
Mean
Standard Deviation
Steel
86.1
62.9
94.1
83.9
82.4
8.9
Fiberglass
6.2
0.0
15.2
6.6
7.1
4.6
Other
3.9
1.5
10.6
5.6
5.1
2.6
Unknown
3.8
0.0
22.3
2.6
6.4
7.6
* Materials were not reported for Indiana, Minnesota, and Texas. Only steel tanks were reported for Montana,
Values reported are percentages of the total tank populations in each state.
Table 5 summarizes the age of the tanks. The average percentages are based on the entire
tank population. The remaining statistics are based upon the percentages reported for each state.
Tank age was found to average 18 years, with approximately 40%-of the tanks being more than
20 years old.
TABLE 5. SUMMARY OF TANK AGE DISTRIBUTIONS COMPILED FROM 14 STATE DATABASES AND
EXPRESSED AS A PERCENTAGE OF THE NUMBER OF CHEMICAL TANKS IN EACH STATE
Range of Tank Age (Years)
Statistical
Parameters
Average
Minimum
Maximum
Median
Mean
Standard Deviation
Oto4
4.9
1.5
22.2
3.4
6.0
5.9
5 to 9
16.2
4.0
23.0
15.2
16.1
7.0
10 to 14
21.6
4.0
26.1
21.1
20.6
9.4
15 to 19
17.6
0.0
25.6
17.1
14.9
7.0
>20
39.7
22.8
86.7
38.7
42.4
16.3
49
-------
CONSEQUENCES FOR RELEASE DETECTION
The results of these tabulations suggest that there is a strong potential for leakage from
tanks containing hazardous substances. The statistics suggest that the tanks are generally old,
made of steel, and fairly large. One would speculate that because the average age of these steel
tanks is 18 years, many are unprotected by rust-resistant coatings and are highly susceptible to
corrosion. As noted in the next section, the analysis performed in [5] suggests that most tank
tightness and automatic tank gauging systems (which are internal leak detection systems) should
be able to test these tanks effectively. Most of these leak detection systems were evaluated on
30,000- or 38,000-L (8,000- or 10,000-gal) tanks, which is consistent with the average capacity
of tanks containing hazardous chemicals. Successful testing of chemical tanks should be
possible, especially because their number is relatively small, approximately 1 to 2% of the total
underground storage tank population. Moreover, a very small number of chemicals (five)
accounts for roughly half of the hazardous substances being stored. External methods of leak
detection can also be used provided that the leak detection system in question has the necessary
specificity.
VOLUMETRIC TANK TIGHTNESS TESTING
The same types of leak detection and monitoring systems used for testing tanks and
pipeline systems containing petroleum products should be applicable to those containing
chemicals provided that the sensors and equipment are compatible with the particular stored
chemical and can be installed and used safely. The performance of these leak detection systems
has, in most cases, been determined through an evaluation based on a single, specific, stored
product. Volumetric leak detection systems, such as tank tightness testing systems and
automatic tank gauges, were developed specifically to test storage tanks containing petroleum
fuels, and any estimates of their performance, therefore, have been based on this class of liquids.
Most performance evaluations of such systems have been conducted in 30,000- or 38,000-L
(8,000- or 10,000-gal) tanks containing either gasoline or diesel fuels. If the tank contains a
chemical that differs, in density, viscosity, or coefficient of thermal expansion, from the product
used in the evaluation of a given leak detection system, the performance of that system when
used to test a tank containing a nonpetroleum chemical will be different from what it was on the
petroleum tank.
An analysis was made of the performance of tank tightness systems (and tank gauging
systems) when applied to tanks containing chemicals other than petroleum fuels [5]. Since
petroleum was the stored product in most evaluations of tightness testing systems, the analysis
50
-------
attempted to determine impact of liquids with viscosities, densities, and thermal properties
different from petroleum. The influence of the viscosity, the density, and the coefficient of
thermal expansion on performance was investigated for the range of chemicals identified in the
14 state databases. The analysis examined the two most important sources of noise: thermal
expansion or contraction of the product stored in the tank and the structural deformation of the
tank resulting from any level or pressure changes before or during a test. Methods of
compensating for thermal expansion/contraction and structural deformation were also
investigated.
The analysis showed that (1) the performance of a volumetric leak detection system is
directly proportional to the coefficient of thermal expansion of the product in the tank and (2) the
waiting period required for the effects of structural deformation to subside is essentially the same
for all values of density (even though higher densities produce greater deformation-induced
volume changes immediately after any product-level change). When a leak detection system is
used with a chemical product having a coefficient of thermal expansion higher than that of the
product used in the evaluation of the system, the system's performance will be lower than it was
in the evaluation. If the performance achieved in the evaluation barely meets the minimum
standards established by the EPA, it is possible that the leak detection system will not meet the
standard when used with chemicals having higher coefficients. Even if the leak detection system
exceeds the rninimum performance standards, it is possible that it will not meet the PFA or PD
requirement; however, if a system has achieved high performance during the evaluation,
judiciously changing the detection threshold can make it possible for the leak detection system to
meet the requirements. Because gasoline has a higher coefficient of thermal expansion than
many chemicals, a system evaluated with a gasoline product can be used with such chemicals
and still maintain a similar level of performance.
This analysis did not examine volume changes due to evaporation arid condensation, or
those due to trapped vapor; the former may be an important source of error in tests conducted on
underfilled tanks, and the latter an important source of error in tests conducted on overfilled
tanks.
CURRENTLY USED APPROACHES TO LEAK DETECTION
An informal survey of the owners and operators of chemical tanks, manufacturers of tank
tightness testing and automatic tank gauging systems, and state and local environmental
regulators was conducted by telephone to determine
51
-------
• what methods of leak detection are being used for underground storage tanks (i.e.,
tanks and associated pipelines) storing hazardous substances,
• the basic characteristics of the chemical tank population to which these methods are
applicable, and
• what inventory practices are being followed by owners/operators of underground
storage tanks containing hazardous substances.
A questionnaire was prepared as a guideline to stimulate discussion. The questionnaire was
designed to shed some light on what methods of leak detection are being applied to single-wall
tanks between 1988 and 1998. The responses were given in confidence, and, as a result, the
organizations discussing their environmental programs will not be disclosed by name. They are
identified only by size and by a general description of the type of business they conduct. The
organizations contacted ranged from small enterprises to large, well-known, Fortune 500
companies. The organizations that were interviewed were located in New York, California and
Illinois.
Two surveys were planned, one to address leak detection practices and one to address
inventory practices. In the initial survey, the survey taker started by contacting tank tightness
testers and organizations that store chemicals to determine (1) which methods of leak detection
are being used and (2) user perceptions as to the effectiveness of these methods. The second
survey was designed to address the inventory practices of tank owners/operators and to collect 30
to 90 days of inventory records for analysis. As a check on the owners' responses, a brief
discussion of inventory practices in the chemical industry was held with a major
inventory/statistical inventory management service.
After the survey taker had contacted only a few organizations using tanks to store
chemicals, it became clear that these organizations were either in the process of or had completed
upgrading their systems to meet the regulatory standards required by 1998. As a consequence,
the emphasis of the questions shifted from the technical details of the types of leak detection
methods being used and the procedures followed for inventory reconciliation in single-wall
tanks, and turned instead to the upgrading approaches. Instead of two separate surveys, only one
was actually conducted.
RESULTS OF DISCUSSIONS WITH TANK TESTERS
Three tank tightness testing services that are well known in the leak detection industry
were asked whether they were capable of testing tanks containing chemicals other than
petroleum and whether they had actually tested such tanks. At the time of the survey, all three
companies had systems that conducted'tests on overfilled tanks, and one had the ability to test
partially filled tanks. (At present, all three firms have the capability to test partially filled tanks,)
52
-------
All three said that they did test tanks containing chemicals. The only constraint on testing was
that the temperature and level (volume) measurement systems inserted into the tank had to be
compatible with the stored chemical. In general, such equipment was constructed of stainless
steel and Teflon. All three firms indicated that up to 5% of their services involved testing tanks
containing products other than petroleum. They all indicated that the performance of their
systems was the same regardless of whether a tank contained petroleum or other chemicals. This
response is consistent with our estimate of the number of tanks containing chemicals and our
previous knowledge of this industry. None of the organizations manufacturing automatic tank
gauges was contacted directly as part of this survey because, based on previous discussions with
several automatic tank gauge manufacturers, it was expected that they would give the same
general response as the tank tightness testing services. Automatic tank gauges are particularly
suited for meeting regulatory requirements in tanks containing chemicals because tests can be
conducted routinely and automatically without adding product to the tank,
RESULTS OF DISCUSSIONS WITH TANK OWNERS AND OPERATORS
The survey taker contacted a total of 19 organizations that use chemicals in their
operations and that own the tanks in which these chemicals are stored. He obtained responses
from 13. The level of response varied considerably, as shown in Table 6, which summarizes the
important aspects of the survey. Six of the firms, which are denoted by an asterisk, responded in
sufficient detail to address all the questions prepared for the survey. A triple dash means that the
organization did not respond to the question or did not know how to respond to the question.
In general, most firms had fewer than 50 tanks containing chemicals, and the median age
of these tanks was approximately 20 years. Two of the firms did not indicate the number or age
of their tanks because they were in the process of replacing all their single-wall tanks with
aboveground tanks, double-wall tanks, or tanks with secondary containment. In all cases the
tanks were used to store chemicals used in company operations. About half of the firms
responded to the question of removal and disposal of the chemicals after the process had been
completed. Waste chemicals were either reclaimed or stored in drams for removal.
None of the firms contacted indicated that they have had their tanks tested with a
volumetric tank tightness testing system; one firm had its tanks tested with an air test, but this
method was discontinued because of inaccuracy and other problems. (Air tests are no longer
recommended, nor are they commonly used.) None of the companies contacted was using or
planning to use automatic tank gauges for monitoring or inventory control purposes.
53
-------
TABLE 6. SUMMARY OF IMPORTANT RESPONSES TO THE SURVEY
Company
Product
Finishing*
Adhesives/Glues"
General
Chemicals*
Adhesives/Glues*
Printing*
Computers*
Chemicals
Cleaning
Chemicals
Adhesives
General
Chemicals
Tank Farm
Finishes/Paints
Computers
Company
Size
Small
Medium
Large
Large
Medium
Large
Small
Medium
Medium
Large
Medium
Medium
Large
Was
inventory
No. of Mean control
Tanks Age used?
23
21
17
53
24
—
11
21
24
13
19
215
—
25+
20
20
15-40
10
—
15-
20
25+
—
*
—
—
Yes
No
Yes
No
Yes
No
No
No
—
—
No
Were
tanks
being
replaced?
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Were Were Were
double-wall aboveground single-wall
tanks being tanks being tanks with
used? used? secondary
containment
Yes
Yes
Yes When
Possible
Yes When
Possible
Yes
When
Possible
Yes
—
When
Possible
Yes
When
Possible
Yes
Yes
being used?
...
—
—
—
...
Single- wall/
vaulted
—
—
„
—
—
—
* Firms that answered all survey questions in detail.
Only three of the firms indicated that they kept inventory records, but these were for
accounting and scheduling purposes only. These firms did not use inventory control data for
leak detection. Based on the discussions with these three organizations, it was determined that
the data being routinely obtained could not be used for inventory reconciliation either because
there was no meter used to indicate the volume of material removed from the storage tank or the
accuracy of this meter was inadequate for such an application.
All of the organizations contacted were replacing or planning to replace their single-wall
tanks with either aboveground tanks or double-wall tanks with interstitial monitors. It was clear
that the use of aboveground tanks was overwhelmingly preferred. The use of aboveground tanks
permits visual inspection for leaks and facilitates maintenance; aboveground tanks also minimize
the cleanup costs associated with an accidental leak.
54
-------
SUMMARY
Even though a diverse cross section of organizations was contacted, the responses obtained
during the telephone survey should not be interpreted quantitatively; the number of organizations
was very limited, and the survey was not statistically designed or statistically analyzed. As a
consequence, the results should be interpreted cautiously, and the temptation to generalize,
particularly about the status of regulatory compliance, should be avoided unless additional data
are gathered. The following observations are noteworthy, however, either because the response
was overwhelming or because it was ambiguous.
First, there is a strong tendency for owners/operators of tank systems to be planning ways
to comply with the "upgraded standards" specified for 1998. There appears to be an emphasis on
replacement of single-wall tank systems with (1) double-wall tanks and pipes equipped with
interstitial monitors (and in some cases combined with external monitors also) or (2) tank
systems mounted completely above ground so that visual inspection is possible. This emphasis
on meeting the upgraded standards has occurred, we believe, because of the potential for serious
environmental damage, the high clean-up costs, and the large liability associated with chemical
contamination of the soil and groundwater. Concern may also stem from the fact that tanks
containing chemicals axe old (averaging 18 years) and constructed of steel (86%). What is not
clear from the survey is how much time will be required for those organizations currently
upgrading their tank systems to complete the process. If the time required for upgrading a tank
system exceeds one year, the regulation requires that the tank system be tested by means of
methods commonly used on tanks containing petroleum.
Second, none of the organizations contacted used inventory control as a means of leak
detection. It also appears that this method of leak detection would be difficult to apply because
of the lack of metering devices or the lack of accuracy in the metering devices being used. This
observation was independently verified by a company that is heavily involved in analyzing
inventory control data for owners or operators of chemical and petroleum tank systems.
Third, the tank testing firms contacted indicated that approximately 5% of their tests were
conducted on tanks containing hazardous chemicals, a figure that is slightly higher than the
estimated percentage of such tanks in existence in the U.S. This is inconsistent with the response
obtained from the 13 tank-owning organizations that responded to the survey. None of these
organizations indicated that they used such services. In addition, the tank testing firms did not
know whether the owners and operators of the tanks they tested employed monthly inventory
reconciliation. (Inventory reconciliation is required by the regulations when the only form of
55
-------
leak detection is an annual tightness test.) The contradictory responses offered by the testing
firms and the owners and operators of tank systems containing chemicals suggest that the
owners/operators who responded to the survey may not be representative of the entire chemical
tank community.
Fourth, additional information is required before any assessment can be made of release
detection practices in effect now and during the next eight years (the time allowed for
owners/operators of chemical storage tanks to upgrade their systems in anticipation of the 1998
EPA deadline).
REFERENCES
1. U.S. Environmental Protection Agency, 40 CFR 280--Technical Standards and
Corrective Action Requirements for Owners and Operators of Underground Storage Tanks.
Federal Register, 53; 185,23 September 1988.
2. U.S. Environmental Protection Agency. Part 302 - Comprehensive Environmental
Response, Compensation, and Liability Act. Federal Register, 11 December 1980.
3, I, Lysyj, R. W. Hillger, J. S. Farlow, and R. Field. A Preliminary Analysis of
Underground Storage Tanks Used for CERCLA Chemical Storage. Proceedings of the
Thirteenth Annual Research Symposium, Hazardous Waste Engineering Research
Laboratory, Office of Research and Development, U.S. Environmental Protection Agency,
Cincinnati, Ohio, July 1987.
4, R. W. Hillger, J. W. Starr, and M. P. MacArthur. Characteristics of Underground Storage
Tanks Containing Chemicals. To be submitted for publication in/, of Oil and Chemical
Pollution, in preparation.
5. J. W. Starr, R. F. Wise, J. W. Maresca, Jr., R. W. Hillger, and A. N. Tafuri. Volumetric
Leak Detection in Underground Storage Tanks Containing Chemicals. Proceedings of the
84th Annual Meeting and Exhibition of the Air and Waste Management Association, to be
held in Vancouver, B. C., Canada in June 1991, in preparation.
56
-------
SUBSURFACE FATE AND TRANSPORT OF PETROLEUM HYDROCARBONS
FROM LEAKING UNDERGROUND STORAGE TANKS;
A BASIS FOR EVALUATING THE EFFECTIVENESS OF CORRECTIVE ACTIONS
by: Warren J. Lyman Chi-Yuan Fan
Patrick J. Reidy U.S. EPA/RREL/RCB
Camp Dresser & McKee Woodbridge Avenue
Ten Cambridge Center Edison, NJ 08837
Cambridge, MA 02142
ABSTRACT
The problems associated with leakage of motor fuels and organic
chemicals from underground storage tanks (USTs) are compounded by a general
lack of understanding of the partitioning, retention, transformation, and
transport of these contaminants in the subsurface environment. The research
material developed in this project is the result of an intensive data
collection and evaluation effort that compiled a very broad range of
knowledge of contaminant behavior in the subsurface into a single document.
The document describes micro-scale fate and transport processes of
contaminants in the subsurface as a means to understanding their larger
scale movement. This, in turn, leads to a more thorough understanding of
the application of corrective measures and remediation techniques. The
micro-scale analysis focuses on 13-loci, each of which represents a location
and condition in the subsurface environment where contaminants may exist
after an UST release.
This technical handbook is a data base of scientific knowledge that can
be drawn upon by environmental scientists, engineers, and managers of
varying levels of technical expertise to define the key parameters and to
increase the level of sophistication in the approach to the problem of motor
fuel leaks from USTs. It serves to strengthen an understanding of the fate
and transport processes vital to effective remediation, and it serves as a
source book of information, data, and equations to support more quantitative
assessments of pollutant fate and transport processes associated with
different remedial technologies.
This paper has been reviewed in accordance with the U.S. Environmental
Protection Agency's peer and administrative review policies and approved for
presentation and publication.
57
-------
INTRODUCTION
BACKGROUND
The U.S. Environmental Protection Agency (EPA) has developed a
comprehensive program for regulating certain underground storage tanks (UST)
that contain regulated substances. ¥ith this program comes the need for
development of guidance to assist those parties involved in complying with
regulatory requirements. One significant portion of the legislation that
has been developed pertains to corrective actions for releases of petroleum
products such as gasoline and other motor fuels. EPA estimates that over 95
percent of the estimated 1.4 million UST systems are used to store petroleum
products.
Current regulations require that owners and operators take corrective
measures to mitigate releases from underground storage tanks. To date,
guidance documents developed by the EPA and other organizations present
various technologies applicable to remediation of the subsurface
environment. However, only minimal amounts of information have been
generated vith regard to evaluating the movement and disposition of motor
fuel contaminants in the subsurface environment, and only recently has
significant research been devoted to this effort. Currently, investigators
have no guidance available to evaluate corrective actions at UST sites that
is based upon the scientific principles governing the behavior and
degradation of motor fuel constituents in the subsurface environment.
Recognizing this, EPA sponsored the comprehensive, scientific literature
research effort resulting in a handbook-style resource document.
The primary focus of this project was to formulate a data base of
comprehensive scientific knowledge that can be drawn upon to define the key
parameters and to increase the level of sophistication in the approach to
the problem of motor fuel leaks from USTs. The focus on organic
contaminants is based on the reality that over 95 percent of materials
stored in underground tanks is some type of petroleum product. As a
starting point, EPA sponsored a two-day seminar where recognized experts and
specialists in eight specific disciplines presented information on the
physical, chemical, and biological properties, and fundamentals of motor
fuel and organic chemical behavior in the subsurface environment. Based on
the information presented at this seminar, an evaluation of available
scientific data was conducted to provide a comprehensive, up-to-date
understanding of the physical and chemical "rules" governing contaminant
fate and persistence in the subsurface environment.
OBJECTIVES OF HANDBOOK
The EPA's Office of Underground Storage Tanks (OUST) is responsible for
establishing the Agency's regulatory program for managing underground
storage tanks. The Risk Reduction Engineering Laboratory (RREL) of EPA's
Office of Research and Development is responsible for providing engineering
and scientific support to OUST. One means of providing this technical
support is through the preparation of guidance materials such as handbooks,
58
-------
manuals, and technical reports. To date, there have been a number of such
documents developed by the EPA and other organizations on various aspects of
the UST corrective action process. A review of these and other currently
available documents indicated that there was no pre-existing guidance which
directly relates fate and transport processes to corrective actions, a
shortcoming deemed serious by EPA.
Therefore, the primary objective of this handbook was to provide a
comprehensive, in-depth research report with detailed coverage of fate and
transport mechanisms that also includes timely discussion of these
mechanisms as they relate to currently used or innovative approaches to UST
releases. The handbook is intended to benefit environmental scientists,
.engineers and managers of varying technical expertise by increasing the
level of sophistication in the approach to the problem of motor fuel leaking
from USTs. It is also intended to serve as a source book of information,
data, and equations, to support quantitative assessments of pollutant fate
and transport.
To further support timely and technically sound responses to leaking
USTs, the information in this handbook provided the technical basis for the
preparation of the desired guidance handbooks recently released by HWERL:
o ASSESSING UST CORRECTIVE ACTION TECHNOLOGIES: Site Assessment and
Selection of Unsaturated Zone Treatment Technologies. Report No.
EPA/600/2-90/011, Risk Reduction Engineering Laboratory, Cincinnati,
OH, March 1990.
o ASSESSING UST CORRECTIVE ACTION TECHNOLOGIES; Early Screening of
Clean-up Technologies for the Saturated Zone Report." No.
EPA/600/2-90/027, Risk Reduction Engineering Laboratory, Cincinnati,
OH, June 1990.
RESEARCH APPROACH
As a starting point, EPA sponsored a two-day seminar where recognized
experts and specialists in eight specific disciplines presented information
on the physical, chemical, and biological properties and fundamentals of
motor fuel and organic chemical behavior in the subsuface environment. Based
on the information presented at this seminar, a concept was developed that a
substance leaking from an UST will be present in and transient between one
or more locations or settings in the subsurface environment. A total of 13
of these locations, referred to as physicochemical-phase loci, were
identified as the focal points for this research report. Each of the 13
loci represents a point in space (or location) and the physical state of the
leaked substance that together describe where and how these contaminants may
exist in the subsurface environment after an UST release. For example,
contaminants may be dispersed as a component of soil gas, or they may be
dissolved in the water film surrounding a wet soil particle in the
unsaturated zone. These are just two examples of the 13 loci.
Collectively, the 13 loci represent all locations/states where and how
leaked material may be present in the subsurface environment. The
distribution of contaminants among these loci is constantly changing, as
59
-------
contaminants tend-to move between different loci at varying rates over time.
Table 1 presents a brief description of each locus. Figure 1 presents a
schematic cross-section of the subsurface environment and identifies where
each locus may exist in terms of the unsaturated and saturated zones.
These 13 loci can be used in reference to any type of contaminant that
enters the subsurface environment. However, for the purpose of this
research-effort, the contaminants of-interest are petroleum products (e.g.,
gasoline) constituents, i.e., hydrocarbons and other chemicals, including
common gasoline additives. As pointed out previously, the reason for
focusing on petroleum product constituents is that these products make up
over 95 percent of the materials managed in underground storage tanks. All
references to contaminants in this handbook are, therefore, with regard to
hydrocarbons and associated organics.
The ultimate objective in conducting research on the fate and transport
of hydrocarbons in the manner presented, i.e., locus by locus, and process
by process from a micro-scale perspective, is to gain a better understanding
of the basis for larger scale contaminant movement. The ways in which the
behavior of hydrocarbons may be altered are directly related to remediation
techniques. Therefore, a comprehensive understanding of the fundamental
"rules" of contaminant behavior engenders a better understanding of how this
behavior may be induced or prohibited to optimize a given remedial strategy.
Each locus, after being initially defined, was researched and evaluated
in terms of the mobilization, immobilization, transformation, and bulk
transport processes that pertain to it, and in terms of how these various
processes affect the locus. Table 2 identifies the different processes that
were considered when evaluating each locus. In general, many, but not all
of these processes apply to a given locus. There are also many processes
which are important to more than one locus. Because of this, it was
necessary to minimize redundancy in the material presented. To accomplish
this, detailed discussion of a given process is usually limited to a single
locus section - that for which the process is of particular importance or to
which it most appropriately belongs. In other loci sections where the
process is discussed, the reader is directed to the section(s) containing
detailed discussions rather than repeat the material.
The one locus that is treated differently from others in the handbook is
locus no. 11 (contaminants sorbed into/onto biota). It is treated
differently in terms of section organization and contents because locus no.
11 is considered to be a transformation process in itself, i.e., it is
considered to be both a locus and a process.
Following discussions of partitioning, transformation, and transport
processes, each handbook section provides guidance on calculating maximum
and average values for the contaminant storage capacity of the locus,
including estimation of parameter values and example calculations. In
addition, example calculations are provided where appropriate for mass
transport processes (e.g., advection, dispersion, and diffusion) and
partitioning equilibrium processes (e.g., calculations for dissolution and
adsorption). Lastly, the relative importance of each locus with regard to
60
-------
TABLE 1
BRIEF DESCRIPTIONS OF THE THIRTEEN
PHYSICOCHEMICAL-PHASE LOCI
Locus
Number Description
1 Contaminant vapors as a component of soil gas in the
. unsaturated zone.
2 Liquid contaminants adhering to "water-dry" soil particles in
the unsaturated zone.
3 Contaminants dissolved in the water film surrounding soil
particles in the unsaturated zone.
4 Contaminants sorbed to "water-wet" soil particles or rock.
surface (after migrating through the water) in either the
unsaturated or saturated zone.
5 Liquid contaminants in the pore spaces between soil particles
in the saturated zone.
6 Liquid contaminants in the pore spaces between soil particles
in the unsaturated zone.
7 Liquid contaminants floating on the groundwater table.
8 Contaminants dissolved in groundwater (i.e., water in the
saturated zone).
9 Contaminants sorbed onto colloidal particles in water in
either the unsaturated or saturated zone.
10 Contaminants that have diffused into mineral grains or rocks
in either the unsaturated or saturated zone.
11 Contaminants sorbed onto or into soil microbiota in either
the unsaturated or saturated zone.
12 Contaminants dissolved in the mobile pore water of the
unsaturated zone.
13 Liquid contaminants in rock fractures in either the
unsaturated or saturated zone.
61
-------
SOIL PARTICLES OR ROCK
LIQUID CONTAMINANT
(Organic Phase)
WATER WITH DISSOLVED
CONTAMINANT
SOIL AIR WITH
CONTAMINANT VAPORS
CONTAMINANTS SORBED ON SOIL OR
DIFFUSED INTO MINERAL GRAINS
MOBILE COLLOIDAL PARTICLES
WITH SORBED CONTAMINANT
SOIL MICROBIOTA WITH
SORBED CONTAMINANT
LOCI NUMBER (SEE TABLE1)
This is a highly schematic representation of the 13 loci outlined in Table 1 as they exist in the
subsurface. A more detailed explanation of each locus is presented in the text.
Figure 1. Locations of Loci in Terms of Unsaturated and Saturated Zones
62
-------
TABLE 2
FATE AND TRANSPORT PROCESSES CONSIDERED IN EVALUATING EACH LOCUS
Partitioning and
Transformation Processes
Bulk Transport Processes
Partitioning
Dissolution of Liquid Contaminant
into ¥ater
Volatilization of Liquid Contaminant
into Soil Air
Sorption of Liquid Contaminant
onto "Dry" Soil
Partitioning between Aqueous and
Vapor Phases
Sorption of Aqueous Phase
Contaminants to Soil
Condensation of Vapor Phase
Contaminants on Dry Soil
Transformation
Chemical Oxidation of Contaminants
Biodegradation of Contaminants
Advection/Dispersion of "Water-Wet"*
Liquid Contaminants
Advection/Dispersion of "Oil-Wet"**
Liquid Contaminants
Advection/Dispersion of Liquid or
Aqueous Phase Contaminants in the
Unsaturated Zone
Advection/Dispersion of Liquid or
Aqueous Phase Contaminants in the
Saturated Zone
Advection/Dispersion of Contaminants
in Soil Air
Transport with Water of Contaminants
Attached to Biota or Colloidal
Particles
Diffusion of Contaminants in Air
Diffusion of Contaminants in Liquids
Diffusion of Contaminants in Solids
* "Water-wet" refers to the condition where water preferentially (vs
petroleum products) wets the surface of the soil particles.
** "Oil-wet" refers to the condition where the petroleum product
preferentially wets the surface of the soil particles.
63
-------
fate and transport of contaminants is discussed in terms of remediation,
Interaction with other loci, identification of critical information gaps,
and recommendations for future research.
Even though remediation is not the subject of this handbook, indirectly
the ultimate purpose of conducting the loci research is to increase the
sophistication of current approaches to site assessment and corrective
action selection. Brief discussions of remediation and corrective measures
are presented in each locus section in three places; in each introductory
subsection, in the subsections discussing effective partitioning processes,
and in the overview of each locus' relative importance presented in the last
subsection.
ORGANIZATION OF HANDBOOK
The main body of the handbook is organized into 13 major sections, one
for each locus. Each section, except for Section 11 (locus no. 11),
consists of the same five main subsections, which, in turn, contain
discussions that are presented similarly. As previously explained, locus
no. 11 is dealt with differently in terms of the structure of its section
because of its nature, i.e., it is both a locus (biota) and a process
(biodegradation). Table 3 presents a generic outline of the locus sections.
Although many sections deviate slightly from the standard outline, the same
types of information and evaluations have been included in all sections.
Brief descriptions of some of the major headings of Table 3 are presented
below.
The title of Section X.I is self explanatory. Section X.2 discusses in
detail the fate and transport mechanisms affecting the locus. The text is
interspersed with discussion of how these mechanisms relate to corrective
measures and how they could be enhanced to support remediation. Section
X.2.2 describes the processes governing contaminant transport, either as
mass movement of the locus (advection/dispersion) or as mass transformation
to other loci (dissolution/volatilization/condensation/sorption). t Section
X.2.3 discusses the mechanisms that influence fixation of the locus by. 1)
enhancing phase exchange or mass transport out of the locus into another
less mobile locus, or 2) fixing the locus as a whole, for possible future
remediation or as a permanent corrective action. Section X.2.4 discusses
transformation that can potentially act on the locus to produce physical
and/or chemical changes, which in turn result in a less complex, less toxic
residual. Among the transformation processes are biodegradation, chemical
oxidation, hydrolysis, elimination, dehydrogenation and redox reactions.
Section X.3 describes methods for quantifying the mass of contaminant
existing in the locus, i.e., storage capacity. A typical section presents
the factors influencing storage capacity and equations to calculate storage
capacity. Guidance is provided on inputs for calculating average and
maximum storage capacity. Sources of data for the various parameters
required in the calculations are cited, many of which are found within the
text.
64
-------
TABLE 3. GENERIC OUTLINE FOE LOCUS SECTIONS
SECTION X - LOCUS NO. X
X.I Locus Description
X.I.I Short Definition
X.I.2 Expanded Definition and Comments
X.2 Evaluation of Criteria for Remediation
X.2.1 Introduction
X.2.2 Mobilization/Remobilization
X.2.3 Fixation
X.2.4 Transformation
X.3 Storage Capacity in Locus
X.3.1 Introduction and Basic Equations
X.3.2 Guidance on Inputs for, and Calculation of, Maximum Value
X.3.3 Guidance on Inputs for, and Calculation of, Average Value
X.4 Example Calculations
X.4.1 Storage Capacity Calculations
X.4.2 Transport Rate Calculations
X.5 Summary of Relative Importance of Locus
X.5.1 Remediation
X.5.2 Loci Interactions
X.5.3 Information Gaps
In Section X.4, example calculations of storage capacity vithin the
locus and other important processes defined by equations within the text
(e.g., transport velocity, dissolution rate, etc.) are worked out. The
examples employ hypothetical site conditions and contaminant characteristics
that might reasonably be encountered in the field to convey a realistic
sense of locus behavior in the subsurface.
Finally, Section X.5 evaluates the locus in terms of its overall impact
in the subsurface and its relation to other loci. Among the factors
considered are mass of the locus relative to total contaminant mass, locus
impact on and transfer to other loci, and the degree to which corrective
action measures can be effectively employed. Also discussed in Section X.5
is the availability of data on the locus in terms of completeness of
research conducted and information gaps needing to be filled to better
understand the locus.
65
-------
THE INCINERATION OF LEAD-CONTAMINATED SOIL
RELATED TO THE COMPREHENSIVE ENVIRONMENTAL RESPONSE
COMPENSATION AND LIABILITY ACT (CERCLA) (SUPERFUND)
by: Howard 0. Wall
Risk Reduction Engineering Laboratory
U.S. Environmental Protection Agency
Cincinnati, OH 45268
ABSTRACT
The fate of lead on incinerated CERCLA soil was evaluated at the USEPA
Incineration Research Facility (IRF) at Jefferson, Arkansas. This facility
houses a pilot-scale rotary kiln incinerator which was fed lead contaminated
soil for the reported study. The analytical results indicate that lead
emissions to the atmosphere after the air pollution control device (APCD)
remained constant at about 7 percent of the lead content in the feed at 816°C
(150Q°F) and 100 percent excess air. When the excess air was reduced to 50
percent, the emissions ranged from 6 to 9 percent of the lead content in the
feed at 816°C (1500°F).
INTRODUCTION
One of the primary purposes of the Incineration Research Facility (IRF)
in Jefferson, Arkansas, is to support the Environmental Protection Agency
(EPA) Regional Offices and the Office of Solid Waste and Emergency Response in
evaluation of incineration as a disposal option for the remediation actions at
Superfund sites. A priority site that fell into this category was the Baird
and HcGuire location at Hoi brook, Massachusetts, which is located in EPA
Region I. Region I requested that the contaminated soil from this location be
burned to generate data that could be used for the evaluation of incineration
as a treatment at this Superfund site. Incineration was proposed to destroy
the pesticides (pp'-DDT, pp'-DDD, pp'-DDE and Methoxychlor), but the fate of
the metals, lead and arsenic, in the soil was also of concern both during and
after incineration. In addition to the destruction of organics, the mission
of the test was to evaluate the different operating conditions and how they
would impact the distribution of the lead and arsenic to the various residual
streams. This paper reports the partitioning to the incinerator discharge
streams: ash, scrubber water and flue gas.
66
-------
The experimental work was conducted at the USEPA Incineration Research
Facility (IRF) (1). The rotary kiln incineration system (RKS) was selected
for this research, with the ionizing wet scrubber as the air pollution control
device. The design characteristics of the kiln system and the ionizing wet
scrubber are:
Length, outside
Diameter, outside
Length, inside
Diameter, inside
Chamber volume
Construction
Refractory
Rotation
Solids retention time
Burner
Primary fuel
Feed System
Liquids
Sludges
Solids
Temperature (max)
Main Chamber
2.61 m (8 ft - 7 in.)
1.22 m (4 ft)
2.44 m (8 ft)
0.95 m (3 ft - 1-1/2 in.)
1.74 m3 (61.4 ft3)
0.63 cm (0.25 in.) thick cold-rolled steel.
12.7 cm (5 in.) thick high alumina castable
refractory, variable depth to produce a
frustroconical effect for moving solids.
Clockwise or counterclockwise 0.2 to 1.5 rpm
1 hr (at 0.2 rpm)
North American Burner, rated at 770 kW (2.6 MMBtu/hr)
with liquid feed capability.
Natural Gas
Positive displacement pump via water-cooled lance.
Moyno pump via front face, water-cooled lance.
Metered twin-auger screw feeder or fiber pack ram
feeder.
1010°C (1850°F)
Characteristics of the AfterburnerChamber
Length, outside
Diameter, outside
Length, inside
Diameter, inside
Chamber volume
Construction
Refractory
Gas residence time
Burner
Primary fuel
Temperature
3.05 m (10 ft)
1.22 m (4 ft)
2.74 m (9 ft)
0.91 m (3 ft)
1.80 m3 (63.6 ft3)
0.63 cm (0.25 in.) thick cold rolled steel
15.24 cm (6 in.) thick alumina castable refractory
1.2 to 2.5 sec depending on temperature and excess air
North American burner rated at 590 kW (2.0 MMBtu/hr)
with liquid feed capability
Natural gas
1200°C (2200°F)
Characteristicsof theIonizing Wet Scrubber APCD
System capacity
inlet gas flow
Pressure drop
Liquid flow
pH control
85 m3/min (3000 acfm) at 78°C (172°F) and 101 kPa
(14.7 psia)
1.5 kPa (6 in. we)
15.1 L/min (4 gpm) at 345 kPa (50 psig)
Feedback control by NaOH solution addition
67
-------
MUFFLE FURNACE EXPERIMENTATION
Prior to the start of the parametric incineration study on the Baird and
McGuire soil, a series of waste composition/Teachability tests were conducted
with the soil in a laboratory-size muffle furnace to determine the optimum
experimental conditions that might be used for the soil incineration tests.
The muffle furnace tests consisted of nine experiments during which various
concentrations of lead in the contaminated soil were heated at 982°C (1800°F)
for one hour. The weight loss (moisture and volatiles) were determined for
each sample and the resulting ash was analyzed for total lead (Pb).
Uncontaminated clay was mixed with the soil to get variable lead
concentrations in the feed. Toxicity Characteristic Leaching Procedure (TCLP)
samples of the feed as well as the ash were also taken and analyzed for lead
and compared with the total waste/ash analysis. Two percent lime by weight
was added to one of the samples and two percent alum (ferric ammonium sulfate,
FeNH/(S04)2) was added to another of the samples to determine if these
-J ,ives affected the distribution of metals in the ash.
INCINERATOR TESTING
A series of five tests were performed using the RKS at the IRF to
determine the relative partitioning of lead to the different waste streams.
This was a part of the testing series to establish that incineration could
effectively destroy the organic contaminants in the soil and to document the
fate of lead distributions as a function of incineration conditions.
The test variables were kiln temperature and oxygen at the kiln outlet.
Kiln temperature was targeted for 816 and 982°C (1500 and 1800°F) and the
oxygen concentration at the kiln exit flue was targeted at 7 and 10 percent,
(50 percent and 100 percent excess air.)
All the soil was fed to the kiln in fiber-pack drums via the RKS ram
feeder system. The fiber-packs, each of which contained about 5.0 kg (11 Ib)
of soil, were fed at the rate of one every 5 minutes resulting in a feed rate
of 60 kg/hr (132 Ibs/hr). The kiln rotation was set to give a nominal solids
residence time of 0.5 hr.
SAMPLES
The samples taken during each test were:
1. Scrubber blowdown water for lead (grab sample),
2. Feed and kiln ash for total lead analysis (grab sample),
3. 02 concentrations at kiln outlet (continuous analyzer),
4. Stack gas particulate matter, lead and POHCs.
5. Feed and ash for TCLP determinations for lead,
6. Scrubber water for TCLP.
68
-------
TEST RESULTS AND DISCUSSION
MUFFLE FURNACE TEST RESULTS
Baird and McGuire soils with varying concentrations of lead were heated
in the muffle furnace at nominal conditions to determine the effect of lead
feed concentration on the resultant lead ash concentration and Teachability.
TCLP leachates of the feeds and ashes were analyzed for lead to observe the
effects on lead mobility.
The test results are summarized in Table 1. These data show the various
concentrations of lead which were derived by mixing the contaminated soil,
which contained 45 mg/kg of lead with uncontaminated clay.
As the concentration of lead decreased, the data suggest that an initial
lead concentration of 45 mg/kg or less will always have less than 0.05 mg/L
TCLP in this matrix and would be well below the guidance level of 5 mg/L for
the ash. The muffle furnace tests on the original soil with a concentration
of 45 mg/kg of lead had an ash content containing 5.1 mg/kg lead, and a TCLP
of <0.05 mg. As the concentration of soil in the mixture (lead in the
mixture) was lowered, the TCLP remained at <0.05 mg/L despite variations of
lead concentrations of up to 5.8 in the ash.
When 2 percent lime was added to the 100 percent soil (one data point),
the lead concentration was 44 mg/kg. After testing in the muffle furnace, the
lead concentration was 6.9 mg/kg in the ash. This was an increase in the
retention of lead in the ash compared to the 5.1 mg/kg retention in the
untreated soil sample and the TCLP was still <0.05 mg/L.
The addition of 2 percent alum to the 100 percent soil increased the
lead content of the^ash to 7.4 mg/kg. This one point determination suggests
that alum would be an addition to a lead contaminated soil that would increase
the retention of the lead in the ash.
Inci neration Test Results
Five incineration tests were conducted on the contaminated soil from the
Baird and McGuire site. The lead concentration of the soil received was not
adjusted as it was for the muffle furnace tests however the lead concentration
varied from 16 to 27 mg/kg. Table 2 presents a summary of the test conditions
and the concentrations of incinerator variables measured during the tests.
The test results in Table 2 show the conditions of operation. Tests 1,
2 and 5 had target temperatures of 816°C (1500°F). A mean average of the
temperatures achieved for these three tests were 832°C (1530°F), 839°C (1540°F)
and 844°C (1551°F). Tests 3 and 4 had a target temperature of 982°C (1800°F)
and operated at 994°C (1821°F) for both tests. Temperatures for all the tests
were close to the target conditions, and operating temperatures were judged
successful. For ease of presentation, the temperatures 816°C (1500°F) will be
used for Tests 1, 2, and 5, and 982°C (1800°F) will be used for Tests 3 and 4,
69
-------
The mean experimental oxygen concentrations achieved in Tests 1, 3 and 5
were 11,3 percent, 10.4 percent and 11.2 percent. Tests 2 and 4 were at 7
percent and achieved concentrations of 6.8 percent and 7.5 percent.
The resulting distributions of lead in the ash kiln exit gases and
scrubber water are contained in Table 3. The ash generation rate was
determined by using the feed rate less the volatiles and moisture. This table
has been arranged by temperature of operation. Tests 1, 2 and 5 are the 816°C
(15000F) and Tests 3 and 4 are the 982°C (1800°F) operating condition. Test 5
was a duplicate of Test 1 to check the analytical data precision.
On the basis of lead retention in the ash, incinerating at 816°C
(1500°F), at 50 percent excess air (Test 2) appeared to give the best
operating condition of the five tests (Figure 1). The retention of the lead
in the ash was 151 percent of the lead feed based on the amount of lead/hr in
the ash versus the feed. The emissions to the atmosphere were 5.7 percent of
the feed (Figure 2). Considering the scrubber waste catches 4.3 percent of
the lead feed and the emissions to the atmosphere 5.7 percent, then the total
emissions before the ionized wet scrubber were 10 percent of the feed (Figure
3). Operating the incinerator at the lower temperature and lower oxygen rate
increased the retention of the lead in the ash and reduced the emissions to
the atmosphere.
Operation at 816°C (1500°F) and 11 percent oxygen was the second best
operating condition that partitioned lead to the ash. The lead retained in
the ash (Tests 1 and 5) averaged 11.4 percent. Atmospheric lead emissions
were 7.1 percent of the feed lead. If the lead emissions to the atmosphere
are the important consideration, rather than retention in the ash, then high
air flow, 11 percent (100 percent excess air) is the better condition
operation. However, the total emissions were 11.5 percent before the
abatement system, indicating that if the air pollution abatement system
operated at the same efficiency at both the higher and lower temperature of
operation, then this was not the best condition of operation.
Incinerating the soil at 982°C (1800°F) indicated that retention of lead
in the ash was almost the same at both oxygen concentrations; 7 percent oxygen
resulted in 34 percent lead retention in the ash and 11 percent oxygen
resulted in 33 percent retention in the ash. The lead emissions were 8.9
percent of the feed at 7 percent oxygen concentration and 6.7 percent at 11
percent oxygen concentration.
The feed and ash, in all tests, had a TCLP value of less than 0.05 mg/L
(the detection limit). A more practical method for summarizing the data for
lead retention in the ash is by decreasing the calculated lead input of the
soil by the scrubber water lead content and the emissions. This indicates
that the ash retention averaged 89 percent of the lead input at 816°C (1500°F)
operating temperature and 83 percent at 982°C (1800°F). These figures do not
change the conclusions made from the analytical results, but it does reduce
the lead retention results of over 100 percent in the ash.
70
-------
The data which indicate that lead is retained in the ash in excess of
100 percent are in line with pilot scale and full scale data previously
obtained for lead. Emissions of lead from sewage sludge incineration range
from less than 1 percent to 20 percent; the 20 percent being when no pollution
control system was used (3).
The EPA Superfund Innovative Technology Evaluation (SITE) program (4)
tested an innovative incineration system where 18 incineration tests with a
feed containing lead were performed. The operating temperatures and oxygen
were varied similar to the test conditions reported in this paper (4).
Results of these tests indicate that 13 out of 18 tests had lead retention
values of greater than 100 percent of the feed.
Metal testing was also done at the IRF (5) with a synthetic mixture
representing a Superfund type soil. The purpose of these tests was to
determine the partitioning of the metals of which lead was included. All nine
tests were conducted at operating conditions of about 900°C (1652°F) and about
a 100 percent excess air at the incinerator outlet. These tests indicated
over 100 percent recovery for lead in three of the nine tests: The average
lead retention in the ash was 91.8 percent, the lowest retention 44.5 percent
and the highest lead retention was 132.8 percent.
Figure 1 shows the calculated results of the lead retention at all
operating conditions. At 6 percent oxygen, 50 percent excess air, and 816°C
(1500°F), 151 percent of the lead was retained in the ash. At 6 percent
oxygen and 982°C (1800°F), 34 percent of the lead in the feed was retained in
the ash. At 11 percent excess oxygen, an average of 114 percent (Tests 1 and
5) retention of lead in the ash was attained at 816°C (1500°F), and 33 percent
of the lead was retained at 982°C (1800°F).
Although mass balances of over 100 percent lead in the incinerator
effluents cannot happen, the lead emissions from the system at the operating
point of 816°C (1500 F) and 7 percent oxygen (50 percent excess air flow)
appeared to be the best operating condition for incineration of a lead
contaminated soil if it is desired to have increased lead retention in the ash
and the lowest stack emissions of lead to the atmosphere.
CONCLUSIONS
1. The greatest amount of lead retained in the ash was (151 percent) at
816°C (1500°F) and 7.5 percent oxygen (about 50 percent excess air
flow).
2. Operating conditions of the lower temperature 816°C vs. 982°C (1500°F
vs. 1800 F) increased lead retention in the ash.
3. Reduced oxygen concentration (lower air flow rates) gave increased
retention of the lead regardless of the operating temperature.
4. The emissions were between 6 and 8 percent of the feed (relatively flat)
at 11 percent excess air regardless of temperature.
71
-------
5. The emissions ranged from about 6 percent at 816°C (1500°F) to about
9 percent at 916°C (1800°F) when at 7 percent oxygen (50 percent excess
air),
6. The TCLP of lead in both the feed and ash was lower than the detection
limit 0.05 mg/L which was well below the 5 mg/L guidance limits.
REFERENCES
1. Acurex Corp., "Pilot-Scale Incineration of Arsenic-Contaminated Soil
From the Baird and McGuire Superfund Site," USEPA Contract No.
68-C9-0038, Work Assignment 0-5, Cincinnati, Ohio, EPA Project Officer:
R. C. Thurnau, Technical Project Manager: M. K. Richards, March 1990.
2. Howard 0. Wall and Marta K. Richards, "The Incineration of Arsenic -
Contaminated Soils Related to the Comprehensive Environmental Response,
Compensation and Liability Act (CERLA), RREL 16th Annual Hazardous Waste
Research Symposium, Cincinnati, Ohio. April 3-5, 1990.
3. Gerstle, Richard W. and Diana Albrinck, "Atmospheric Emissions of Metals
From Sewage Sludge Incineration," Journal of the Air Pollution Control
Association, Volume 32, No. 11, Pages 1119-1123, November 1982.
4. EPA/540/5-89/007A, Technology Evaluation Report "SITE Program
Demonstration Test Shirco Pilot-Scale Infrared Incineration System at
Rose Township Demode Road Superfund Site," Volume I, Howard Wall, EPA
SITE Program Manager, April 1989.
5. Acurex Corporation "Pilot-Scale Evaluation of the Fate of Trace Metals
in a Rotary Kiln Incinerator with Ionizing Wet Scrubber, USEPA
Cincinnati Contract No. 68-C9-0038, Work Assignment 0-3, Project
Officer: Robert C. Thurnau, Technical Task Manager:
Gregory J. Carroll, Cincinnati, Ohio, April 1990, Draft Report.
72
-------
200
% Pb RETENTION IN ASH
180-
160-
140-
120 -
100-
80-
60-
40-
20
FEED BASED
(7% Oxygen)
(11% Oxygen)
0 i 1 1 1 1 1 1 1 1 r
800 820 840 860 880 900 920 940 960 980 1000
KILN TEMPERATURE, C
7% OXYGEN
11% OXYGEN
Figure 1. Lead retention in ash.
73
-------
10
% Pb EMITTED
9-
8-
7-
6-
5-
4-
3-
2-
1 -
0
(7% Oxygen)
(11% Oxygen)
800 820 840 860 880 900 920 940 960 980 1000
KILN TEMPERATURE, C
7% OXYGEN
11% OXYGEN
Figure 2, Lead emission in stack.
74
-------
100
% Pb RETENTION IN ASH
95 -
90-
85-
80
75 -
EMISSIONS BASED
(7% Oxygen)
(11% Oxygen)
70 1 1 1 T i 1 r
800 820 840 860 880 900 920 940 960 980 1000
KILN TEMPERATURE, C
7% OXYGEN
11% OXYGEN
Figure 3. Lead retention in ash.
75
-------
TABLE 1. MUFFLE FURNACE TEST RESULTS
ction
minated
oil
%
100
80
60
50
40
20
0
100
100
Fraction
Background
Soil
%
0
20
40
50
60
80
100
(2% lime, added
(2* alum added
Calculated
Concentration
Soil
mg/kg
lead '(Pb)
45
39
33
30
26
20
14
) 44
} 44
Ash
Concentration
c
mg/kg
lead (Pb)
5.1
5.1
5.5
4.3
5,2
5.8
3.7
6.9
7.4
TCLP
leacha
roncentra
mg/L
<0.05
<0.05
<0.05
<0.05
<0.05
<0.05
<0.05
<0.05
<0.05
76
-------
Table 2. Operating Conditions and Lead Results
Test
Date
Kiln exit temperature, °C
Kiln exit temperature, °F
Length test, hours
Feed rate, kg/hr
Scrubber water recirculating
tank volume, liters*
Kiln exit oxygen, percent
Scrubber blowdown, L/min
Average flue gas flow rate,
acm
dsni
acfm
Loss on ignition, percent
(moisture and organics)
Lead concentrations
Feed (mg/kg)
Feed TCLP (mg/L)
Ash (mg/kg)
Ash TCLP (mg/L)
Scrubber blowdown (mg/L)
IWS exit flue gas
1
(9-26-89)
832
1529
3.17
54
1425
11.3
1.90
56
31.3
450
17
21
<0.05
31
<0.05
0.09
0.053
2
(9-29-89)
844
1552
3.16
55
1425
6.8
1.90
41.7
22.4
429
17
16
<0.05
30
<0.05
0.07
0.045
3
(9-27-89)
994
1822
2.92
56
1425
10.4
1.90
56
28.4
199
17
27
<0.05
11
<0.05
0.11
0.061
4
(9-28-89)
994
1822
3.16
56
1425
7.5
1.90
55.6
27.6
610
17
17
<0.05
7
<0.05
0.11
0.050
5
(10-5-89)
839
.1541
3.83
56
1425
11.2
1.90
54.6
29.3
407
17
20
<0.05
26
<0.05
0.08
0.037
(mg/dscm)
Total emissions, kg/hr 0.017 0.002 0.015 0.002 0.031
*There is a precipitate in the scrubber water, quantity and identity not determined.
77
-------
TABLE 3. INCINERATOR TEST RESULTS
lead retention
Total Emissions
CO
Test
1
5
2
3
4
Incinerator exit
kiln tenp.. °C
816 (1500°F)
816 (1500°F>
816 (1500°F)
982 (1800°F)
982 (1800°F)
% oxygen content
of flue gas at
kiln exit
11
11
7
11
7
Lead retention
in ash, X of
feed mass
calculated
122
106
151
33
34
in ash, %
Calculated
based on
Emissions
87.3
90.3
90
81
64.5
Lead retention
in scrubber
water, % of
feed massT
4.6
3.5
4.3
2.3
6.6
Lead in stack
gas. X of feed
8.1
6.2
5.7
6.7
8.9
Lead balance
% of feed
135
116
161
42
49
before Scrubbe
(Scrubber and
Sis)
X of feed
12.7
9.7
10.0
9.0
15.5
-------
FULL-SCALE POHC INCINERABILITY RANKING AND SURROGATE TESTING
by: Mr. Andrew Trenholm
Midwest Research Institute
Gary, NC 27513
Dr. C. C. Lee
U.S. Environmental Protection Agency
Risk Reduction Engineering Laboratory
Cincinnati, OH 45268
Capt. Helen Jermyn
U.S. Air Force, HQ AFESC
Tyndall Air Force Base, FL 32403
ABSTRACT
The U.S. Environmental Protection Agency, Office of Research and
Development (EPA/ORD), and the Headquarters Air Force Engineering and Services
Center (HQ AFESC) are interested 1n evaluating the incinerability ranking of
principal organic hazardous constituents (POHC's) and surrogate compounds that
are used for destruction and removal efficiency (ORE) tests at hazardous waste
incinerators. The results of these evaluations will aid development of a more
effective and cost-efficient trial burn and performance monitoring process.
As a part of the evaluation process, a test was conducted at a full-scale
hazardous waste incinerator.
There were two test objectives. The first was to evaluate an
incinerability ranking system commonly used by EPA. This system ranks organic
compounds based on the gas-phase thermal stability under oxygen-starved
conditions and is based on available experimental data. The second objective
was to evaluate sulfur hexafluoride (SF6) as a surrogate for POHC destruction.
Sulfur hexafluoride is one of the most stable compounds known with respect to
thermal decomposition. Thus, it has been hypothesized that the SF^ DRE would
represent a lower bound to other POHC DRE's in an incinerator. This paper
presents the results of the test relative to the two test objectives.
This paper has been reviewed in accordance with the U. S. Environmental
Protection Agency's peer and administrative review policies and approved for
presentation and publication.
79
-------
INTRODUCTION
The EPA/ORD and HQ AF£SC contracted to Midwest Research Institute (MRI)
to evaluate the incinerability ranking of POHC's and surrogate compounds that
are used for ORE tests (trial burns) at hazardous waste incinerators. The
incinerability ranking evaluated is a thermal stability system developed for
EPA by the University of Dayton Research Institute (UDRI). The surrogate
compound evaluated is SF*. The results of these evaluations should aid
development of a more effective and cost-efficient trial burn and performance
monitoring process.
Incinerability of POHC's has been measured by a variety of ranking
systems, the most common being based on the heat of combustion (He) of POHC's,
with lower He indicating more difficult destruction. The method that has best
correlated with field data is the thermal stability ranking system developed
by UDRI for EPA. This system ranks POHC's based on their gas-phase thermal
stability under oxygen-starved conditions. The ranking is based on a large
amount of laboratory-scale experimental data. The experimental data is
extrapolated through comparisons of compound structure and properties to
obtain a ranking for all POHC's. Evaluation of this ranking system was one
objective of this study.
The second objective of this study was to evaluate SF, ORE as a
conservative indicator for the ORE of POHC's. Since promulgation of the
hazardous waste incinerator performance standards in January 1981, there has
been a continuing Interest in a real-time surrogate compound to measure
incinerator compliance with the DRE performance standard. One such possible
surrogate is SF6. Sulfur hexafluoride is one of the most stable compounds
known with respect to thermal decomposition. Thus, it has been hypothesized
that the SF6 DRE would represent a lower bound to other POHC DRE's.
Furthermore, SF6 can be measured at very low concentrations in stack gas on a
real-time basis using onsite gas chromatpgraphic techniques. For these
reasons, using SF6 injection with real-time DRE measurement represents a
potentially attractive surrogate to determine compliance with DRE performance.
Previous field tests conducted to gather data on SF* destruction in hazardous-
waste combust.ion devices have shown the potential or this compound as a
surrogate for POHC destruction, but have not answered all concerns about its
use.
The remainder of this paper presents the approach to achieving the
experimental objectives, a discussion of the project results, and brief
conclusions.
APPROACH
Conducting a full-scale experimental test requires identifying and
obtaining permission to test from an operating hazardous waste incinerator.
Eastman-Kodak Company's chemical waste incinerator located in Rochester, New
York, was identified as a suitable site, and permission was obtained to
conduct the test. This incinerator was representative of many currently
operating hazardous waste incinerators and was amenable to the variety of
80
-------
sampling and waste spiking necessary to achieve the project objectives. Kodak
personnel provided invaTuable flexibility and assistance in operating the
incinerator during the test.
The Kodak incineration system treats a variety of combustible liquid and
solid wastes generated at the Kodak Park manufacturing site in Rochester. It
consists of a rotary kiln, mixing chamber, and secondary combustion chamber,
followed by a quench chamber and venturi scrubber. The thermal capacity of
the kiln and secondary chamber are a nominal 90 million British thermal units
per hour (Btu/h).
An overview of the design of the test matrix conducted at this
incinerator is presented 1n Figure 1. The test was conducted over four days,
with the incinerator operating at two different conditions. Condition A and
Condition B. Each test condition included 12 sampling periods. Test
Condition A entailed operating the incinerator at a high-temperature, low-
oxygen condition while firing liquid waste to the kiln. Test Condition B
involved operating the incinerator at a low-temperature, high-oxygen condition
while firing only liquid and solid wastes to the kiln. The spread in
temperature and oxygen concentration between the two conditions was maximized
to the extent possible while remaining within acceptable combustion
conditions. Temperature was the primary independent variable, while oxygen
levels varied as necessary to achieve the desired temperatures.
Test Day
Measured
Parameter
VOST and
SF6 Samples
Stack Velocity,
Water, Temperature
Measurement
Waste
Grab Samples
SF6 FeeeJ
Location
Solid Waste
Liquid Waste
Process
Condition
A. :ii Temp/Low Oj
B. Lo Temp/Hi Oj
Contingency,
• Change
Process
Conditions
• Contingency
•. — Represent sample data points
Figure 1. Overview of test matrix.
The wastes used during the test included two liquid organic waste feeds
to the kiln and secondary chamber and a surrogate solid waste feed to the
kiln. The physical and chemical characteristics of each stream were
maintained as consistent as possible throughout the test. Liquid wastes were
composed primarily of waste solvent compounds, which were blended to as
homogeneous a feed as possible. Solid wastes were bulk sawdust, contained in
81
-------
SB-gallon (gal) flberpaks.
used during the tent.
Some auxiliary fuel (No. 2 diesel fuel) was also
Six volatile POHC's were spiked to or were native to the liquid waste
stream, and one volatile POHC was spiked to some of the solid waste drums.
The spiked liquid waste was fed via the kiln liquid waste burner. Table I
lists the spiking compounds, their thermal stability classification, and the
matrix into which they were spiked. These compounds were chosen to represent
the five highest of the seven class rankings. Class 1 is the most thermally
stable.
TABLE 1. TARGET POHC COMPOUNDS
Compound
Monochl orobenzene
Methyl ene chloride
Tetrachl oroethyl ene
Methyl ethyl ketone
Chloropropene
1 , 1 , 1-Trichloroethane
Toluene
UDRI class
1
2
2
3
4
5
2
Spiking matrix
Liquid waste
Liquid waste
Liquid waste
Liquid waste
Liquid waste
Liquid waste
Solid waste
The SF6 was spiked into both the liquid and solid wastes during different
portions of the test. During days 1, 2, and 4, SF6 was spiked into the liquid
waste fed to the kiln. During day 3, SF6 was spiked alternately into the
solid and liquid waste streams as defined in Figure 1. The SF< spiked to the
solid waste was microencapsulated, the capsules packaged Into ISO-mill 11iter
(ml) plastic bottles, and the bottles placed into drums of waste fed to the
Incinerator. These same drums were the ones spiked with toluene. The SF6
spiked to the liquid waste was injected as a gas into the liquid waste feed
line. The injection point was located upstream of the kiln burner nozzle.
Measurements of the emissions of POHC's and SF6 were made in a transition
duct located between the secondary combustion chamber and the quench. A
volatile organic sampling train (VOST), as described in SW-846 Method 0030,
was used to collect samples of the POHC's. These samples were analyzed by gas
chromatography/mass spectrometry (GC/MS) using selected 1on monitoring (SIM).
The SIM analysis provided a very low detection limit that allowed
quantification of very high DRE's. The SF, samples were collected 1n tedlar
bags and analyzed onsite with a gas chromatograph equipped with an electron
capture detector. Composite samples of the waste feeds were collected for
analysis of POHC's by GC/MS.
RESULTS
This section presents a brief discussion of the process operation
followed by the project results. The results are organized by the two project
objectives described earlier.
82
-------
PROCESS OPERATION
Process operation was replicated closely from day to day during the test
except for the planned variations in waste feed rates and combustion
temperature. Table 2 shows the average values for the key process variables
for each of the two test conditions. The waste feed rates were changed to
achieve the desired combustion temperature change of about 200*F between test
conditions.
TABLE 2. KEY PROCESS OPERATING DATA
Parameter
Kiln
Liquid waste feed rate
Solid waste feed rate
Temperature
Combustion air flow
sec
Liquid waste feed rate
Temperature
Combustion airflow
SCC exit
Oxygen
Carbon monoxide
Units
Ib/h
Ib/h
*F
acfm
Ib/h
•F
acfm
%
ppm
Condition A
1,760
0
1730
7,200
2,350
1920
4,330
15.1
4
Condition B
1,240
830
1510
7,220
2,130
1,700
3,180
15.7
1
The concentration of each individual POHC in the waste feeds was kept
relatively constant throughout the test. The concentration varied between
POHC's from 2 to 25 percent, depending on the amount that was native to the
waste. Sulfur hexafluoride was fed at a rate of 2 to 10 pounds per hour
(Ib/h) in the liquid waste feed line and 0.15 Ib/h in the solid waste.
POHC INCINERABILITY RANKING
This section presents the data on POHC DRE's gathered during this study
and an evaluation of the ranking of these DRE's compared to the ranking
predicted by EPA's thermal stability ranking system. One of the seven POHC's
selected for the study (methyl ethyl ketone) was present in the waste feeds at
unexpectedly low concentrations; therefore, the following discussion addresses
only the other six POHC's.
Table 3 shows the average DRE's for each of the two process operating
conditions. These two conditions differed principally in that Condition A had
combustion temperatures about 2QO*F higher than Condition B. The ORE data
show little difference between these two conditions, with the possible
exception of methylene chloride. Methylene chloride DRE's appeared to be
slightly higher for Condition A. In general, daily averages differed as much
as the condition averages, and the differences were within expected sampling
and analysis accuracy. Because these differences between the conditions were
83
-------
small and appeared random, data for both conditions were combined for the
evaluation of the ranking discussed below.
TABLE 3. AVERAGE POHC ORE'S
POHC
Chlorobenzene
3-Chloropropene
Methyl ene chloride
Tetrachloroethene
Toluene
1,1, 1-Tri chl oroethane
Average ORE, percent
Test Condition A
99.999955
99.99984
99.999969
99.99988
99.999985
99.999991
Test Condition B
99.999923
99.999904
99.99990
99.99986
99.999991
99.999991
Figure 2 displays the ORE results compared to the predicted thermal
stability ranking. The measured DRE's are shown on the figure as the range of
individual sample values, the middle 50 percent of the values (boxes on the
figure), and the median values for each POHC. The data are arrayed on the
figure by POHC with data for the hardest-to-destroy POHC to the left,
increasing to easier-to-destroy POHC's to the right. The thermal
stability/incinerability index is shown at the bottom of the .figure. If the
ranking of the measured DRE's matched the ranking predicted by the thermal
stability index, the DRE's should increase from POHC to POHC from left to
right on the figure.
Considering only the median value for each POHC, four of the six POHC's
ranked correctly relative to one another. The two that did not rank as
predicted, chlorobenzene and toluene, were easier to destroy than predicted.
Considerable overlap occurs, however, in the ranges of values for each POHC.
This data scatter is greater than typically observed for VOST trial burn
results for two reasons. First, each data point is the analytical result for
a single pair of traps, while trial burn results are expressed as an average
of several pairs of traps. This averaging reduces the apparent scatter in the
data. Second, the results of the SIM analysis had more scatter than typical
VOST analysis results because of the extremely low levels that were
quantified.
The measured ORE values were subjected to nonparametric (rank order)
statistical tests for further evaluation. The statistical tests showed a high
degree of consistency in the measured ranking from sample to sample throughout
the test. They also confirmed the discussion above relative to the agreement
with the predicted ranking.
84
-------
KXXOOOOO
99.99995-
^ 99.99990-
o^
0)
| 99.99985-j
UJ
* 99.99980
99.99975
99.99970
F
[
1399.999990 =
199.999904
'• 99.999901
] 99.999952
1 ffi 99.999993
99.999959
Tho box encloses
Ihe middle 50% of
Ihe ORE values
and the median
ralue is given.
19 35 36 65-66 120 201
(Chloro- (Toluene) (tetra- (Methylene (3-Chloro- (1,1,1-
benzene) chloro- chloride) propane) Trichloro-
ethylene) ethane)
Thermal Stability Incinerability Index
Figure 2. ORE for each POHC.
Two factors that related to how the data were reduced were investigated
to determine if they affected the observed ranking of the POHC's. These
factors were blank correction of the data and the relationship between POHC
concentration in the waste and ORE. Blank corrections are discussed first
below.
Because of the low levels of detection achieved with the SIM analysis,
blank trap levels were a problem for some compounds, most noticeably for
toluene, 3-chloropropene, and 1,1,1-trichloroethane. When blank levels are
significant, they bias the concentration and emission rates high and the DRE's
low. To check the effect on ranking of the calculated DRE's, a simple
subtraction of blank levels was used to correct the data. Figure 3 shows the
unadjusted and blank-corrected DRE's. Blank correction increased all the
DRE's, but the ranking trend did not change.
Figure 3 also shows the feedrate-normalized DRE's. An earlier study
identified the effect of POHC concentration in the waste feed on measured
DRE's. Destruction and removal efficiency increases as the POHC concentration
increases. A similar relationship was found using the data from this study,
which was used to calculate the feedrate-normalized DRE's plotted in Figure 3.
This correction of the calculated DRE's increased some values, decreased
others, and did not change others. However, as happened for the blank-
corrected DRE's, the ranking trend did not change.
Figure 3 also shows DRE's calculated considering both of the above
corrections together. Again, the ranking trend is the same. The effect of
these corrections, however, was to decrease the differences in ORE from POHC
to POHC. The lowest and highest corrected DRE's differed by a factor of
only 20.
85
-------
JOO.OOOOOO-
9S.999990-
99,999980-
99,999970-
s?
*•*• 99,999960-
J
§ 99,999950-
uu
DC
Q 99.999940-
99.999930-
99.999920-
99,999910-
99.999900-
r
,
I
199.999991 .
' 99.999985 (
199.999969 '
-99.999964
199.999997
-99.999990
199.999986
t
-99.999969
;
1
1
199,999979
I
:99.999964 •
-
J99.999908
r
199.999982 ;
199.999972
=99.999964
-99.999955
t
399.999994 -j
199.999985
-=99.999982
-99.999959
199.999999
;99.999993
Ctiloro- Toluene Telrachtoro- Metylene 3-CtiIoro- 1,1,1-Tti-
bonzcne etliylena Chloride propene chloro-
athane
— Unadjusted ORE
11 Blank CorrcclHd ORE W Feedrate Normalized ,C?3 Blank Corrected and
ORE Fecdrate Normalized ORE
Figure 3. Comparison of corrected ORE values.
SF6 DESTRUCTION
The study objective related to SF6 focused on two questions about the use
of this compound as a surrogate compound to measure POHC ORE. The first
question concerned the effect on SF, ORE of the method of introducing the SF6
to the incinerator. Conceivably, the method of introduction could affect what
combustion conditions the SF6 experiences. Historically, studies of SF. as a
surrogate compound have introduced the SF6 by spiking it into the liquid waste
line or combining it with the combustion air. No study was identified where
SF6 was introduced with solid waste.
During one day of the test, SF$ was spiked alternately into the solid and
liquid waste streams. For the first half of the day, SF6 was spiked as a gas
into the liquid waste feed line. During the second half of the day, the
microencapsulated SF6 was placed into the drums of solid waste to be fed to
the incinerator. No other process parameters were changed on this day. The
average DRE's measured when the SF6 was fed with the liquid and solid wastes
were 99.989 and 99.986 percent, respectively. Thus, the method of feeding SF6
into the incinerator via liquid or solid waste to the kiln did not affect the
ORE.
The second question v/as whether SF6 gives a conservative value of ORE
relative to the POHC DRE's, Earlier studies have generally shown that SFx ORE
is conservative (has a lower ORE) relative to POHC DRE's. Figures 4 and 5
show that the data from this study confirm the earlier studies. Figure 4
shows that the SF- DRE during Condition A (the higher-combustion-temperature
condition, where the highest SF6 DRE's were measured) was an order of
86
-------
magnitude or more lower than the POHC DRE's. Figure 5 shows the sime result
during Condition Bt but the SF6 DRE's were even more conservative.
EC
Q
100.0000-
QQ QQQU-
yy.yyyo
QQ QQQft-
yy.yyyo
QQ QQQA-
yy,yyy*+
99.9992-
99.9990-
QQ QQR«-
E
P
] 99.999770
!)99.999964"~ "•999990|
1 C
99.999901 =
?99.999952f
f
ihe i
and \
- are fi
jox encloses the middle 50% of the DRE values
he median value is given. The SF6 DRE values
om Condition A only.
9 S99.999993
J99.999959
4 19 35 36 65-66 120 201
(SF6, (Chloro- (Toluene) (Tetra- (Methylene (3-Chloro- (1,1,1-
Aonly) benzene) chloro- chloride) propene) Trichloro-
ethylene) ethane)
Thermal Stability Incinerability Index
Figure 4. SF6 and POHC DRE values.
3.
cc
100.000
99.9954
99.990
99.985
99.980-
99.975
99.970
Condilion A
99.999770 99,999954 • 99.999990 99.9999O1 99.999952 99,999959 99.999993
iCondillon B
- 99.993816
I he value given is the median UHt lor tne PUhCs in
both conditions and for SF6 during the specified condition.
SF6 Chloro- Toluene Tetra- Methylene 3-Chloro- 1,1,1-
benzene chloro chloride propene Trichloro-
-ethylene • ethane
Figure 5. SF6 and POHC ORE values.
87
-------
As can be seen on Figure 5, the SF< DRE's measured during Condition B
were considerably Tower than the SF, DRE's measured during Condition A. These
data indicate a dependence of SF6 ORE on combustion temperature. Combustion
temperatures during Condition A were 1730*F and 1920*F in the kiln and
secondary chamber, respectively. During Condition B, they were 1510*F and
1700*F, respectively. Thus, the SF6 ORE decreased from 99.9998 to 99.994
percent, with a 20Q'F drop in combustion temperature. Earlier discussions
showed that a similar dependence on combustion temperature was not observed
for the POHC's. As a result, SF6 becomes a more conservative indicator as
combustion temperature decreases.
CONCLUSIONS
Below are brief statements of the primary conclusions from this study.
1. The ORE for POHC's ranked highest by the thermal stability ranking system
(those easiest to destroy) generally followed the ranking order predicted
by the system.
2. The DRE's for the two POHC's studied that were ranked lowest by the
system (chlorobenzene and toluene) did not agree with the predicted
ranking order. They were easier to destroy than predicted.
3. The 200*F change in combustion temperature between test conditions did
not affect the POHC DRE's or the observed ranking order.
4. The measured ranking order was very consistent from test period to test
period (sample to sample).
5. The ORE measured for SF6 was lower than the DRE's for all the POHC's
studied in this project; thus, SF6 was a conservative indicator of POHC
ORE.
6. The ORE for SF6 had a distinct dependence on combustion temperature. A
decrease in temperature of about 200" F caused a decrease in SF6 ORE of
about one order of magnitude.
7. The method of feeding SF< to the incinerator (in the liquid waste versus
solid waste feed to the Kiln) did not-affect the ORE.
88
-------
EPA's MOBILE VOLUME REDUCTION UNIT FOR SOIL WASHING
by: Hugh Masters
Releases Control Branch
Risk Reduction Engineering Laboratory
US Environmental Protection Agency
Edison, New Jersey 08837
Bernard Rubin
Roger Gaire
Porfirio Cardenas
Foster Wheeler Enviresponse, Inc.
Livingston, New Jersey 07039
ABSTRACT
This paper discusses the design and initial operation of the U.S.
Environmental Protection Agency's (EPA) Mobile Volume Reduction Unit (VRU) for
soil washing. Soil washing removes contaminants from soils by dissolving or
suspending them in the wash solutions (which can be later treated by conven-
tional wastewater treatment methods) or by volume reduction through simple
particle size separation techniques. Contaminants are primarily concentrated
in the fine-grained (<0.063 mm, 0.0025") soil fraction. The VRU is a pilot-
scale mobile system for washing soil contaminated with a wide variety of heavy
metal and organic contaminants. The unit includes state-of-the-art washing
equipment for field applications.
The VRU equipment was originally conceived by the EPA. It was designed
and fabricated by Foster Wheeler Enviresponse, Inc. under contract to EPA's
Risk Reduction Engineering Laboratory (RREL) in Edison, New Jersey, with the
following objectives:
1. To make available to members of the research community and to the
commercial sector the results of government research on a flexible,
89
-------
multi-step, mobile, pilot-scale soil washer capable of running treatabi-
lity studies on a wide variety of soils; ' '"'.,..
2. To demonstrate the capabilities of soil washing; and
3. To provide data that facilitate scaleup to commercial size equipment.
The design capacity of the VRU is 100 Ib/hr of soil, dry-basis. The VRU
consists of process washing equipment and utility support services mounted on
two heavy-duty semi-trailers. The process trailer equipment accomplishes"
material handling, organic vapor recovery, soil washing, coarse soil screen-
ing, fine particle separation, flocculation/clarification, and steam genera-
tion via a boiler. The utility trailer carries a power generator, a process
water cleanup system, and an air compressor. The VRU is controlled and
monitored by conventional industrial process instrumentation and hardware.
Shakedown operations are currently in progress and future plans include
testing EPA-produced synthetic soil matrix (SSM) spiked with specific chemical
pollutants. The addition of novel, physical/chemical treatment processes,
such as sonic/ultrasonic cleaning and acid leaching, will expand the VRU's
extraction capability in soil decontamination.
90
-------
INTRODUCTION
Section 121(b) of the Comprehensive Environmental Response, Compensa-
tion, and Liability Act (CERCLA) mandates the EPA to select remedies that
"utilize permanent solutions and alternative treatment technologies or
resource recovery technologies to the maximum extent practicable" and to
prefer remedial actions in which treatment "permanently and significantly
reduces the volume, toxicity, or mobility of hazardous substances, pollutants,
and contaminants as a principal element."
In most cases soil washing technologies are used in conjunction with
other remedial methods for the separation/segregation and volume reduction of
hazardous materials in soils, sludges, and sediments. In some cases, however,
the process can deliver the performance needed to reduce contaminant con-
centrations to acceptable levels and, thus, serve as a stand-alone technology.
In treatment combinations, soil washing can be a cost-effective step in
reducing the quantity of contaminated material to be processed by another
technology, such as thermal, biological, or physical/chemical treatment. In
general, soil washing is more effective on coarse sand and gravel; it is less
successful in cleaning silts and clays.
A wide variety of chemical contaminants can be removed and/or con-
centrated through soil washing applications. Removal efficiencies depend on
both the soil characteristics (e.g., soil geology and particle size) and the
processing steps contained within the soil washer. Experience has shown that
volatile organics can be removed with 90+% efficiency. Semivolatile organics
are removed to a lesser extent (40-90 percent). They usually require the
addition of surfactants to the wash water. Surfactants are surface-active or
wetting agents, that reduce the surface tension at the interface between the
hydrophobic contaminants and the soil, thereby promoting release of the
contaminants into the aqueous extraction medium.
Metals which are less soluble in water, often require acids or chelating
agents for successful soil washing. A chelating agent, such as ethylenedi-
aminetetraacetic acid (EDTA), bonds with the metal and facilitates
91
-------
solubilization in the extraction medium.
The VRU process can be applied to the treatment of soils contaminated
with hazardous wastes such as wood-preserving chemicals (pentachlorophenol,
creosote), electroplating residues (cyanides, heavy metals), organic chemical
production residues, and petroleum/oil residues. The applicability of soil
washing to general contaminant groups and soil types is shown in Table 1.
This table has been reproduced from an EPA report, "Treatment Technology
Bulletin - Soil Washing," dated May 1990.
The EPA has developed the VRU to meet the following objectives:
1. To make available to members of the research community and to the
commercial sector the results of government research on a flexible,
multi-step, mobile, pilot-scale soil washer capable of running treatabi-
lity studies on a wide variety of soils;
2. To demonstrate the capabilities of soil washing; and
3. To provide data that facilitates scaleup to commercial size equipment.
The EPA plans to investigate other extraction processes which may be
added to the VRU at a later data. The addition to the VRU of novel physi-
cal/chemical treatment processes, such as sonic/ultrasonic cleaning and acid
leaching, will expand its overall extraction capability in soil decontamina-
tion.
SYSTEM DESCRIPTION
The VRU is a mobile, pilot-scale washing system for stand-alone field
use in cleaning soil contaminated with hazardous substances. The VRU is
designed to decontaminate certain soil fractions using state-of-the-art
washing equipment. The total system consists of process equipment and support
Utility systems mounted on two heavy-duty, semi-trailers.
92
-------
TABLE 1. APPLICABILITY OF SOIL WASHING TO GENERAL
CONTAMINANT GROUPS FOR VARIOUS SOILS
Contaminant Groups
u
c
o
O
Inorgank
1
«M
ll
V
Q
Halogenated volatiles
Halogenated semivoiatiies
Nonhalogenated volatiles
Nonhaiogenated semivoiatiies
PCBs
Pesticides (halogenated)
Dioxins/Furans
Organic cyanides
Organic corrosives
Volatile metals
Nonvolatile metals
Asbestos
Radioactive materials
Inorganic corrosives
Inorganic cyanides
Oxidizers
Reducers
Good to Excellent Aooiicabilitv: Htahc
successful
Matrix
Sandy/ SOty/Chy
Gravelly Soils Soik
m
V
m
V
T
V
V
V
V
m
m
Q
V
V
V
V
V
V
V
V
T
T
V
V
T
T
T
T
a
V
V
T
V
V
•robabiHty that technology will b«
Moderate to Marginal Applicability. Extras* care in choc
Not Applicable: Expert opinion that technology wtU not
Ming technology
work
93
-------
Figure 1, General Block Diagram, shows the VRU basic pilot plant
subsystems as follows:
1. Soil handling and conveying
2. Organic vapor recovery
3. Soil washing and coarse screening
4. Fines/floatables gravity separation
5. Fines flocculation/water clarification and solids disposal
6. Water treatment
7. Utilities - electric generator, steam boiler, and compressed air unit
The generator, air compressor, water heater, water filters/carbon
adsorbers, recycle water pump, gasoline tank (for the generator) and deli sting
tank are located on the utility trailer. All remaining equipment is located
on the process trailer. The VRU system is controlled and monitored by conven-
tional industrial process instrumentation and hardware, including safety
interlocks, alarms, and shutdown features.
PROCESS DESCRIPTION
Figures 2, 3, and 4 present the Process Flow Diagram for all VRU
subsystems in terms of their process equipment functions.
1. SoilHandlingandConveying
Raw soil is delivered from battery limits to a vibrating grizzly that
separates the particles greater than +f" into a drum for redeposit and
collects the smaller particles (-!" +0) for transfer to the feed surge
bin. (One half-inch is the maximum particle size that can be handled in
the mini-washer, but smaller screen sizes may be selected.) From this
bin, the -|" soil is conveyed through a steam-jacketed screw conveyor
where the volatile organics and water are vaporized. Both live steam
and jacketed steam can be introduced so that the efficiency of the steam
extraction can be determined. The conveyor flow is adjusted by a speed
94
-------
Raw
contaminated
soil
>
i
,2-
Organic vapor
recovery
Fines To posttreatment
FT eatables To posttreatment _
i i
-1-
^ Soil handlina
and
conveying
.3. .4.
^ Soil washinq m Fines/floatables ^
and gravity separation |
coarse screening ~' — '"•" "" •••— ~— J i
__..,... ' ' * f ^ ' ' *
M
r
w
-5-
Fines flocculati
water clarificat
and solids dispo
l To redeposit or
ake up/ -1/2" +10-mesh (0.079"/2mm) solids further treatment
ecycle "" " " *-
ater -10 +60 mesh (-0.079"[2mm] + 0.0098"[0.25 mm]} To redeposit or
further treatment
f Makpun wafpr
i*^
an/ 1 1 Water To delistinq/disposal
fon treatment Slowdown or posttreatment
sal 1 * """ '" " :B"
Clay/silt sludge To posttreatment^
.7.
Utilities
- Electric generator
- Boiler
- Compressed air
Figure 1. General Block Diagram - The U.S. Environmental Protection Agency
Volume Reduction Unit (VRU) for Soil Washing
-------
IO
CTl
-------
P "-TO f-l/J
Is*
oii/ROM«ais
ID D1SKSIL BEOCLE «
IMUCtlON
UNIT
(H
HUOCE
rtua cut
10 OiSPOSll
FI01RE 3
PMCESS FLO» DUGRW
EPA
MOBILE VOLUME
REDUCTION UNIT
FOB SOIL iASHINC
-------
OD
FIGURE 4
PROCESS FLO* DIACRMI
EPA
MOBILE VOUie
WUCTIQN UNIT
FOR SOIL WSHING
-------
controller on the conveyor motor. The solids pass through a motor-operated
rotary valve (which prevents air infiltration), then into the feed hopper of
the mini-washer.
2. Organic Vapors Recovery
Volatiles stripped from the soil in the screw conveyor are either
collected in the volatile organic compounds (VOC) condenser and fall by
gravity into the process condensate seal tank, or are adsorbed in
vapor-phase activated carbon containers located upstream of the vent
blower.
The spent carbon will be periodically replaced based on vent gas
analyses. The vapor train is maintained under vacuum by an induced
draft blower. The vacuum level is adjusted by manual admittance of
atmospheric air upstream of the blower to maintain a slight negative
pressure on the vapor system. Clean vapors, leaving the blower, vent to
the atmosphere.
3. Soil HashingandCoarse Screening
Soil is fed to the mini-washer at a controlled rate of approximately 100
Ib/hr by the screw feeder. Filtered wash water, which can be heated to
150*F (maximum), is added to soil in the feed hopper and also sprayed
onto an internal slotted trommel screen (with a 10-mesh (0.079") slot
opening) mini-washer. Five manually controlled meters can control the
flow up to approximately 10:1 overall weight ratio water to soil. Hot
water should be more efficient in extracting contaminants, but heating
is optional. When required, dilute surfactant/detergent, and/or caustic
can be metered at a controlled rate into the feed hopper.
Two vibrating screens, equipped with anti-blinding devices, are provided
to continuously segregate soil into various size fractions. These
screened fractions can be collected to measure the effectiveness of con-
taminant removal for each soil fraction recovered, and to determine the
99
-------
effectiveness of soil washing in cleaning a particular contaminated soil
fraction to achieve sufficient volume reduction.
Mini-washer overflow, containing the coarser solids, falls onto the
first 10-mesh (0.079"/2 mm) vibrascreen. First vibrascreen overflow
(-i" + 10 mesh (0.079"/2 mm)} solids flow by gravity down to a recovery
drum. The underflow is pumped at a controlled rate, using a progressing
cavity pump, onto the second 60-mesh (0.0098"/0.25 mm) vibrascreen where
it is joined by the Mini-Washer underflow.
The overflow from the second vibrascreen (- 10-mesh (0.079") + 60-mesh
(0.0098")), is gravity fed to another recovery drum. Second vibrascreen
underflow (a fines slurry) drains into an agitated tank. The VRU is
designed with the following flexibility:
a. The mesh sizes for both the mini-washer and vibrascreens can be
varied (i.e., the screen size could be 20- or 30-mesh (0.033" or
0.023").
b. Additional soil cleaning by use of water sprays or steam sprays
will be evaluated for each vibrascreen.
c. Screened soil fractions, collected in the recovery drums, can be
redeposited if sufficiently cleaned or further cleaned by addition
of rinse water, followed by reslurrying and pumping the slurry
back over the screens (recycle mode). In the future these soil
fractions will be sent for treatment by various extraction units
currently under development by EPA's RREL in Edison, New Jersey.
4. Fines/FIeatables GravitySeparation
Slurry from the second screen (fines slurry) tank, containing particles
less than 60-mesh (0.0098"/0.25 mm) in size, is pumped to a Corrugated
Plate Interceptor (CPI). Material lighter than water (fleatables such
as oil) will overflow an internal weir, collect in a compartment within
100
-------
the CPI, and drain by gravity to a drum for disposal. CPI-settled
solids (soil particles - 60- to about 400-mesh (0.0098" to about
0.0015") will be discharged by the bottom auger to a recovery drum.
The VRU has the flexibility to redeposit or further clean these settled
soils, if required, by addition of rinse water followed by pumping the
slurry back through the CPI. As mentioned above, these soils could also
be sent, in the future, to an extraction unit.
5. Fines Flocculation, Water Clarification, andSolids Disposal
Aqueous slurry, containing fines less than about 400-mesh (34 urn/
0.0014"), overflow the CPI and gravity feed into an agitated tank. The
slurry is then pumped to a static flash mixer located upstream of the
floe clarifier's mix tank. Flocculating chemicals are introduced into
this static flash mixer. Typically, liquid alum and aqueous polyelec-
trolyte solutions are metered into the static flash mixer to neutralize
the repulsive electrostatic charges on colloidal particles (clay/humus)
and promote coagulation. The fines slurry is discharged into the floe
chamber which has a varispeed agitator for controlled floe growth (sweep
flocculation). Sweep flocculation refers to the adsorption of fine
particles onto the floe (colloid capture) and continuing floe growth to
promote rapid settling of the floe and its removal from the aqueous
phase. The floe slurry overflows into the clarifier (another corrugated
plate unit). Bottom solids are gravity fed by an auger to a drum for
disposal, or to the sludge slurry tank (depending on solids concentra-
tion) for subsequent concentration in a filter package unit. Con-
centrated cake from the filter is discharged to another drum for
disposal. This system has the ability to clarify the process water and
dewater the sludge. The efficiency of solids dewatering can be deter-
mined, and cost savings estimated, for trucking waste sludge to a dis-
posal/treatment site.
6. Water Treatment
Clarified water is polished with the objective of reducing suspended
101
-------
solids and organics to low levels that permit recycle of spent wash-
water. Water is pumped from the floe settler overflow tank at a
controlled rate through cartridge-type polishing filters operating in
parallel, in order to remove soil fines greater than 10-um (3.94xlO"4").
One urn (3.9xlO"5") cartridges are available, if required.
Water leaving the cartridge filter flows through activated carbon drums
for removal of hydrocarbons. The carbon drums may be operated either in
series or parallel, and hydrocarbon breakthrough monitored by sampling.
A drum will be replaced when breakthrough has been detected.
In order to recycle water and maintain suitable dissolved solids and
organic levels, aqueous bleed (blowdown) to the boiler delisting tank
may be initiated at a controlled rate. Delisted material will be sealed
in drums and sent for disposal in accordance with respective state and
local regulations.
Treated recycle (recovered) water is sampled for analysis before it
flows into the process water storage tank. Supplementary water is fed
into this tank from a tank truck. Recovered and added water is pumped
by the water recycle pump (and optionally fed to the water heater) for
subsequent feed to the mini-washer. A side stream from the water
recycle pump is utilized as cooling water in the VOC condenser and
either returned to the process water storage tank or sent to the
sewerage system.
7. Utilities Systems
The VRU is equipped with a steam boiler, electric generator, and a
compressed air system.
Field Operations
While in the field, the VRU would be supported by a decontamination
trailer, a mobile treatability lab/office, and a storage trailer for supplies,
102
-------
spare parts, miscellaneous tools, etc.
Summary of VRU Features
1. The VRU is a mobile, pilot-scale washing system for field use in
cleaning soil contaminated with hazardous materials, using state-of-the-
art washing equipment and support utilities.
2. The unit has the ability to remove VOCs by steam heating and stripping.
3. It is capable of washing with water (in combination with surfac-
tants/detergents) up to a 10:1 water to soil ratio while also varying
water temperature from ambient to 150°F.
4. The mini-washer screen and vibrascreens can be varied in mesh size.
Additional use of soil cleaning by water or steam sprays on the vibra-
screen decks can be evaluated.
5. Three screened soil fractions (including CPI-settled solids) can be
further cleaned by slurrying with the addition of rinse water and
recycling the slurry over the vibrascreens or the CPI.
6. The floc-clarifier system has the ability to clarify the process water
and dewater the sludge.
7. Additional treatment of the clarified process water through polishing
filters and activated carbon should allow, in most cases, reuse of this
water for recycle to the washing circuit.
8. Side streams from the VRU will be treated using various physical/chem-
ical extraction units currently under development by EPA.
9. The VRU offers a unique method for conducting treatability studies on
various contaminated soils.
103
-------
REFERENCES
1. Foster Wheeler Enviresponse, Inc., "Cleaning Excavated Soil Using
Extraction Agents: A State-of-the-Art Review," January, 1990, EPA/600/
$2-89/034.
2. Foster Wheeler Enviresponse, Inc., "Workshop of Extractive Treatment of
Excavated Soil," December, 1988.
3. EPA Treatment Technology Bulletin, "Soil Washing," Draft issued May,
1990.
4. Traver, R.P., "Development and Use of the EPA's Synthetic Soil Matrix
(SSM/SARM)." U.S. EPA Releases Control Branch, Risk Reduction Engineer-
ing Laboratory, Edison, N.J., 1989.
104
-------
GUIDANCE FOR TREATABILITY TESTING UNDER CERCLA; AN UPDATE
by: David Smith, Edward Bates, Malvina Milkens
Risk Reduction Engineering Laboratory
US Environmental Protection Agency
Cincinnati, Ohio 45268
James Rawe
Science Applications International Corporation
Cincinnati, Ohio 45203
ABSTRACT
The Risk Reduction Engineering Laboratory is working with EPA's Office
of Emergency and Remedial Response to prepare guidance documents on the
subject of treatability testing for Superfund sites. This paper describes a
recommended approach to treatability testing in stages or tiers. Remedy
screening, remedy selection and remedy design treatability tests are
described. The paper also discusses technology specific treatability guidance
documents that are in preparation. Technology specific documents address
solidification of inorganics, soil washing, aerobic biodegradation (screening
scale only), soil vapor extraction, and chemical dehalogenation. The
documents present a structured, tiered approach unique to testing each
technology and identify critical factors which must be evaluated during
testing of these potential cleanup technologies.
This paper has been reviewed in accordance with the
U.S. Environmental Protection Agency's peer and
administrative review policies and approved for
presentation and publication.
105
-------
INTRODUCTION
EPA's Office of Research and Development (ORD) has a long history of
involvement in treatability study research for toxic chemicals. Much of that
work has been in support of wastewater treatment regulations for both
industrial and municipal systems. The goals of those studies can be
classified as follows:
studies to determine the effectiveness of a specific technology/waste
combination
studies to determine the effectiveness of a specific process(es) against
one or more toxic chemicals (not from any particular source)
studies to optimize the performance of a particular technology/waste
combination.
All of these goals apply to treatability testing under CERCLA. ORD's
treatability experience is currently being applied and extended to the
Superfund program in evaluation of remedial technologies for potential
application at NPL sites. In fact, the primary goal of the Technical Support
Branch of RREL is to provide expert advice on treatment technologies for
specific NPL sites. Much of the assistance being given involves advice on use
of treatability studies for evaluation of treatment alternatives.
This paper discusses the concept of treatability testing as applied to
the Superfund program. The paper discusses the background of increased
emphasis on treatability testing under CERCLA, the current status of
treatability testing, and some of the problems in applying treatability tests
in the Superfund scenario. It will give some specific examples of
treatability tests and other related tools available to Regional Superfund
programs.
INHIBITIONS TO TREATABILITY TESTING
One of the criticisms of the EPA Superfund program has been that
treatability studies are not conducted during the Remedial
Investigation/Feasibility Study (RI/FS) time period of site remediation (1).
The reasons for this situation can be easily understood. The strict deadlines
on conducting RI/FS's do not encourage conducting treatability studies, which
by their nature take time to plan and complete. The goal for completion of
the RI/FS is 22 months. Several months are frequently necessary to get a
preliminary idea of the problems presented at a site. It could be a year or
more before adequate site contaminant data are available to enable a site
manager to be in a position to commit to a treatability study. Furthermore,
guidance on what constitutes adequate treatability testing under the Superfund
program has been lacking. Among technology experts (both within and outside
of EPA) there are differences of opinion as to how technologies should be
evaluated.
106
-------
An internal EPA review of the Superfund program has produced several
recommendations regarding the use of innovative technologies and treatability
testing:
Program guidance should ensure that treatment technologies are given
stronger emphasis
EPA should establish technology support teams within the Office of
Research and Development.
EPA should establish a treatability assistance program to perform
treatability tests, develop standard testing protocols, and maintain a
database of testing results. Guidance on how to use treatability tests
in selecting clean-up technologies should be provided.
ORD should provide easier access to information on technology
performance.
The Superfund Innovative Technology Evaluation (SITE) program (and other
research programs) should be expanded to provide evaluations of
innovative technologies and provide for rapid dissemination of results.
The focus of these recommendations is that innovative technologies
should be investigated more thoroughly at Superfund sites and the use of
treatability tests should be increased to evaluate treatment technologies
scientifically before they are chosen for site remediation.
TREATABILITY GUIDANCE
One way that RREL has responded to the need for treatability study
assistance is by production of guidance documents on the conduct of
treatability studies. In December of 1989, RREL published "Guidance for
Conducting Treatability Studies Under CERCLA" (2). This guide resulted from
input from numerous individuals representing EPA's regional offices,
contractors and technology vendors as well as from EPA's Office of Research
and Development (ORD). The document describes the steps necessary to conduct
treatability testing, from selecting technologies, issuing work assignments,
to preparing workplan and QA plans. The document also describes how to use
treatability results in the FS.
Remedial alternatives, under the CERCLA program, are evaluated against 9
criteria:
Overall protection of human health and the environment
Compliance with 'Applicable or Relevant and Appropriate
Requirements' (ARARs)
Implementability
Reduction of toxicity, mobility, or volume
Short-term effectiveness
Cost
Long-term effectiveness
107
-------
Community acceptance
State acceptance
The 'guide' describes how treatability studies can help to address the
first seven of these criteria. Traditionally, treatability studies have been
used to determine the effectiveness of the technology in reducing contaminant
levels or in meeting regulatory levels (which relate to Compliance with
ARARS). However, the Superfund remedy evaluation process asks that
treatability studies be formulated to answer much more comprehensive questions
relating to issues such as cost, short and long term effectiveness and
potential problems with implementability.
Recognizing the unique requirements of evaluating and selecting remedies
under CERLCA, this guide presents a tiered testing strategy for conducting
treatability tests and for incorporating treatability results into the remedy
evaluation process (figure 1). Prescreening consists of literature reviews
and consultation with technology experts to obtain an indication of whether
the technology has been successfully applied in similar situations. In
general, 'screening' level testing is used to provide a qualitative evaluation
of the potential effectiveness of a technology. The remedy screening tier of
testing is used to indicate whether further testing is warranted to more
thoroughly evaluate a treatment technology or whether the technology should be
screened out. In order to begin treatability testing as early in the RI/FS
process as possible, remedy screening treatability tests are intended to be
run based on minimal site information. In general, it is not necessary to
have detailed data on contaminant levels and spatial distribution. Nor is it
necessary to have in-depth data on site characteristics such as soil
permeability or soil particle size distribution. This type of detailed data
is usually not available in the early stages of site investigation but where
needed, can often be generated as a part of treatability testing. Remedy
screening treatability testing is designed so that a series of simplified
tests are run which represent a range of process options within a technology
class, such as 'biodegradation' or 'thermal treatment'.
'Prescreening' is an important step in the evaluation of treatment
technologies and should precede treatability testing. The purpose of
prescreening is to obtain an indication of whether the technology is
potentially applicable to the situation in question and what scale of
treatability testing, if any, is warranted. Literature reviews, database
searches, vendors literature, and consultation with technology experts are all
potentially valuable information sources which can be utilized during remedy
prescreening.
Remedy screening treatability testing is characterized by relatively low
cost and short time for completion. Remedy screening testing may not be
necessary for situations where the performance of a technology is well
documented with contaminants and matrices similar to those being considered.
Remedy screening testing does not necessarily have to be conducted by a
technology vendor. In general, remedy screening treatability tests are
generic to a class of treatment technology and therefore can be done by any
suitably equipped laboratory.
108
-------
o
ID
"fl
H
O
a
he)
M
o
H)
O
H
(D
o
w
o
p
(D
ft
cr
F-
arre
CHAMCTEHZAtlOM
i
lECHNOtOQY
SCHKWNO
MAMAQEMENT DECMTOW FACTORS:
* SWtindCoimnunftyAonpte'Ki*
• Schiduh CoMraM*
MdthxwIOM*
(D
Cfi
ft
H-
3
-------
The remedy selection tier of testing is used to provide data to support
evaluation of a specific remedy in the FS. The remedy selection treatability
test should provide performance data which will indicate whether ARARs or
cleanup goals can be met at the site by the technology. Remedy selection
treatability tests should also allow for estimation of costs associated with
implementation of the remedy to the accuracy required for the FS (+50/-30%).
Remedy selection treatability testing requirements vary depending on the
technology being evaluated and on site specific factors. For some
technologies, testing only at a laboratory bench scale may be sufficient to
provide performance data adequate to meet the needs of the FS. In other
cases, testing at both a bench and pilot scale may be required. Pilot scale
testing will usually be necessary where it is difficult to simulate field
conditions in the laboratory (e.g. in-situ treatment technologies). Where the
types of experiments and equipment involved in remedy selection treatability
tests are very specific to the treatment process, remedy selection testing
will probably have to be conducted by the technology vendor. In other cases,
where the treatment process could be carried out by a number of vendors and
the treatment equipment is more commonly available (e.g. some types of
incineration), remedy selection treatability testing could be conducted by any
suitably equipped facility.
Remedy design treatability testing will usually be required after a
record of decision has been issued. The purpose of remedy design testing is
to optimize the selected treatment process and to obtain detailed cost and
performance data. Remedy design testing is highly vendor specific and will
usually be conducted by the vendor as a step in remedy implementation.
TREATABILITY TEST TIMING AND THE RI/FS PROCESS
This scenario for treatability testing has been devised to fit into the
overall Superfund remedy evaluation process. For a specific site, numerous
technologies can be pre-screened early in the RI/FS process based upon
available information. Where adequate performance data are available for a
treatment technology with similar waste material, remedy screening testing may
not be necessary. Where there are significant concerns regarding the
applicability of treatment technologies, screening tests can be conducted in a
relatively short time period.
The Superfund program has targeted a 22-month RI/FS process. Included
in this time period is initial site scoping, the field investigation (probably
in at least two phases), preparation and evaluation of the remedial
investigation report and the feasibility study, detailed analysis of remedial
alternatives, and issuance of the record of decision.
Treatability tests have been formulated in a tiered approach that can
be fit into such a schedule. Figure 2 illustrates where the treatability
study tiers fit into the RI/FS process. To accomplish this, remedy screening
treatability testing (if appropriate) must be started very early in the 22-
month time period. However, usually at this early stage in site
investigation, there are significant data gaps regarding site characteristics.
Many of the site characteristics and measurement parameters which would enable
experts to recommend potential treatment technologies will not be available.
110
-------
RI/FS
I dent if Icat ion
of Alternatives
=> ROD
Remedy
Se Iect i on
RD/RA
Site
Character Izat i on
and Technology
Screen!ng
EvaluatIon
of AIter nat i ves~
freatablIFty Study
Scop i ng
Remedy Screening to
Validate Technology
Remedy Selection to
Develop Performance
Data
immiimiiiiimiiiimiimmiiwmiiiiiHiwimiiimm
I trp I ementat i on
of Remedy
Remedy Des i gn Test i ng
to Develop
Performance, Cost and
Design Data
mmiiiiiiimmiiiiittmiiitiffltiiimmiiiiiinmii
Figure 2 - Tiers of Treatability Testing in the RI/FS Process
Hence, in the past site managers were asked to conduct remedy screening
treatability studies for a number of technologies and make selections based
upon minimal information. Remedy screening treatability studies are being
designed so that they are relatively inexpensive and reasonably quick to
perform. Evaluation of a number of technologies at the screening level will
provide a more scientifically supported selection of treatment technologies on
which to conduct detailed testing.
When conducting screening treatability tests there is also a greater
risk of both 'false positives' (deciding to conduct further testing on an
inappropriate technology) and 'false negatives' (deciding that a technology is
not appropriate for a site when in fact it is appropriate). By conducting a
number of relatively inexpensive screening tests for a specific site, these
risks of'inappropriate decisions regarding treatment technologies at the
screening level are acceptable when balanced against the savings of time and
money.
The results of screening tests, coupled with additional information
obtained during the RI may indicate that a treatment technology should proceed
to remedy selection treatability testing. Although these tests are more time
111
-------
consuming to conduct, in general, it should still be possible to conduct them
within the RI/FS timeframe. The cost of remedy selection testing is justified
considering the typical costs associated with site cleanups.
TECHNOLOGY SPECIFIC TREATABILITY GUIDANCE DOCUMENTS
'A Guide for Conducting Treatability Studies Under CERCLA' presents a
generic approach to conducting treatability studies for treatment
technologies. The process of treatability testing consists of more than the
actual experimental details of testing. In the Superfund program,
treatability testing includes pre-screening of alternatives, planning
treatability studies, the actual conduct of studies, as well as interpretation
of treatability data and incorporation of that data into the analysis of
alternatives. The guide defines treatability testing in the broad sense;
considering steps necessary from initial site scoping through selection of a
treatment technology in a Record of Decision.
Although the guide is comprehensive in dealing with a number of
technical and administrative topics relating to treatability studies, by its
nature it cannot deal in any depth with the technical details involved in
testing a specific technology. For this reason, RREL is in the process of
developing a number of guidance documents which deal with the testing of
specific treatment technologies.
Shown in Table 1 are the documents which are currently under
preparation. With the exception
of the biodegradation guide, these
documents are designed to assist
TABLE 1
TREATABILITY GUIDES IN
PREPARATION
Technology '' ;' Publication Expected
Soil Washing Harch 1991
Stabilization of inorganics Dec 1991
Aerobic Biodegradation * Feb 1991
Soil Vapor Extraction Jgne 1991
Chemical Dehalogenation April 1991
Solvent Extraction Dec 1991
Thermal Desorption Dec 1991
*-.
* Screening level only " -
in the conduct of treatability
testing, from initial site scoping
through final remedy selection.
The biodegradation guide is only
designed to include the steps
necessary to screen aerobic
biological degradation as a
potential treatment remedy.
The guides are intended as
management tools rather than
cookbook 'protocols' for
conducting treatability testing.
While these guidance documents
contain technical information
necessary to conduct acceptable treatability experiments on a potential
treatment technology, they also contain steps of a more administrative nature
that are necessary in the conduct of treatability testing. Such steps
include:
pre-screening the technology to determine whether it is potentially
applicable at a site,
recommended content of a workplan that would be prepared to plan
treatability testing,
112
-------
interpretation of treatability testing data and incorporation of that
data into the feasibility study analysis of alternatives.
All of the technology
specific guides being
prepared are being formatted
similarly. A common format
(Table 2} was designed for
these guides so that the
user can consult that
portion of the guide which
is needed at a particular
point in time (i.e. when
preparing or reviewing a
workplan for a treatability
study).
TABLE 2 - OUTLINE FOR TECHNOLOGY
SPECIFIC TR1ATABILITY GUIDANCE
DOCUMENTS
INTRODUCTION ... , • "'
* Background
• Purpose and scope'
• Intended audience , .
• » Use of the guide •. :
TECHNOLOGY DESCRIPTION & PRELIMINARY SCREENING .... :
* Technology description
« Pre-screening the technology. •
THE USE OF TREATABILITY TESTS IN REHEDY"SELECTION
* The Process of treatability testing in selecting
a rertiedy - •
• Applicability of treatability tests
TREATABILITY STUDY WORK PLAN
SAMPLING AND ANALYSIS PLAN; QUALITY ASSURANCE PROJECT PLAN
TRIATABILITY DATA INTERPRETATION •
'•' Technology evaluation ' • •• ' ...
» Estimation of costs •" • . •
In developing these
guidance documents, a common
approach has proven useful.
Early in the development
stages of each document, a
workgroup was conducted to
get input on how
treatability tests are
conducted with the
technology chosen. For this
workgroup meeting, a 'strawman' treatability guide was prepared. This
'strawman' document was useful in giving workgroup attendees something to
focus constructive comments. We attempted to get workgroup participation from
a number of interested sources including EPA's Regions, EPA's Office of Solid
Waste and Emergency Response (OSWER), ORD/RREL, academia, consulting firms and
technology vendors. Participation by Regional RPMs and technology vendors has
been particularly useful in learning the experiences gained in the actual
conduct of treatability studies in CERCLA and other programs. In getting
technical reviews of the various drafts of each treatability guide, we
attempted to solicit comments from as broad a group as possible.
During the development of these documents, a number of facts became
apparent and shaped the thinking on the 'prescriptiveness' of these documents.
First, in the expert workgroup meetings which were held to gain input for
these documents, it was apparent that those experts who are currently doing
treatability testing on treatment technologies have different (and equally
valid) methods of conducting that testing. Although experts generally agreed
on the critical parameters that need evaluation for a particular technology,
their approach to investigation of those parameters varied.
Secondly, the database of treatability study results is not of
sufficient magnitude to be able to recommend a 'standard treatability
protocol' which would lead to optimum technology evaluation in all cases.
The
113
-------
experts have agreed that a considerable amount of professional judgement is
necessary in formulating a treatability test which incorporates site specific
factors and circumstances.
Lastly, the details of treatability testing are highly matrix and
process dependent. In order to make the guidance documents broad enough to
cover more than one specific process and matrix, a degree of flexibility in
the details of experimentation is needed. As an example, biological treatment
can be applied to various media such as liquid, sludges, soils, or sediments.
Biological treatment can also be applied via a number of processes such as
slurry biodegradation or composting. The details of experimentation with each
of these process/matrix combinations are different.
Therefore, the final versions of these documents have been constructed
to be flexible enough to allow professional judgement in formulating
treatability test plans and to allow for variations in treatability test
methodology. The documents concentrate on formulating a tiered approach to
testing with that technology and on identifying critical factors which need to
be investigated in each of those tiers. The documents, in many cases, present
options for investigating those critical factors, but the selection of the
appropriate option depends on the specific circumstances at a site and should
be left to those more familiar with the site.
AEROBIC BIODEGRADATION - REMEDY SCREENING
This guide describes the screening
would not by itself lead to selection
of a biological treatment remedy.
Such remedy selection guidance will
be forthcoming at a later date.
The main determining factor in
pre-screening site/matrix
combinations for potential for
aerobic biodegradation is literature
review for the degradability of the
compounds (or similar compounds) at
the site. There are site factors
which may ultimately preclude
biodegradation at a site (such as
extreme pH or concentrations of
contaminants which are toxic to the
degrading organisms), but it may be
possible to negate these factors by
various pretreatment methods. These
methods would be explored in the
remedy selection tier of testing.
This guide discusses the
important aspects of screening
testing. Aliquots of soil from a
site are placed in soil reactors.
tier of testing only and, therefore,
^:; ;PEGRAD^IP;N;^^
??P'f esereenf hg::::*--! Li teratur'e'.- rJeMetitf ofe'
;biodegradabi.tr'ty:o|;,j;orjt;amiriaf)ts '.of;jcorifcerp'-of.'• • *.;
"-sjn)i^rJcpffli»i^si::.'5'So^:;:::s^ ' '
H.:'pr,:iactfVity';Tr^"ii)i,tin9;':
•':"±~""""ScriJeniina';*!:::'jaifi:::w:; 1^jtisk%rcorii't|iirjfng;';;'
:':'6r":iJnsaturat^d."v
Mut r I ents". ^and/or'
. . .....
cio'r%>£3i^l<^/^
'
if iu Str J Ie:*conr'cs.'iised "to
.assiess^;-^/.of:#7.'gft5v'sai^t:i:rt9^
"
iSiilC^ife'K'J:
^n?||:!t:|;^l:;IIi||A||i^l^
114
-------
One aliquot is initially analyzed for a baseline contaminant level and
additional aliquots are analyzed at later times to estimate the degree of
degradation. 'Sterile controls' (it is difficult to sterilize soil) are
recommended so that degradation due to abiotic mechanisms can be taken into
consideration. If volatile compounds are present, analysis of off-gas can be
conducted, but is not routinely recommended for the screening level of
testing. The main question to be answered at a screening level of testing is
whether further testing is warranted to evaluate a potential biological
treatment remedy in more detail. Therefore, the goal of screening biological
testing is not whether contaminants are reduced to a 'cleanup level', but
rather to demonstrate that biodegradation is taking place. The guide proposes
that if 20%-50% degradation is detected, then further testing at a remedy
selection tier may be warranted.
SOIL VAPOR EXTRACTION
The screening tier of treatability testing for SVE involves the use of
soil columns composed of soil from the site. This guide draws heavily from
guidance and experience developed as a part of Leaking Underground Storage
Tank research. This guide discusses
pre-screening site characteristics to
determine whether the vapor pressure
of contaminants of concern are
greater than 0.5 mm Hg.
;;t^
Screening tests with soil
columns are run for a short time to
determine whether contaminants are
being removed. In many cases,
screening testing may not be
necessary for SVE if the vapor
pressure of the compounds of concern
are significantly above 0.5 mm Hg.
Screening testing may be warranted
where the vapor pressure of the
compounds of concern is near 0.5 mm
Hg or where the soil from the site
appears to be very non-permeable.
The remedy selection tier of
testing for SVE consists of at least
3 parts: 1) column tests, 2) field
air permeability measurements, and 3)
mathematical modelling. In some
cases (i.e. complex sites or where
bedrock contamination exists), pilot
scale testing may be necessary for
remedy selection. Column testing for
SVE remedy selection treatability studies provides information regarding the
ultimate cleanup level that can be expected from the technology. It also
provides an estimate of the effective Henry's law constant for use in
mathematical modelling. Field air permeability measurements are used to
jj.;.|i.i^
''
.fief#e^
«*33 Bl*at:8;f f he ^f^ji
"j>otj|rtt| a 1 1 y}; a.ch leva(3 1 e S
' *
:i
' ' '
.
:;:; :|(Bi»eciy ;;Se I ectf
^ssss^xxrx:;;1
3M»f|lcpi:
ii5I|?x^ill
-&p£'?p!;
Kiio-fe'tiier
115
-------
obtain data on the permeability of the site so that air movement underground
can be modelled. Mathematical models use data from both the column tests and
air permeability tests to estimate the time for site remediation and
associated construction and operating costs.
CHEMICAL DEHALOGENATION
Chemical dehalogenation involves addition of a chemical reagent to
contaminated material (soil, sludge, liquid). The reagent effects the removal
of one or more halogen atoms from the contaminant molecule. The technology is
applicable to dioxins, PCBs, some
chlorinated pesticides and other
halogenated carbon compounds. The
technology is generally best suited
to situations where contaminant
concentrations are at ppm levels or
greater. For soils and sludges, the
total waste volume should be greater
than -1000 M3 to be cost effective.
CHEMICAL DEHALOGENATION
TREATABILITY TESTING
Prescreemng - Applicable to halogcnated
aromatics and aliphatScs (PCBs, dioxins/furans,
chlorophenols, ethylene dibromlde). Host cost
effective for volunes > 1000 H.
Remedy Screening - Bench scale test under
conditions of most likely success. Success
indicated by 80-90% contaminant reduction.
Remedy Selection - Optimization of
process/reagent at bench scale. Toxicity
testing recommended on treated soil.
Remedy Screening
Remedy Selection
tab Test Time
"2-5 days
"3-6 weeks
Cost
$10-$50K
S20-S100K
Remedy screening for this
technology involves treatment tests
on a 'worst case' sample of the
matrix to be treated. Remedy
screening tests are conducted in the
laboratory under the most favorable
conditions for dehalogenation. Since
chemical dehalogenation processes
vary among vendors, the most
favorable conditions will be process
dependent. Typical conditions would
involve an excess of reagent,
relatively high temperatures and
process treatment times. The goals for remedy screening testing are to
achieve -90% reduction (or the cleanup goals if known) in the contaminant
concentration.
Remedy selection treatability testing for chemical dehalogenation
involves the same physical scale of test equipment. However, the goal of
remedy selection treatability testing is to optimize the treatment conditions
to the contaminant/matrix combination(s). Treatment residuals are analyzed to
verify that they meet the cleanup goals. Toxicity testing is also done on
treatment residuals to verify that overall toxicity is reduced. The optimized
treatment conditions allow for estimation of costs for the full scale remedy
(sufficient to meet RI/FS cost criteria).
SOLIDIFICATION/STABILIZATION OF INORGANICS
Stabilization of inorganics (toxic metals and some inorganic compounds)
is an accepted developed technology for many metals. In some cases a
screening level treatability test can be waived if sufficient data exists for
treatment of the same form of the specific target metals in similar soil
116
-------
matrices. However, for other
inorganics (i.e. arsenic, mercury,
and inorganic compounds) or for
differing ionic forms of compounds, a
screening level study is recommended.
The screening level study is
generally done using several mixtures
of generic reagents including
Portland cement, fly ash, and clays.
Results are evaluated using both
destructive (TCLP) and non-
destructive (ANS 16.1) leaching tests
along with measurements of
permeability and unconfined
compressive strength.
STABILIZATION OF INORGANICS
TREATABILITY TESTING
Presereening * Applicable to.inetals for which a
substantial data base on successful application
exists,
Remedy Screening - Bench scale test for metals-
and inorganic compounds for.which a substantial
data base of succesful treatment does not
exist. Can be skipped for commonly stabilized
contaminant/soiI types.
Remedy ffeiection - Tiered optimization of
reagent and mixture ratios at bench scale using
vendor specific technology for the more
difficult to stabilize inorganics. ;-.
.labTest Time Cost
Remedy Screening
Remedy Selection
2-3 months
3-6 months
$10-$50K
$50-$150K
Remedy selection treatability
testing is conducted in a similar
manner but involves more intensive
testing including more samples,
quality assurance and more analysis
of stabilized product. A tiered
approach is preferred. Results from the first round of tests on various
mixtures and ratios of reagents leads to a second round of more intensive
testing on the particular mixture and ratio of reagents that performed well In
the initial tier. For more difficult to stabilize metals (i.e. arsenic or
mercury) or inorganic compounds, vendor specific reagents and procedures may
be necessary to adequately evaluate the effectiveness of the technology.
Although not included in this particular document, it appears that
stabilization of many semi-volatile organic compounds can also be accomplished
by use of vendor specific treatment technologies. A treatability guide
specific to evaluating stabilization of organic compounds is being considered
for a future publication.
SOIL WASHING
The screening tier of treatability testing for Soil Washing involves the
use of jar tests with soil from the site. Screening tests are run for a short
time period (less than 1 week) to determine whether contaminants are being
removed. In some cases, screening tests may not be necessary for soil washing
if the identities of the compounds of concern and the physical and chemical
characteristics of the soil indicate the technology may be successful at a
site.
The remedy selection testing tier for soil washing consists of bench-
scale wet-sieve tests. They yield data which verify that the technology can
meet expected cleanup goals, provide information in support of the detailed
analysis of alternatives, and give an indication of optimal operating
conditions. Toxicity testing is also performed on treatment residuals. The
optimized operating conditions allow for estimation of cost for the full-scale
remedy.
117
-------
SUMMARY
A tiered approach has been
described that can be used in
evaluating alternative technologies
applicable to Superfund wastes. This
approach will aid in obtaining
defensible data in support of
remedies selected at Superfund sites.
In addition to general guidance on
the conduct of treatability studies,
EPA-RREL is preparing technology
specific guidance documents which
will help in evaluating the
applicability of those technologies
at sites. These documents describe
how to prescreen the technology for
potential applicability, important
factors to consider in conducting
tests and how to interpret and use
treatability data in the Superfund
remedy evaluation process.
SOIL WASHING TREATABILITY
'TESTING
Remedy Screening - Laboratory Jar tests can
give a preliminary indfcation of removal
efficiencies, but are not normally used.- The
contaminant identity and physical and chemical
characteristics of the soil can usually be used
to predict the potential success of soil
hashing. ' •
Remedy Selection - Bench-scale wet-sieve tests
indicate ultimate cleanup levels potentially
achievable. Optimization of process/reagent Is
done at the bench level.
tab Test Time Cost
Remedy Screening "2-5 days $10-$50K
Remedy Selection "3-6 week $20-$100K
REFERENCES
1. U.S. Environmental Protection Agency, A Management Review of the
Superfund Program.
2. Guide for Conducting Treatability Studies Under CERCLA. EPA/540/2-
89/058, U.S. Environmental Protection Agency, Cincinnati, Ohio, 1989,
118
-------
S.I.I.E. DEMONSTRATION OF A SOIL WASHING SYSTEM
BY BIOTROL INC., AT A WOOD PRESERVING SITE
IN NEW BRIGHTON, MINNESOTA
by: William D. Ellis Mary K. Stinson
Science Applications EPA Risk Reduction
International Corporation Engineering Laboratory
McLean, VA 22102 Edison, NJ 08837
ABSTRACT
A pilot scale demonstration of BioTrol, Inc.'s Soil Washing System
(BSWS) was conducted under the U.S. EPA's Superfund Innovative Technology
Evaluation (SITE) Program, at a Superfund site in New Brighton, Minnesota.
The BSWS treated soils contaminated with wood treating wastes, including
creosote-fraction polynuclear aromatic hydrocarbons (PAHs) and
pentachlorophenol (penta).
In the SITE Program test, the component technologies of the BSWS were
the Soil Washer (SW), Aqueous Treatment System (ATS), and Slurry Bio-
Reactor (SBR), The demonstration determined the contaminant reduction
efficiency of all three BioTrol technologies. Also, a material balance was
determined for the penta and PAHs in the Soil Washer. A 2-day test used
200 ppm penta soil, with similar total PAH concentrations, and a 7-day test
used 1000 ppm penta and PAH soil. The equipment ran 24 hrs/day.
The SW separates the relatively clean sand fraction (roughly two-
thirds by weight of the soil) from the highly contaminated silt and clay
fractions, by intensive scrubbing and size classification. The remaining
contaminated silt/clay slurry was treated in the SBR. Contaminated process
water from the SW was treated in the ATS. The ATS is a microbiological
system for degrading toxic organics in water. It consists of naturally
occurring microbes growing on a plastic matrix in tanks, which degrade
penta and PAHs into harmless carbon dioxide, water, and inorganic chloride.
The SBR is a three-stage microbiological system for degrading organics on
slurries of soil fines, consisting of three upright, continuously-stirred
reactors. The results will be available once laboratory analyses are
complete.
This paper has been reviewed in accordance with the U.S. Environmental
119
-------
Protection Agency's peer and administrative review policies and approved
for presentation and publication,
INTRODUCTION
A pilot scale demonstration of the BioTrol Soil Washing System (BSWS)
was conducted under the U.S. Environmental Protection Agency's (EPA's)
Superfund Innovative Technology Evaluation (SITE) Program from September 25
to October 16, 1989, at the MacGillis & Gibbs Superfund site in New
Brighton, Minnesota. In this test, the three component technologies of the
BSWS were: the Soil Washer (SW), Aqueous Treatment System (ATS), and
Slurry Bio-Reactor (SBR). Soils were treated that were contaminated with
wood treatment process wastes, including creosote-fraction polynuclear
aromatic hydrocarbons (PAHs) and pentachlorophenol (penta).
SITE Demonstration of the BioTrol Soil Washing System (BSWS)
This SITE demonstration evaluated the soil washing system developed
by BioTrol, Inc., of Chaska, Minnesota. The object of this demonstration
test was to determine if the concentrations of penta and creosote-fraction
PAHs in the soil entering the SW, the water entering the ATS, and in the
slurry entering the SBR, could each be reduced by at least 90 percent in
the respective effluents. A second objective was to determine the fate of
penta and PAHs by material balance calculations for the SW, to assure that
the measurements of reduction efficiencies were not affected by
uncontrolled material losses to the environment.
Science Applications International Corporation (SAIC) assisted the
EPA in conducting the demonstration test, handling community relations,
disposing of residuals generated by the test, and preparing the final test
report. The analytical laboratories of Acurex Corporation, Radian
Corporation, and SAIC analyzed samples from the demonstration test.
BioTrol worked with the EPA and SAIC by providing input to the design of
the demonstration test, and providing all necessary equipment and manpower
for demonstration of the BSWS.
TheBioTrol Soil Washing System
The SW operates on the principle that a significant fraction of the
chemicals in a contaminated soil are either physically or chemically bound
to the silt, clay or humic particles, and removal of these fine particles
leaves the bulk portion of the soil (mostly sand) relatively clean. Figure
1 is a simplified diagram of the SW, Excavated soils are screened to
remove debris and mixed with water to form a slurry. The slurry is
120
-------
subjected to a series of intensive scrubbing and physical classification
steps to scour the contaminants and silt and clay fines from the sand
particles. The washed sand is separated from the slurry, and the remaining
contaminated fines can be treated in the SBR. Contaminated process water
from the SW can be sent to the ATS to remove penta and PAHs.
The ATS is a microbiological system for degrading toxic organics in
contaminated water. It consists of naturally occurring microbes growing on
plastic support material in tanks. Figure 2 is a simplified diagram of the
ATS. The SW effluent process water's pH is adjusted and nutrients are
added to optimize the performance of the ATS microbes. Next, BioTrol adds
a specific, naturally occurring microorganism to the microbes already
present in the SW effluent. This combination of microbes rapidly degrades
the penta and PAHs into carbon dioxide, water, and inorganic chloride,
which are harmless products .
The SBR is a three-stage microbiological system for treating
degradable organic contaminants associated with the fine soil particles.
The equipment used in the demonstration was a pilot-scale EIMCO BioLift
reactor system manufactured by the EIMCO Process Equipment Company.
BioTrol uses the SBR to remove contamination from the clay and silt
discharged by the SW. The SBR consists of three upright, continuously-
stirred reactors, each with a capacity of 60 L . Figure 3 is a simplified
diagram of the SBR. The silt and clay slurry enters the first reactor
where the degradation of organic contaminants by the pre-inoculated microbe
population begins. As the slurry flows to each successive tank, the
contaminants are further degraded to inorganic products.
The Test Site
The MacGillis & Gibbs Company has operated a wood treatment facility
on this site since about 1920. Contaminants present at the site include
creosote, penta, and chromated copper arsenate (CCA) . Creosote was used
as a wood preservative from about 1920 until about 1950. During the late
1940s the MacGillis & Gibbs Company began using a 5-percent mixture of
penta in fuel oil. Penta was phased out in the mid-1970s and replaced by
CCA. The MacGillis & Gibbs site was placed on the National Priorities List
(NPL) in conjunction with the neighboring Bell Lumber and Pole site in
September 1983 because of surface and groundwater contamination.
The site is underlain by the New Brighton Formation, consisting of
silty, fine to medium grained sands with intermediate and laterally
discontinuous silt and sand lenses. Processed penta waste had been placed
into the disposal area, from which contaminated soil was obtained for the
121
-------
demonstration test. The soil is particularly amenable to soil washing
because of the high proportion of sand,
DEMONSTRATION TEST DESIGN
Test Plan __...••.
Two demonstration tests were conducted. For the first test, the SW
ran for 2 days, using the low penta concentration soil (100-200 ppm), and
the ATS ran for about 4 days, but starting a day later than the SW. The
second test included all three BSWS technologies, using high-penta soil •
(500-1000 ppm) from the disposal area, and lasted 7 days. On the fourth
day of the high-penta soil test, the SW continued to operate, but 2,000 L
of fine particle slurry from the thickener was diverted to a holding tank
as feed for the two-week SBR test.
During the low-penta test, contaminated SW process water from the low
concentration soil test was treated by the ATS. Automatic composite
samples of the ATS system influent and effluent were collected. During the
high-penta test, the contaminated water from the SW was treated in the ATS
and recycled to the SW. Sampling of the ATS influent and effluent
continued until all process water generated by the SW was treated. Solids
released from the ATS on the last day of testing were,collected in a bag
filter. A carbon canister collected exhausted mists and vapors from the
ATS during the test.
For the SBR test, manual composite samples of the input and output
streams were collected beginning about two system retention times (2 x 5
days) after the start of continuous operation. The SBR effluent slurry was
collected in a drum for eventual dewatering and disposal. A carbon
adsorption canister was used to collect any fugitive emissions from the
SBR.
RESULTS
TheSoil Washer
Low-Penta Soil Test--
Table 1 shows the concentrations of penta in the feed soil and the
effluent streams from the soil washer. The mass balance of solid and water
streams for the low-penta soil test showed a 4 percent deficit in the
output relative to the input, a minor discrepancy. The penta mass balance,
however, showed a 40 percent gain in the output streams relative to input
122
-------
streams. The possible reasons for this large discrepancy, in which the
input mass of penta was probably undervalued, are discussed in the
Conclusions section. The following performance results are corrected based
on the Total Effluent penta as correct, and assume no gain relative to
Total Influent penta.
With 176 mg/kg penta (corrected) in the feed soil, the washed soil
showed 12.3 mg/kg5 a 93 percent decrease in penta concentration for that
soil fraction relative to the whole feed soil. The penta in the fine
particle cake was 310 mg/kg. Thus, 3 percent of the effluent mass
contained 37 percent of the penta. In comparison, the washed soil
contained only 8 percent of all penta in output streams. The fine oversize
and coarse oversize efflue'nt streams together accounted for 16 percent of
the penta, and the combined dewatering effluent accounted for the rest, 39
percent (treated in the ATS).
Seven PAHs were detected in the feed soil, at levels of 5-140 mg/kg,
and totaling 240 mg/kg. Levels of 25 mg/kg total PAHs were found in the
washed soil, while 700 mg/kg of PAHs was found in the fine particle cake.
The contamination reduction for the washed soil was 90 percent.
High-Penta Soil Test--
Table 2 shows the concentrations of penta in the feed soil and the
effluent streams from the soil washer. The mass balance of solid and water
streams for the high-penta soil test showed a 4 percent increase in the
output relative to the input, a minor discrepancy. The penta mass balance,
however, showed a 50 percent gain in the output streams relative to input
streams. The possible reasons for this large discrepancy, in which the
input mass of penta was probably undervalued as in the low-penta test, are
discussed in the Conclusions section. The following performance results are
corrected based on the Total Effluent penta as correct, and assuming no
gain or loss relative to Total Influent penta.
With 980 mg/kg penta (corrected) in the feed soil, the washed soil
showed 85.7 mg/kg, a 91 percent decrease in penta concentration for that
soil fraction relative to the whole feed soil. The penta in the fine
particle cake was 1290 mg/kg. Thus, 3 percent of the effluent mass
contained 29 percent of the penta. In comparison, the washed soil
contained only 11 percent of all penta in output streams. The fine
oversize and coarse oversize effluent streams together accounted for 30
percent of the penta, and the combined dewatering effluent accounted for
the rest, 30 percent (treated in the ATS).
123
-------
Nine PAHs were detected in the feed soil, at levels of 15-70 mg/kg.
The total concentration of PAHs in the feed soil was 340 mg/kg. The PAHs
showed 83-93 percent decrease in the washed soil. The total concentration
of PAHs in the fine particle cake was 970 mg/kg, or 285 percent higher than
in the feed soil.
The Aqueous Treatment System
The ATS operated at a 11 L/min rate during the test, corresponding to
a 3 hr residence time. The ATS showed a mean decrease of penta in the Low
Soil Test from 13 mg/L to 1.4 mg/L, an 89 percent removal of penta.
For the High Soil Test, penta levels decreased from a mean of 41 mg/L
to a mean of 2.2 mg/L, a 94 percent removal of penta.
The levels of all PAHs were below detection limits for most influent
and effluent samples in both the Low and High Soil Tests, so a percent
removal efficiency could not be calculated.
The Slurry Bio-Reactor
The SBR showed an increasing efficiency in penta degradation, both in
the aqueous phase and in the wet solids, up to the end of the 14-day test.
The penta degradation in the aqueous fraction was 91 percent on day 11
(relative to day 6 influent, since the SBR residence time was 5 days), and
97 percent on days 12-14. The solids cake, from lab filtration of the
influent and effluent slurries, showed the following efficiencies for days
11-14: 65, 61, 79, 92 percent.
Only three PAHs were detectable in both influent and effluent, and
they showed a trend toward higher removal efficiencies near the end of the
test. Comparing day 14 effluent to day 9 influent, removals for chrysene,
fluorene, and pyrene were 86-99 percent.
CONCLUSIONS
The BSWS is a combination of three effective treatment processes for
organic hazardous waste. The SW uses water for leaching some of the water
soluble penta, which is subsequently mineralized biologically in the ATS.
The SW also accomplishes separation of the large, relatively clean sand
fraction from the highly contaminated fine particles (mostly silt and clay)
and from the coarse and fine oversize particles by scouring and size
classification processes.
124
-------
The efficient separation of the relatively clean washed soil fraction
(predominantly sand) from the other solid fractions is a useful waste
volume reduction process for contaminated soil cleanup. The fine and
coarse oversize were woody material, mainly chips and sawdust, with a
significant BTU value if dried before incineration. The fine particle cake
was a concentrated, highly contaminated waste which can also be treated by
incineration, or treated in the SBR. The SBR was quite efficient in
treating a slurry of the fine particle fraction, removing both penta and
PAHs. The ATS efficiently degrades penta in the combined dewatering
effluent, removing the penta solubilized in the SW, and allowing the water
to be reused.
The increase in the total mass of penta in the effluents relative to
influents may be caused by poor extraction recoveries of penta during the
analysis of the feed soil. The particle size separation and abrasion in
the presence of surfactants during the soil washing may free penta from the
matrix, making it more extractable in the effluent samples.
The MacGillis & Gibbs site is one of 54 wood preserving sites
currently listed on the NPL. The BioTrol soil and water treatment
technologies are potentially applicable to cleaning up these sites, as well
as other sites with soil and water contaminated by organics. For sites
where the contaminants would not be readily biodegraded, the SW could
perform waste volume reduction, to reduce the cost of subsequent treatment
and disposal options for the contaminated aqueous and fine fractions.
REFERENCES
1. Crawford, R. L., and W. W. Mohn, Microbiological Removal of
Pentachlorophenol From Soil Using a Flavobacterium; Enzyme Microbiology
and Technology Z, 617-620; 198i.
2. Saber, Diane L., and R. L. Crawford, Isolation and Characterization of
Flavobacterium Strains That Degrade Pentachlorophenol; Applied
Environmental Microbiology 50, 1512-1518; 1985.
3. Frick, Thomas D., Ronald L. Crawford, Michael Martinson, Tom Chresand,
and George Bateson, Microbiological Cleanup of Groundwater Contaminated
by Pentachlorophenol; Reducing Risk from Environmental Chemicals
Through Biotechnology, G. Omenm and Others, Editors; Plenum Press, New
York, 1988.
4. Twin City Testing Corporation, 1987; Remedial Investigation Report,
HacGillis and Gibbs Company, New Brighton. Minnesota. February 17, 1987,
125
-------
Table 1. Soil Washer Mass Balance Results: Low-Penta Soil Test
Influent &
Cf #!i irtrit &Af*&e*
INFLUENTS;
Feed Soil
Municipal Water
Flocculant Solution
TOTAL INFLUENTS
kg
11,210
61,710
4,914
77,830
Percent
14%
79%
6%
100%
Concen-
tration
(mg/kg)
126
0
0
CORRECTED TOTAL PENTA*
CORRECTED FEED SOIL
EFFLUENTS:
Washed Soil
Fine Particle Cake
Fine Oversize
Coarse Oversize
Combined Dewatering
Effluent
TOTAL EFFLUENTS
Gain or Loss
PENTA**
12,940
2,374
653
1,594
56,850
74,410
- 3,420
17%
3%
1%
2%
76%
100%
• - 4%
176
12.3
308
101
162
13,4
. Panto
Total
(g)
1,412
0
0
1,412
1,976
159
731
66
258
762
1,976
+ 564
Percent
of Total
in or Out
100%
8%
37%
3%
13%
39%
1 00%
4- 40%
Based on total effluent penta.
Based on corrected total influent penta and Feed Soil mass.
126
-------
Table 2. Soil Washer Mass Balance Results: High-Penta Soil Test
Influent &
WlllUdll IVIdOO
INFLUENTS:
Feed Soil
Municipal Water
ATS Treated Water
Flocculant Solution
TOTAL INFLUENTS
kg
17,590
4,182
73,550
13,450
108,800
Percent
16%
4%
68%
12%
100%
Concen-
tration
(mg/kg)
650
0
2.65
0
CORRECTED TOTAL PENTA*
CORRECTED FEED SOIL
EFFLUENTS:
Washed Soil
Fine Particle Cake
Fine Oversize
Coarse Oversize
Combined Dewatering
Effluent
TOTAL EFFLUENTS
Gain or Loss
PENTA**
22,860
3,931
1,007
3,132
82,300
113,200
+ 4,400
20%
3%
1%
3%
73%
100%
+ 4%
980
85.7
1290
932
1370
62.8
Total
(g)
11,430
0
195
0
11,620
17,240
1,959
5,071
939
4,291
5,168
17,430
+ 5,810
Percent
of Total
In or Out
98%
2%
100%
11%
29%
5%
25%
30%
100%
+ 50%
* Based on total effluent penta, with ATS Treated Water penta subtracted.
* * Based on corrected total influent penta and Feed Soil mass.
127
-------
ro
CO
f Municipal ^
Water
^ Supply J
J Oversize | [ Coarse |
1 Debris J I Oversize J
n /l
i
I Contaminated L^. SCIOM & Mixing ^ Scraon to» |:rottl
I Soil J °"OT" Trommel "^ "*"" "^ Flotation
4
* '
..
4
i
Aqueous Combined
Treatment «*-• Dawatering ^-_ Thickef,jna
System Effluen' , ,
See Fig 2 Holding Tank
K&y* i '
i ) UnitC^«r«iion /^ Sludge A
I Underflow or I.
(~~) Procas* Stream or 1 Fine Particle I
— Output I Surry )
-*• Piracttonotnow ^ .
~- " w
•*"" Optional Direction y
of Flow
*— Dewatering
j Fine j
I Oversize I
I ,
1 •
Multi-Stage
^ Countercurrent
Attrition/Classifcatio
Circuit
4
r
| Fine Silts, ]
1 Clays, and 1
1 Organics J
i
t
Slurry
See Fig. 3
*
«
fc[ Fine Particle]
1 Cake J
^" Washed "
^ So'ri
" Product
1 Cationic 1
I Polymer I
Figure 1. Flow Diagram of the Soil Washer (SW)
-------
To Atmosphere
1
Carbon
Canister
i
Off Gas
V J
vo
Contaminated
Water from
Soil Washer
K«y:
i J Unit Operation
on
imorOutf
Nutrition Addition
and
pH Adjustment
Ml
Heat
Exchan
iTir
'
ger
^
Heater
Bag
Filter
(Optional)
*
*
r ^
Waste
Solids
V _>
Aqueous
Treatment System
t
Blower
Recycle to
Soil Washer
Treated Water
To Carbon Filter
andPOTW
(End of Test Only)
Figure 2. Flow Diagram of Aqueous Treatment System (ATS)
-------
CJ
o
Fine Particle
Slurry (FPS)
from Soil
Washer
Key:
| \ Unit Operation
( } Process Stream or Output
Nutrients
I
Air
To Atmosphere
I
Carbon
Canister
Off Gas
t
Feed Storage
Tank
V _>
i.
Slurry
Bio-Reactor
(SBR)
/
f
Treated
Slurry
V
Figure 3. Flow Diagram of the Slurry Bio-Reactor (SBR)
-------
BACKGROUND INFORMATION ON CLEAN PRODUCTS
RESEARCH AND IMPLEMENTATION
by: Beverly J, Sauer, Robert s. Hunt, and
Marjorie A. Franklin
Franklin Associates, Ltd.
Prairie village, KS 66208
ABSTRACT
The concept that products can be made "environmentally
friendly" or "clean" has been attracting much attention.
However, there is as yet no accepted definition of what is
meant by "environmentally friendly," nor any agreement on how
to achieve clean products. This paper provides information on
the current state of research and implementation on clean
products and identifies issues to be resolved. The focus is on
consumer products, although the same criteria and methodologies
can be used for any product or process.
This paper has been reviewed in accordance with the U.S.
Environmental Protection Agency's peer and administrative
review policies and approved for presentation and publication.
131
-------
THE LIFE CYCLE OF PRODUCTS
The manufacture, use, and disposal of products can impact
on resources and the environment at every stage in the
product's life cycle. The life cycle of a product moves from
extraction of raw materials to processing stages and on through
manufacture. The product then goes through the distribution
channels to the consumer. Finally the product is consumed,
disposed of, or perhaps recycled.
While there is general agreement that it is desirable to
minimize a product's overall impact on resources and the
environment, it is not so easy to determine what the impacts
really are and how one product compares to another. Many
claims that a given product is "environmentally friendly" are
based on only one of the many possible points or types of
impact on the environment.
EXISTING AND PROPOSED CLEAN PRODUCTS PROGRAMS
Existing labeling programs (Figure 1) range from well-
controlled national programs and simple shopping guidelines
recommended by various consumer/environmental groups to
labeling claims with undefined technical basis made by
manufacturers and retailers trying to cash in on consumers'
rising environmental concerns.
GERMANY: BLUE ANGEL
The Federal Republic of Germany is clearly the pioneer in
the field of national environmental labeling. Its "Blue Angel"
program has been in existence since 1978 and is used by other
countries as a model. Over 3,000 products in 57 product
categories now carry the Blue Angel label. The program defines
clean products as those which:
"when compared with other products fulfilling the same
function and when considered in their entirety, taking into
account all aspects of environmental protection (including
the economical use of raw materials), are as a whole
characterized by a particularly high degree of
environmental soundness without thereby significantly
reducing their practical value and impairing their safety."
It is claimed that a cradle-to-grave approach is used in
evaluating products for the label? however, it appears that,
lacking any outstanding environmental impacts in other areas,
differentiation of products in a given category is usually made
on a single criterion. This criterion may be recycled content,
reusability, or some other environmental concern.
132
-------
Figure 1
ENVIRONMENTAL LABELS USED
IN OTHER COUNTRIES
WEST GERMANY - Blue Angel
CANADA - Environmental Choice
JAPAN • Ecomark
Nordic Environmental Label
133
-------
Criticisms of the program include; 1) failure to update
(tighten) criteria; 2) not enough emphasis on quality and
usability of labeled products; 3) use of a single
environmental criterion; 4) failure to provide broader
labeling opportunities; 5) no guarantee that an unlabeled
product may not be equally as environmentally sound as a
labeled one, or even superior; and 6) because of exclusion
from consideration of some product categories, some
manufacturers use their own labels, with resultant consumer
confusion.
CANADA: ENVIRONMENTAL CHOICE
Canada's Environmental Choice program produced its first
three guidelines in summer, 1989. As of October 1990, 26
product category guidelines had been submitted for public
review or finalized and put into use.
The program literature discourages use of the term
"environmentally friendly" in favor of referring to products
which "reduce the burden on the environment." A product which
is a good environmental choice is "any product which is made,
used or disposed of in a way that causes significantly less
harm to the environment than other similar products."
JAPAN: ECOMARK
Japan's environmental labeling program was launched in
February 1989. The program aims to promote "clean" innovation
by industry, heighten consumers' environmental awareness,
recommend products which contribute to environmental protection
and conservation, and symbolize an ecological lifestyle.
"Clean" products considered for labeling are those which
cause little or no pollution when used or discarded, improve
the environment in use, or otherwise contribute to conservation
of the environment. The logo's use will also be applied to
environmentally favorable activities such as recycling
programs.
To qualify for the Ecomark, products must have been
manufactured with attention given to preventive measures being
taken against environmental pollution in manufacturing, product
disposal not involving difficult processing, opportunity being
available for conserving energy or resources through use of the
product, demonstration of compliance with quality and safety
laws, standards, and regulations, and price not being
excessively higher than that of comparable products.
134
-------
NORDIC COUNTRIES
In November 1989 the Nordic Council of Ministers agreed to
implement a voluntary environmental labeling program. Common
criteria developed with the cooperation of all participating
Nordic countries and a common label will be used. The
environmental performance of selected product groups will be
assessed in terms of such factors as raw material extraction
involved, production processes used, and disposal methods
available, and a set of minimum requirements will be
established. In some cases, the label will be granted to the
least harmful product in a group, while in other cases the
label will be granted to products that represent an
alternative, more environmentally sound means of satisfying
consumer needs.
Participation of individual Nordic countries will be
voluntary. Norway, Sweden, Iceland, and Finland have all
indicated that they will participate, while Denmark is waiting
to see whether the European Economic Community will adopt a •
labeling program before it decides whether to participate.
EUROPEAN COMMUNITY
In 1989, a feasibility study on an EC environmental
labeling system was conducted for the Commission of European
Communities by the Danish Technological Institute in
cooperation with the University of Lund, Sweden. The summary
report was published in January 1990.
The EC has proposed a plan for an environmental labeling
program in which companies apply to their national government
for the label, which would be awarded by an independent jury
set up at the EC level. Once established, the program would be
taken over by the planned European Environmental Agency.
AUSTRALIA: GREEN SPOT
Australia is preparing to launch a labeling program late in
1990. The Green Spot program is proposed to identify and label
consumer products which are environmentally sound in terms of
four broad impacts: 1) they cause substantially less pollution
than other comparable products; 2) they are recycled and/or ,
recyclable; 3) they make a significant contribution to saving
non-renewable resources or minimizing use of renewable
resources; and 4) they contribute to a reduction of adverse
environmental health consequences. Types of products which are
considered universally environmentally benign are not to be
included in the labeling program.
135
-------
U.S. FEDERAL, STATE, AND REGIONAL ACTION
A task force of Attorneys General of eight states held a
public forum on environmental marketing in March 1990. A
report of the findings and recommendations, released in
November 1990, calls on the Federal Trade Commission and the
EPA to work jointly with the states to develop uniform national
standards for environmental marketing claims. The EPA, FTC,
and White House Office of Consumer Affairs have already begun
meeting to address this issue. The Attorneys General task
force, now representing ten states, will hold hearings on its
environmental marketing recommendations in December.
Environmental labeling legislation has been introduced on
the federal level, with Senator Frank Lautenberg's
Environmental Claims Act of 1990. The Act calls on the EPA to
provide uniform and accurate standards and definitions for
environmental marketing claims.
In the absence of a nationally authorized environmental
labeling program in the United States, individual states and
regional organizations have begun to attack the issue of
"environmental friendliness" and labeling. New York, Rhode
Island, Connecticut, and New Hampshire have passed legislation
governing use of a recycling logo, while California recently
passed an Environmental Advertising Act. Legislative efforts
primarily have been directed at defining and banning
environmentally unacceptable goods, rather than promoting
"clean" products. Judgments of whether or not goods are
environmentally friendly are usually based on recyclability,
degradability, and reusability.
Perhaps more than any other, the issue of degradability
illustrates the differences in perceptions of what is better
for the environment. While many states have bills seeking to
ban nondegradable plastics, many bills have also been
introduced to ban degradable plastics because of the lack of
information on the identity and effect of products which may be
mobilized by the breakdown of the material and the possibility
of contamination of plastics recycling operations. Also,
proposed legislation often does not specify a preferred or
optimum substitute material for banned materials, or does not
indicate that the environmental effects of substitute products
have been thoroughly considered.
CONEG (Coalition of Northeastern Governors) has focused its
attention on the issue of environmental responsibility in
packaging. Its preferred packaging guidelines, in order of
preference, are: no packaging, minimal packaging, consumable,
returnable, refillable/reusable packaging, and recyclable
packaging or recycled material in packaging. CONEG is also
136
-------
supporting requirements for removal of toxic agents such as
lead, cadmium, and mercury from packaging.
LABELING ORGANIZATIONS
Scientific Certification Systems of Sacramento, California,
is one of the most recent entrants into the field of product
certification and labeling, with its "Green Cross" program.
The company certifies environmental claims in specific areas
which can be scientifically documented and awards a seal to
recognize products with outstanding achievements in these
areas. As of September 1990, Green Cross had certified various
products on the basis of recycled content, including products
in the categories of paper products, glass containers, and
recycled plastic bags. In the future Green Cross plans to look
at products on the basis of biodegradability (soaps,
detergents), energy savings (appliances), and production from
sustainable, renewable resources. Eventually the organization
may issue an overall environmental seal of approval.
Green Seal is another labeling organization with a
different approach. Rather than focus on individual
characteristics, Green Seal plans to evaluate products'
environmental impacts on a cradle-to-grave basis. Public
comments have been solicited on recycled paper criteria, and
the first seals are expected to be issued in early 1991.
ENVIRONMENTAL GROUPS
The Pennsylvania Resources Council (PRC) sponsored an
environmental shopping seminar in March of 1990. They also
publish an environmental shopping guide, which recommends
buying items packaged in recycled or recyclable materials or
reusable containers and avoiding mixed material packaging and
excessive packaging.
The New York Public Interest Research Group has put out a
pamphlet which is similar to PRC's guide in its recommendations
on packaging. Consumers are urged to avoid single-use,
disposable items, difficult-to-recycle or non-recyclable
packaging, and toxic packaging and to look for reduced, reused,
and recycled products.
Among environmentalists, enthusiasm for green marketing is
high. In an informal telephone survey, many environmental
organizations were eager to hear of any developments in this
area, particularly regarding the possibility of the beginning
of a standardized approach to environmental labeling claims.
Many environmental shopping guides are now widely
available, as well as "save the world" books that contain
product/packaging recommendations. Shopping and environmental
137
-------
action guides include books and pamphlets published by various
individuals and environmental groups intended to provide the
reader with information to use in making purchasing decisions,
investments, etc., that will have the least effect on the
environment. Most of these publications do not claim to
provide the answer of what is environmentally best, but rather
aim to help readers make informed choices or modify their
habits in order to minimize waste and pollution and conserve
resources. The following are only a few of the guides
available: Shopping for a Better World; The Green Consumer;
The Green Consumer's Supermarket Shopping Guide; 50 Simple
Things You can Do To Save the Earth; How to Hake the World a
Better Place; and Save Our Planet—750 Everyday Ways You Can
Help Clean Up the Earth.
These guides generally appear to use a single-criterion
approach to evaluating products. None attempted even a
simplified life cycle assessment. When more than a single
criterion is used, the presentation of data is often one-sided.
Data on a single topic may vary considerably from book to book
depending on its source.
PRIVATE ORGANIZATIONS (COMPANIES, SUPERMARKETS, ETC.)
Many manufacturers are eager to respond to environmental
concerns by labeling their products "environmentally friendly."
Many manufacturers are sponsoring evaluations of the
environmental implications of their products. The results of
these evaluations may be published in private reports, in
informational pamphlets, or as advertisements.
Manufacturers' environmental labels typically are based on
a single environmental criterion, providing no clue as to
whether any other environmental impacts were considered.
Common criteria for labeling claims are those with high public
interest or visibility, such as recycled content,
degradability, and lack of CFCs in content or manufacture.
In the various programs discussed, criteria have been
developed for various product groups, groups selected for one
or more of the following reasons: 1) the product is a major
constituent of the waste stream; 2) the product has a
significant impact on the waste stream due to toxicity, etc.;
3) product use provides a substantial environmental benefit;
4) the product meets safety and quality requirements for normal
use; 5) product requirement levels for the label are high to
challenge industry to meet or exceed current levels of clean
technology; 6) the product is easy to evaluate; 7) the
product is commonly used; and 8) the product does not shift
environmental impacts from one area only to create problems in
another.
138
-------
CRITERIA THAT HAVE BEEN USED TO EVALUATE PRODUCTS
Criteria that have been used to evaluate products are
discussed below. Most "clean products" recommendations are
based on one or a few of these criteria rather than a total
environmental impact evaluation.
RECYCLED CONTENT
Recycled content is the most popular and widespread
criterion used. It is a popular criterion because ,of wide
recognition and support by consumers? however, different groups
may use different definitions or requirements for recycled
content.
RECYCLABILITY/REUSABILITY
In the United States, recyclability and reusability are
widely used as criteria in legislation although definitions may
vary. For example, proposals in Massachusetts and Oregon to
ban environmentally unacceptable packaging differed in their
definitions of "recyclable." Obviously, standard definitions
of terms would be a step in the right direction. In addition,
proposed legislation is not specific on the materials or
containers that are to replace those deemed unacceptable.
DEGRADABILITY
Degradability is a popular and widely disputed criterion.
It has been heavily used as an advertising point, but is
currently being questioned or even denounced by many
environmentalists. Many manufacturers and retailers focus on
degradability as a positive characteristic; however, at least
one "environmentally conscious" mail order company has
temporarily withdrawn its biodegradable plastic bags for re-
evaluation of their environmental effects. Some legislative
proposals have called for bans on nondegradable plastics, while
others have attempted to eliminate degradables.
HAZARDOUS/TOXIC MATERIAL CONTENT
This criterion can be used as justification for .the
necessity of environmental labeling or can be used to
disqualify products from eligibility for labeling.
WATER POLLUTION IMPACTS
Water pollution has not been a major criterion although it'
is given specific attention in several environmental guides,
particularly those having to do with phosphates and bleaches in
detergents and biodegradability of various household products.
139
-------
SOIL POLLUTION IMPACTS
This is another criterion that is an integral part of a
cradle-to-grave analysis, but is not a popular single criterion
for labeling, except perhaps in the case of organically grown
foods. Soil pollution is given some attention as an issue
associated with the disposal of batteries and concern about the
fecal wastes in disposable diapers. It is also being mentioned
as a concern regarding degradable plastics due to lack of
knowledge about the identity and effect of degradation
products.
AIR POLLUTION IMPACTS
The most popular air pollution issue in the past few years
has been CFC content or "ozone friendliness," which has been
covered by labeling programs, environmental shopping
guidelines, and proposed legislation. Labeling programs and
shopping guides generally focus on aerosol products, while
legislation tends to focus on foam plastics. Air pollution
effects are also used as a criterion when discussing disposal
of products by incineration.
NOISE POLLUTION IMPACTS
This criterion is little used in the United States;
however, it has been used as the primary criterion in labeling
certain West German products, e.g., lawn mowers, car mufflers.
PRODUCTION PROCESSES USED
The draft Canadian guidelines for re-refined oil and
recycled cellulose construction materials specified acceptable
processes for oil demetallization and hydrotreating and for use
of a dry process to produce recycled paper products. These
specifications were removed from the final guidelines.
USES OF RESOURCES (INCLUDING ENERGY)
This criterion can be subdivided into use of energy and use
of resources. Unless energy usage is the primary evaluation
criterion, it is hard to tell whether it has been addressed in
assessing environmental impact. It is difficult, for example,
to determine whether the increased energy usage for collection,
transportation, cleaning, and distribution of reusable
products, such as refillable glass bottles, has been considered
by legislative bodies seeking to ban certain disposable
products.
Product recommendations on the basis of resource
conservation are most often directed at plastics as a user of
140
-------
petroleum (considered a non-renewable resource) and paper as a
user of wood (considered a renewable resource).
OTHER CRITERIA
Other criteria that have been used are: use of more benign
products/processes, general requirement of safety/usability,
amount or type of packaging, provision of information for the
consumer, overall corporate reputation regarding social/
environmental issues, effect on rainforest, longer lasting or
repairable products, weight or volume contribution to landfills
or waste streams, or disposal problems.
METHODOLOGIES THAT HAVE BEEN USED TO EVALUATE PRODUCTS
PRODUCT LIFE CYCLE ASSESSMENT
Environmental problems potentially can be alleviated by
either direct or indirect means. A direct means would include
bans of specific materials, processes, etc., or economic
incentives or disincentives (such as grants or taxes) which
have an immediate effect. Examples of indirect means include
the banning of a product or the substitution of one product for
another in order to correct some problem not as clearly linked
to the product. For example, the global warming problem could
be addressed by banning products whose manufacture generates
large carbon dioxide emissions.
The worth of either a direct or an indirect approach can
only be assessed by a life cycle assessment which examines the
entire complex of operations associated with a product. The
theory behind the use of product substitution or banning as a
means for environmental benefit is that if the product is not
purchased, then the manufacturing and processing will cease
and, along with it, the environmental consequences will cease.
However, the substitute product also produces environmental
consequences that need to be evaluated. A narrow focus
analysis can greatly err in assessing the actual impact of any
action that affects purchasing habits.
In comparing a given product or a set of products on two or
more environmental issues, even life cycle assessment may not
be enough to give adequate guidance. The reason is that there
are no weighting factors that tell how to compare environmental
impacts, for example, of one pound of toxic heavy metal sludge
to one gallon of water usage or consumption of one Btu of
energy.
Up to this time, life cycle assessments have focused on
performing an "inventory" (listing and quantification) of the
materials and energy used and environmental releases (air,
141
-------
water, solid waste) from all stages in the life of a product
from raw material acquisition to ultimate disposal (Figure 2).
However, a group of life cycle assessment experts at an August
1990 workshop in Vermont sponsored by The Society of
Environmental Toxicology and Chemistry agreed that a more
sophisticated approach is needed. In addition to the inventory
of energy, materials, and releases, the associated
environmental impacts should be analyzed (entering into the
areas of risk and hazard assessment), and the changes needed to
bring about environmental improvements should be analyzed as
well. Methodology for this more comprehensive type of life
cycle assessment is currently being developed for the EPA.
MATRIX APPROACH (PASS/FAIL)
While the life cycle assessment approach is the most
comprehensive methodology, a more common approach is use of a
matrix with "pass/fail" ratings. For example, a widely used
consumer guide in England lists a variety of packaging
materials with a series of environmental criteria (recyclable?,
degradable?, etc.,) with yes/no answers and some comments.
WEIGHTING SYSTEMS
A problem with either the life cycle assessment approach or
the matrix approach is that decisions as to which criteria are
most important are left to the reader or consumer of the
product. For example, two products (say aluminum cans and
glass bottles) can be compared using a life cycle assessment.
One product may "win" based on air pollution impacts and the
other may "win" based on the amount of solid waste to be
disposed. Which is more important?
ISSUES/TECHNICAL PROBLEMS TO BE RESOLVED
SELECTION OF PRODUCTS TO Bl EVALUATED
Several of the criteria used in product selection are
somewhat vague or controversial, such as selecting products
that are a significant factor in the waste stream (the issue
here is definition of the term, "significant"), selecting
products that are simplest to do, or selecting products that do
not contain hazardous components.
COMPLETE LIFE CYCLE ASSESSMENT VERSUS EASIER, QUICKER
METHODOLOGIES
Decisions must also be made as to whether a complete life
cycle assessment should be made as opposed to use of easier,
quicker methodologies. The main concerns here are time, and
expense involved in analysis versus environmental benefit, and
142
-------
U)
Energy
Energy
Raw materials
Acquisition
I
Wastes
Wastes
Energy
Energy
i
I
Wastes
|
i
I
Wastes |
Package Recycling
Energy
Materials
Processing
Packaging
Manufacture
I t I :
Bottling or
Filling and
Distribution
1 '
| !
t
Soft Drink
Consumption
Reuse
/*
Final Disposal,
Recycle,
or Reuse
Wastes
.*
Figure 2. General materials flow for "cradle-to-grave" analysis of soft drink distribution system.
-------
consumer loyalty and possible disillusionment or confusion. A
true cradle-to-grave life cycle assessment is time-consuming,
expensive, and raises difficult questions about weighting the
relative importance of various environmental impacts. Also, .
once environmentally committed consumers have embraced an idea,
it may be difficult to change their minds with the facts.
HANDLING TRADE-OFFS
In complete life cycle-assessments, summaries of
environmental impacts, such as total energy usage or water
usage associated with one product', can be 'directly compared to
the same impacts associated with another. Problems arise over
weighting different impacts relative to each other. There are
no established scientific methodologies for deciding which is
more important. Some ways to handle these trade-offs are;
weighting systems (this involves subjective judgment as to
which components are least desirable or'most harmful);
pass/fail systems (using quantitative comparisons with minimum
or maximum allowable levels); letting the consumer decide
(abandon a simple logo and present environmental impact
information, letting the consumer decide what is
environmentally preferable); or using only one easily
determined criterion (the advantage is easy evaluation by
consumers, but the disadvantage is oversimplified and possibly
erroneous conclusions on environmental impacts).
IMPLEMENTATION ISSUES
There are several options as to who should implement
labeling programs. Some Clean product programs 'are in effect
at the national level," but not in the United States. A number
of states, however, are moving in the direction of some kind of
"environmentally friendly" product regulations. The states
that are in CONEG have been particularly active in this regard.
Environmentalist or other nonprofit groups also are involved in
studying clean product/source reduction issues. Finally, many
private companies have been carrying out their own initiatives,
sometimes in connection with groups like CONEG.
POLICY IMPLICATIONS
There would be some advantages to implementing a clean
product program nationwide. However, having a federal program
would not necessarily preclude states or other organizations
having their own programs as well. There would be the need to
determine which agency(ies) would implement the program,
research would be required, and an implementation mechanism
would have to be developed and administered.
Another issue is the necessity to update the criteria used
to measure products. These criteria can change with time as
144
-------
research provides new information on environmental phenomena.
Also, new products are continually being developed, and
processes used to manufacture products also evolve over time.
RECOMMENDATIONS
Both manufacturers and consumers generally appear to
recognize the potential benefits of labeling. As yet, however,
no universally accepted and supported course of action has been
identified. Current efforts by various individual groups may
be well-intentioned, but do not adequately address the
comprehensive environmental impacts associated with a product's
entire life cycle and, therefore, may offer consumers misguided
direction.
Additional effort in several areas could aid in the
development and support of clean products programs. These
areas include: standardized definitions and usage of
environmental impact terminology, survey of consumers to find
out what types of information/education would be most useful,
further development of methodologies to thoroughly and
effectively evaluate products on a life cycle basis,
development of a standardized environmental labeling program,
and other reward incentives for manufacturers who provide
cleaner products.
Additional measures that could minimize environmental
impacts of consumer products might include: education on
proper use and disposal of products, elimination of high
environmental impact products for which acceptable, less
damaging alternatives exist, elimination of excess packaging,
and efforts to reshape today's convenience-oriented consumer
perspective to a more environmentally responsible attitude.
The benefits to the environment, and consequently to
mankind, that may be gained by support of clean products are
considerable. The information provided and issues raised in
this report can serve as a starting point.
145
-------
INDUSTRIAL POLLUTIONPREVENTION STRATEGY;
RESEARCH PRIORITIES AND OPPORTUNITIES FORTHE 1990«S
ivars Licis
Pollution Prevention Research Branch
Risk Reduction Engineering Laboratory
U.S. Environmental Protection Agency
26 W. Martin Luther King Drive
Cincinnati, OH 45268
ABSTRACT
This paper is based on research performed under contract to
EPA for the purpose of helping the Agency (Pollution Prevention
Research Branch) establish a data base .for use in identifying
pollution prevention priority research areas. This research study
determined what constitutes the more serious pollution problems
within the industrial sector, what solutions have been tried, and
what opportunities exist for technology transfer or basic research.
The project was structured around the existing Standard Industrial
Classification (SIC) system as a point of departure while also
considering the information generated by the Toxics Release
Inventory as reference. Direct knowledge and experience of people
in specific industry and pollution prevention activities were also
incorporated into the study.
On this basis, a resulting list of 17 high priority industries
were identified, and for each, a more detailed assessment was
performed to define major problems and pollution prevention
opportunities.
The paper also discusses other developments that come into
play as part in defining the course of research in this area such
as the recent announcement by the Agency to pursue goals for
serious reductions in 17 High Priority Contaminants. A list of
recommendations is offered for improving project selection as well
as future research prioritization.
This paper has been reviewed in accordance with
the U.S. Environmental Protection Agency's peer
and administrative review policies and approved
for presentation and publication.
146
-------
INTRODUCTION
The 1984 Amendments to RCRA, the Hazardous and Solid Waste
Amendments (HSWA), specifically mandates waste minimization as an
objective for the nation's environmental management program (1).
A means of implementing this directive has been the encouragement
of source reduction and recycling approaches by both industry and
the public. During the intervening years this policy has evolved
into the Pollution Prevention Program that includes all waste
generated, energy consumed and resulting relative risk to the
environment and impact on ecological systems posed by human
activity. The basic precepts of the program are the full
understanding and voluntary cooperation of all parties (producing
industry and consuming public) with regulatory incentives applied
only when needed.
OBJECTIVE
The objective of this research was to provide a data base that
could be used as guidance by the EPA for the development of a
research strategy for pollution prevention in the industrial area.
More specifically, the objective was to identify a short list of
industries, or industry segments, that present significant problems
in terms of waste generated and/or opportunities for waste
reduction through source reduction and recycling. Once identified,
each of the industries or industry segments were to be studied in
more detail to gather available information within that industry
segment and to discuss the pollution prevention problems and
opportunities with the various sets of personnel affiliated with
each segment both in the public and private sectors, for the
purpose of providing a basis for defining pollution prevention
research projects.
APPROACH
The study was designed to use the existing SIC system (2) .
A long list of SIC's (approximately 200) was to be developed along
with a list of selection criteria (Table 1) . The list is
intentionally all inclusive and does not require the simultaneous
satisfaction of all criteria.
This list was distributed to number of experts in the
pollution prevention arena (state pollution prevention programs,
various EPA departments involved with pollution prevention, the
pollution prevention offices of the Regional EPA offices, other
federal government departments, industry technical associations,
and a small number of specific industry personnel). The
participants were asked to review the long list of SIC's, and, in
147
-------
the light of the selection criteria and their expertise, pick out
20 industries or industry segments and arrange them in priority
order l through 20.
This, approach was largely dictated by the lack of readily
available, good, quantitative data, defining the pollution hazard,
the amount of pollution generated for each industry, beyond the
Toxics-Release Inventory (TRI) reports (3) for hazardous waste, and
the lack of rigorous definition of industries (SIC's or SIC
segments) as they relate to pollution prevention.
TABLE 1. SELECTION CRITERIA FOR SIC's
1. Importance of the industry to nation or society.
2. Significance of all or certain waste streams in toxicity,
volume or both.
3. Large frequency of small and mid-sized firms that would
benefit from government participation.
4. Significant benefits that would be derived from waste
minimization efforts that reduce toxicity and/or volume.
5. Waste minimization is not expected to adversely impact
product quality or marketability.
6. Waste minimization would offer cost benefits, at least in
long run.
7. Waste minimization in this industry would be readily
transferable to other industries.
8. Industry has exhibited an interest in waste minimization.
9. Waste minimization appears to be technologically achievable.
10. Industry would benefit from government involvement because
of lack of direction, capital, or technical sophistication.
11. Industry would be receptiv.e to waste minimization studies.
12. The industry will not be viable in the long run without
massive changes.
This approach, has the potential weakness of leaving out a few
priority areas while overstating others to a degree. This problem
was not considered a serious flaw at this^ stage however, because
the utilization of this information would only comprise one of the
sources of data used for research planning by the EPA, with other
148
-------
sources providing a system of checks and balances. This approach
was taken as a starting point upon which further refinements would
be made.
IMPLEMENTATION
The development of the long list of SIC's, using informed
judgement, resulted in a 175-item. mixture of SIC industries and
industry segments that were selected to generally fit the needs of
the pollution prevention research criteria.
This list and the selection criteria of Table 1 were
distributed' among the organizations participating in the research
prioritization, as discussed above, and the results compiled to
define a short list of.17, high priority SIC's.
Once the short list of SIC's was determined, a significantly
more detailed investigation was performed on each of the 17 by
gathering available information, and discussions with the various
experts in each field, technical associations, government and
industry. For each of the 17 items on the short list, the results
were compiled and also summarized in tabular format. Excerpts from
these data are included under Results, below. The full project
report and appendixes (4) are in process of being published as an
EPA Project Report to be available from the National Technical
Information Service in FY 91.
RESULTS
The resulting short list of 17 industries is presented in
Table 2. Presentation of . the complete set of problems and
opportunities is beyond the scope of this format. Sample results
are discussed for two.items on the list of 17 SIC's; |226-Textile
Dyes Dyeing, and #'753- Automotive Repair and Refinishing,
TABLE 2. SHORT LIST OF PRIORITY INDUSTRIES*
Textile Dyes and Dyeing.... '..".,..' 226
Wood Preserving. , 2491
Pulp and Paper. ......'..... .v .. .. 26
Printing.- . . . 271-275
Chemical Industry. .....' , , 281
Plastics. . . '. ." 2821
Pharmaceuticals... .. . .'. ..... . '..«•' 283
Paint Industry. , ..'.......'. .'. '..'..".'. 285
149
-------
Ink Manufacture 2893
Petroleum Industry 291
Steel Production Industry. 331
Non-ferrous Metals. 333-334
Metal Finishing (Electroplating) 3471
Electronics/Semiconductors 3674
Automotive Manufacture/Assembly 371
Laundries/Dry Cleaning, 721
Automobile Refinishing/Repair. 753
* The list of 17 has been normalized (i.e., some of the
industry segments have been aggregated in an attempt to
present industry segments of equal pollution prevention
significance. This list is in order of SIC number and not
in order of priority.
The results presented per topic are far from exhaustive, due
to business confidentiality and the limitations of the scope of
this effort.
TEXTILE DYES AND DYEING (SIC-226)
Largely because of environmental concerns, the dye industry
has undergone extensive change in the last few decades. Many of the
dyes originally used (e.g., coal tar dyes, SIC 2865) are now
considered toxic and have been replaced with material perceived to
be less dangerous. New classes of fiber reactive dyes (e.g.,
triazine based) can reduce the use of azo dyes and contribute to
lower concentrations of dyes during washing and rinsing. However,
due to business confidentiality, specific technical information was
not made available for this study (as was the case with several
other industries).
In the textile dyeing and finishing industry, extensive
changes have also been occurring, possibly as a result of changes
in the fiber blends being produced, Chromates used for oxidation
of vat dyes have been replaced by other chemicals; formaldehyde,
used in dyeing and in durable press finishes, has been reduced or
eliminated. The industry still requires a large amount of water
and generates a large amount of wastewater. Progressively more
automation is being introduced and this appears to contribute to
better control and smaller releases of pollution to all media.
This is still a developing area and improved process control
software is needed to achieve further reductions.
Specific processes seem especially attractive for waste
reduction opportunities. For example, wool scouring generates
caustic wastewater. Processes such as hyperfiltration can be used
150
-------
to recover the caustic from spent solutions for reuse/recycling.
More cost/efficient membranes are needed to reduce the payback
period and increase applicability.
Some dyes can be recovered and recycled. Economies of scale
produce a significant barrier. The capital outlay to recover
various types of dyes used are beyond present economics for other
than large mills. Less expensive, smaller scale equipment is
needed to fit the requirements of smaller mills, or mills that use
a large assortment of dye types.
Solvent finishing is an idea that was explored approximately
20 years ago, but has now largely fallen into disuse. With rising
costs, and stricter regulatory requirements, this approach could
be re-examined. For example, where degreasing of wool fabric with
caustic is currently practiced, solvent degreasing, with liquid and
vapor solvent recovery may provide an environmentally attractive
alternative. Such a "reverse" approach has been used outside the
US (Compendium, ENV/WP.2/5/Add.85) . A close look has to be made
at the total picture so that short range economics are not traded
for longer range environmental costs.
AUTOMOBILE REPAIR SHOPS (SIC-753)
The auto repair industry is a significant source of waste.
Primarily, it is a collection of small shops. As a consequence,
there is little structured pollution prevention research or
allocation of staff dedicated to pollution prevention.
The major pollutants generated by the industry are waste
paints, VOCs from spray painting, and metal-bearing dusts from
paint overspray, grinding and sanding of finishes, degreasing
solvents (often chlorinated) and oils and other automotive fluids
removed during repair and used batteries.
The current means of managing liquid wastes is usually by off-
site disposal, with the hydrocarbons either recovered by
distillation, used for their fuel value, or simply destroyed by
incineration. Small scale, on-site distillation equipment that
would allow reuse of solvents is available but has not achieved a
significant degree of use, partly because of regulatory
requirements and lack of significant economic incentives. A
demonstration of such an application would help to prove the cost-
effectiveness of recovery, and make operational data available to
the industry. The dusts generated by sanding and grinding can be
collected by existing and innovative equipment, but at this time
there have been no uses found for these mixed materials that could
be considered pollution prevention.
151
-------
In the area of spray painting, the major source of VOC
emissions by the industry, there currently is no cost-effective
technology for the significant reduction of these wastes for small
operations. The spray booths being installed to comply with OSHA
and environmental regulations, while maintaining surface finish
quality, address only particulates (overspray solids) but not VQCs.
There is a need for small scale solvent vapor recovery for this and
other industries. EPA assistance in stimulating research in this
area could result in major pollution reductions, both in this and
other, small industries.
In principle, and based on a number of studies, it can be
concluded that the application equipment and techniques used for
spray painting are highly inefficient in terms of waste produced
per unit of product due to overspray, atomization of volatile
constituents into the air and/or water and the additional wastes
resulting from equipment clean up, soiled protective clothing, etc.
More sophisticated coating technologies are and have been
investigated, and are in use for certain applications. However,
technology such as electrostatic painting, dip coating, etc. that
are now making inroads in the facilities of Original Equipment
Manufacturers are not yet practical or even appropriate for the
refinisher. Recently, a low pressure/high volume (LPHV) spray gun
has been introduced which markedly improves transfer efficiency and
thus reduces VOC and particulate emissions.
While still to be fully developed, there is a significant list
of other coating technologies intended to replace solvents, such
as high-solids paints, U-V Curables, Ultrasonic activated, hot
melt, etc., that could be advanced by research leading to
application. Additionally, a number of these technologies are
process specific, potentially limiting their wide' adoption.
A half-way approach, using high pressure carbon dioxide in
place of a portion of the solvents in paint formulations could
become available in a few years and would contribute to significant
reductions in VOC emissions.
The recharging/repair of automotive air conditioners may be
an opportunity to recover Freon. Equipment is now available for
such recovery and some types are in use.
Similar technology may be applicable to Freon recovery from
commercial and residential refrigeration and air conditioning
equipment, including the foam insulation panels. It should be noted
in both areas that leakage of the unit, with loss of the Freon to
the atmosphere BEFORE the unit arrives at the repair facility, is
common. Investigation of improvements for recovery may be
worthwhile. The substitution of organic and inorganic blowing
agents and vacuum panels for insulation are some of the other
approaches that could be evaluated.
152
-------
Waste oil recovery is widely practiced by the industry, with
the recovered oil usually being burned in commercial boilers or
incinerators. Reprocessing of oil has had an on/off history over
the past several decades, largely due to regulatory impacts.
Innovations in waste oil reprocessing could return more of this
resource to the consumer network. Some research is being carried
out (Alaska, California) to extend the useful life of hydrocarbon-
based oils by monitoring of deterioration and high performance
filtering. Synthetic motor oils are another approach but do not
appear to have attracted widespread consumer or vehicle
manufacturer attention at this time. Evaluation of the waste/
pollution associated with this product could be of interest.
Systems comparable to Freon or oil do not exist for waste
antifreeze reuse and the bulk of this material probably ends up in
POTWs or in mixed solvent wastes destined for incineration or fuel
blending. The recovery and recycling of ethylene glycol
contaminated with metals and chlorinated hydrocarbons (from solder
and gasoline), plus other trace elements, needs additional
research. Processes are available to remove solids by filtration,
and additive (anti-rust) packages are available for reformulation
or reconstitution. The present problem seems to be the development
of a method/procedure for collection from small generators. The
incentives that would make gas stations, repair shops and private
residences participate in the activity in a much greater way do not
seem to be present. An area of interest for research could be the
characterization of the amount of antifreeze material discarded
annually, its contaminants, the current management methods and
associated problems. The resulting information would form the
basis for design and decision-making regarding hardware and
operating requirements for antifreeze recovery.
Batteries are another major source of waste from the
automotive repair industry. Collection and recovery of the lead
from the plates and suspended mixture of lead oxides and lead
sulfate is extensively practiced. However, reuse of the waste
sulfuric acid is not. Investigations and discussions with
representatives of battery manufacturers have indicated that under
proper operating conditions spent acid drained from batteries can
be filtered to remove iron and copper contamination and the acid
then refortified and recycled for battery use. The suggested
research here would be to identify the specific problems and look
for improvements leading to better technologies and associated
economics.
Degreasing of vehicle parts can be considered integral to any
repair or maintenance operation. Chlorinated and non-chlorinated
hydrocarbons traditionally have been used in such operations.
Recycling is widely practiced, often through off-site, contracted,
services. Aqueous cleaning and degreasing solutions have been
proposed and are being considered by some segments of the industry.
Other technologies, such as blasting with solid carbon dioxide has
153
-------
been tried. Careful evaluation of such products and procedures may
produce pollution prevention answers useful to this industry.
GENERIC TECHNOLOGY
Investigation of industries within the short list resulted in
the identification of generic pollution prevention needs applicable
to many industries. These came in the form of unsolicited
suggestions offered by a significant number of experts in the
various fields of industry. These were finalized into a list of
13 generic research needs (see Table 3).
As indicated by various industry spokesmen, a number of
generic or "core" research areas were identified where pollution
prevention advances are needed. These would be applicable across
a large number of industries or industry segments. Because of the
large potential for improvements, it is recommended that these
research areas receive significant priority in formulating a
research program.
TABLE 3. LIST OF 13 GENERIC TECHNOLOGY IMPROVEMENTS NEEDED
VOC Control(Recovery technology)
CFC Substitutes
Oil-Water separation
Improved seals for pumps and valves
Equipment modifications
Improved operational testing (process baths, etc.)
Small-scale recovery for recycling"
Inventory control techniques for Pollution Prev.
Metal degreasing
Acid recovery
Boiler waste reduction
Adsorption systems for regeneration and recovery
Industrial process scrap metal waste reductions
VOC Control
With the current emphasis on air quality, it is somewhat
surprising that more is not being done to develop technology for
chemical vapor recovery. Many governmental agencies and industries
appear to be satisfied to destroy vapors by incineration processes,
at best recovering energy from the degradation of valuable
chemicals. It is suggested that a large effort would be beneficial
toward developing recovery/reuse approaches for such solvents.
154
-------
This would affect industries as diverse as printing, painting
(furniture, auto finishing), baking, etc. Currently, systems are
available based on condensation by chilling or carbon adsorption;
however, such systems are costly and apparently do not give the
removal achievable with incineration techniques. And, where mixed
solvents are used in a process, recovery must be coup'led with
reformulation or separation to produce usable solvents.
Consequently, industry often chooses the destructive approach to
satisfy regulatory constraints.
CFC Substitutes
The quest for substitutes for the ozone layer damaging
chlorofluorocarbons is well underway and several companies are
heavily committed to the development of suitable alternatives. EPA
may have opportunities to evaluate specific alternatives under
development for specific applications; some of these will be
chemically similar chemicals but others could be completely
dissimilar chemicals or alternate processing methods. Alternate
technologies that do not require fluorocarbons (e.g., for air
conditioners, degreasing, plastic foam) would achieve major source
reductions.
Oil-water Separation
Many industries generate waste oil that is contaminated with
water, or, in many cases, the water is the predominant species.
For many processes, the current oil/water separation techniques do
not produce reusable separate constituents. In some cases this
accounts for large volumes of wastewater being generated. While
some of the wastewater is suitable for reuse/recycle, many
companies are still perceiving it more cost-effective to treat and
discharge or dispose of it. Research on oil-water separation, such
as, for example, emulsion breaking by either physical or chemical
means would be widely useful. Activities involving cutting and
cooling fluids, fluids such as those in metalworking or machining,
petroleum refining and drilling are some of the sources of waste
that could be reduced by research leading to improvements.
Improved Seals for Valves, Pumps, etc.
A large plant has numerous valves which may be leaking at any
one time. Improving the design or the seal material could conserve
the materials being lost both as vapor and as liquid while
minimizing the discharge to surface runoff, the air, or to
wastewater collection systems.
155
-------
Significant waste may be generated during start-up and shut-
down of processes, both routine and unanticipated. Some of these
discontinuities in operation are a result of premature failure of
valves, seals, etc. Thus, improvements in longevity or
predictability of seal failures could reduce waste such as spills,
off-spec products, waste reagents or any feed and product material
of a process. For example, the start-up of a printing press
produces considerable waste paper as colors and registration are
adjusted; the industry is now devoted considerable effort toward
developing automated equipment that will minimize such start-up
losses and allow change-overs "on the fly."
EquipmentModifications
In a wide range of industries the same equipment is used today
as has been used for decades. Significant processes are operated
because they were "always operated that way." While it is
difficult to identify specific industries or processes with a quick
review, the potential for improvements after a focused study is
significant. Resulting changes may lead to improved yields,
decreased by-products, etc., and have a major impact on waste
production. For example, a redesign of a reaction kettle or the
use of a new design a baffle or stirrer could accelerate a desired
reaction and/or improve yield. Even exhaust pipe sizing can affect
the slight overpressure at which a reaction may be occurring.
Incorporation of ultrasonic agitators or high pressure gas lances
can improve the efficiency of reactions as well as the efficiency
of reactor clean-out between batches, thus minimizing
chemicals/solvents needed to achieve a desired level of
cleanliness. While these types of improvements are, for the most
part, practiced as routine improvements for reducing costs,
increasing profits and staying competitive, the'approach from a
pollution prevention perspective (while keeping track of economics)
offers new opportunities.
Bath Testing (Manual Process Control, Small-Scale Operations)
Simple, convenient, quick tests are needed for operators to
determine when a process bath, reaction mixture, or rinse water has
reached its safe loading and thus help to determine when discharge
is necessary. Certain of these tests do exist, but often they are
not relied on by operators. Instead, discharges or disposal of
baths, rinses, etc., are done on an arbitrary, routine schedule
that may be exceeding required frequency and produce significantly
larger volumes of waste. Recommendations are for feed-back and
feed-forward control loops which allow optimization.
156
-------
Small Scale Recovery
Distillation, evaporation, carbon adsorption and regeneration,
etc., while in common use, do not exist in widespread use at small
scale for the purpose of recovering solvents from paints,
degreasers and reaction vessels. Evaluation of hardware and
economics on an impartial basis could provide needed information.
Inventory Control
Automated inventory control has been shown to produce
significant decreases in waste. Needed is software that is
tailored to the small firm for tracking their raw feed materials,
products and wastes generated. In addition to forming a convenient
method by which to evaluate actual costs, it would also serve in
placing focus on waste and liability costs and therefore create an
environment for waste reduction incentives.
Metal Degreasing
Vapor or liquid degreasers are widely used in industries
ranging from semiconductors to auto refinishers. While
considerable progress has been made in designing these to minimize
solvent loss and carryover, indications are that solvent recovery
still amounts to only about 60%. Total redesign of degreasers or
consideration of many of the dragout control concepts used in the
electroplating industry may be the necessary next step, as well as
careful consideration of non-solvent alternatives such as aqueous
or physical degreasing. Degreasing with aqueous solutions coupled
with ultrasonics has received some attention for degreasing metal
parts in Europe. Cleaning by blasting of various substances from
sand to walnut shells has also been investigated. Another research
area is to design processes that avoid needing the degreasing
altogether.
Acid Recovery
Recovery of strong acids (e.g., sulfuric, hydrofluoric,
.nitric, hydrochloric) has long been recognized as a desirable route
to minimizing waste. However, capital cost for corrosion-resistant
acid stills has usually limited their application to large
centralized facilities - which then face the problems of
transportation risks and costs. With the exception of hydrofluoric
acid, recovery has not usually been cost-effective. An
electrodialytic bipolar membrane technology has been commercialized
which allows recovery of concentrated acids and re-conversion of
salt products to the acids (and bases) (Chem Eng., Dec 1989 p81).
Such technology could have major impact on the steel industryj
157
-------
chemical, dye, explosives, and other industries can also be
investigated.
Boiler WasteReduction
Industrial power-generating boilers are a significant source
of wastes, particularly during cleaning operations. California,
in a summary of its 1988-1989 Waste Minimization Grants, noted that
up to 11,000 tons of solid toxic waste is produced annually in that
state from this source. While extensive research may be underway,
the emphasis has not been on waste minimization. Reconsideration
of this segment of industry with a source reduction viewpoint may
elicit novel means of preventing the formation of water treatment
sludges, etc.
Adsorption Systems
Carbon adsorption is widely used for waste treatment and, less
frequently, for chemical recovery. Other adsorbents have also been
developed over the years (resins, zeolites, etc.). A research
program examining cost and technical effectiveness of such
different adsorbents - including regeneration - AND at developing
chemical selectivity that might be achievable with one or more of
these materials and that would allow systems to segregate waste
components (from air, water, etc.) into reusable chemicals could
be very productive. For example, water creates problems in carbon
adsorption but certain hydrophobic zeolites do not readily adsorb
water; consequently organics can be desorbed and recovered in an
anhydrous state rather than as water/organic mixtures requiring
further treatment. Newer, proprietary products with higher
adsorption capacities are now being developed (Chem Eng Nov/89
p!7). Support for development could be fruitful, particularly if
applied to the recovery of more expensive solvents such as
fluorocarbons, specialty esters, etc.
Scrap Metals
A number of industries, including the finishing of castings,
machinery fabrication, and auto^refinishing, generate scrap metals
as cuttings and turnings, grinding dusts, Damaged parts, etc.
These materials, often contaminated with cutting fluids, are
usually discarded as solid waste or, at best, are sold to scrap
dealers for reprocessing. Improved casting, forging, and machining
processes and equipment would simultaneously reduce trhe waste loads
produced and the amount of raw material used to fabricate the
product. Such changes in production practices are usually brought
about for reasons other than environmental concern, such as
significant economics factors or regulatory pressures (e.g. worker
safety). Where significant capital investment is involved, these
158
-------
changes are very slow. Added stimulus could be productive via
research that brings about a fore-runner or working example. The
tool and die industry could be an area of focus.
RECOMMENDATIONS
Among a number of broadly-based needs recognized by this
effort is the acute need for a classification tool, similar to the
SIG used for this study. Instead of being based on a value of
production/receipts/revenues basis (such as the SIC), it would be
designed specifically for use as part of the pollution prevention
work to serve as a means of comparing various data on a common
ground. Approaches under consideration are concepts such as
relative risk, true cost and life-cycle analysis. Some work has
been started under these concepts but much effort is still needed
before these can become practical working tools.
Within the industries investigated as part of this study,
several appeared to lend themselves to EPA assistance for research
more readily than others for the short range. In addition, there
are many indications that rapid changes are, or will be taking
place and that regular updating of this information should be
scheduled to keep pace with developments, as well as to make
improvements. The following recommendations emphasize those SIC's
where it appears that EPA participation would be most productive.
There is no priority indicated by position on the list.
o Textiles - recovery of dyes and scouring agents from
wastewater.
o Wood preserving - investigations of new, less toxic preserving
agents.
o Pulp and paper - improved recovery of coated stock?
restoration of fiber strength in recycled paper, deinking.
o Printing - improvements in pre-press photographic chemistry
through the use of computer technology; solvent recovery.
o Chemical industry - substitution of less toxic solvents,
solvent reuse/recycling. Process changes to eliminate use of
toxic constituents.
o Plastics - segregation of scrap plastics; compatibilization.
Toxic solvent/cleaner substitution or reduction.
o Pharmaceuticals - solvent reuse; substitution.
o Painting - low and non-VOC painting techniques; improved
application technology; substitution of less hazardous coating
technology.
159
-------
o Ink manufacture - low and non-VOC inks? elimination of
metallic pigments? substitution of less toxic inks, solvents
and cleaners.
o Petroleum exploration/refining - improved recovery of usable
oil from drilling muds and processing wastewaters.
o Steel industry - reuse of tar decanter sludge and electric arc
furnace dust; reuse of recovered calcium fluoride.
o Non-ferrous metals - isolation of arsenic contamination to
allow reuse of stack dusts; improved hydrometallurgical
processes minimizing sulfur oxide emissions.
o Metal finishing - non-cyanide plating systems? improved
chemical recovery from cyanide plating processes, toxic
constituent substitution or reduction.
o Electronics - "clean" fabrication techniques that eliminate or
minimize degreasing solvent use. Mechanical cleaning
technologies, reduction of toxic contaminants.
o Automobile refinishing/repair - reductions in solvent losses
in various operations, solvent substitution.
o Laundries/dry cleaning - improved solvent recovery, process
improvements.
This is the list recommended for priority pollution prevention
research in the industrial area. Within each industrial segment
considered a priority area, there are one or more suggestions,
concepts, or problems. It is recommended that, with further
refinement and updating, these can serve as one basis for the
development of EPA research projects for the short term future.
Crossing the industries are basic, common needs that have
potential for making significant pollution prevention impact.
These were identified as the "generic" technology, above. These
are a mixture of short-term improvements as well as longer-term,
basic, or ."core" research.
o Improving chemical reaction rates or making reactions more
product-specific, such as by improved catalysis, use of
ultrasonics, use of microwave heating, improved reactor
designs, etc.
o Improved equipment/system reliability to reduce the need for
start-ups and shut-downs that generate wastes.
160
-------
o Improved volatile organic chemical control, by development of
equipment affordable by smaller companies to minimize losses
from valves, pumps, etc. during use.
o Material substitutions in product fabrication that avoid or
minimize waste generation steps such as degreasing,
electroplating and painting, etc.
Future refinements to the prioritization procedure should
consider using a simplified list of criteria. A number of the
expert participants stated that selecting candidate industries
while keeping in mind a list of 12 criteria was counterproductive.
Additionally, the Toxic Release Inventory (TRI) data and similar
data available for providing amounts/toxicities/relative risk
information should be incorporated into refinements as practical.
REFERENCES ' '
1. Hazardous and Solid Waste Amendments of 1984. Public Law 98-
616, November 8, 1984.
2. Executive Office of the President, Office of Management and
Budget. 1987. Standard Industrial Classification Manual.
3. U.S. Environmental Protection Agency. June 1989. The Toxics-
Release Inventory: A National Perspective. EPA 560/4-89-005
4. U.S. Environmental Protection Agency. Draft project report in
publishing.
161
-------
THE EVALUATION OF AN ADVANCED REVERSE
OSMOSIS SYSTEM AT THE SUNNYVALE. CALIFORNIA
HEWLETT-PACKARD FACILITY
by: Lisa M. Brown
U.S. Environmental Protection Agency
Risk Reduction Engineering Laboratory
Cincinnati, Ohio 45268
Robert Ludwig
California Department of Health Services
Sacramento, California 95814
Ilknur Erbas-White
Science Application International Corporation
Santa Ana, CA 92705
ABSTRACT '
The effectiveness of an Advanced Reverse Osmosis System
(AROS) in the recovery of nickel plating bath solutions and rinse
water was technically and economically evaluated at the Hewlett-
Packard (HP) Facility in Sunnyvale, California under the
California/EPA Waste Reduction Innovative Technology Evaluation
(WRITE) Program'.
The AROS is basically a reverse osmosis (RO) unit that
provides zero discharge capability. This system has specially
adapted membranes that do not require pH adjustments to neutral,
a microprocessor control to manage the RO membranes, and a
continuous monitoring system that monitors the influent,
permeate, and concentrate for temperature, flow rate, and
conductivity.
HP determined that the permeate and concentrate from the
AROS unit could be recycled into the process; however, the
payback period of 4.4 years was insufficient for capital purchase
under HP's corporate purchasing policy.
This paper has been reviewed in accordance with the U.S.
Environmental Protection Agency's peer and administrative review
policies and approved for presentation and publication.
162
-------
INTRODUCTION
This study was performed under the California/EPA Waste
Reduction Innovative Technology Evaluation (WRITE) Program, and
was a cooperative effort between EPA's Risk Reduction Engineering
Laboratory (RREL), the Alternative Technology Division of the
Toxic Substances Control Program within the Department of Health
Services (DHS) of the State of California, Hewlett-Packard (HP),
and Water Technologies, Inc. (WTI). Science Applications
International Corp. (SAIC) provided technical support on this
WRITE project.
The WRITE Program is part of the RREL's pollution prevention
research program. Under the WRITE Program, the cooperative
efforts of the USEPA and state or local environmental programs
are used to identify, develop, demonstrate and evaluate
innovative pollution prevention techniques. Specifically, the
WRITE Program provides engineering and economic evaluations plus
information dissemination for methodologies that have the
potential of reducing the quantity and/or toxicity of waste
produced at the source of generation, or to achieve practicable
on-site reuse through recycling.
, In this project the effectiveness of an Advanced Reverse
Osmosis System (AROS) in the recovery of nickel plating bath
solution and rinse water was evaluated and the costs were
compared with that of an existing chemical precipitation
treatment system at the Hewlett-Packard Facility in Sunnyvale,
California.
The plating operation that HP tested the AROS on was a
nickel plating system consisting of,two plating baths followed by
a "dirty" rinse tank and then .a "clean" rinse tank. The rinse
water flows countercurrent to the flow of the items being plated.
The overflow, 4 to 5 gpm, from the "dirty" rinse tank is a
wastewater.
HP's existing wastewater treatment system for plating wastes
(Figure 1) involves precipitation of metals as hydroxide salts.
Chemical sludges are pumped to a recessed plate filter press
system for dewatering. Dewatered sludge is disposed to an
offsite disposal facility in California. Effluent from the
precipitation tanks is pumped to ultrafilters for final polishing
prior to discharge. Solids collected by the filters are then
pumped to a filter press and then shipped and disposed offsite.
The AROS (Figure 2) is basically a reverse osmosis (RO) unit
with specially adapted membranes that do not require pH
adjustments to neutral. The unit includes a microprocessor
control using proprietary software to manage the RO membranes.
The unit contains a continuous monitoring system that monitors
the influent, permeate, and concentrate for temperature, flow
rate, and conductivity.
163
-------
NICKEL
PLATING
RINSE
JP.
COPPER
PLATING
RINSE
TIN
PLATING
5 RINSE
OTHER
PLATING
RINSES
FeS04 H2S04
t
COPPER
REDUCTION
UNIT
pH = 2
FILTRATE
FILTER PRESS
DEWATERING
NaOM
I
COMPLETELY
MIX REACTOR
pH= 11
RECIRCULATION/
SETTLING
TANKS
SLUDGE
ULTRAF1LTRATION
UNITS
FILTRATE
DISCHARGE
TO CITY
SEWER
(75,000 gpd)
RECIRCULATED
FILTER SOLIDS
(2-3% SOLIDS)
DEWATERED
SLUDGE
TRANSPORT AND
DISPOSAL (11.5
(TONS/MONTH)
Figure 1. Schematic Diagram of Existing Wastewater Treatment System at Hewlett-Packard
-------
CLEAN RINSE
TANK 2
Sp«c/3//zed Rovers* osmosis
Concantrat/Ofi
An*
f U
Control
Valve
TO PLATING BATH 1
DIRTY RINSE
TANK1
Figure 2. Diagram of the Advanced Reverse Osmosis System (AROS)
* Courtesy of Water Technologies, Inc,
-------
The RO membranes clean rinses to pre-specifled standards and
concentrate plating salts, in order to recycle both rinse water
and plating salts. An AROS can reconcentrate dilute solutions to
at or near bath strength (typically a concentration of 40% to 70
% is accomplished) without any evaporation or additional
concentration technology.1
The AROS unit was installed at HP in November 1989. After
initial installation and debugging, the system was test run from
about November 21, 1989 to December 18, 1989. The system was
temporarily taken off line at the end of 1989, to allow Hewlett-
Packard to test and evaluate the plating bath quality and to
create a baseline of comparison for bath contents and
performance. Results were considered acceptable and the AROS
unit was restarted in January 1990, and was operated on-line
during most of 1990.
OBJECTIVES
This project includes additional sampling to establish a
one-day snap shot of the AROS unit operation at the Sunnyvale
facility. Removal efficiencies obtained based on the actual data
were used to prepare a technical evaluation of the system.
Economic analysis is based on data obtained from Hewlett-Packard.
Details about the design of the sampling program are provided in
the Quality Assurance Project Plan (QAPP).
SAMPLING
Streams in and out of the AROS unit were sampled on October
17, 1990 to obtain a one day snap shot of the system's operation.
On Wednesday, 17 October 1990, SAIC observed sampling of the
AROS unit installed at Hewlett-Packard in Sunnyvale. Hewlett-
Packard staff conducted sampling while SAIC personnel supervised.
The sample containers were prepared by Western Analytical
Services Laboratory; labels were filled out and chain of custody
maintained? and samples were placed in the bottles as explained
in the QAPP. A representative from WTI, who manufactures the
AROS unit, and Robert Ludwig of DHS were also present to observe
sampling.
Pour liquid streams were sampled as shown in Figure 3:
1) Influent to the AROS treatment unit which is the nickel
plating rinse water from "dirty" rinse tank No. 1
2) Deionized water used as makeup water to the AROS unit
3} Permeate from the AROS unit which is clean water
produced by the unit that is returned to the clean
"rinse" tank No. 2
166
-------
FLOW OF PARTS IN THE
PLATING OPERATION
CD
PORT
SAMPLE
1 1
NiS04 NiSO4
PLATING PLATING
BATH BATH
1 2
EMERGENCY ^
BYPASS TO
WPS
INPUT
, TmggC-.,,,.,
•f
V
I
TOAR
s
r
i
OSUNIT
CONCENTRATE
DIP SAMPLE
!
"DIRTY" "CLEAN"
RINSE RINSE
TANK TANK
1 2
DEIONIZED WATER
MAKE-UP (DIP
SAMPLE FROM
SOURCE TANK)
*
AROS
UNIT
CONDUCTIVITY
.„ . A Mf\ t f^WI^I
AMU LbvhL
MEASUREMENTS
CX] *•
PORT SAMPLE
PERMEATE
Figure 3. Schematic Diagram of the Advanced Reverse Osmosis System (AROS) for the Nickel Plating
Operation and Sample Locations
-------
4} Concentrate, consisting of approximately 50% nickel
plating solution produced by the AROS unit and returned
to the plating bath No. 1
Streams l, 3, and 4 were collected as composites. Stream 2
was a one-time sample. Upon collection, all samples were stored
on ice, with the exception of the concentrate, which would have
crystallized if put on ice. At the end of the day, samples were
poured into the prepared bottles for shipment to the laboratory.
Samples from streams l, 2, and 3 were shipped to the laboratory
in a cooler with blue ice. The concentrate sample was shipped
separately as a hazardous material and was not maintained on ice.
The laboratory confirmed receipt of all samples in good condition
the following morning at 10:00 am.
ANALYSIS OF SAMPLING RESULTS
Sample analysis includes nickel, chloride, sulfates, pH,
total dissolved solids, conductivity, color, and total organic
carbon. Sampling results are shown in Table 1. As can be seen
in the table the AROS unit produced a composite permeate that was
satisfactory as clean rinse water makeup, its intended purpose.
Similarly, the concentrate was of quality (40% to 50% plating
bath concentration) that could be used as nickel plating bath
solution makeup.
ECONOMIC ANALYSIS OF THE AROS SYSTEM
The Hewlett-Packard Corp. (HP) maintains detailed cost
records for their existing plating wastewater treatment system
and deionized water production operations. These costs serve as
a comparison for HP in their assessment of the cost effectiveness
of the trial AROS unit. At HP the AROS unit only treated a small
fraction, e.g. about 3 percent of the total plating wastewater
flow.
At HP the savings from use of the AROS unit were directly
related to the incremental reduction in spending for the
following cost items:
Sewer discharge fees and water cost, estimated by HP at
$0.004/gal. or $4 per 1000 gal.
Deionized water production cost, estimated by HP at
$0.0064/gal., or $6.40 per 1000 gal.
Plating wastewater treatment costs, estimated by HP at
$0.0062/gal., or $6.20 per 1000 gal.
168
-------
Table 1. Sampling Results from Evaluation of the AROS Unit at Hewlett-Packard
CHEMICAL
UNIT INFLUENT
CONCENTRATE
PERMEATE
DEIONIZED DETECTION PERCENT
WATER LEVEL REMOVAL
10
Nickel
Chloride
Sulfate
pH
TDS
Conductivity
TOC
ppra
ppm
ppm
_
ppm
umhos/cm
PPM
650
120
1100
6
2575
1985
30.8
52700
7800
79000
4.1
171500
53750
1625
20.5
29
18
5.65
165
139.5
7.01
0.19
ND
0.11
6.1
ND
3.75
0.74
0.01
0.1
0.1
__
6
0.5
0.5
96.8
75.8
98.4
__
93.6
93.0
77.2
Note: In case of duplicate analysis results, arithmetic averages are used
Percent removal is calculated based on the ((Influent-permeate)/Influent}*100 formula
-------
It is understood that these plating wastewater treatment
costs include:
(a) Labor
(b) Power
(c) Chemicals
(d) Expendable parts and supplies replacement
(e) Monitoring, e.g. analysis of influent and effluents
(f) Sludge treatment, transport and disposal
Purchase of new plating chemicals to make up for
plating solution drag-out losses, estimated by HP at
$5.00/gal.
The above listed cost items are the major incremental cost
savings resulting to HP from use of the AROS system as shown in
Table 2, HP estimates the annual savings listed above to total
$26,250/year. .
TABLE 2
Estimated Annual Incremental Savings From Use
of the AROS Unit as Reported by
Hewlett-Packard Corporation, 1990 Costs
ITEM
NO.
1
2
3
4
DESCRIPTION
Sewer Discharge Fees
and Water Costs
Deionized Water
Production Cost
Plating Wastewater
Treatment Costs
Purchase of New
Plating Chemicals at
an 85 Percent
Reduction
ESTIMATED
SAVINGS .
$/GAL.
0.004
0.0064
0.0062
5.00 ' •
QUANTITY,
GAL.
1,275,000
1,275,000
1,275,000
1260 x 0.85
TOTAL
ANNUAL
SAVINGS
$5100
$8160
$7905
$5355
TOTAL ESTIMATED ANNUAL SAVINGS:
$26,520
This incremental cost savings is balanced against the annual
expenditure for owning and operating the AROS system, as follows:
Electrical Power
R.O. Membrane Replacement
Labor and Expendable Parts
Carbon Filters
Telephone Modem Contact With AROS Mfg.
TOTAL ANNUAL OPERATING COSTS
$1629
$2200
$5000
$ 90
$ 500
$9419
170
-------
Subtracting $9,419/Yr. from $26,250/yr., HP estimates that
the net annual savings from use of the AROS unit would be
approximately $17,100/yr. The AROS unit costs approximately
$75,000, which represents approximately $63,000 for the AROS unit
plus another $12,000 for making the installation permanent and
training of operating personnel. The payback period is 4.4
years.
CONCLUSIONS
As shown in the economic analysis section of this paper, the
Hewlett-Packard evaluation showed an estimated net annual savings
of approximately $17,000/year through use of the AROS unit.
Under company policy this savings was insufficient to justify the
capital expenditure of approximately $75,000 ($62,600 for the
unit, plus installation and training costs). Hewlett-Packard has
decided not to purchase the AROS unit.
Because the AROS unit treated such a small increment of the
wastewater flow at HP it was difficult for the AROS unit to be
cost effective; however, in a different setting the AROS unit
might be very cost effective. Economy of scale worked against
the AROS unit at HP. The purchase decision might have been
different if the AROS unit been installed at a smaller facility
where it would have treated a larger fraction, or even all, of
the plating wastewater. Also, the AROS unit would be more
economically competitive for a new facility that did not already
have an amortized wastewater treatment facility in place.
The AROS unit performance was considered excellent.
REFERENCES
1. Rich, R.R. and T. von Kuster, Jr., "Recovery of Rinse Water
and Plating Bath from Process Rinses Using Advanced Reverse
Osmosis", Conference on Metal Waste Management Alternatives;
Pasadena & San Jose, California; September, 1989.
2. Science Applications International Corporation, "Quality
Assurance Project Plan for the Evaluation of an Advanced
Reverse Osmosis System at the Sunnyvale, California Hewlett-
Packard Facility," USEPA Contract No.68-C8-0062, Work
Assignment 1-18, Cincinnati, Ohio, July 1990.
171
-------
THE BEHAVTOR OF TRACE METALS IN ROTARY KILN INCINERATION;
RESULTS OF INCINERATION RESEARCH FACILITY STUDIES
by: D. J. Fournier, Jr., L. R. Waterland, and J. W. Lee
Acurex Corporation
Incineration Research Facility
Jefferson, Arkansas 72079
Gregory J, Carroll
U.S. Environmental Protection Agency
Risk Reduction Engineering Laboratory
Cincinnati, Ohio 45268
ABSTRACT
Two series of pilot-scale tests evaluated the fate of trace metals fed to a rotary kiln incinerator.
A venturi/packed-column scrubber was used in one series, and a single-stage ionizing wet scrubber
was used in the other. Test variables were kiln and afterburner temperatures and feed chlorine
content. Results show that bismuth and cadmium were relatively volatile, averaging less than
40 percent recovered in the kiln ash. In contrast, more than 75 percent of the arsenic, barium,
chromium, copper, magnesium, and strontium were recovered in the kiln ash. Lead behaved as a
volatile metal in one series, but as a nonvolatile metal in the other. Relative metal volatilities
generally agreed with expectations based on vapor pressure/temperature relationships, although
arsenic was much less volatile than predicted. Increased kiln temperature caused bismuth,
cadmium, and lead to become more volatile, but did not affect the remaining metals. Increased
chlorine content caused increased volatility of copper and lead. Metal fate was not affected by
changes in afterburner temperature.
Increased kiln temperature caused the average flue gas paniculate metal distributions to shift
from roughly 20 to 60 percent less than 10 urn for all test metals, except chromium. Metal
enrichment correlated with relative metal volatilities, with the more volatile metals most affected.
Increased chlorine content from 0 to 4 percent caused cadmium, copper, and lead distributions to
shift from 20 to 55 percent less than 10 pm. Increased chlorine to 8 percent had no further effect.
For chromium, increased chlorine content from 0 to 4 to 8 percent caused a shift of 2 to 20 to
50 percent in particulate less than 10 \im,
Average metal collection efficiencies for the venturi/packed-column scrubber ranged from 31
to 88 percent; the overall average for metals was 57 percent. For the single-stage ionizing wet
scrubber, average metal collection efficiencies ranged from 22 to 71 percent; the overall average for
metals was 43 percent.
This paper has been reviewed in accordance with the U.S. Environmental Protection Agency's
peer and administrative review policies and approved for presentation and publication.
172
-------
INTRODUCTION
The hazardous waste incinerator performance standards, promulgated by EPA in January 1981,
established direct particulate and HC1 emission limits and mandated 99.99 percent destruction
removal efficiency (DRE) for principal organic hazardous constituents (POHCs). These
performance standards indirectly control emissions of hazardous constituent trace metals by limiting
particulate emissions to 180 mg/dscm. Risk assessments conducted subsequent to the promulgation
of the standards suggest that trace metal emissions from some incinerators treating waste streams
with high levels of metals could pose unacceptable risks to human health and the environment. In
response to this potential risk, the EPA has proposed regulations to establish limits on metal
emissions from hazardous waste incinerators (1).
Despite the importance of metal emissions, there are only limited field data available to validate
risk assessments and to assist regulatory development efforts. Data describing the effects of
incinerator operation and waste composition on trace metal emissions are particularly lacking.
In response to these data needs, two series of tests were conducted at the U.S. EPA Incineration
Research Facility (IRF) to determine the fate of trace metals fed to a rotary kiln incinerator. The
primary difference between the two series was the primary air pollution control system (APCS) used
for particulate and acid gas control. One test series was performed with a venturi/packed-column
scrubber. The other test series was performed with a single-stage ionizing wet scrubber. The
purpose of these test programs was to extend the data base on emissions and residual discharges
of trace metals during incineration of hazardous wastes. Test data will support the EPA's Office
of Solid Waste in its continuing development of regulations for hazardous waste incinerators. Test
data will also be used to evaluate the predictive capabilities of a numerical metal partitioning model
and to perhaps guide further model refinement (2).
The primary objective of these test programs was to determine the fate of five hazardous
constituent and four nonhazardous constituent trace metals fed in a synthetic solid waste matrix to
a rotary kiln incinerator. Of specific interest was the distribution of the metals as a function of
incinerator operating temperatures and waste feed chlorine content. The five hazardous trace
metals investigated were arsenic, barium, cadmium, chromium, and lead. The four nonhazardous
trace metals investigated—bismuth, copper, magnesium, and strontium—were included primarily to
supply data to support the model evaluation.
TEST PROGRAM
All tests were conducted in the pilot-scale rotary kiln incinerator system at the IRF, illustrated
in Figure 1. The main components of the system are the rotary kiln chamber, afterburner, flue gas
quench, primary APCS, and secondary APCS. As shown, the quenched flue gas can be directed to
either of the two primary APCSs installed in parallel. The flue gas is further treated by a secondary
APCS consisting of a demister, carbon bed and HEPA filter before exiting to the atmosphere.
SYNTHETIC WASTE MIXTURE
Similar synthetic waste feeds were used for both test series. The test waste contained a mixture
of organic liquids added to a clay absorbent material. The trace metals were incorporated by
173
-------
SINGLE-STAGE IONIZING
WETSCRUBBER
txKT >~-t>4
CD
ATMOSPHERE
STACK
CARSON BED HEPA ,
ADSORBER FILTER |
oocr
ROTARY
KILN
ROTARY KILN
INCINERATOR
I
MODULAR PRIMARY AIR
POLLUTION CONTROL
DEVICES
I REDUNDANT AIR i
1 POLLUTION CONTROL '
I SYSTEM I
Figure 1. Schematic of the IRF rotary kiln incinerator system.
spiking an aqueous mixture of the metals onto the organic liquid-containing solid material. The
resulting synthetic feed represented a solid hazardous waste containing both organics and metals.
The waste material was continuously fed to the rotary kiln via a twin-auger screw feeder at a
nominal rate of 63 kg/hr (140 Ib/hr).
The organic liquid base consisted of toluene, with varying amounts of tetrachloroethylene and
chlorobenzene added to provide a range of chlorine contents. Synthetic waste chlorine was varied
from 0 to nominally 8 percent of the waste feed. The analyzed organic fractions for the three waste
feed mixtures are given in Table 1. Table 2 summarizes the average metal concentrations in the
combined waste feed,
TEST CONDITIONS
The test matrix was the same for both test series. Table 3 summarizes the target and average
achieved values for the three test variables. Each variable was varied over three levels, with the
other variables held nominally constant. Target kiln exit temperatures were 816°, 871°, and 927°C
(1500°, 1600°, and 1700°F). Target afterburner exit temperatures were 982°, 1093°, and 1204°C
(1800°, 2000°, and 2200°F). Target concentrations for chlorine in the synthetic waste feed were
0, 4, and 8 percent. Both test series included replicate testing of one test condition to provide
information on data variability. All tests were conducted under excess air conditions. Oxygen
174
-------
TABLE 1. ORGANIC COMPOUND CONCENTRATIONS IN THE SYNTHETIC WASTE
Test series
Venturi/
packed-
column
scrubber
Single-stage
ionizing wet
scrubber
Test
1
2 through 7
(average)
8
1
2 through 8
(average)
9
Toluene
23.2
18.8
14.6
23.1
17.8
11.6
Weight % in
Tetrachloroethylene
0
3.3
7.1
0
3.1
6.0
mixture
Chlorobenzene
0
3.3
6.9
0
3.0
5.6
Chlorine
content*
0
3.9
8.3
0
3.6
6.9
*Calculated based on measured tetraehloroethylene and chlorobenzene concentrations.
TABLE 2. AVERAGE INTEGRATED FEED METAL CONCENTRATIONS
Concentration (mg/kg)
Metal
Arsenic
Barium
Bismuth
Cadmium
Chromium
Copper
Lead
Magnesium
Strontium
Venturi/packed-column
scrubber test series
44
,53
150
8
87
470
52
17,200
280
Single-stage ionizing wet
scrubber test series
48
390
330
10
40
380
45
18,800
410
175
-------
TABLE 3. TARGET AND AVERAGE ACHIEVED TEST CONDITIONS
Test
scries
Vcnturi/
packed-
column
scrubber
Single-
stage
ionizing
wet
scrubber
Feed mixture
Cl content (%)
Test
1
2
3
4
5
6
7*
8
1
2
3
4
5
6
It
8t
9
Date
9/14/88
8/25/88
9/16/88
8/30/88
9/7/88
9/9/88
9/20/88
9/22/88
8/17/89
8/2/89
8/4/89
8/1/89
8/16/89
8/15/89
8/9/89
8/11/89
7/28/89
Target
0
4
4
4
4
4
4
8
0
4
4
4
4
4
4
4
8
Actual
0
3.7
4.2
3.8
3.6
3.4
4.6
83
0
3.5
3.5
3.5
3.7
3.6
3.6
3.8
6.9
Kiln exit temperature
•C ("F)
Target
871
816
927
871
871
871
871
871
871
816
927
871
871
871
871
871
871
(1600)
(1500)
(1700)
(1600)
(1600)
(1600)
(1600)
(1600)
(1600)
(1500)
(1700)
(1600)
(1600)
(1600)
(1600)
(1600)
(1600)
Average
874
825
928
878
871
875
873
870
900
819
929
877
885
887
881
879
881
(1606)
(1517)
(1702)
(1612)
(1599)
(1607)
(1603)
(1599)
(1652)
(1507)
(1704)
(1610)
(1625)
(1629)
(1618)
(1615)
(1617)
Afterburner exit temperature
•C (°F)
Target
1093
1093
1093
1093
1204
982
1093
1093
1093
1093
1093
1093
1204
982
1093
1093
1093
(2000)
(2000)
(2000)
(2000)
(2200)
(1800)
(2000)
(2000)
(2000)
(2000)
(2000)
(2000)
(2200)
(1800)
(2000)
(2000)
(2000)
Average
1093 (1999)
1071 (1959)
1092 (1998)
1088 (1991)
1196 (2184)
983 (1803)
1094 (2000)
1092 (1998)
1088 (1990)
1095 (2002)
1092 (1998)
1096 (2006)
1163 (2125)
1017 (1863)
1103 (2018)
1098 (2008)
1087 (1988)
•Test point 7 is a duplicate of test point 4.
fTest points 7 and 8 are replicates of test ]
replicates of test point 4,
concentrations were nominally 11.5 and 7.5 percent in the kiln and afterburner exit flue gas,
respectively. Estimated solids residence time in the kiln was 1 hr.
SAMPLING AND ANALYSIS
The sampling and analysis protocols were designed to track the discharges of the test metals in
the incinerator residuals and flue gas. For each test, composite samples of the kUn ash and
scrubber blowdown were collected. Flue gas sampling for metals at the scrubber exit was performed
with a Method 5 train modified for metals capture (3). Samples of the clay and the aqueous metals
spike solution were collected and the feedrates of each noted during each test. Kiln ash weights,
scrubber blowdown, and scrubber liquor volumes were also determined for each test.
Flue gas particulate samples were collected at the afterburner exit during the single-stage
ionizing wet scrubber test series. Using a variation of a Method 17 sampling train (4), at least 1 g
of particulate was collected during each test. After obtaining the total particulate weight, these
samples were size fractionated using a centrifugal classifier in accordance with the procedures in
ASME Power Test Code 28 (5). The resulting size cuts were weighed, digested, and analyzed for
metals.
176
-------
Samples were digested following the procedures of Methods 3010 and 3050 (6). All metals
analyses for the single-stage ionizing wet scrubber test series were by inductively coupled argon
plasma (ICAP) spectroscopy per Method 6010 (6). For the venturi/packed-column scrubber test
series, arsenic and lead analyses were by graphite furnace atomic absorption (6). Bismuth and
strontium analyses were by flame atomic absorption (6). Analyses for the remaining metals were
by ICAP.
RESULTS AND DISCUSSION
AVERAGE TRACE METAL DISCHARGE DISTRIBUTIONS
When subjected to incineration conditions, metals are expected to vaporize to varying degrees,
depending on their relative volatilities. To characterize a metal's volatility, equilibrium analyses can
be performed to identify the metal's volatility temperature for a given set of incinerator conditions.
The volatility temperature is the temperature at which the effective vapor pressure of a metal is 10'6
atm. The effective vapor pressure is the combined equilibrium vapor pressures of all species
containing the metal. It reflects the quantity of metal that would.vaporize under a given set of
conditions, A vapor pressure of 10'* atm is selected because it represents a measurable amount of
vaporization. The lower the volatility temperature, the more volatile the metal is expected to be.
Volatility temperatures are a major parameter in the partitioning model used to predict metal
behavior in an incinerator (2).
One objective of these tests was to identify the discharge distributions of the test metals relative
to each other. To address this objective, metal discharge distributions have been summarized for
each test program and presented in Figures 2 and 3. These figures show the amounts of metal
found in each discharge stream normalized as a fraction of the total found in the three discharge
streams—kiln ash, scrubber exit flue gas and scrubber liquor. In these figures, the bar for each metal
represents the range in the fraction accounted for by each discharge stream over all tests of the
respective test series. The average fraction for that test series is noted by the midrange tick mark.
Metal discharge distribution data are plotted versus the metal volatility temperatures calculated
assuming oxidizing conditions. For both test series, these figures indicate a correlation between
observed volatility and volatility temperature for all the metals tested, except arsenic. With the
exception of arsenic, the average normalized kiln ash fraction generally increased with increased
volatility temperature.
Arsenic was much less volatile than expected, but behaved consistently for all tests. This
observation is consistent with other studies at the IRF (7) and with studies at cement kilns (8). It
is possible that arsenic forms a thermally stable compound in the incineration environment or
becomes physically bound in the solid matrk.
Bismuth and cadmium were relatively volatile compared to the other test metals. On average,
less than 40 percent of the bismuth and cadmium was recovered in the kiln ash, compared to an
average of greater than 75 percent of the arsenic, barium, chromium, copper, magnesium, and
strontium. Lead volatility behavior differed between the two test series. For the venturi/packed-
column test series, the average fraction of lead recovered in the kiln ash was 20 percent. For the
single-stage ionizing wet scrubber test series, the average fraction of lead recovered in the kiln ash
was 82 percent.
177
-------
KILN ASH
z
5 i)
z <
— iu
Z S
Su.
CJ
(? ~
u.
1UU
80
60
40
20
n
i As
_
_
—
«
T
_L _j-
_ T MgT^.
Bi
-Cd
Ir,, »f -*-^'
Ba _.Cu -+-
Pb
^ 1 1 ~L 1 1 1 t 1
»
S
-------
I
Z LLJ
-J CC
2r>
-5» W
? <
LU
Z S
HO"
cc ^
u_
cc
LU
f*Q ^^,
m CO °
irgo:
° UJ CO
zEjui
O H LL
pxO
b UJ v°
< 2^
CC
LU
100
80
60
40
20
o
KILN ASH
I"
- — .
_.
r
Bi
• Cd •
11 t
0 200 400 600
VOLATILITY
SCRUBBER
10n
80
60
40
20
0
Cd
-
~
-
Bi
"
"AS
I , ,
tPb
J
0 200 400 600
VOLATILITY
$ t * Mg i
If Sr cr
Ba 4-
Cu
i i i i i
ESDI 32-90
800 1000 1200 1400 1600
TEMPERATURE (-C)
EXIT FLUE GAS
Cu
Ba | Sr M ^.r
. =1= , + 1 1 1 : **? ±
800 1000 1200 1400 1600
TEMPERATURE (-C)
SCRUBBER LIQUOR
cc
O
^>
a
_j ^
m cc
m *~i
gm
to 2
S^1
^— '
I
100
80
60
40
20
0
™
—
Bi
Cd
f ,-
i i
r
Bt3 Cu Sr Cr
, J , ± , , 3= Mp *
200 400 600 800 1000 1200
VOLATILITY TEMPERATURE ('C)
1400 1600
Figure 3. Distribution of metals in the RKS discharge streams in the single-stage ionizing wet
scrubber tests.
179
-------
The inconsistent behavior observed for lead may be because the volatility temperature of lead
is within the range of kiln temperatures tested or may be related to increased lead volatility in the
presence of chlorine. The volatilities of some metals may be altered through reactions in the
incineration system. For example, chlorine can react with some metals to form new compounds that
volatilize more readily. Many metal chlorides are more volatile, as indicated by a lower volatility
temperature than their corresponding oxide or elemental forms. For example, equilibrium
calculations for a mixture containing 5-percent chlorine result in PbCl4 as the principal vapor phase
species. The corresponding volatility temperature of -15°C (5eF) is considerably less than the
volatility temperature of 627° C (1160°F) for lead under oxidizing conditions. Test data presented
in the next section suggests that lead became more volatile with increasing kiln temperature and
with increasing chlorine. Although the tests were performed under the same nominal conditions,
minor differences between the two programs may have combined with the sensitivity to both of
these test variables to cause the wide variation in lead discharge distributions.
EFFECTS OF INCINERATOR OPERATING CONDITIONS ON METAL DISTRIBUTIONS
Test results from the single-stage ionizing wet scrubber test series show that increasing the kiln
temperature from 816° to 927° C (1500° to 1700°F) caused a noticeable increase in the volatility
of cadmium, bismuth, and lead. Figure 4 shows that as the kiln temperature increased there was
a significant decrease in the kiln ash fraction of these metals, with corresponding increases in the
scrubber exit flue gas and scrubber liquor fractions. Although the volatility of lead increased with
higher kiln temperature, lead still remained relatively nonvolatile and was found primarily in the
kiln ash for this test series. Changes in kiln temperature had no significant effect on the discharge
distributions of any of the remaining metals. Data from the venturi/packed-column test series
showed that kiln temperature had less pronounced effects on metal volatility.
Changes in afterburner temperature did not significantly affect the distributions of any of the
metals among the scrubber exit flue gas and scrubber liquor discharge streams for either test series.
Variations in feed chlorine content did not affect metal discharge distributions within the limits
of data variability established by replicate test conditions during the single-stage ionizing wet
scrubber test series. However, as shown in Figure 5, increased feed chlorine content did cause
increased volatility of copper and lead during the venturi/packed-column scrubber test series. As
noted, the calculated lead volatility temperature is reduced significantly for cases that consider the
presence of chlorine. The calculated volatility temperature for copper also decreases significantly
when chlorine is considered (from 1116° to 1276C (1975° to 260° F)). However, because neither
metal was as volatile as would be expected if its volatility temperature were that of the metal
chloride, it is suspected that only part of the lead and copper reacted with the chlorine to form the
more volatile metal chloride species. The absence of a similar clear relationship between chlorine
and the volatility of these metals during the single-stage ionizing wet scrubber test series also
suggests that only partial reactions occurred.
METAL FLUE GAS PARTICLE SIZE DISTRIBUTIONS
Metals find their way into flue gas particulate via two pathways. In one pathway, the metal
remains in a condensed phase through the entire incinerator system and is carried out of the system
with entrained ash in the combustion gas. In the second pathway, the metal vaporizes at some point
in the incinerator, then recondenses when the flue gas cools. Both vaporization and condensation
can occur locally under proper conditions.
180
-------
CADMIUM DISCHARGE DISTRIBUTIONS
70
KILN ASH SE FLUE GAS LIQUOR
BISMUTH DISCHARGE DISTRIBUTIONS
KILN ASH
SE FLUE GAS
LIQUOR
LEAD DISCHARGE DISTRIBUTIONS
-30
KILN ASH
SE FLUE GAS
LIQUOR
Figure 4. Effects of kiln temperature on the discharge distributions of cadmium, bismuth, and
lead in the single-stage ionizing Vet scrubber tests.
181
-------
C3- 100
LEAD DISCHARGE DISTRIBUTIONS
o
UJ
a:
^
co
55
80
60
40
y 20
O
f I 4%
Eiii"?
— 100
Q
UJ
DC 80
CO
55
5 60
H 40
UL
O
2
Q 20
KILN ASH SE FLUE GAS LIQUOR
COPPER DISCHARGE DISTRIBUTIONS
0%
4%
8%
KILN ASH
SE FLUE GAS
LIQUOR
Figure 5. Effects of feed chlorine content on the discharge distributions of copper and lead in the
venturi/packed-column scrubber tests.
182
-------
Vaporized metals can condense homogeneously into condensation nuclei that grow into a very
fine fume, or they can condense heterogeneously onto existing flue gas partieulate. In both
mechanisms the tendency is to enrich (be found at higher per mass concentration) in fine
partieulate. In the former mechanism, fume particles are very fine (1 um or less). In the latter
mechanism, the surface-to-mass ratio is higher for fine particles than for coarse particles. Because
condensation onto an available surface is a per surface area event, this also leads to enrichment in
fine partieulate.
Via the above mechanisms, the distribution of a given metal among flue gas particle size ranges
is strongly influenced by the extent to which the metal vaporizes in the incineration system.
Refractory metals that do not vaporize significantly tend to be relatively evenly distributed in the
flue gas partieulate size ranges on a per mass (mg/kg partieulate) basis. Volatile metals tend to
enrich in the fine paniculate fractions, with enrichment tendency increasing with increasing
volatility.
Figure 6 shows the fractions of the partieulate metal found in the less than 10 |im size range
during the single-stage ionizing wet scrubber test series. The effects of increased kiln exit
temperature are shown. The fractions of the total partieulate sample in this size range are also
shown. Values for the three replicate test conditions were averaged and plotted as a single point.
Metal partieulate distributions are plotted against the volatility temperatures to facilitate comparison
of relative metal behavior.
With the exception of chromium, the average metal distributions in the flue gas particle size
range less than 10 |im shifted from roughly 20 percent to an average of 60 percent as the kiln
temperature was increased from 816° to 927"C (1500° to 1700°F). In addition, the redistribution
of metals to this size range generally correlated with the relative volatilities of the metals, with the
volatile metals most affected. Interestingly, arsenic in the flue gas behaved as the most volatile
metal with respect to particle size redistribution; more than 80 percent of the arsenic partieulate
was found in the less than 10 \im size fraction at a kiln exit temperature of 927°C (1700°F).
Although most of the arsenic remained in the kiln ash, the fraction that exited with the flue gas
became significantly enriched in the fine partieulate fractions during tests at higher temperatures.
Observed enrichment of the test metals in the less than 10 }im paniculate suggests that some metal
vaporization occurred in the system, even though many of these metals were predominantly
nonvolatile as indicated by their tendency to remain predominantly in the kiln ash.
Although a relatively small fraction of the metal fed may escape the kiln, a propensity to
concentrate in the finer paniculate fractions may increase the risk posed by these emissions. These
data are significant given the greater challenge to air pollution control devices posed by smaller
paniculate, the ability of smaller particles to penetrate the deep lung, and the toxicity of many
metals. However, the data also suggest that metal enrichment in the fine partieulate fractions can
be controlled operationally by limiting incinerator temperature.
The effects of the waste feed chlorine content on total partieulate and metal-specific size
distributions are shown in Figure 7. When feed chlorine was increased from 0 to 4 percent, the
fraction of total partieulate in the less than 10 pm fraction increased from 20 to approximately
35 percent. This is expected if the presence of chlorine in the feed serves to increase the volatility
of some feed inorganic constituents. When reviewing the data, the effects of chlorine were taken
to be most significant when the metal distributions were shifted more than the distributions of the
183
-------
1UU
CO
r-
s
•| 80
o
«*
fe 60
LU
O
DC
UJ
0- 40
UJ
£»
|_
5 20
s
o
ft
0816"C(1500T)
|
j
/
3AS .
L
f ^
*871°C(1600*F)
iCd D927'C(1700T)
f oPb
Bin
v •*
_ 4
c
fr DBa ^
4
pCu c
f Srn ^
4
* *
1 1 <
<
pMg
r
4
'
S SAMPLE
icr
op
i
Q
83
500 1000 1500
VOLATILITY TEMPERATURE (*C)
2000
Figure 6. Effect of kiln temperature on the distribution of metals in the afterburner exit flue gas
participate size fractions in the single-stage ionizing wet scrubber tests.
80
v>
I
60 --
LU
040
UJ
DU
UJ
20 -
O
¥As *
0 I
4
fCd
Bit
<
0 0%
* 4%
0 8%
a
>
*
<»
1
1 Pb y Cu Mg %
p
VBa
T
1
Sr*f J
T £
T
^
F pCr
I
0
4 4 '
1 1 '
! SAMPLE
1
O)
Si
o
Q
w
UJ
500 1000 1500
VOLATILITY TEMPERATURE (*C)
2000
Figure 7. Effect of feed chlorine content on the distribution of metals in the afterburner exit flue
gas particulate size fractions in the single-stage ionizing wet scrubber tests.
184
-------
total particulate sample. Thus, flue gas particulate size distributions for barium and strontium and,
to a lesser extent, arsenic, bismuth, and magnesium, were considered not to be significantly affected
by waste feed chlorine concentrations. For these metals, the magnitude of the shift to the finer
particulate fractions was about the same as the shift for the total particulate sample, primarily
reflecting the shift in the particulate sample size distribution.
Chlorine had a more pronounced effect on the particulate size distributions of cadmium,
chromium, copper, and lead. For cadmium, copper, and lead, the shift to finer particulate occurred
with the initial feed chlorine content increase from 0 to 4 percent. The distribution of these metals
in particulate of less than 10 fim increased from approximately 20 to approximately 55 percent. No
additional redistribution occurred with the further feed chlorine content increase to 8 percent.
Chromium distribution in particulate of less than 10 fim increased with both feed chlorine content
increases, from 2 to 20 to 50 percent with chlorine increased from 0 to 4 to 8 percent. The impact
on copper and lead particulate distributions is expected based on their reduced volatility
temperatures in the presence of chlorine. Cadmium and chromium redistributions to finer
particulate with increased chlorine are not similarly predicted by reduced volatility temperatures.
APPARENT SCRUBBER COLLECTION EFFICIENCIES
The apparent scrubber efficiency for collecting flue gas metals was determined for each test.
The apparent scrubber efficiency represents the ratio of the normalized metal fraction measured
in the scrubber liquor to the sum of the normalized metal fractions measured in the scrubber liquor
and scrubber exit flue gas. Figures 8 and 9 summarize the collection efficiencies for the
venturi/packed-column scrubber and single-stage ionizing wet scrubber test series, respectively. The
bar for each metal represents the range of scrubber efficiencies over the respective test series, with
the average noted by the midrange tick mark.
For the venturi/packed-column scrubber test series, average metal-specific collection efficiencies
ranged from 31 to 88 percent; the overall average for all metals was 57 percent. For the single-
stage ionizing wet scrubber test series, average metal-specific collection efficiencies ranged from 22
to 71 percent; the overall average for all metals was 43 percent. Figures 8 and 9 show that the
collection efficiencies for each metal varied significantly during each test series. However, average
efficiencies were generally higher for the less volatile metals.
CONCLUSIONS
The following conclusions are based on the results of the two completed trace metals test
programs in the pilot-scale incinerator at the IRF:
• In the rotary kiln incinerator, cadmium and bismuth were relatively volatile. Over all tests,
the average fractions of these metals recovered in the kiln ash was less than 40 percent.
• In the rotary kiln incinerator, arsenic, barium, chromium, copper, magnesium, and strontium
were relatively nonvolatile. Over all tests, the average fractions of these metals recovered
in the kiln ash was greater than 75 percent.
185
-------
_. lUU
£
o
g 80
o
fe
£ 60
LU
CQ
m
Ji 40
O
CO
z
a 20
a.
Sc „
— .
-As
""•*"**
Bi-
-Cd
"""-""
I
4-Ba
Sr-
-Pb
fc
Ifl
-Cr
-Cu
I !
m
C\J
o
Q
CO
LU
500 1000 1500
VOLATILITY TEMPERATURE (*C)
2000
Figure 8. Apparent collection efficiencies for metals achieved by the venturi/packed-column
scrubber.
O
1 80
o
u.
u_
LLJ
CD
CO
3
O
CO
60
40
20
As
Mg
Cd
•- Pb
Ba
Sr__
--Cr
--Cu
500 1000 1500
VOLATILITY TEMPERATURE ('C)
2000
Figure 9. Apparent collection efficiencies for metals achieved by the single-stage ionizing wet
scrubber.
186
-------
* Lead behavior in tlie rotary kiln differed between the two test series. For the
venturi/packed-column test series, the average fraction of lead discharged to the kiln ash
was 20 percent. For the ionizing wet scrubber test series, the average fraction of lead
discharged to the kiln ash was 82 percent. This inconsistent behavior may be related to the
strong relationship between lead volatility temperature and chlorine or to the sensitivity of
lead volatility to temperature over the range tested.
» Relative metal volatilities in the kiln generally agreed with expectations based on metal
volatility temperatures, with the exception of arsenic, which was much less volatile than
expected. It is possible that arsenic forms a thermally stable compound in the incineration
environment or becomes physically bound in the solid matrix.
» Results from the ionizing wet scrubber test series showed that increased kiln temperature
caused increased volatility of bismuth, cadmium, and lead. There was a significant decrease
in the kiln ash fraction of these metals, with corresponding increases in the scrubber exit
flue gas and scrubber liquor fractions. Discharge distributions of the remaining metals were
insensitive to changes in kiln temperature.
» Afterburner temperature did not affect metal distribution to the scrubber exit flue gas and
scrubber liquor discharge streams.
» Increased feed chlorine content caused increased volatility of copper and lead during the
venturi/packed-column test series. There was a significant decrease in the kiln ash fraction
of these metals, with corresponding increases in the scrubber exit flue gas and scrubber
liquor fractions. Discharge distributions of the remaining metals did not vary conclusively
with changes in feed chlorine content. Also, variations in feed chlorine content did not
conclusively affect any metal discharge distributions during the single-stage ionizing wet
scrubber test series.
» Both kiln temperature and feed chlorine content affected the distributions of at least some
of the metals among the flue gas particulate in the less than 10 \im size range. Size
distributions of the metals most nearly reflected the overall entrained particulate size
distribution for the tests with the lowest kiln temperature and no chlorine in the waste feed;
very little redistribution among the particulate was observed. For these two tests,
approximately 20 to 25 percent of each metal and of the total particulate sample were in the
less than 10 pm particulate.
« Increasing the kiln temperature to 927° C (1700°F) caused the average size distribution to
shift to approximately 60 percent less than 10 jim for all test metals except chromium. The
test data suggest that increased kiln temperature over this range caused the flue gas
particulate metal distributions to shift to the finer particulate size fractions. Additionally,
the redistribution of metals to this size range generally correlated with the relative
volatilities of the metals, with the volatile metals most affected. Interestingly, arsenic in the
flue gas behaved as the most volatile metal, becoming most enriched in the less than 10 jim
particulate size range. Test data show that even metals that are classified as relatively
nonvolatile based on their behavior in the kiln undergo some vaporization and
recondensation, with resulting concentration in the finer particulate size range.
187
-------
The addition of chlorinated compounds to the waste feed primarily affected cadmium,
chromium, copper, and lead distributions in the flue gas particulate less than 10 jim. For
cadmium, copper, and lead, the increase in waste feed chlorine content from 0 to 4 percent
caused the distributions to shift from roughly 20 percent to approximately 55 percent less
than 10 |im. No further effects with feed chlorine increased to 8 percent were observed for
these metals. For chromium, increased chlorine content from 0 to 4 to 8 percent caused
redistributions of 2 to 20 to 50 percent in the particulate less than 10 Jim.
Average metal collection efficiencies for the venturi/packed-column scrubber ranged from
31 to 88 percent; the overall average for all metals was 57 percent. For the single-stage
ionizing wet scrubber, average metal collection efficiencies ranged from 22 to 71 percent;
the overall average for all metals was 43 percent. In general, collection efficiencies were
higher for the less volatile metals.
188
-------
REFERENCES
1. Federal Register, Vol. 55, No. 82, April 27, 1990.
2. Barton, R.G., Clark, W. D., and Seeker, W. R., Fate of metals in waste combustion systems.
Paper presented at the First Congress on Toxic By-Products of Combustion, Los Angeles,
California. August 1989. Accepted for publication: Combustion Science and Technology.
3. Proposed methods for stack emissions measurement for CO, O2, THC, HC1 and metals at
hazardous waste incinerators. U.S. Environmental Protection Agency, Office of Solid Waste,
November 1989.
4. 40 CFR Part 60, Appendix A.
5. Determining the properties of fine paniculate matter. ASME Power Test Code 28.
6. Test methods for evaluating solid waste: physical/chemical methods. EPA SW-846,3rd edition,
November 1986.
7. King, C. and Waterland, L.R. Pilot-scale incineration of arsenic contaminated soil from the
Baird and McGuire Superfund site. Acurex draft report prepared under EPA Contract 68-C9-
0038, March 1990.
8. Gossman, D., Black, M. and Ward, M, The fate of trace metals in the wet process cement kiln.
Paper presented at the 1990 Specialty Conference on Waste Combustion in Boilers and
Industrial Furnaces, Air & Waste Management Association, Kansas City, Missouri. April 17-20,
1990.
189
-------
SOIL HEATING TECHNOLOGIES FOR IN SITU TREATMENT; A REVIEW
By: Janet M. Houthoofd, John H. McCready and Michael H. Roulier
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
ABSTRACT
In the remediation of soils contaminated with hazardous compounds,
the costs, logistical concerns, and regulatory requirements associated with
excavation, ex situ treatment, or off site treatment make in situ treatment
a highly attractive alternative. Because temperature affects a number of
physical, chemical, and biological processes in soils, the effectiveness
and efficiencies of some in situ treatment technologies can be improved by
controlling soil temperature. Technologies such as bioremediation and gas
phase removal of organic compounds, for example, can be enhanced through
heating of the soil. Consequently, there is great interest in the
development and testing of methods of soil heating to be used in
conjunction with in situ treatment technologies. This paper reviews
temperature effects and various ways that heating is being incorporated
into in situ remediation of contaminated sites.
This paper has been reviewed in accordance with the U.S.
Environmental Protection Agency's peer and administrative review policies
and approved for presentation and publication.
190
-------
INTRODUCTION
The U.S. EPA's Superfund research program is developing methods for
in place (in situ) removal of contaminants from soils and for in place
treatment of contaminated soils. This work is motivated by the high cost
of managing large volumes of soil with low levels of contamination and by
the need to comply with provisions of the Superfund Amendments and
Reauthorization Act (SARA) and the Resource Conservation and Recovery Act
(RCRA).
A recent U.S. EPA report (U.S. EPA, 1990a) on in place treatment
describes a large number of chemical and physical processes (e.g.
oxidation, reduction, precipitation) that could potentially be used in situ
to immobilize or detoxify contaminants in soils. The majority of these are
conceptual or have been tested only in the laboratory. Although in place
chemical treatments involving these processes have been proposed, the major
developments have been in biodegradation, stabilization/ solidification,
and removal of contaminants 1n the gas phase. Many chemical/ physical/
biological processes in soil are affected by temperature to some degree.
Consequently, U.S. EPA is investigating methods for heating soil in place
to affect these processes and make in situ treatments more effective.
TEMPERATURE EFFECTS
Temperature dependent chemical/physical phenomena that affect the
efficiency of in situ treatment processes include vapor pressure, mass
transfer coefficients, equilibrium constants, and reaction velocity/rate
constants. Co-distillation of immiscible compounds and movement of water
and gases in unsaturated soils are other relevant phenomena affected by
temperature.
It is well known that vapor pressure of liquids increases with
increasing temperature (Maron and Prutton, 1965). This allows increases in
temperature to increase the rate of removal of individual organic compounds
in the gas phase. It also allows gas phase removal of compounds that do
not have a sufficiently high vapor pressure at normal soil temperatures.
Increasing temperature also increases the diffusion coefficient and, hence,
the rate of diffusion in the gas phase as well as the liquid phase (Call,
1957; Elhers et al., 1969; Lavy, 1970). In situ radio frequency heating
and in situ vacuum-assisted steam stripping both take advantage of this
effect for removal of organic compounds in the gas phase.
The mass transfer coefficient for oxygen across the air-water
interface increases approximately linearly with increasing temperature in
the range 0 to 35* C (Downing and Truesdale, 1955). Temperature also
affects the magnitude of coefficients for transfer of other chemicals
across this interface (Thibodeaux, 1979).
Compounds with low water solubility (e.g. PCB, pentachlorophenol)
will volatilize in the presence of boiling water at temperatures lower than
their boiling points as pure compounds (Maron and Prutton, 1965). This
191
-------
phenomenon, called steam distillation or co-distillation, enhances the
effectiveness of in situ radio frequency heating and in situ vacuum-
assisted steam stripping for removal of organic compounds in the gas phase.
Chemical treatment of contaminated soil in situ has been envisioned
(U.S. EPA, 1990a) but has not yet been applied to any significant extent.
As in situ chemical treatments are developed it will be possible to take
advantage of the increases in reaction velocity/rate constant and the
changes in equilibrium constants for most chemical reactions with
increasing temperature (Maron and Prutton, 1965). The only application of
this effect in situ has been the investigation of radio frequency heating
for use in soil to facilitate potassium polyethylene glycolate (KPEG)
treatment of chlorinated organics such as PCB (Dev, 1986).
Temperature gradients in unsaturated soil have a small but
appreciable effect on the rates of movement of water (Letey, 1968). Water
moves along the temperature gradient, i.e. from areas of high temperature
to areas of lower temperature. It may be possible to use the temperature
gradient effect to aid in distribution of treatment chemicals in situ.
Temperature has a profound effect on the activity of microorganisms
in soil and this effect can be used to great advantage during in situ
bioremediation. Host microorganisms grow only within a limited temperature
range; maximum and minimum growth temperatures are about 30° C apart. The
optimum temperature for growth of mesophiles is about 38" C; the optimum
for thermophiles is about 62° C (Tortora et al., 1989). Soil heating can
be used to improve the growth rates of these classes of organisms because
their optimum temperatures are well above normal soil temperatures. This
makes it possible to shorten the time required for in situ bioremediation
by heating the soil. In areas of colder climate it may not be possible to
conduct in situ bioremediation during much of the year because of low soil
temperatures. Artificially warming the soil can extend the time during
which bioremediation is effective and thus shorten the overall time for
remediation. Finally, it may be possible to favor the activity of specific
organisms if their optimum temperature is higher than the optimum for
competing organisms.
RADIO FREQUENCY HEATING
The in situ radio frequency (RF) heating method allows relatively
rapid and uniform in place heating of large volumes of soil in the vadose
zone. The heating is performed by the application of electromagnetic
energy in the radio frequency band. The temperature rise occurs due to
ohmic or dielectric heating mechanisms and does not rely on thermal
conductivity of the soil matrix. As shown in Figure 1, electrodes are
inserted into the contaminated soil through drilled bore holes. The
electrode array consists of three rows of electrodes inserted to the depth
of the zone to be decontaminated. RF power is applied to the center row of
electrodes, and the two outer rows serve to confine the energy within a
192
-------
Figure 1. Radio Frequency Heating
defined volume of soil. The soil is heated volumetrically and uniformly to
temperatures between 150° C and 300° C. In this temperature range many
volatile and semivolatile hazardous organic compounds can be vaporized.
Vaporized native soil moisture and contaminants are collected and treated
on site through a soil vapor extraction and treatment system. The
contaminant vapors and boiled water are recovered by the application of a
vacuum to selected hollow electrodes, which have been specially designed to
allow both the application of RF power to the soil and the collection of
vapors from the soil. The collection electrodes are connected to a vacuum
manifold which transports the gases and vapors to an on-site treatment
system where vapors may be condensed and uncondensed gases may be treated
by carbon adsorption, combustion, or scrubbing. An additional component of
the system is generally a rubber sheet barrier laid out over the soil
surface. It serves to prevent fugitive emissions and to provide thermal
insulation to avoid excessive cooling of the near surface zone.
In addition to the usual advantages of a process that works in situ
as opposed to ex situ (above ground), RF heating has several other
attractive points. For instance, the process produces only one one
thousandth the amount of fluids, gases, and liquids produced by
incineration. Uniformity of heating allows more uniform decontamination
than some alternate technologies. Furthermore, researchers claim the
technique can handle variations in the soil matrix, such as soil moisture
and clay stringers present in sandy soil. It may also be more effective
than certain other technologies in remediating clayey soils, which are
often more difficult to decontaminate than sandy soils.
A number of bench- and pilot-scale RF tests as well as limited field
testing have been performed (Dev and Downey, 1988; Dev et al.. 1984; Dev et
aL., 1986; Dev et al., 1988; Dev et a!.. 1989; Anonymous, 1988), A field
test was conducted on hydrocarbons at the Volk Air National Guard Base (Dev
193
-------
and Downey, 1988; Dev e.t al.. 1989; Anonymous, 1988). Table 1 summarizes
various test results (Sresty et al.. 1990). Removal efficiencies of up to
99% have been obtained. Further field demonstration is planned (Sresty et
al_._) 1990). The in situ RF heating decontamination technology is also
being offered on a commercial basis (Anonymous, 1989).
TABLE 1. SUMMARY OF DECONTAMINATION TESTS USING RADIO
FREQUENCY HEATING TECHNOLOGY
Soil
Type
Sandy
Sandy
Sandy
Sandy
Clayey
Contaminants
Tetrachloro-
ethylene
Chlorobenzene
Jet fuel,
solvents
Aroclor 1242
Jet fuel,
solvents
Cone.
Range, ppm
10, 1000
10, 1000
1-5000
1000
1-1000
Treatment
Time, hr.
4
16
4
16
16-24
12 days
16
16-48
Type of
Exper.
Bench
Pilot
Bench
Pilot
Pilot
Field
test
Pilot
Pilot
Results
% Removal
94 to 98
94 to 98
94 to 98
94 to 98
91 to 99
94 to 99
70 to 99,8
70 to 98.3
Source: Sresty et.at..199Q
STEAM AND HOT AIR STRIPPING
In situ steam stripping or hot air stripping or a combination of the
two have also been used to recover highly to moderately volatile organics
from contaminated soil. Contaminants with boiling points less than 250" C
are most amenable. Heat from injected steam or hot air assists in
vaporization of the compounds, while the gas flow carries the contaminants
to the soil surface. A vacuum can accelerate the rate of volatilization
and speed the transport of contaminants to the surface as well as guard
against leakage to the outside environment. The off gases are then pulled
through an above-ground treatment train, for example, condensers followed
by carbon adsorption. Figure 2 represents a generalized typical system.
Several diverse studies and tests of the technology have been
reported. In one case, a field pilot test was performed to investigate the
advantages of piping waste heat from a catalytic incinerator to a site
contaminated with hydrocarbons (DePauli, 1990). Tests indicated an
improvement of 70% in total hydrocarbon removal due to hot air injection.
In New Mexico a hot air injection system is being used to increase
biodegradation rates for in situ remediation of leaking underground storage
194
-------
Injection Wei i
G
\
SffiSSSS^SSSSSjfe^-gj^sSSSSg;::
i?s i
.. - * *
*• w 8S
I
Steam
snerator
::r?
T
1
UUV^WMI^^ v«^l |^ U^^M ||
xi
Off-gas cover'' !
a
\_
Blower f Condenser f Adsorption
1 1 1
f
T
t
A4
ii
J V
f
r
r
i
i
i
1 ^. Injection Wei 1
*S**
1
i
1
1
i
1
./
Figure 2. Vacuum-Assisted Steam Stripping
tanks (Hinchee and Smith, 1990a), Lord et al. (1988, 1989, 1990) conducted
laboratory studies of vacuum-assisted steam stripping of organic
contaminants from soils. An analytical model was developed to determine
some basic parameters. Small-scale experiments were run employing a
geosynthetic cap to facilitate the confinement and collection of steam and
contaminants. Effects of varying steam pressures on contaminant removal
efficiencies were also investigated. Nunno et al. (1989) reported that
several field tests of an in situ steam stripping process were conducted on
contaminated soil in the 1980s in the Netherlands with varying levels of
success. Looking at a possible variation to the steam stripping process,
Jackson et al. (1990) performed a literature review to investigate the
technical feasibility of using surfactants to enhance in situ steam
stripping processes for removing organic compounds from contaminated soils.
Laboratory-scale experiments to investigate the use of surfactants with
steam stripping were planned. Steam injection combined with vacuum
extraction was tested at pilot-scale at a site in San Jose, California
(Udell and Stewart, 1989; Baum, 1988). Soil at the site contaminated with
several organic solvents was successfully cleaned to a low part per million
level. Some contaminants were recovered in quantities sufficient to make
recycling of them back into commercial chemical use feasible. A
combination steam and hot air stripping technology was demonstrated in situ
in U.S. EPA's Superfund Innovative Technology Evaluation (SITE)
Demonstration Program in 1989. This system used rotating augers to break
up and mix the soil as well as to inject steam and hot air. At the site in
San Pedro, California, twelve blocks of soil approximately 7 ft. by 4 ft.
were treated to a 5 ft. depth. Removal efficiencies for volatile organic
compounds averaged 85%. Semivolatile organic compounds were also removed,
but at a lower efficiency (U.S. EPA, 1990b; de Percin, 1990}. This same
195
-------
system is capable of injecting remediation agents, such as stabilizers or
chemicals, and mixing them into the soil (La Mori, 1989). A process being
studied in U.S. EPA's SITE Emerging Program uses enhanced oil recovery
technology to recover a major portion of the organic liquid phase in
subsurface oily waste accumulations. The oily waste is mobilized by in
situ heating with steam and is displaced to production wells by sweeping
with hot water. Residual contaminants not recovered are subsequently
treated by in situ microbial degradation. Laboratory-scale studies of the
technology have been performed (Johnson and Guffey, 1990; U.S. EPA, 1989).
VITRIFICATION
In situ vitrification (ISV) is a soil heating technology in which a
powerful current of electricity is transferred within a square array of
electrodes that are inserted into contaminated soil at the desired
treatment depth. A layer of graphite is then applied within the area of
the electrodes on the ground surface to act as a starter path. Because dry
soils are not electrically conductive, the graphite acts as a starter to
initiate a melting zone in the contaminated soil. The melting zone
(generally about 1600°C to 2000°C ( 29QO°F to 3600°F)) caused by this
electrical current gradually works its way downward and outward (past the
electrodes) melting the contaminated soil. As the melting action is taking
place, the pyrolized by-products migrate to the surface of the vitrified
zone and burn in the presence of air. Evolved gasses are trapped under an
off-gas cover placed on the top of the treatment area and sent to a
treatment unit to ensure that emissions are within regulatory limits. When
the process is terminated the contaminated soil cools into a stabilized
crystalline block, with about ten times the strength of unreinforced
concrete (see Figure 3). It is not affected by either wet/dry or
freeze/thaw cycling and passes EP-Tox and TCLP leach testing criteria for
priority pollutant metals (Buelt et al.. 1987).
A large scale ISV system, with a maximum electrode array of 30 ft.
width by 30 ft. depth, can produce a soil melt rate of 4 to 6 tons per hour
and can encompass a total melting zone of 1000 tons of contaminated soil.
Generally, the melt grows outward to form a melting zone approximately 50%
wider than the electrode array. Recent testing has shown that organic and
other vapors are not driven outward by the heat of the ISV process but
rather migrate toward the melt and toward the surface of the melt (Hansen
et al... 1990). As the void volume in the soil can account for 20 to 40% of
the volume, a proportionate volume reduction occurs when the melting zone
is vitrified. This subsidence can then be covered with clean backfill and
vegetated.
The contaminated soil conditions have a direct bearing on the ISV
melting process and the quality of the vitrified block. This technology is
most effective on low moisture soils. If groundwater is present and soil
permeability is less than 1 X 10 cm/sec, additional energy must be
expended in order to drive off the water to heat the soil. Typical soils
require 800 to 1000 kwh/ton total energy input, which can be supplied by
local utilities or be diesel-generated in remote locations. ISV processing
196
-------
Vltrlffed Soll/Vfeste
Figure 3. In Situ Vitrification
requires relatively high levels of glass forming materials, such as silica
(50-80%) and alumina (5-12%), in the soil to form and support a high
temperature melt. Sufficient levels (2 to 5%) of monovalent alkali
cations, such as sodium and potassium, must also be present to provide the
electrical conductivity needed to advance the heat rate in the contaminated
soil. Most of the soils tested for ISV processing (in 30 out of 32 cases)
have been found to contain adequate levels of these materials (Hansen ejt
al.. 1990).
This technology can potentially treat a wide range of contaminants.
It is therefore possible that mixtures of organic, inorganic, and
radioactive contaminants in a solid media could be processed
simultaneously. Organic contaminants are destroyed by pyrolysis and are
expected to be most effectively treated at concentrations in the 5 to 10
percentage-of-weight range. Inorganic contaminants are then incorporated
into a vitrified residual product and are expected to be most effectively
treated at the 5 to 15 percentage-of-weight range. Site conditions that
may limit applicability of the ISV process are 1) individual void volumes
in excess of 150 cubic feet, 2) metals in excess of 5 percent of the melt
weight or continuous metal occupying 90 percent of the distance between the
electrodes, 3) rubble in excess of 10 percent by weight, and 4) combustible
organics in the soil or sludge (U.S. EPA, 1989).
The ISV technology was originally created and applied to stabilize
radioactive and radioactive mixed wastes, and has been applied as a large-
scale test at the Department of Energy's (DOE) Hanford site. It has
however, recently gained interest for the treatment of hazardous waste.
The first such application of ISV is currently being performed at a private
site contaminated with PCBs. Superfund sites slated for ISV processing
this year include, Parsons/ETM (EPA), Denver Radium (EPA), Ionia City
Landfill (EPA/PRPs), Northwest Transformer (EPA/PRPs), M-l Ponds (Army),
and Site 10 (Air Force).
197
-------
Cost estimates for this technology range from $275 to $600 per ton of
treated soil, depending on the site characteristics.
ELECTRICAL RESISTANCE HEATING
Electrical resistance heating technology heats soil by passing an
electrical current through electrodes that are inserted into ground (Heath,
1990). Unlike in situ vitrification, however, no melting action of the
soil takes place to stabilize contaminants. Enough current is applied
through the electrodes to remove most of the soil moisture and volatile
contaminants. Once the soil moisture is driven off by the heat
(approximately 65°C (150°F)), the voltage is increased to stimulate in
place oxidation of any nonvolatile organics which are then transformed into
lighter components that volatilize. During the process, all water vapor
and volatiles are collected using conventional vacuum extraction
techniques.
While electrical resistance heating has been employed at the pilot
scale for the melting of permafrost, it has only recently been explored at
the bench scale for in situ treatment of contaminated soils. During
experimentation, however, there was an approximately 90% removal of
volatile and semivolatile/nonvolatile materials tested which makes this
technology a potentially efficient way to treat contaminants in situ
through soil heating (Hinchee and Smith, 1990b).
SOIL PROPERTY MODIFICATION
Heating through modification of the soil properties can produce only
small temperature increases but the methods are simple and no additional
energy inputs are required, making this a "low-tech" approach to soil
heating. Modification of surface properties regulates the incoming and
outgoing energy, increasing the heat absorption of the soil during warm or
sunny periods and reducing heat loss during cold or dark conditions. This
is achieved by the use of plastic sheeting, stripping surface vegetation,
or applying organic mulches to the ground surface (Hinchee and Smith,
1990c). The soil can also be irrigated to increase thermal conductivity
and the net heat transfer into the soil.
Clear polyethylene sheets applied to a stripped soil surface are used
to increase radiation collection during the day and reduce convective and
conductive heat loss at night. In Alaska, where this technology has been
tested at the pilot scale level for applications to improve crop growth and
melting permafrost, clear polyethylene has produced positive results. A
soil temperature increase of 16.7 C (30°F) at a depth of an inch was
achieved using clear polyethylene to improve increased crop yield and seed
germination (Dinkel, 1966). In testing to melt permafrost (Nicholson,
1978), small plots were tested with a variety of soil surface applications
and produced the following soil temperature increases:
o 2°C (3.6°F) for stripping vegetation
198
-------
o 4.5°C (8°F) for clear polyethylene over natural vegetation
o 5.5°C (100F) for clear polyethylene over stripped vegetation.
Increasing soil water content by irrigation increases the heat
capacity of the soil, raises the humidity of the air, lowers air
temperature over the soil, and increases thermal conductivity. The
cumulative result is a reduction in daily soil temperature variations and
an increase in the net heat transfer into the soil (Baver et al., 1972).
Soil property modification can provide in situ heating at a low cost
and is a viable supplement for bioremediation, particularly in colder
climates. Because temperature increases are nominal and can often take
days and weeks to attain, this concept could not achieve the high
temperatures required to significantly increase vapor pressures and promote
removal of contaminants in the gas phase. To date, however, none of the
soil surface modification concepts have been demonstrated for remediation
at contaminated sites,
SOLAR HEATING WITH OPTICAL FIBERS
The sun is a low intensity source of energy for soil heating; under
ideal conditions the power density available from the sun is about one
kilowatt per square meter of collector. Parabolic collectors coupled to
optical fibers are another method that is being tested for using solar
energy to heat soil in situ. Collectors have been used in the past for
direct heating of air and water. There are significant heat losses during
transmission that would make this process undesirable for in situ heating
of soil. An approach to overcoming these heat losses during transmission
is being tested in a research project (Brown and Murdoch, 1990) that was
started in July 1990.
Compound parabolic concentrators, which have the advantage of
collecting scattered sky light as well as direct sunlight, are being
modified for coupling to a cable of optical fibers. The optical fiber has
the potential to transfer the solar radiation with high efficiency over
long distances. By conducting the solar energy as light, rather than as a
hot fluid, thermal losses along the transmission line are eliminated. The
collector surfaces also remain relatively cool and radiate away little
energy because conversion to heat occurs at the radiator or the end of the
optic cable, rather than at the collector.
Two problems must be resolved to allow development of a workable and
efficient system. The coupling between the optical cable and the solar
collector must transmit energy without causing sufficient temperature
increase to damage the cable. The shape of the collector must also be
designed to capture the maximum energy without requiring an elaborate
tracking system. The shape of the collector also affects the transfer of
energy to the optical cable through the coupling; some parabolic
collectors are so sharply focused that local melting of the cable cannot be
prevented with any of the available couplings.
199
-------
In addition, the project is collecting and evaluating information on
heat transfer in soils to determine whether projected power inputs from
solar collectors coupled to optical fibers will be capable of achieving
satisfactory heating rates and maximum temperatures in soil.
CONCLUSIONS
There are a number of physical, chemical, and biological processes in
soils that are affected by temperature; this temperature effect can be used
to advantage for improving the efficiency of in situ treatment
technologies, A number of methods for heating soil have been developed or
adapted from other applications; most of these have been tested in
conjunction with one or more in situ treatment technologies. Soil heating
has been used to enhance bioremediation and gas phase removal of organic
compounds. In situ soil heating could potentially be applied to enhance
the performance of other technologies such as chemical treatment but has
not yet been used for this purpose.
ACKNOWLEDGMENTS
The authors wish to thank Robert E. Hinchee and Lawrence A. Smith of
Battelle, Columbus, OH, who contributed to the literature review for this
paper under work assignment 0-09, Contract 68-CO-0003.
REFERENCES
Anonymous. Radio frequency zaps waste sites. Midwest Engineer.
October 1988, p. 4.
Anonymous. Weston to offer in situ radio frequency hazwaste treatment.
WasteTechNews. July 3, 1989.
Baum, R. Low-cost cleanup of toxic petrochemicals. Chemical and
Engineering News. 66(50): 24, 1988.
Baver, L.» Gardner, W. H. and Gardner, W. R. Soil Physics. John Wiley and
Sons, Inc., NY, 1972. pp.273-74.
Brown, H. and Murdoch, L. Feasibility of using solar energy and optical
fibers to heat subsurface soils. Work Assignment 1-17, Contract 68-
C9-0031, U.S. Environmental Protection Agency, Risk Reduction
Engineering Laboratory, Cincinnati, OH, 1990.
Buelt, J. and Westsik, J. In situ vitrification - preliminary results from
the first large scale radioactive test. Pacific Northwest
Laboratories, Richland, Washington, 1987.
Call, F. Soil fumigation. V. Diffusion of ethylene dibromide through
soils. J. Sci. Food and Aar. 8:143-150, 1957.
200
-------
DePauli, D. Field demonstration of in situ soil venting of JP-4 jet fuel
spill site at Hill Air Force Base (Draft). ESL-TR-90-21, Vol. 2,
U.S. Air Force, 1990.
de Percin, P. Demonstration of in-situ steam/hot stripping technology.
in: Proceedings of the 83rd Annual Meeting & Exhibition. Air & Waste
Management Association, Pittsburgh, PA, 1990. Paper/session 90-14.5.
Dev, H., Bridges, J., and Sresty, 6. Decontamination of hazardous waste
substances from spills and uncontrolled waste sites by radio
frequency in situ heating. Jn: Proceedings of the 1984 Hazardous
Materials Spills Conference. Government Institutes Inc., Nashville,
TN, 1984. p. 57.
Dev, H. Radio frequency enhanced in-situ decontamination of soils
contaminated with halogenated hydrocarbons. In: Proceedings of the
Twelfth Annual Research Symposium on Land Disposal, Remedial Action,
Incineration and Treatment of Hazardous Waste (EPA/600/9-86/002),
U.S. Environmental Protection Agency, Cincinnati, OH, 1986. p. 402.
Dev, H., Condorelli, P., Bridges, J., Rogers, C., and Downey, D. In situ
radio frequency heating process for decontamination of soil. In: ACS
Symposium Series No. 338, Solving Hazardous Waste Problems: Learning
From Dioxins. American Chemical Society, New York, New York, 1986
(published 1987). p. 332.
Dev, H., Sresty, G., Bridges, J., and Downey, D. Field test of the radio
frequency in situ soil decontamination process. In: Proceedings of
the 9th National Conference: Superfund '88. The Hazardous Materials
Control Research Institute, Washington, D.C., 1988. p. 498.
Dev, H. and Downey, D. Zapping hazwastes. Civil Engineering. August
1988. p. 43.
Dev, H., Enk, J., Sresty, G., Bridges J., and Downey, D. In situ soil
decontamination by radio frequency heating - field test. ESL-TR-88-
62, U.S. Air Force Engineering & Services Laboratory, Tyndall Air
Force Base, FL, 1989. 175pp.
Dinkel, D. Polyethylene mulches for sweet corn in northern latitudes.
American Society for Horticultural Science, vol 89:40, 1966.
Downing, A. and Truesdale, G. Some factors affecting the rate of solution
of oxygen in water. J. Applied Chem. 5:570-581, 1955.
Elhers, W., Letey, J., Spencer, W. and Farmer, W. Lindane diffusion in
soils. I. Theoretical considerations and mechanism of movement.
Soil Sci. Soc. Amer. Proc. 33:501-504, 1969.
201
-------
Heath, W. In situ heating to destroy and remove organics from soils. The
Department of Energy/Air Force Joint Technology Review Meeting, Soil
and Groundwater Remediation, Atlanta, Georgia, 1990.
Hansen, J., Timmerman, C. and Liikala, S. Status of in situ vitrification
technology: a treatment process for destruction and/or permanent
immobilization. Geosafe Corporation, Kirkland, Washington, 1990.
Hinchee, R. and Smith, L. Innovative in situ soil heating concepts: hot
air injection. Technology review, U.S. Environmental Protection
Agency, Cincinnati, OH, 1990a. 3pp.
Hinchee, R. and Smith, L. Innovative in situ soil heating concepts:
electrical resistance heating. Technology review, U.S. Environmental
Protection Agency, Cincinnati, OH, 1990b. 2pp.
Hinchee, R. and Smith, L. Innovative in situ soil heating concepts: soil
surface modifications. Technology review, U.S. Environmental
Protection Agency, Cincinnati, OH, 1990c. 3pp.
Jackson, D., Weinburg, A., Bisson, D. and Wevill, S. Literature review:
surfactant enhancement of in-situ steam and vacuum reclamation of
hazardous waste sites. Interim report, U.S. Environmental Protection
Agency, Cincinnati, OH, 1990. 54pp.
Johnson, L. and Guffey, F. Contained recovery of oily wastes (CROW). Draft
final report, U.S. Environmental Protection Agency, Cincinnati, OH,
1990. 97pp.
La Mori, P. In-situ hot air/steam extraction of volatile organic
compounds. In: H. Muralidhara (ed.), Solid/Liquid Separation: Waste
Management and Productivity Enhancement. Battelle Press, 1989. p.
386.
Lavy, T. Diffusion of three chloro s-triazines in soil. Weed Sci. 18:53-
56, 1970.
Letey, J. Movement of water through soil as influenced by osmotic pressure
and temperature gradients. Hilgardia, 39(14):405-417, 1968.
Lord, A., Koerner, R., Murphy, V. and Brugger, J. Laboratory studies of
vacuum-assisted steam stripping of organic contaminants from soil.
In: Proceedings of the Fourteenth Annual Research Symposium on Land
Disposal, Remedial Action, Incineration and Treatment of Hazardous
Waste (EPA/600/9-88/021), U.S. Environmental Protection Agency,
Cincinnati, OH, 1988. p. 65.
Lord, A., Koerner, R., and Hull ings, D. and Brugger, J. Laboratory studies
of vacuum-assisted steam stripping of organic contaminants from soil.
In: Proceedings of the Fifteenth Annual Research Symposium on
Remedial Action, Treatment and Disposal of Hazardous Waste
202
-------
(EPA/600/9-90/006), U.S. Environmental Protection Agency, Cincinnati,
OH, 1989 (published 1990), p. 124.
Lord, A., Hull ings, D., Koerner, R. and Brugger, J. Vacuum-assisted steam
stripping to remove pollutants from contaminated soil - a laboratory
study, in: Proceedings of the Sixteenth Annual Hazardous Waste
Symposium on Remedial Action, Treatment and Disposal of Hazardous
Waste (EPA/600/9-90/037), U.S. Environmental Protection Agency,
Cincinnati, OH, 1990. p. 377.
Maron, S. and Prutton, C. Principles of Physical Chemistry, 4th Edition.
Collier-MacmiHan Canada Ltd., Toronto, Ontario, 1965. 886p.
Nicholson, F. Permafrost modification by changing the natural energy
budget. In: Proceedings of the Third International Conference on
Permafrost, Alberta, Edmonton, Canada, 1978.
Nunno, T., Hyman, J., Spawn, P., Healy, J., Spears, C. and Brown, M.
Assessment of international technologies for Superfund applications -
technology identification and selection. EPA/600/2-89/017, U.S.
Environmental Protection Agency, Cincinnati, OH, 1989.
Sresty, 6., Dev, H., Gordon, S. and Chang, J. Methodology for minimizing
emissions by remediation of environmental samples containing wood
preserving chemicals. Draft progress report, U.S. Environmental
Protection Agency, Cincinnati, OH, 1990. 69 pp.
Thibodeaux, L. Chemodynamics; Environmental Movement of Chemicals in Air,
Water, and Soil. John Wiley & Sons, NY, 1979. p. 211.
Tortora, 6., Funke, B. and Case, C. Microbiology: an introduction. The
Benjaminim/Cummings Publishing Company, Inc., NY, 1989. pp. 145-146.
Udell, K. and Stewart, L. Field study of in-situ steam injection and
vacuum extraction for recovery of volatile organic solvents. UCB-
SEEHRL Report No. 89-2, University of California, Berkeley, CA,
1989.
U.S. EPA. The Superfund innovative technology evaluation program:
technology profiles. EPA/540/5-89/013, U.S. Environmental Protection
Agency, Cincinnati, OH, 1989. 119 pp.
U.S. EPA. Handbook on in situ treatment of hazardous waste-contaminated
soils. EPA/540/2-90/002, U.S. Environmental Protection Agency,
Cincinnati, OH, 1990a. 168pp.
U.S. EPA. SITE demonstration bulletin: in-situ steam/hot-air soil
stripping. EPA/540/M5-90/003, U.S. Environmental Protection Agency,
Cincinnati, OH, 1990b. 2pp.
203
-------
THE UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
MUNICIPAL WASTE COMBUSTION RESIDUE
SOLIDIFICATION/STABILIZATION PROGRAM
Carlton C. Wiles
Risk Reduction Engineering Laboratory
United States Environmental Protection Agency
Cincinnati, Ohio 45268
David S. Kosson
Rutgers, The State University of New Jersey
Piscataway, New Jersey
Teresa Holmes
United States Army Corps of Engineers
Waterways Experiment Station
Vicksburg, Mississippi
ABSTRACT
Vendors of solidification/stabilization (S/S) and other technologies are
cooperating with the U.S. Environmental Protection Agency's (U.S. EPA's)
Office of Research and Development (ORD), Risk Reduction Engineering
Laboratory to demonstrate and evaluate the performance of the technologies to
treat residues from the combustion of municipal solid waste (MSW).
Solidification/Stabilization is being emphasized in the current program. This
technology may enhance the environmental performance of the residues when
disposed in the land, when used as road bed aggregate, as building blocks, and
in the marine environment as reefs or shore erosion control barriers.
The program includes four S/S process types: cement, silicate, cement kiln
dust and a phosphate based process. Residue types being evaluated are fly
ash, bottom ash and combined residues. An array of chemical leaching tests and
physical tests are being conducted to characterize the untreated and treated
residues. This paper discusses program design, status and preliminary
results.
The S/S evaluation program is the first part of ORD's Municipal Solid
Waste Innovative Technology Evaluation (MITE) program.
204
-------
INTRODUCTION
During the past two years there has been a significant concern expressed
about the management of the residues from the combustion of municipal solid
waste. Much of this concern has centered on the fact that when the
residues are subjected to the Extraction Procedure for Toxicity (EP tox) and
the Toxicity Characteristics Leaching Procedure (TCLP) they will fail for lead
and cadmium a significant portion of the time. This occurs more often for the
fly ash, less for the combined fly ash and bottom ash, and least often for the
bottom ash alone. Because of this, a controversy exists as to whether or not
the residues should be considered and regulated as a hazardous waste or
exempted because they originated from burning municipal solid waste. Several
states are requiring that these residues be disposed into landfills with
designs and operating procedures as, or more, stringent than those for
hazardous waste. Municipal Waste Combustion (MWC) ash characteristics are
extremely variable as is the leachate from these ashes. Ranges of metal
concentrations observed in bottom and fly ashes from many sources are
presented in Table 1(1). Detailed descriptions of the chemical and physical
characteristics of MWC residues are available"'3' ' .
TABLE 1. RANGES OF TOTAL AND LEACHABLE METALS IN UNITED STATES MSW COMBUSTOR
ASH AS DETERMINED BY RESEARCHERS05
Com- Bottom Ash
pound
mq/kq
Pb
Cd
As
Cr
Ba
Ni
Cu
ND =
31 -
0.81
0.8
13 -
47 -
ND(1
40 -
36,600
- 100
- 50
1,500
2000
.5) - 12,910
10,700
Not Detectable; ()
Bottom Ash
Leachate
mq/1
0.02 - 34
0.018 - 3.94
ND(O.OOl) - 0.122
ND(0.007) - 0.46
0.27 - 6.3
0.241 - 2.03
0.039 - 1.19
» Detection Limit
Fly Ash
mq/kq
2.0 - 26,000
5 - 2,210
4.8 - 750
21 - 1,900
88-9000
ND(1.5) - 3,600
187 - 2,300
Fly Ash
Leachate
mq/1
0.019
0.025
ND(0.
0.006
0.67
0.09
0.033
- 53.35
- 100
001 - 0.858)
- 0.135
- 22.8
- 2.90
- 10.6
Because of the growing concern about the residues and anticipating the
need for appropriate treatment techniques, the Office of Research and
Development designed and implemented a program to evaluate the use of
solidification/stabilization technologies for treating the residues. The
program was formally announced on September 19, 1989. Originally known as the
U.S. EPA MWC Ash Solidification/Stabilization Evaluation Program, it is now
the Municipal Innovative Technology Evaluation program (MITE). This paper
presents the design and status of the current program.
205
-------
THE MITE PROGRAM
The MITE program is an Office of Research and Development (ORD) program
designed to conduct demonstrations of technologies for managing municipal
solid waste. The objective is to encourage development and use of innovative
technology for municipal solid waste management. The program is patterned
after the Superfund Innovative Technology Evaluation program (SITE). It is,
therefore, a cooperative program in which the technology developer and/or
vendor pays the cost of conducting the demonstration. U.S. EPA pays the cost
of testing and evaluation, including analytical cost. U.S. EPA will report
the results of the evaluations in an unbiased manner, thus providing a means
for assisting municipalities and others to better evaluate and select
technologies more appropriate for their given situation.
The current program is demonstrating and evaluating alternatives for the
treatment of residues from the combustion of municipal waste. While it is
uncertain if treatment will be required prior to disposal, it is most likely
that treatment will be necessary for any utilization option. Solidification/-
Stabilization technology was selected for initial evaluations based upon
experience and knowledge of the technology for treating hazardous waste and
experimental studies on solidifying municipal waste combustion (HWC)
residues . Solidification/Stabilization, in general terms, is a technology
where one uses additives or processes to transform a waste into a more
manageable form or less toxic form by physically and/or chemically
immobilizing the waste constituents. Most commonly used additives include
combinations of hydraulic cements, lime, pozzolans, gypsum, silicates and
similar materials. Other types of binders, such as epoxies, polyesters,
asphalts, etc. have also been used, but not routinely. More detailed
descriptions of S/S technology are availablem. The program objective is to
provide a credible data base on the effectiveness of S/S technology for
treating the residues.
Preliminary design of this program was completed by the U.S. EPA.
Because U.S. EPA believed it important to have results completely unbiased and
as scientifically credible as possible, a panel of international experts was
assembled to provide oversight to the program. This Technical Advisory
Panel (TAP) consists of experts from academia, industry, state and federal
governments, and environmental groups.
PROGRAM ORGANIZATION AND DESIGN
Organization - The program involves the participation of several
different organizations with separate roles. The Risk Reduction Engineering
Laboratory (RREL) is managing and directing the program. The TAP is providing
valuable peer review, oversight and technical design. This service is donated.
Staff at the U.S. Army Corps of Engineers Waterways Experiment Station (WES) are
coordinating and observing the demonstrations at WES facilities located in
Vicksburg, Mississippi. WES 1s also responsible for performing the physical
testing and some of the extraction/leaching tests. A laboratory experienced
1n MWC residue analysis 1s performing the majority of the analytical work.
Specialized analyses, testing and modeling is being performed by the
University of Illinois and the Netherlands Energy Research Center. Rutgers
206
-------
University in conjunction with the New Jersey Institute of Technology is
assisting in the coordination of the various activities and participants.
Vendors are participating by providing valuable time and money.
Tests and Analyses - The program was conceived by U.S. EPA and the basic
design was based on the testing and evaluations performed on hazardous and
other waste treated by solidification/stabilization technologies in various
research and evaluation programs of U.S. EPA. At the request of U.S. EPA, the
TAP reviewed and modified this preliminary design. The tests and analytical
protocols included in the program are provided in Tables 2, 3, 4, 5, 6 and 7.
The purpose for conducting the test and analysis listed is also included.
Methods listed in the Tables are either approved U.S. EPA or ASTM methods.
TABLE 2. CHEMICAL ANALYSIS PERFORMED ON TREATED AND UNTREATED ASH
Assay Met hod Purpose
Total Extractable Metals 3050, 6010 See Metals Analysis List
(Table 6)
Dioxins/Furans 8280 Community Concern
(Untreated Only)
pH, Anions, Total 9045, 300.0, Salts and Ionic Species
Available Dissolved 160.1, 350.2
Solids, and Ammonia
Loss on Ignition 209D Residual Organic Matter
(typ. 2-5%) and Water of
Hydration
Chemical Oxygen 508A Reduced Inorganic and
Demand Organic Matter
Total Organic Carbon Residual Organic Matter
Summary of Leach Tests - For information, following is a summary of
three leach tests used:
• Availability (static pH) - The availability leach test was
developed by the Netherlands Energy Research Foundation to
quantify the maximum amount of a species which could be released
to the environment under assumed worst case conditions during the
lifetime of the material. The test does not provide information
on release rate or anticipated natural leachate concentrations.
The test is carried out using two serial extractions of the
material to be tested. The sample to be tested is crushed to less
than 300 microns and extracted at constant pH 7 for the first
extraction and pH 4 for the second extraction at a liquid to solid
ratio of 100:1 for each step. Extraction pH is maintained through
use of a pH controller delivering dilute nitric acid.
207
-------
• Distilled Water Leach Test - The distilled water leach test is
designed to estimate species release during contact with
uncontaminated natural waters. The test is carried out by
serially contacting material crushed to less than 2 mm with
distilled water four times. The liquid to solid ratio for each
contacting is 10:1. The pH of the system is established for the
material being tested (typically alkaline for solidified MWC
residue). The first and second extracts (extracts 1 and 2) were
combined for analysis, as were extracts 3 and 4, to minimize
analytical costs.
t Monolith Leach Test - The monolith leach test was carried out to
assess the rate of species release from solidified/stabilized MWC
residues. A cylindrical sample (4 cm dia by 4 cm) is contacted
with distilled water for up to 64 days. Contacting water is
replaced at 1, 2, 4, 8, 16, 32 and 64 days and is analyzed.
Modeling of release data will establish effective diffusion
coefficients for estimating long term species release rates. This
leach test is a modified version of ANSI 16.1.
Ash Types Tested - Residue selected for testing was limited to that collected
from a modern state-of-art waste to energy facility (i.e., high burn out, lime
scrubber with fabric filter, etc.). There were several reasons for limiting
the number of residues included in the program. The prime objective is to
evaluate solidification/stabilization for treating the residues, rather than
determine how characteristics of different residues may affect the performance
of the technology. In addition the apparent variability of MWC
residues is becoming less of an issue, especially with the newer combustion
facilities. Proper sampling and analysis, changes in air pollution controls
and similar factors will play more important roles in the variability of
residues. The program currently includes four different S/S process types
plus one control. Because of the extensive list of tests being performed, the
analytical cost for the program is the major U.S. EPA expense. For each
additional source of residue added these costs must be duplicated. This would
have reduced the number of processes which could be evaluated to an
unacceptable number. The program is also developing and evaluating testing
protocols that can be used to evaluate selected S/S processes on different
residues if required in the future.
These considerations quickly led to the conclusion that the program would
test the residue from only one facility. The residue types are the fly ash
(including the scrubber residue), the bottom ash and the combined ash. The
MWC facility has the following process sequence: (i) primary combustor with
vibratory grates, (ii) secondary combustion chamber, (iii) boiler and
economizer (iv) dry scrubber with lime, and (v) particulate recovery using
baghouses (fabric filters). Bottom ash sampled was quenched after exiting
from the combustion grates. Fly ash sampled was mixed residuals from the
scrubber and baghouses. The fly ash was screened to pass a 0.5 inch
square mesh. The bottom ash and combined ash were screened to pass a 2 inch
square mesh at the MWC facility. Materials not passing through the 2 inch
mesh were rejected. After shipment to the WES, each ash type was dried to
less than 10% moisture, crushed and screened to pass a 0.5 inch mesh
(nominally 3/8 inch after clogging), and homogenized.
208
-------
TABLE 3. PHYSICAL TESTS CONDUCTED ON TREATED AND UNTREATED ASH
Physical Test
Purpose
Moisture Content
Loss on Ignition
Modified Proctor Density
Bulk Density
Particle Size Distribution
Cone Penetrometer
Pozzolanic Activity*
Porosity/Surface Area
Permeability
Unconfined Compressive
Strength (UCS)
UCS after Immersion
Freeze/Thaw**
Wet/Dry**
Useful general data
Residual/Organic Matter and
Hydrated Water
Compressibility
Volume and Similar Physical Changes
Potential Use as Aggregate
Curing Rate and Hardness
Untreated S/S Potential
Potential for Liquid-Solid Contact
and Diffusion Effects
Resistance to H20 Transmission;
Assist in Determining Contaminant
Release Mechanisms
Load Bearing Capacity
Hydration Effects and Swelling
Physical Weathering Effects
Physical Weathering Effects
* Untreated Ash Only
** Treated Ash Only
Processes Selected - Process types selected in the program are cement
based, silicate based, cement kiln dust and phosphate based. A non-vendor
cement process is being performed by experienced staff of WES and U.S. EPA in
Vicksburg, MS.
Process selection was competitive based upon evaluation of proposals
submitted by parties interested in participating. A formal Request For
Participation was issued by U.S. EPA which provided information required
to respond. Under direction of U.S. EPA, the TAP developed evaluation
criteria which was used to make final selections.
209
-------
TABLE 4. LEACHING TESTS FOR TREATED AND UNTREATED ASH
Leach Test
Purpose
TCLP (1 extract)
Distilled Water Leach Test
(4 extracts)
Acid Neutralization
Capacity (10 extracts)
Monolith Leach Test
(7 extracts)
Static pH e pH » 4.0
with HN03 Liquid:Solid
Ratio is 100:1
Regulatory Leach Test
Extended Extraction in a Well-Mixed
System without Acid
Buffering Capacity of Solid and pH
Dependence of Metals Release
Estimate Potential Release Rates
Through Diffusion
Total Species Available for Release
Under "Worst Case" Scenario
TABLE 5. CHEMICAL ANALYSIS PERFORMED ON LEACH TEST EXTRACTS
Assay
Method
Purpose
Metals
Chemical Oxygen
Demand (COD)
Total Suspended Solids
Total Dissolved Solids
pH
3020
508A
160.2*
160.1
150.1
See Metals Analysis List
(Table 6)
Surrogate for Leachable
Organic Species
Physical Erosion of Solid
Leachable Total Salts
* Monolith leach test only (ANSI 16.1)
210
-------
TABLE 6. LIST OF METALS SUBJECTED TO ANALYSIS
Metal
Untreated and Treated
Ash (Solid)
Neutron
ICP or AA Activation
Aluminum
Antimony
Arsenic
Barium
Beryl 1 i urn
Boron
Cadmium
Calcium
Chromium
Cobalt
Copper
Iron
Lead
Lithium
Potassium
Magnesium
Manganese
Mercury
Molybdenum
Nickel
Selenium
Sodium
Silicon
Silver
Strontium
Thorium
Tin
Titanium
Vanadium
Zinc
X
X
X
X
X
. --
X
--
X
X
X
--
X
X
X
_-
.-
X
__
__
X
__
__
X
TABLE 7. ADDITIONAL METALS ANALYSIS
Untreated
Metal
Cesium
and
Dysprosium
Gallium
Hafnium
Indium
Rubidium
Scandium
Uranium
X X
X
X
X
- ._
X
X
X
X
X
X
%
X
X
X
X
X
X
X
X
X
Extracts
ICP or AA
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
—
X
X
X
X
X
X
X
X X
USING NEUTRON ACTIVATION
Treated Ash (Solid)
Neutron
Activation Only
X
X
X
X
X
X
X
X
211
-------
Twenty-one responses were received and evaluated. The responses were
divided into 11 S/S processes, 6 vitrification processes and 4 other
miscellaneous processes. Based upon the evaluation criteria, the S/S process
proposals were judged to be superior. In order not to select similar S/S
process types (e.g., two cement based) with the limited resources available,
the decision was made to select the best proposal out of the different types
available. The vitrification process proposals were generally incomplete and
failed to address some major issues. This, in conjunction with the potential
high quantities of residues required for most of these processes, resulted in
the decision not to select one for evaluation. Alternatives for evaluating
vitrification processes are being pursued. Proposals in the other
miscellaneous category were not acceptable and were rejected.
During the request for participation, evaluation and selection process,
provisions were made for maintaining confidentiality of information so marked
by the responders.
Following is a brief description of each of the processes selected.
Cement Based Process - This process involves the addition of polymeric
adsorbents to a slurry of MWC ash prior to the addition of portland cement.
The final product is soil-like rather than monolithic.
Silicate based process - This is a patented process using soluble
silicates as an additive with cement. The additives are used to promote
several types of reactions with the polyvalent metal present to produce
insoluble metal compounds, gel structures, and promote hydrolysis, hydration
and neutralization reactions. The process immobilizes heavy metals through
reactions involving complex silicates. The final product is clay-like
material.
. CKD process - This is a patented process involving mixing the MWC ashes
with quality controlled waste pozzolans and water. Good quality control on
the reagents is required because they are secondary materials derived from
processing other materials. Therefore, the pozzplanic characteristics
critical to the process are subject to change. The finished product is
similar to moist soil, but hardens to a concrete-like mass within several
days.
Phosphate process - A water soluble phosphate is used in this patented
process to convert lead and cadmium to insoluble forms. The process is
designed such that fly ash is mixed with lime, then this material can be mixed
with the bottom ash and the mixture treated with a source of water soluble
phosphate. The process does not alter the physical state of the ash.
DEMONSTRATIONS
Process - The procedures for conducting the demonstrations were
established so that the process vendors could review data from
characterizations of the various ash prior to the demonstration. Samples of
the ashes were also furnished to the vendors so that they have the opportunity
to pretest their process prior to the demonstration. This permitted them to
make modifications if desired. Vendors were responsible for providing any
specialized equipment or ingredients required. Each agreed to permit
212
-------
observation by U.S. EPA selected observers if it was necessary to conduct the
demonstration at the vendor's facilities. Otherwise the demonstrations were
to be conducted at a U.S. EPA selected facility and observed by U.S. EPA
designated staff.
During the process demonstration, each vendor was requested to carry out
three replicate batches for each ash type. A total of between 50 and 100
gallons of each ash type is being treated for each process. Numerous molds
and samples are prepared from these batches. All molds and sample containers
are provided by WES and U.S. EPA. Each vendor provides enough process
additives for analysis and archiving. Most equipment and laboratory
facilities required for the demonstrations are provided by WES.
Scale - The processes are being demonstrated at bench scale. Reasons for
this include the technologies being tested, resources required for full scale
demonstrations and the desire to include as many different processes as
possible within available resources. The program plan was to conduct a full
scale field demonstration of a selected process if deemed necessary. Because
of the nature of S/S technologies, U.S. EPA and the TAP believed that bench
scale demonstrations were adequate to prove if the technology is an effective
treatment for MWC residues. Sufficient experience is available for conducting
the engineering and design required for scaling to a specific situation.
Furthermore, the bench scale permitted much more detailed testing to be
completed and thus more exploration of the basic mechanisms involved in the
process. This in turn will assist in the determination of expected long-term
behavior. A drawback with this scale however, is the difficulty in sampling
and variability associated with bottom ashes.
Status - The S/S process demonstrations have been completed. Because of
the nature of the S/S process, sample curing requirements (i.e., 28 days) and
other specific test requirements, the physical testing, chemical testing and
analytical procedures were delayed and just recently completed. The very
large volume of data generated is still being compiled, organized and
interpreted. The final report is expected by the end of June 1991.
Future MITE Demonstrations - Future MITE demonstration candidates have
been solicited by notice in the Commerce Business Daily, through appropriate
MSW trade organizations, interested developers and similar means. At this
time, emphasis for these demonstrations is expected to be on processes for
recovering marketable products from the MSW stream. Additional industry and
state cooperative evaluations of MWC ash treatment and/or utilization
processes are planned under separate programs.
RESULTS
Complete results from the various physical and chemical tests are not
available at this writing. Samples of the data and results being generated
are provided in Tables 8 through 11,
Strengths after Water Immersion - One objective of conducting the leach
testing and physical tests was to evaluate the effectiveness of the various
processes to retain physical durability and the metals of concern when exposed
to different stresses. One potentially valuable observation is how well the
213
-------
physical structure can be expected to withstand degradation under exposure to
wet conditions (e.g., marine environment, road base, construction blocks,
etc). If one assumes that physical durability will improve the capability of
the treated form to resist leaching, then strength before and after immersion
will provide insight about this characteristic. Table 8 compares the
unconfined compressive strength (UCS) before and after immersion in a water
solution of the residues (i.e., bottom, air pollution control (ARC), and
combined) treated with the various processes.
Compared to the other processes, the control process (i.e., cement)
showed somewhat better effectiveness in retaining physical strength after
immersion. Also of interest is the general trend that the ARC residues (as
compared to bottom and combined) appeared to be more difficult to treat as
measured by lower strengths. While this confirms observations from other
researchers, one must note that UCS measurements results often have wide
ranges. However, data presented in Table 8 had reasonable variations with
exception of WES combined ash after 28 day cure time and the bottom ash after
28 day immersion.
In comparing strengths of solidified waste forms, one must note that
there is little scientific evidence that directly relates increased strength
with decreased release rates of pollutants of environmental concern. Also
increased strength may not be important in the case of placement in a
landfill. In such cases strength concerns deal with sufficient load bearing
capacity necessary to support equipment and landfill covers, etc. In some
cases, these may be as low as 12 to 15 psi or lower. These relatively low
strengths often can be easily achieved through routine compaction.
Additionally many HWC residues contain sufficient pozzolanic properties which
when combined with the excess lime from wet scrubbers will result in some
hardening of the ashes without additional additives .
Strengths greater than those shown in Table 8 would be required for
potential uses such as shore erosion control and some construction
applications. Higher strengths have been routinely achieved10"
CONTAMINANT RELEASE AND COMPARISON OF LEACH TEST
One of the most critical concerns associated with management of MWC
residues is the potential environmental damage from release of heavy metals
such as lead and cadmium. Several leach tests (or extractions) were performed
to help assess the expected release under different conditions. Tables 9, 10,
and 11 summarize some preliminary results and provide examples of the types of
comparison one might make when all the results are available. Table 9
compares the concentrations of constituents indicated in the leachates
(extracts) when the WES control treated combined ash was subjected to the
Toxicity Characteristics Leaching Procedure (TCLP) and the distilled water
leach (DL) test. Also provided are the TCLP regulatory limits. Note in all
cases, the concentrations of the TCLP regulatory metals of concern were below
the regulatory limits. In the case of the distilled water leach test, barium
leached at levels above the TCLP limit. The principal species leached using
the distilled water leach test were Ca, K, Na, and Cl. Of interest to note is
that the total dissolved solids (TDS) were approximately 2300 mg/1,
significantly greater than the primary drinking water standard of 500 mg/1.
214
-------
TABLE 8. COMPARISON OF UNCONFINED COMPRESSIVE STRENGTH (UCS) OF STABILIZED
HWC RESIDUES (after 28 days of curing, and after 28 day immersion in water
solution (0,10 g lime/L distilled water) subsequent to 28 days of curing.
Curing was carried out at 20° C and 98 percent relative humidity)
Process
Bottom Ash
UCS fpsiq)
ARC Residue
UCS (psig)
Combined Ash
UCS fpsig)
UntreatedP
28 day cure 7
28 day immersion CD
after 28 day cure
WES Control
28 day cure 1152
28 day immersion 1075
after 28 day cure
Vendor 1
28 day cure 1081
28 day immersion 4324
after 28 day cure
Vendor 2
28 day cure 150
28 day immersion 149
after 28 day cure
Vendor 3
28 day cure 350
28 day immersion CD
after 28 day cure
Vendor 4
28 day cure 55
28 day immersion 197
after 28 day cure
5
CD
555
434
136
2245
175
CD
154
CD
unconsolidated6
unconsolidated
10
CD
441-
531
252
29
228
152
275
18
83
176
1 - Untreated MWC residue prepared at optimum moisture content and compacted
with Modified Proctor Compaction Effort.
2 - CD - Cube Disintegrated - Cube disintegrated from free
standing monolith to unconsolidated form.
3 - Wide variation in data (Coefficient of Variation = 0.54)
4 - Wide variation in data (Coefficient of Variation = 1.3)
5 - Cubes had spongy consistency after immersion.
6 - Unconsolidated - Following treatment product remained
unconsolidated.
215
-------
Table 9. COMPARISON OF LABORATORY EXTRACT CONCENTRATIONS FOR
LEACHING TESTS ON TREATED MWC RESIDUES.
PROCESS: WES Control
ASH TYPE: Combined Ash
Metals (ug/i)
Aluminum
Antimony
Arsenic
Barium
Beryllium
Boron
Cadmium
Calcium
Chromium
Cobalt
Copper
Iron
Lead
Lithium
Magnesium
Manganese
Mercury
Molybdenum
Nickel
Potassium
Selenium
Silicon
Silver
Sodium
Strontium
Tin
Titanium
Vanadium
Zinc
DL E(1+2)
4566.7
40.4 U
10.0 U
2626.7
2.5
NA
1.0 U
226666.7
12.0 B
11.3
253.3
19.7
28.7
56,7
13.0
3.7 U
0.2 U
63.0
20.0 U
136666,7
5.0 U
1900.0
7.4 U
120000.0
2700.0
55.0 U
12.0 U
10.8
36.0
DL E(3+4)
8600.0
40.4 U
10.0 U
1200.0
2.5
NA
1.0 U
136666.7
13.7
7.0 U
75.3
12.7
12,7
15.3 B
8.0
3.7 U
0.2 B
43.0
20.0 U
3200.0
5.0 U
1466.7
7.4 U
2666.7
546.7
S5.0 U
12.0 U
8.9
37.3
TCLP (E2)
9266.7
40.4 U
13.0 B
500.0
3.0
NA
620.0
1966666.7
23.8
60.7
2890.0
3066.7
1333.3
66.7
105666.7
7466.7
0.7 A
18.3 B
293.3
66000.0
5.0 U
30666.7
7.4 U
93333.3
2366.7
55.0 U
29.0 A
7.4
30333.3
TCLP
REG. LIMIT
50
1000
1000
5000
200
5000
U=undetected, A=U(1 of 3), B=U{2 of 3), NA= not analyzed
216
-------
Table 9. COMPARISON OF LABORATORY EXTRACT CONCENTRATIONS FOR
LEACHING TESTS ON TREATED MWC RESIDUES (continued).
PROCESS: WES Control
ASH TYPE: Combined Ash
Anlons (mg/l)
Bromide
Fluoride
Chloride
Sulfate
DL E{1+2)
13.43
NA
881.33
5.32
Nitrogen Species:
Nitrite
Nitrate
Ammonia
Phosphorous
Other Assays
pH
IDS
COD
TOC
NA
3.61
NA
1.00 U
[mg/l)
11.59
2313.33
58.60
23.90
DL E{3+4)
1.34
NA
75.17
7.26
NA
1.13
NA
1.00 U
10,88
507.33
15.60
3.83
TCLP (E2>
NA
NA
NA
NA
NA
NA
NA
NA
11.21
NA
NA
NA
U=undetected, A=U(1 of 3), B=U(2 of 3), NA= not analyzed
DL E(1+2)=distllled water leach test, extracts 1 and 2 combined for analysis
DL E(3+4)=dIstINed water leach test, extracts 3 and 4 combined for analysis
TCLP {E2)=TCLP leach test, extractant 2
217
-------
TABLE 10. COMPARISON OF SPECIES RELEASE FOR LEACHING TESTS ON TREATED MWC
RESIDUES (mg released/kg treated product, dry solid).
PROCESS: WES Control
ASH TYPE: Combined Ash
DL E(1..4)
Metals (mg/kq ds)
Aluminum
Antimony
Arsenic
Barium
Beryllium
Boron
Cadmium
Calcium
Chromium
Cobalt
Copper
iron
Lead
Lithium
Magnesium
Manganese
Mercury
Molybdenum
Nickel
Potassium
Selenium
Silicon
Silver
Sodium
Strontium
Tin
Titanium
Vanadium
Zinc
286.6
1.8 U
0.4 U
83.1
0.1
NA
0.0 U
7905.9
0.6
0.4 U
7.2
0.7
0.9
1.6 B
0.5
0.2 U
0.0 B
2.3
0.9 U
3043.7
0.2 U
73.3
0.3 U
2669.6
70.6
2.4 U
0.5 U
0.4
1.6
TCLP (E2)
205.7
0.9 U
0.3 B
11.1
0.1
NA
13.8
43840.3
0.5
1.3
64.4
68.1
29.6
1.5
2354.0
166.2
0.0 A
0.4 B
6.5
1471.3
0.1 U
680.8
0.2 U
2081.3
52.7
1.2 U
0.6 A
0.2
675.0
Avail.
4255.1
8.9 B
3.5
155.0
0.6
NA
20.4
103279.8
4.8
3.5
281.0
152.7
281.0
3.8 A
3753.3
302.6
0.1
9.0
13.7
4766.7
1.1 U
7203.8
1,6 U
4405.5
161.5
11.9 U
2.6 U
-1.4 U
1750.8
Total
(SW846)
27054.3
NA
25.6
750.1
15.5
NA
25.3
NA
84.7
NA
1107.1
NA
1263.9
11.5
NA
NA
7.1
NA
114.5
NA
5.4 U
NA
2.3
NA
NA
155.2
NA
NA
3102.7
Total
(NAA)
49089.9
243.9
NA
NA
NA
NA
34.1
142810.3
465.8
36.6
1573.4
92392.4
NA
NA
NA
1609.9
NA
NA
NA
NA
0.8
NA
6.8
26315.5
NA
NA
5414.1
41.7
4830.7
Usundeteeted, A=U(1 of 3), B=U(2 of 3), NA= not analyzed
218
-------
TABLE 10, COMPARISON OF SPECIES RELEASE FOR LEACHING TESTS ON TREATED MWC
RESIDUES (mg released/kg treated product, dry solid) (continued).
PROCESS: WES Control
ASH TYPE: Combined Ash
DL E(1..4)
Anions (mg/kg ds)
Bromide
Fluoride
Chloride
Sulfats
321.6
NA
20814.0
274.2
Nitrogen Species:
Nitrite
Nitrate
Ammonia
Phosphorous
Other Assays
pH (S.U.)
IDS
CCD
TOG
NA
103.1
NA
43.4 U
(mg/kg ds)
11.59
61382.66
2371.13
603.51
TCLP (E2)
NA
NA
NA
NA
NA
NA
NA
NA
11.21
NA
NA
NA
Avail.
3584.2 U
NA
31619.7
29620.8
NA
NA
NA
215.3 U
4.00
NA
NA
NA
Total
(SW846)
151.5
NA
15522.1
125.2
NA
3.4
3.9
0.1 A
11.6
23467.1
4114i.O
6576.4
Total
(NAA)
736.2
NA
16543.6
NA
NA
NA
NA
NA
NA
NA
NA
NA
U=undetected, A=U(1 of 3), B=U(2 of 3), NA= not analyzed
DL E(1..4)=dlsti!Ied water leach test, extracts 1, 2, 3 and 4
TCLP (E2)=TCLP, extractant 2
Avail.=availabllity leach test
Total (SW846)=total analysis as per USEPA methods SW-846
Total (NAA)=total analysis by neutron activation
219
-------
TABLE 11. COMPARISON OF RELEASE FOR UNTREATED AND TREATED
RESIDUES CORRECTED FOR PROCESS DILUTION.
PROCESS: WES Control
ASH TYPE: Combined Ash
TCLP
Untreated
Metals (mg/kg ash, da)
Aluminum
Antimony
Arsenic
Barium
Beryllium
Boron
Cadmium
Calcium
Chromium
Cobalt
Copper
Iron
Lead
Lithium
Magnesium
Manganese
Mercury
Molybdenum
Nickel
Potassium
Selenium
Silicon
Silver
Sodium
Strontium
Tin
Titanium
Vanadium
Zinc
2091.2
0.6 A
0,3 A
19.9
0.0 U
34.0
11.1
39589.1
1.5
2.3
62.6
715.9
47,3
1.4
2038.7
208.2
0.0
0.3
10.6
2557.6
0.4 U
725.4
0.1 U
2792.8
60,5
0.4 U
1.0
0.1 U
858.2
Treated
246.8
1.1 U
0.3 B
13.3
0.1
NA
16.6
52608.3
0.6
1.6
77.3
81.7
35.6
1.8
2824.8
199.4
0.0 A
0.5 B
7.8
1765.6
0.1 U
817.0
0.2 U
2497.5
63.2
1.5 U
0.8 A
0,2
810.0
DL E(1.
Untreated
2425.7
1,0
0.1
38.2
0.0
1.0
0.0
10676.3
0.3
0.2
12.9
1.2
0.2
1.1
2.7
0.1
0.0
0.9
0.5
4323.8
0.1
14.7
0.1
4463.1
70.7
0.7
1.8
0.1
1.1
-4)
Treated
344.0
2.1 U
0.5 U
99.7
0.1
NA
0.1 U
9487.0
0.7
0.5 U
8.6
0.8
1.1
1.9 B
0.5
0.2 U
0.0 B
2.8
1.0 U
3652.4
0,3 U
87,9
0.4 U
3203.6
84.8
2.9 U
0.6 U
0.5
1.9
U=undetected, A=U(1 of 3), B=U(2 of 3), NA= not analyzed
220
-------
TABLE 11. COMPARISON OF SPECIES RELEASE FOR UNTREATED AND TREATED
RESIDUES CORRECTED FOR PROCESS DILUTION (continued).
PROCESS: WES Control
ASH TYPE: Combined Ash
TCLP
Untreated Treated
Anions (mg/kg ash, ds)
Bromide
Fluoride
Chloride
Sulfate
299.7 NA
7.6 NA
14705.1 NA
14759.7 NA
Nitrogen Species:
Nitrite
Nitrate
Ammonia
Phosphorous
Other Assays
pH (S.U.)
TDS
OCD
TDC
0.5 NA
1748.4 NA
NA NA
26.5 NA
(mg/kg ash, ds)
5.19 11.21
NA NA
NA NA
NA NA
DL E(1..4)
Untreated Treated
305.2 385.9
4.1 NA
28511.5 24976.8
2153.6 329.0
0.7 NA
38.1 123.7
NA NA
3.0 52.1 U
10.97 11.59
59900.0 73659.2
2489.4 2845.4
648.3 724.2
U=undetected, A=U(1 of 3), B=U(2 of 3), NA= not analyzed
DL E(1..4)=distil!ed water leach test, extracts 1, 2, 3 and 4
TCLP (E2)=TCLP, extractant 2
221
-------
Table 10 compares species release for several leaching tests and total
concentrations as determined by SW 846 and Neutron Activation Analysis (NAA).
Values indicated are based on dry weight of treated product. Several points
of interest are:
• total analysis of species by SW 846 was poor relative to those
determined by neutron activation. Compared to NAA, Method SW 846
recovered 55% of Al, 74% of Cd, 18% of Cr, 70% of Cu and 64% of
Zn.
Following are observations of the relative amounts of the indicated
species that leached using the indicated leach test when compared to either
the NAA or SW 846 analysis:
• Using the availability leach test:
•• less than 1% of Ca, Fe and Ti leached,
•• between 1% and 5% of Sb, Be, and V leached,
•• between 5% and 10% of Al, Co, and Sn leached,
•• between 10% and 20% of As, Cu, Mn, Ni, and Na leached, and
•• greater than 20% of the following leached (actual numbers
are indicated), Ba(21%), Cd(60%), Ca(72%), Pb(22%), Zi(33%),
and Zn(36%).
• Using the TCLP:
•• less than 1% of Al, Be, Cr, Fe and V leached,
•• between 1% and 5% of As, Ba, Co, Cu, and Pb leached,
•• between 5% and 10% of Mn, Ni, and Na leached,
•• between 10% and 20% of Li and Zn leached, and
•• 40% of Cd and 31% of Ca leached.
• Using the distilled water leach test:
•• less than 1% of Al, As, Be, Cr, Co, Cu, Fe, Pb, Mn, Ni, V,
and Zn leached,
•• between 5% and 10% of Ba, Ca, Li, and Na leached.
Compared to total analysis, Pb can be used as an example to compare to
the TCLP, the distilled water and the availability tests as follows:
• Availability - 22% Pb leached
• TCLP - between 1% and 5% Pb leached
222
-------
• distilled water - less than 1% leached
In the case of the available leach test results, note that the amounts
of sulfate and chloride released were greater than the amounts present as
indicated by the total analysis, indicating that these procedures may be poor
for the species indicated. Also note that the total dissolved solids
recovered were greater than total analysis indicated.
Species released by the TCLP and the DL test from untreated combined ash
as compared to the WES treated ash is provided in Table 11. Values presented
are corrected for dilution effects of the additive. For this particular case,
treatment resulted in no significant change in the release of Ba, Ca, Cu, Pb,
Li, Mg, Mn, Ni, K, Si, NA, Sr, Zn, Cd, and IDS. Treatment did reduce release
of Fe, Al, and S04.
CONCLUSIONS
This paper presented preliminary results from evaluating several S/S
processes for treating MWC residues. Conclusions must wait until all analysis
and data interpretation are complete. Results presented, however, do indicate
the potential value of the data and the comparison which will be facilitated
when all results are available.
Preliminary data do indicate potential problems with relying on selected
leach tests and analyses or relying on only one leach test for assessing
effectiveness of treatments. Data presented also indicates the value of
strength testing before and after immersion as a comparative tool to judge
different processes. The use of strength, however, as a measure of
effectiveness to reduce release of contaminants is not proven.
References
1. Wiles, C. C. "Characterization and Leachability of Raw and Solidified
U.S.A. Municipal Solid Waste Combustion Residues" ISWA 86 Proceedings of
the 5th International Solid Waste Conference, Copenhagen, Denmark.
September 1988.
2. U.S. EPA (Environmental Protection Agency) Characterization of MWC Ashes
andLeachates from MSW Landfills, Monofills and Co-Disposal Sites.
EPA 530-SW-87-028A, Office of Solid Waste. October 1987.
3. U.S. EPA (Environmental Protection Agency) Addendum to Characterization
ofHWCAshes and Leachates from MSW Landfills, Honofills and Co-Disposal
Sites. Office of Solid Waste, June 1988.
4. J. L. Ontiveros, T. L. Clapp and D. S. Kosson. "Physical Properties and
Chemical Species Distributions Within Municipal Waste Combustor Ashes."
In Environmental Progress. Vol. 8, No. 3, pp 200-206, August 1989.
5. H. A. van der Sloot, et. al. "Leaching Characteristics of Incinerator
Residues and Potential for Modification of Leaching." In Proceedings of
the International Conference on Municipal Waste Combustion, Vol. 1,
p 2B-1, April 1989.
223
-------
6. D, R. Jackson, "Evaluation of Solidified Residue from Municipal Solid
Waste Combustors." U.S. Environmental Protection Agency, EPA/6QO/S2-
89/018, February 1990.
7. Wiles, Carlton C., "A Review of Solidification/Stabilization
Technology." Journal of Hazardous Materials, 14:5-21, 1987.
8. Federal Register, 40CFR Part 261 et. al. "Hazardous Waste Management
System; Identification and Listing of Hazardous Waste; Toxicity
Characteristics Revisions; Final Rule, Environmental Protection Agency,
March 29, 1990.
9. R. W. Goodwin, Ph.D., P.E., "Utilization Applications of Resource
Recovery Residue" Proceedings. First U.S. Conference on Municipal Solid
Waste Management, Washington, D.C. pp. 898 - 915. June 13-16, 1990.
10. F.J. Roethel, V.T. Breslin, "Interactions of Stabilized Incineration
Residue with the Marine Environment". Proceedings of the First
International Conference on Municipal Solid Waste Combustor Ash
Utilization. October 13-14, 1988, Philadelphia, PA. (Eds. T. Eighmy
and W, Chesner).
224
-------
WEATHERING OF SELECTED DEGRADABLE PLASTIC MATERIALS
UNDER OUTDOOR AND LABORATORY EXPOSURE CONDITIONS
by: Anthony L. Andrady
Research Triangle Institute
P. O. Box 12194, Research Triangle Park, NC 27709
ABSTRACT
i
Enhanced photodegradable plastic films show a faster loss in ultimate tensile elongation on
exposure to the outdoor environment relative to a comparable regular plastic film. The accelera-
tion factor obtained was 4-13 depending on the type of photodegradable polymer tested. The
factor was markedly dependent upon location of exposure with Phoenix (Arizona) showing the
fastest rate of breakdown of the five outdoor sites tested. This geographic variation appears to
be mainly a result of varying amounts of sunlight received at the different locations.
A functional set of definitions and a classification of Enhanced Degradable Plastics has been
developed.
This paper has been reviewed in accordance with the U. S.
Environmental Protection Agency's peer and
administrative review policies and approved for
presentation and publication.
225
-------
INTRODUCTION
Responding to a consistently high per capita consumption of plastics amounting to about
190 Ibs/person,1 the US chemical industry annually produces in excess of 50 billion pounds of
plastic resin.2 Of this, nearly a fourth is used by the packaging sector^ to fabricate products
which often have a very short useful lifetime (or are even single-use items). The familiar plastic
bags, foam cups, fast food packages and drinking straws belong to the latter group of products
and invariably are highly visible components of municipal solid waste as well as urban litter.
The litter problem usually associated with the urban areas, has in recent times turned out to
be even more severe in the case of marine environment.* Since the industry-wide switch over
from natural to synthetic gear in the 1940's,5 the fishing industry has become the largest single
user of plastics at sea. Accidental losses and dumping of gear-related plastics as well as pack-
aging materials from fishing vessels is a major source of plastic pollution of the world's oceans.
Discharge from merchant vessels and passenger liners further exacerbate the situation. Plastic
debris at sea is not merely an aesthetic problem, but is also a serious ecological threat. A variety
of marine animals including different types of sea birds, turtles, fish, marine mammals and ceta-
ceans are reported to suffer from entanglement in or ingestion of plastic debris.6 Recently
observed decreases in population of seals, for instance, have been primarily linked to the effect
of plastics debris in the oceansJ
Along with public education, the use of enhanced degradable plastics, often referred to as
"degradable plastics", has been proposed as a corrective strategy.8 Degradable plastics are those
plastic materials chemically modified or otherwise formulated to deteriorate at an accelerated rate
in the outdoor environment. The more rapid environmental deterioration of degradable plastics
compared to regular plastics, is expected to drastically reduce the lifetimes of debris and thereby
remove plastic waste from the environment at an observable rate outdoors. In view of the
exceptional durability and extended lifetime of regular plastic debris, any small improvement in
the rate of degradability will have a significant impact on the waste management
pegradable Plastics
The term "Degradable Plastics" is strictly a misnomer suggesting the existence of non-
degradable plastics; all polymers are of course environmentally degradable. In the case of syn-
thetic organic polymers, the rate of biologically mediated degradation in the environment is too
slow to be of any practical consequence. It is more appropriate to use the term "enhanced
degradable plastics" or "rapidly degradable plastics" for those plastics designed (or selected) for
relatively faster breakdown in the environment.
Enhanced degradable plastics technology strives to accelerate the breakdown of commodity
plastic materials by several different approaches; chemical modification of the polymer, synthe-
sis of new rapidly biodegradable thermoplastics, and by the incorporation of additives into the
plastic materials, to promote faster breakdown in the environment. There is no generally
accepted level of enhancement in rate needed to classify a material as an enhanced degradable
plastic. One such standard might be the naturally produced biopolymers which are presumably
compatible with the environment If a synthetic polymer approaches comparable rates of break-
down in the environment, it might be argued that the material also has a similar level of
environmental acceptability.
226
-------
Several different classes of such plastics have been developed over die years and these
claim enhanced degradability in sunlight, under soil, in sea water (or marine mud), and under
composting conditions. From a practical standpoint, use of Enhanced Degradable Plastics will
have the following impact on solid waste management.
(a) Litter reduction. The lifetime of Htter will be reduced thus reducing the cost of litter col-
lection and disposal. In this role, the enhanced degradable plastics might be regarded as
being equivalent to source reduction if the lifetime of litter will be shortened to an extent
to make collection unnecessary.
(b) Marine plastic waste. Unlike with land litter, there is no mechanism by which plastic
waste is removed from the marine/estuarine environments. In the absence of alternative
strategies, the enhanced degradable plastics might play a role in addressing this need.
(c) Composting. Mechanical sorting of the solid waste prior to composting is both costly
and time consuming. If plastic films and containers can be left in the composting stream,
it represents a substantial operational cost saving. Enhanced degradable plastics may
therefore be desirable in a composting operation. The same is true of anaerobic digestion
processes as well.
(d) Landfills and sewers. These essentially anaerobic environments do not affect plastics.
Consequently even natural products such as food wastes and yard wastes usually con-
tained in plastic bags undergo very slow breakdown under dry landfill conditions. Con-
tainment of organic waste capable of ready degradation in enhanced degradable plastic
bags might in some instances hasten their breakdown in a landfill.
The main goal of the present study was to assess the performance of selected common
degradable plastics materials under several typical exposure conditions. Plastics were tested as
films, and in the case of polystyrene as extruded foamed sheets, the forms in which these
plastics occur in the post-consumer waste.
A major drawback in discussing enhanced degradable plastics is the lack of adequate defini-
tions and a practical classification system. These are currently under development by the ASTM
sub committee D20.96 and are expected to be available shortly. The following definitions apply
to the present work.
Disintegration: The loss of integrity, embrittlement, or breakdown of a material on exposure to
the environment.
Deterioration: Disintegration of a material predominantly due to physical changes, (e.g..
damage to materials due to freeze-thaw cycles, damage from thermal expansion,
dissolution, and damage from rodent and insect attack on plastics)
Degradation: Disintegration of a material primarily due to chemical processes, (light-induced
degradation of polymers, hydrolysis, microbial assimilation of polymers.)
Note that degradation alters the chemical nature of the material while deterioration does not.
The reader is cautioned that the above terms are used rather loosely, and in a widely different
sense from above in the literature.9 Both Deterioration and Degradation might be further clas-
sified according to the agency bringing about the disintegration. Table I illustrates the use of the
227
-------
above terminology with the various types of breakdown. Examples of these processes and
relevant enhanced degradable plastic products are shown.
Table I. Proposed Definitions of Environmental Breakdown Processes and
Enhanced Degradable Plastics.
DEGRADATION .
Adistategratbn caused
predominantly by chemical
changes
/ *
DISINTEGRATION
Bwiktefown (size reduction)
of material Mo smal fractions,
Including Bmbrlttlement
\
DETERIORATION
A eSslntegfatlon caused *
predominantly by physical
changes
PHOTODEGRADATION
light-Induced
BIODEORADATION
Brought about by living animals,
plant, particularly microbe*
OXIDATIVi DEGRADATION
Caused by therrnooxkJativt
road Ions
HYDROLYSIS
Caused by reaction with water
BIODETIRIORATION
Brought about by living animals
end plant
DISSOLUTION
(HydrodBWrloratlon) Brought
about by water
THERMAL DETERIORATION
Caused by t reeza-thawtng or
thermal cycling forces
Examples
a
b
C
d
e
f
§
Product
Ecolyte
ECO resin
Plastigone
Biopol - ICI
*~
--
ADM,
Ecostar
PCUPE-
Union Carbide
Belland
a Polymers with ketone groups in main chain or as a side chain.
b Metal-compounds additives for PE
c Poly(hydroxybutyrate valerate)
d Metal-compounds and inorganic pigments in polyolefins
e Acid catalyzed hydrolysis of cellulose
f Starch-polymer composites; biodegradable polymer - other polymer composite
g Soluble acrylic copolymers; soluble poly(vinyl alcohol) files
228
-------
EXPERIMENTAL
MATERIALS:
All plastic materials used in the study was obtained from the manufacturers as extrusion
blown films, These were:
(a) Ethylene-carbon monoxide eopolymer(~l% CO) HiCone Division
[ECO copolymer] ITW Company
(b) LDPE film containing metal compound pro-oxidant Plastigone Company
[LDPE/metal compound]
(c) LDPE/starch granule blend, with metal compound pro-oxidant Archer Daniels
[LDPE/starch/metal compound] Midland Company
Each manufacturer also submitted samples of plastic film not containing any degradable
additive or not chemically modified to serve as a reference in the exposure experiment.
4" • '
EXPOSURE: * '
Results discussed here pertain to outdoor land and Weather-Ometer® exposures. Outdoor
exposures were carried out according to ASTM D1435 with the samples exposed with a ply-
wood backing at an angle of 45 degrees facing south. The following exposure locations were
used.
(a) Miami, FL
(b) Cedar Knolls, NJ
(c) Chicago IL
(d) Wittmann, AZ
(e) Seattle, WA
ASSESSMENT OF DEGRADATION:
Tensile property determinations, particularly elongation at break, were used for assessing
the extent of degradation of exposed plastic materials in the case of all samples except for foamed
polystyrene. Tensile testing was carried out according to ASTM D882 standard method for
plastic laminates.
The materials were also exposed in an Atlas Ci-65 Weather-Ometer® equipped with a
borosilicate-filtered xenon lamp as the light source. The exposure was conducted in accordance
with ASTM G26 and the cycle used was 102 minutes light/18 minutes light and water
spray/63 °C black panel temperature. This exposure cycle generally yields data which correlate
well with Florida outdoor exposure, •
RESULTS AND DISCUSSION
The non-uniformity of outdoor exposure conditions is well known to lead to variability in
weathering test results.10 Unlike in the case of establishing permanence properties of regular
plastic films, the outdoor exposure of enhanced degradable plastic films requires a much shorter
period of exposure, often 2-3 weeks outdoors. As such, the short term (day to day) fluctuation
229
-------
in key factors such as average temperature, rainfall ,and the available sunshine, is likely to affect
the variability of test data to a greater extent compared to longer exposures of months or years.
In an attempt to overcome this difficulty a "duplicate exposure" protocol was developed during
this work. This protocol, however, was used in selected exposure sites only due to cost
constraints.
Figure 1 illustrates the basis of the "duplicate exposure" procedure. Essentially, each time a
sample is removed from the original set of plastic films, it is replaced by a fresh sample. Thus,
two complete sets of samples are available for testing at the end of the test period; one with
1,2 nth week, of exposure and another with nth, (n-l)th, 1st week of exposure.
With no drastic changes in weathering conditions both sets should yield very similar data.
Inconsistent data between the two sets would suggest non-uniform exposure conditions during
the short duration of exposure.
SET1
ADD
REMOVE
TOTAL
n
0
n
0
1
n-1
0
1
n-2
0
1
n-3
* • *
* * *
* * *
0
1
2
0
1
1
0
1
0
SET 2
ADD
REMOVE
TOTAL
XPOSUKETIME
1
0
1
1
0
2
1
0
3
1
0
4
• * *
* * *
* » •
1
0
n-3
1
0
n-2
0
n-1
n-1
I I I I ill
0 1 2 3 •• * n-2 n-1 n
Weeks
Figure 1. Double Exposure Protocol for an Exposure for n Weeks
with Weekly Sampling of Set 1 '
Figure 2 shows typical data obtained for loss in tensile elongation at break due to outdoor
exposure of ethylene-carbon monoxide (1%) copolymer. The shape of the curve suggests a
logarithmic relationship between the elongation at break and duration of exposure. The
enhanced degradable plastic material as well as the control material which is the same plastic
without the degradable additive or modification show a similar dependence. However, in the
case of the reference samples only minimal changes in elongation at break was obtained during
the relatively short period of exposure and the data showed a high degree of scatter.
Kinetics of the change in tensile elongation at break can be conveniently studied by fitting
the data with an empirical equation of the following form.
E = a +
where E is elongation at break, d is the duration of exposure and
a and b are constants
230
-------
o
o
W
g
B
600
400
200
0
• ••
• 6P
O 6PC
0
20 40 60
Exposure Time (days)
80
Figure 2. Percent elongation at break vs. exposure time for ethylene-carbon monoxide
copolymer exposed outdoors in New Jersey. Code 6P refers to the copolymer
samples, and 6PC refers to the low-density polyethylene control.
Plotting the data in the form suggested by the equation yields a pair of empirical parameters
a and b. The parameter b is a measure of the rate of disintegration of the plastic films and a
study of the change in value of b with geographic location can yield information as to variability
that might be expected on the basis of climatic differences.
(a) Agreement between "Duplicate Exposure" Sample Sets
Figure 3 shows a typical set of data obtained for two sets of enhanced degradable samples
exposed at a single location. The two sets of data agreed very well and a statistical analyses of
the data showed them to belong to the same population with a high degree of confidence. This
implies that the trend in data is not a result of sharp changes in key climatic factors (possibly
available sunshine, and temperature) during the period of exposure.
While the scatter associated with data for enhanced degradable plastic materials at the dif-
ferent locations was minimal, that for the control plastic film often showed considerable scatter.
The tensile elongation at break of regular polyethylene material did not change significantly due
to degradation during the short duration of exposure monitored. However, some change due to
"annealing" under exposure conditions leading to small changes in modulus, strength, and
231
-------
ultimate elongation may often take place during initial stages of outdoor exposure. These lead to
scatter in the data for control samples.
1000
• PG
• PC Duplicate
O PGC
Q PGC Duplicate
10
20 30 40 50
Exposure Time (days)
60
Figure 3. Percent elongation at break (log tensile elongation at break) vs. exposure time
for LDPE films containing added metal compound pro-oxidant films exposed
outdoors in Miami. Code PG refers to photodegradable polyethylene and PGC
refers to LDPE control.
Table II lists the values of regression coefficients obtained for the "Duplicate Exposure"
sample sets at various locations. The gradient of the line, b, which is an indicator of the rate of
degradation agree well for the two sets of data at each location.
(b) Geographic Variability
A list of regression coefficients obtained for data on the three types of enhanced degradable
plastics exposed at five outdoor locations is given in Table n. The data clearly indicate that the
enhanced degradable plastics underwent a faster rate of environmental photodegradation com-
pared to control plastic material. Ratio of the gradients for enhanced degradable and control
polymer might be used as a measure of the enhancement in degradation (or an enhancement fac-
tor). As both the control and the enhanced-degradable samples were exposed concurrently at all
locations, the enhancement factor would be expected to be constant Average acceleration fac-
tors obtained at the different locations are shown in Table HI.
232
-------
TABLE II. Regression Coefficients for Fit of Test Data to Equation 1.
Sample
Location -bxlO3 (days'1)
a
1. LDPE Control
HiCone photodegradable
six-pack ring material
2. LDPE Film
Control
Plastigone material (LDPE/MX*)
AZ
IL
FL
NJ
WA
AZ
IL
FL
NJ '
WA
AZ
IL
FL
NJ
WA
AZ
IL
FL
NJ
WA
25
4
14
10
4
257
87
65
52
40
33
13
22
12
16
122
49
72
50
54
3. Starch-polyethylene blends
1047
648
953
708
823
550
246
184
139
97
607
742
917
139
97
885
1091
960
1033
1004
LDPE Film Control
ADM, photodegradable
starch/LDPE blends
(LDPE/starch/MX)
AZ
EL
FL
NJ
WA
AZ
IL
FL
NJ
WA
25
11
9
6
10
391
148
129
137
102
300
354
371
340
395
89
67
41
66
54
Note: MX = transition metal compounds (pro-oxidant additive).
TABLE m: Enhancement Factors for Three Types of Enhanced Degradablc Materials
for Different Geographic Locations.
Enhancement Factors
Location
ECO Copolymer
LDPE/MX*
LDPE/Starch/MX*
AZ
IL
FL
NJ
WA
10
22
4
5
10
4
4
3
4
3
16
14
14
23
10
Average
(with standard error)
10 + 3
4 ±0.2
13 + 2
*MX = transition metal compounds (pro-oxidant additive).
233
-------
ECO copolymer and the polyethylene/starch blends containing metal compounds showed
about the same degree of enhancement, approximately by an order of magnitude. LDPE/MX
material, somewhat similar to the latter plastic material in that it too contains metal compound
pro-oxidants, showed a consistent factor of 4 for all locations of exposure.
It is of interest to compare the rankings of different geographic locations on the basis of the
b parameter. Such a ranking might be made for both the control plastic films as well as for the
enhanced degradable samples. However, the usefulness of the former is somewhat limited due
to the considerable scatter associated with the data. Rankings obtained with the three types of
enhanced degradable plastic films are as follows.
ECO copolymer AZ>IL>FL>NJ>WA
LDPE/MX AZ>FL>WA>NJ>IL
LDPE/starch/MX AZ>IL>NJ>FL>WA
Except with the LDPE/metal compound pro-oxidant system (Plastigone material), the rank-
ing for the control polyolefin did not agree with that for the enhanced degradable material. The
difference in the above rankings is not surprising as the materials tested are able to undergo both
photodegradation as well as thermal oxidatiye degradations. These processes may have different
activation energies and the different climatic conditions at the test sites would have affected the
rates of two degradation processes differently. The relevant activation energies are not reliably
known for the three systems.
Assuming that the primary mechanism of degradation involved in outdoor exposure is
photodegradation, it is reasonable to expect at least an approximate correlation between the rate
of degradation with average available sunlight The available light at different exposure locations
varied widely. However, the spectral quality of sunlight at the five locations was also different
depending upon the solar zenith angle and the season of the year. Any difference in the spectral
irradiance distribution of sunlight will, of course, have a very significant impact upon the effi-
ciency of degradation, its magnitude being determined by the activation spectrum for the particu-
lar polymer. A plot of the value of b parameter versus the average available sunlight for three
types of degradable plastics is shown in Figure 4.
Surprisingly good correlations are obtained with the three exposure locations. The
LDPE/stareh/metal compound pro-oxidant system (ADM material) shows the highest degree of
sensitivity (i.e. dependence of the rate of loss in extensibility on outdoor exposure upon the
average amount of sunlight available at that location). However, the sensitivity of the plastic
material to light-induced degradation is a function of formulation. By alternating the
concentration of chromophores or the concentration of the pro-oxidant, the sensitivity of any of
the tested films could be varied
Weather-Ometer® exposure data were available for two types of enhanced degradable
plastics only - six-pack ring material and Plastigone material. The loss of tensile elongation
obtained in the Weather-Ometer® exposures were faster compared to the outdoor exposures.
The relevant regression coefficients are shown in Table IV.
The acceleration factor obtained for Plastigone material is about 7 for Weather-Ometer®
exposure. In the case of six-pack ring materials, the control sample was of different thickness;
the acceleration factor, however, was about 3. Both are different from the average values
obtained on the basis of the outdoor exposures. This is to be expected because of the widely dif-
ferent exposure conditions used. The Weather-Ometer® exposure involves continuous exposure
234
-------
to light at a temperature much higher than the ambient. Temperature dependence of the
photodegradation in the case of six-pack ring material and the contribution of thermooxidative
degradation in the case of the Plastigone material can modify the results in such an accelerated
study. However, the Weather-Ometer® exposure is an accelerated exposure compared to
outdoor weathering as the b values (expressed in Table IV) in hours'1 are much higher than
those for outdoor exposure.
400
300-
i
§
•a
03
•o
O
Q
200-
100-
0
• 6P
O PG
A ADM
10 20
Average Daily Total Radiation (MJ/m2)
30
Figure 4. Rate of outdoor degradation versus the total global solar radiation for three types of
Enhanced Photodegradable Plastics. 6P: ethylene-carbon monoxide copolymer
(~ 1 percent CO), PG: LDPE containing added metal-compound pro-oxidants, and
ADM: LDPE/starch (~6%) blends containing added metal compound pro-oxidants.
235
-------
TABLE IV. Data from Weather-Ometer® Exposure of Enhanced Degradable Plastics.
Material
Six-pack ring material
Control LDPE film
Plastigone material
Plastigone control
a
235
1463
927
599
-b x 103
(hrs-1)
8.74
2.59
6.35
0.94
CONCLUSION
The three types of enhanced photodegradable plastic film samples tested showed the tensile
elongation to be a property suitable for assessing the outdoor weathering rates for this class of
plastics. This is a useful conclusion as the ASTM D20.96 committee on Degradable Plastics is
considering the use of this test.
Plastic materials showed considerable amount of location-dependent variation in per-
formance as measured by the rate of loss of ultimate elongation exposure. Arizona was the
harshest outdoor exposure environment while New Jersey and/or Washington states were the
mildest sites. The ratio of the breakdown rate of enhanced degradable materials relative to that
of the base (unmodified) plastic, is considered as an enhancement factor. This factor shows
variability from location to location. The location-dependent variability in enhancement factor
can be explained in terms of different average sunlight levels available at these locations.
REFERENCES
1. Plastic Waste Management Institute, Tokyo, Japan. /«: A. Hershkowitz and E. Salerni
(ed.) Garbage Management in Japan. Inform, Inc., New York, 1987. p 51.
2. Modern Plastics. Resin Report Jan. 1989. McGraw Hill, 1989.
3. Franklin Associates. In: Characterization of Municipal Solid Waste in the United States.
1960-2000. Franklin Associates, Ltd. 1988.
4. Interagency Task Force on Persistent Marine Debris. 1988 Report. Chair: Department of
Commerce, National Oceanic and Atmospheric Administration. May 1988.
5. Uchida, R. N. The types and estimated amounts of fish net deployed in the North
Pacific. Honolulu, Hawaii. In: Proceedings of the Workshop on the Fate and Impact
of Marine Debris. U. S. Department of Commerce, Honolulu, Hawaii, 1984. p. 37-
108.
6, CEE. Plastics in the Ocean: More than a litter problem. Center for Environmenal
Education. February, 1987.
236
-------
7. Laist, D. Overview of the Biological Effects of Lost and Discarded Plastic Debris in the
Marine Environment. Marine Poll. Bull., IS.: 305,1987.
8. Proceedings of Symposium on Degradable Plastics. Society of Plastics Industry.
Washington, DC. June 1987.
9. S. A. Barenberg, J. L. Brash, R. Narayan, and A. E. Redpath. In_: Degradable
Materials. Perspectives, Issues and Opportunities, CRC Press, Boston, 1990.
10, A. David and D. Sims (Eds.), "Weathering of Polymers". Applied Science Publishers,
New York, 1983.
237
-------
POLLUTION PREVENTION RESEARCH WITHIN THE FEDERAL COMMUNITY
by: James S. Bridges and Emma Lou George
Pollution Prevention Research Branch
Risk Reduction Engineering Laboratory
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
ABSTRACT
One of the primary ongoing programs for promotion and encouragement of pollution
prevention research is a cooperative program between the United States Environmental Protection
Agency (EPA) and the Federal community at large. EPA's Waste Reduction Evaluations at Federal
Sites (WREAFS) Program supports pollution prevention research through joint assessments of
problematic areas at selected sites. The three primary objectives of the WREAFS Program are to:
1) conduct waste minimization assessments and case studies; 2) conduct research and
demonstration projects jointly with other Federal activities; and 3) provide technology and
information transfer of pollution prevention results.
This paper describes the WREAFS Program support of pollution prevention research
throughout the Federal community and provides current status on all projects to date. These include
joint efforts with the Departments of Agriculture, Defense, Energy, Interior, Transportation,
Treasury, and Veteran Affairs. Seven of eleven projects are with the Department of Defense under
the Branches of the Air Force, Army and Navy. Two projects are with the Coast Guard under the
Department of Transportation. At present there are six waste minimization opportunity assessments
(WMOA) in various stages of completion, four on-going and two recently initiated projects. Under
the WREAFS cooperative umbrella, two other research and development projects are ongoing within
the Departments of Agriculture and Defense.
These assessments have identified case study and research opportunities to implement
pollution prevention for a range of military and industrial operations including metal cleaning, solvent
degreasing, spray painting, vehicle and battery repair, ship bilge cleaning, torpedo overhaul, buoy
restoration, lens grinding, hospital operations and other industrial processes.
The waste minimization recommendations are source reduction methods including technology,
process and procedural changes and recycling methods of reuse or recycling. The WMOA consists
of four steps: Planning and Organization, Assessment, Feasibility Analysis, and Implementation.
The fourth step, implementation, is conducted at the discretion of the Federal facility.
This paper has been reviewed in accordance with the U.S. Environmental Protection Agency's
peer and administrative review policies and approved for presentation and publication.
238
-------
INTRODUCTION
The objectives of the WREAFS Program are to identify new technologies and techniques for
reducing wastes from industrial processes used by Federal agencies and to enhance the adoption of
pollution prevention through technology transfer. New techniques and technologies for reducing
waste generation are identified through waste minimization opportunity assessments and are
evaluated through joint research, development and demonstration (RD&D) projects. The
information and data from these projects are then provided to both the public and private sectors
through various technology transfer mechanisms, including project reports, project summaries,
conference presentations and workshops.
The waste minimization opportunity assessments are conducted by an assessment team that
is composed of personnel from EPA, staff from the federal facility that is cooperating in the program
and others who can provide technology and processing expertise. The assessments follow the
procedures described in the EPA Report, Waste Minimization Opportunity Assessment Manual1
(EPA/625/7-88/003)1. This manual provides a systematic procedure for identifying ways to reduce
or eliminate waste generation. The development of this procedure was supported by the Risk
Reduction Engineering Laboratory, U.S. Environmental Protection Agency, Cincinnati, Ohio.
As a result of joint waste minimization opportunity assessments (WMOA's), RD&D projects
are identified with recommendations for pollution prevention under the implementation phase. The
demonstration projects are conducted under interagency agreements with joint funding by EPA and
the cooperating Federal agency. Waste minimization workshops and other technology transfer
methods are used to communicate the results of these projects to the Federal community and the
private sector.
WREAFS PROGRAM PROCEDURES
The WREAFS Program procedures used for conducting the waste minimization assessments
are closely related to the WM procedures presented in the Manual. Figure 1 describes the course
followed by a typical WREAFS project. The assessments consist of four major phases:
(1) Planning and Organization: organization and goal setting;
(2) Assessment: careful review of a facility's operations and wastestreams and the
identification and screening of potential options to reduce waste;
(3) Feasibility Analysis: evaluation of the technical and economic feasibility of the options
selected and subsequent ranking of options;
(4) Implementation: procurement, installation, implementation and evaluation.
Waste Minimization Opportunity Assessment Manual (EPA/625/7-88/003) is available free from CERI, Cincinnati.
239
-------
Many of the WM opportunities identified during WREAFS projects involve low-cost changes
to equipment and procedures presently employed at other Federal facilities or within private
industry. These WM opportunities can often be implemented by the facility without extensive
engineering evaluations. Some other WM opportunities identified during these projects will require
further study before full implementation can be realized. Typically, opportunities requiring further
evaluation are those that have the potential for affecting the process and/or require the use of new
procedures or equipment. In such cases it may be necessary to conduct demonstration projects.
DEMONSTRATION AND EVALUATION PROJECTS
The types of research projects that can be pursued under the WREAFS Program are those that
are expected to advance the knowledge and practice of waste minimization technologies and
methods, and have broad applicability to Federal facilities and private industry. Depending on the
nature and state of development of the WM option selected for demonstration and evaluation, these
projects may include: (1) process design, (2) detailed design and specification, (3) system
procurement, (4) installation and start-up, (5) monitoring and (6) reporting. Some projects may
require bench-scale and/or pilot testing prior to or as a part of the demonstration project. Other
projects may utilize full-scale equipment directly on the production line.
COMPLETED ASSESSMENTS (WMOA'S)
Five WREAFS assessments are now completed, three at DOD facilities, including two with
the Navy and one with the Army; one at the Veteran's Medical Center in Cincinnati, Ohio; and one
with the Coast Guard at Governor's Island, New York. With these five, the assessment surveys have
been completed and waste minimization options have been identified for implementation. A
description of each project follows.
PHILADELPHIA NAVAL SHIPYARD
This project was conducted in cooperation with the Environment, Safety and Health Office
of the Philadelphia Naval Shipyard (PNSY). The shipyard has an ongoing program for waste
minimization. With their guidance, several industrial operations were selected for application of the
new waste minimization procedures. The shipyard plans to use these results as guidance for
evaluating other pollution prevention/waste minimization activities.
FACILITY DESCRIPTION
The Philadelphia Naval Shipyard, the nation's oldest continuously operating naval shipyard,
is located in South Philadelphia on 1,000 acres of land. Since its inception, 127 ships have been
constructed, with last ship launched in 1971. It now specializes in revitalizing and repairing ships
already in fleet. The Service Life Extension Program (SLEP) is the shipyard's largest program and
it's comprehensive keel-up restoration arid modernization overhaul extend the life of aircraft carriers
by half, at approximately one-third of the cost of a new carrier.
240
-------
AREAS OF WM OPPORTUNITY EVALUATIONS
The industrial activities selected for this project included:
• Aluminum cleaning and spray painting
The aluminum cleaning is performed to remove oil and other materials from the surfaces of
aluminum sheets prior to welding and is critical to the welding if not properly cleaned. The cleaning
line consists of two process and two rinse tanks. The tanks become diluted and contaminated with
waste liquids and are disposed. Spray painting involves solvent degreasing and a water curtain booth
for painting. Paint solids from overspray and sludge residues are removed and drummed with booth
water discharged into the sewer. For the discharge, one organic polymer is used for colloidal
precipitant dispersion and a second for coagulation into a sludge for screen removal.
• Spray painting of steel parts including structural columns
Spray painting of steel parts requires an area for shot blasting and painting. A dry air
filtration system, a booth for blasting and painting steel columns and a water curtain booth are
required. All are used for epoxy spray painting of steel surfaces.
• Citric acid bilge derusting operations in drydock
This is a chemical cleaning process for ships' tanks, bilges and void spaces performed in
drydock. It employs the use of a citric acid/triethanolamine (TEA) solution to remove oxides from
metal surfaces with subsequent neutralization followed by rinsing. The volume of spent solutions
from a derusting/neutralization/rinse operation is typically about 3000 gallons. It generally has a pH
below 4.0, contains toxic metals and is hauled to a treatment/disposal facility.
WASTE MINIMIZATION OPTIONS AND RECOMMENDATIONS
After the assessment and feasibility analysis phases were completed, seven options were
evaluated and ranked. The best options for implementation were:
• Awareness and Training for personnel/procedure-related options
Paint and paint wastes comprise the second largest hazardous waste stream generated at the
shipyard. A program emphasizing operator involvement and responsibility could reduce waste paint
in overspray, paint remaining unused in cans and paint solidification prior to use.
• Dragout Reduction and Bath Maintenance and • Two Stage Rinsing
A hand-held spray rinse applied over process tanks would return 90% of the dragout back
to the process tank. A bath maintenance system employing an oil skimmer for floating oil/grease
removal and a cartridge filter for suspended solids removal would extend the usable life of process
tanks from 3 months to a year. Employing a two stage rinse with sequencing of first and second
would prolong the life of the rinsing tanks.
241
-------
RESEARCH, DEVELOPMENT AND DEMONSTRATION
Recovery of concentrated citric acid solution represents a viable candidate for further
research, development and demonstration. It would involve the implementation of equipment for
the recovery of the citric acid/TEA solution. This process would employ an electrodialytic membrane
unit for separation and removal of dissolved metals. This technology has been applied to similar
chemical solutions but its application to this waste has not been previously demonstrated.
FORT RILEY, KANSAS
This project is an assessment conducted at the U.S. Army Forces Command (FORSCOM)
maintenance facilities at Fort Riley, Kansas. It was initiated from data received from the Fort Riley
Environmental Office on amounts of hazardous waste streams generated onsite. There are ten other
U.S. Army FORSCOM installations providing potential for application of similar waste minimization
options.
FACILITY DESCRIPTION
Fort Riley is a permanent U.S. Army Forces Command (FORSCOM) installation that
provides support and training facilities for the 1st Infantry Division (Mechanized), Non-Divisional
Units, and tenant acitivities. It is owned and operated by the U.S. Government and occupies
approximately 121,000 acres in north central Kansas. Fort Riley provides the U.S. Army with the
capability to house and train an Army division and associated land combat forces, as well as to
service Army functions in the midwest area.
AREAS OF WM OPPORTUNITY EVALUATIONS
Large hazardous waste streams are generated here consisting of spent automotive cleaning
solvents and various RCRA listed wastes including waste battery acid, waste caustic cleaners and
spent parts wash water. Currently these hazardous wastes are handled as follows:
• Waste battery acid collected in 15-gallon plastic drums
• Caustic cleaners collected in 55-gallon metal drums
Both are classified wastes and sent by truck to the hazardous waste storage facility
• Waste water from the automotive parts washer is discharged to an onsite nonhazardous
waste evaporation pond system. These will eventually be reclassified under RCRA regulations as
D002 and D008 waste due to alkalinity and lead content, respectively.
The areas selected for evaluations were the battery repair and service shop which generates
waste battery acid, the automotive subassembly rebuild area that generates waste automotive wash
water and the radiator repair shop where spent caustic cleaner is disposed as a hazardous waste.
242
-------
Battery Repair Shop
The battery service area drains about 7,200 gallons per year of battery acid for disposal. The
drained batteries are inverted, and either disposed if determined unusable, or repaired and refilled
with fresh 37 percent sulfuric acid, recharged and reused.
Automotive Subassembly Rebuild Shop
Prior to rebuilding various automotive subassemblies, e.g., engines, clutches, transmissions,
these units are disassembled and placed in a specially designed high pressure hot-water washer for
cleaning. This washer continuously circulates hot alkaline solution through high-pressure jets to
provide the cleaning action. The washer wastewater contains chromium and lead which must be
disposed of as a RCRA characteristic waste.
WASTE MINIMIZATION OPTIONS AND RECOMMENDATIONS
After the assessment phase the following recycling options were recommended for the two
major wastes, waste battery acid and the washer wastewater.
Recycling of Waste Battery Acid
After collection, the waste battery acid would be transferred to an acid-resistant tank and
mixed by recirculating pumps. Based on specific gravity the acid strength would be adjusted to 37-
38% H2SO4 with 78% sulfuric acid to yield standard battery acid. This refortified acid would be
pumped through an acid-resistant filter to remove particulates and collected in drums for use in the
battery repair operation. Other treatment operations would be required for more extended battery
life. The installation of an expensive acid-resistant cooler can be avoided with the use of 78%
sulfuric acid (60° Baume) rather than 93-98% (66° Baume) due to heat evolved during mixing.
Rccirculation of Washer Wastewater
Instead of disposing of this wastewater as a hazardous waste, it would be recycled following
filtration through an in-line filter to remove discrete particulates, followed by de-emulsification and
removal of oils. Additional detergent would be added as needed and water recycled to the parts
washer.
Implementation of recycling both the battery acid and the waste water would cost about
$35,000 with an annual operating cost savings of about $149,000. Payback period for simultaneous
implementation of options would be less than five months. Adequate testing of refortified, recycled
battery acid and monitoring of the detergency of the washer wastewater would be required.
RESEARCH, DEVELOPMENT AND DEMONSTRATION
Proposed particulate filtration and de-emulsification processes would qualify as research,
development and demonstration projects. In-plant experimentation of filter element types, necessity
for multiple filters and cleansing effectiveness of recylcled wash water would require further
determination.
243
-------
NAVAL UNDERSEA WARFARE ENGINEERING STATION
KEYPORT, WASHINGTON
The purpose of this project was to identity waste minimization opportunities for two
industrial units at the Naval Undersea Warfare Engineering Station (NUWES) in Keyport,
Washington. This project was conducted in cooperation with the Naval Energy and Environmental
Support Activity (NEESA) of Port Hueneme, California, in coordination with the Environmental
Division (Code 075) of the NUWES Keyport Civil Engineering Department. Several departments
at NUWES Keyport are involved in an ongoing program to further the process of waste minimization
at NUWES.
FACILITY DESCRIPTION
NUWES Keyport is located within the central Puget Sound area of northwestern Washington
State. The Keyport property was acquired by the Navy in 1913 and first used as a quiet water range
for torpedo testing. Later the Station was used for torpedo repair. The facility acquired its current
name in 1978 in recognition of the inclusion of various undersea warfare weapons and systems
engineering and development activities.
AREAS OF WM OPPORTUNITY EVALUATIONS
The principal activities currently conducted at NUWES Keyport are the design and testing
of torpedoes. These activities generate a variety of potentially hazardous wastes, including fuel, oil,
hydraulic fluid and grease, various metal and plating bath liquids, paint and thinner, Freon®, alcohol,
mineral spirits and other solvents, resins, acids and caustics, chromate and cyanide salts, pesticide
residues, wastewater treatment sludge, waste dye and detergent.
The major component of waste management involves the use of Otto Fuel II (Otto fuel)
which is used for propelling torpedoes. Otto fuel is composed of propylene glycol dinitrate with
lesser amounts of 2-nitrodiphenyl-amine and di-n-butylsebacate. Otto fuel is a monopropellant in
that it burns without oxygen. The Navy currently treats all Otto fuel-contaminated solid waste as an
explosive, reactive waste. Two grades of usable Otto fuel are distinguished:
« Condition-A pure, virgin Otto fuel used for torpedoes in fleet
« Condition-B collected from defueled torpedoes and reused for proofing (testing)
purposes.
The two areas selected by the Navy for evaluation were torpedo maintenance shops, both
having similar operations, processes, and waste streams. The major activities of these two shops are
as follows:
• Weapons Depot Maintenance of unproofed torpedos, i.e. ones that have exceeded shelf
life of 8 years. The Otto fuel is drained and the torpedoes are disassembled and restored.
In-water testing of torpedoes with subsequent breakdown and restoration.
• Advanced Capability tests the newest version of Mark 48 torpedo, assembling and
proofing, disassembling, cleaning and reassembling.
244
-------
* Research & Development assemble, proof, tear down, clean and reassemble torpedoes.
• Mark 46 Shop defuels, disassembles, cleans, reassembles and refuels Mark 46 torpedoes.
Diethylene glycol (DEG) is used to clean fuel tanks after draining.
Wastes generated during these processes include:
« cyanide-containing liquid wastes and sludges that are Otto fuel combustion
byproducts
Otto fuel-contaminated solvents and oils generated during cleaning of parts
Otto fuel-contaminated wastewaters
Otto fuel-contaminated solids, primarily clothing and rags
Used oil
Used hydraulic fluid
DEG (diethyiene glycol) and Otto fuel contaminated rinse waters
The Otto fuel sump is an underground stainless steel holding tank for Condition-B torpedo
fuel contaminated with seawater, fresh water and alcohol. A cyanide sump is located within the Otto
fuel sump and contains sludges and wastes contaminated with cyanide as well as detergents, oils,
grease and alcohol. Fuel is recovered from these tanks by separation and treatment.
All the wastestreams are RCRA reactive wastes.
WASTE MINIMIZATION OPTIONS AND RECOMMENDATIONS
NUWES Keyport has performed well in the handling, storage and minimization of waste
materials on-base. During this assessment, no major waste minimization options were identified that
NUWES has not already implemented or plans to implement. However, the following five options
were identified to aid in the process:
• Volume Reduction of Otto Fuel-Contaminated Clothing
This includes a) segregating used clothing and b) removing uncontaminated portions. This
would significantly reduce the amount that needs to be disposed as a hazardous waste. Otto fuel has
a distinctive yellow color that would facilitate recognition of contaminated clothing. Often only small
areas are involved and these would be removed by cutting off the contaminated portions or by
substituting disposal sleeves and leg cuffs. All other uncontaminated parts would be discarded as
non-hazardous materials. This would require a minimal capital outlay and savings would be realized
in reduced disposal costs.
• Automated Cleaning of Parts and Fuel Tanks
Automated cleaning of parts and fuel tanks would result in more efficient and faster cleaning,
smaller amounts of hazardous waste liquids and smaller amounts of contaminated clothing. Three
dip tanks were to be replaced soon with automatic parts washers using biodegradable cleaning
liquids. More extensive or complete automation of cleaning operations within the two shops would
aid in reducing wastes and would involve a cleaning media of water and detergent in an agitator or
jet system or an ultrasonic cleaner. While this option would require capital outlay for the purchase
245
-------
of a cleaning unit, it would allow for reduced disposal of cleaning solutions and reduced raw
materials purchased. Payback period for this option is 0.4 years and is one of the two fastest
payback periods (total capital investment/net operating cost savings) represented by these
recommendations.
• Automated Fuel Tank Draining
Automated fuel tank disassembly by robotics has been in use at the Mark 46 Shop for 3J/2
years, resulting in more efficient and faster operations and smaller amounts of waste liquids and
contaminated clothing. Similar equipment would be installed in the Mark 48 Shop and would have
a relatively short payback period due to decreased costs of labor, contaminated clothing disposal and
spill cleanup. The decreased demand on manpower could be refocused and utilized in other areas.
• Modify the Deep Sink Draining Schedule
Switching from automatic weekly draining to an "as needed" non-automatic schedule would
result in reduced cost due to a smaller purchasing volume of cleaning solvents, reduced volume of
hazardous waste disposal and less man-hours. This is the other option with the fastest payback
period, only requiring schedule modification and no capital outlay.
• Recycling of Mineral Spirits
Mineral spirits used for the cleaning of parts are currently treated as a RCRA hazardous
waste, combined with other liquid waste streams and sent to an official off-site facility for
incineration. Recycling of the mineral spirits could be used to recover up to 86% of the spent
solvent. This also has a short payback period, involving moderate to high capital outlay for
equipment, but savings in both the decrease of disposal costs and purchase of mineral spirits, making
it an appealing option.
RESEARCH, DEVELOPMENT AND DEMONSTRATION
The following five research needs were identified during the course of this assessment:
• Evaluate cost effectiveness of clothing with disposable sleeves and cuffs.
• Develop a test for determination of spent deep sink cleaning liquids.
• Evaluate the cost feasibility of robotics for draining, defueling and rinsing topedoes.
» Identify potential recycling options for waste hydraulic fluid.
• Evaluate current practices for used torpedo engine oil.
246
-------
U.S. COAST GUARD SUPPORT CENTER NEW YORK
GOVERNORS ISLAND, NEW YORK
This project was conducted at U.S. Coast Guard facilities on Governors Island, New York
in cooperation with Coast Guard officials. It was initiated through contacts with the Hazardous
Waste Office in the Industrial Division and conducted through them with interviews of Coast Guard
Headquarters officials in Washington. This project attempted to develop both management initiatives
as well as technical changes that can be made at Governors Island for waste minimization purposes.
Technical waste minimization evaluations conducted at this site centered on paint removal by
blasting, painting and solvent recovery.
FACILITY DESCRIPTION
Governors Island is located off the southern tip of Manhattan and is accessible primarily by
a Coast Guard operated ferry. The island encompasses 175 acres and consists solely of Coast Guard
facilities which are grouped together under the name "Support Center New York". The island serves
as a support center for Coast Guard activities conducted within the New York area, for tenant
commands located on the island and is the home port for a number of Coast Guard cuttefs. There
are 22 different commands represented on the island, each of which reports to Headquarters in
Washington or to an off-site location. Support Center New York supports all activities on the island,
however it does not have authority over all commands.
AREAS OF WM OPPORTUNITY EVALUATIONS
Management Activities
During the early stages of the assessment it was determined that a review of the hazardous
waste management activities on Governors Island was a useful study area. Three areas were
reviewed:
• Successful waste minimization steps currently in practice include use of lead-free paint
throughout the Coast Guard, development of a new paint with lower volatile organic compound
(VOC) content, use of solar batteries in aids-to-navigation requiring battery power and elimination
of engine coolants containing dichromate additives. A hazardous waste compactor for paint cans has
been purchased for decreasing volume of waste. Blasting grit used for paint removal has been
reduced 50% by installation of a new baghouse and recycling system. Disposable brushes usage
reduces the waste from brush cleaning thinner.
• Waste minimization problem areas identified included governmental issues concerning lack
of motivation for compliance, funding and procurement practices. End-of-year spending results in
purchase of new paint when current reserves are sufficient. Lack of proper storage for preservation
of paint quality contributes to excessive waste. Turnover of military personnel contributes to
problems of hazardous waste handling.
• Site organization and facilities are major contributing factors in the problems of waste
generation and disposal. The presence of tenant commands with no centralized procurement nor
accountability causes poor management of supplies. Use of multiple disposal contractors by tenant
commands causes significant hazardous waste generation, primarily of paint. Storage facilities
247
-------
throughout the island are unheated and promote degradation of paints resulting in large quantities
for disposal. It has been estimated that 50% of paint disposal is unnecessary.
• Potential solutions for problems are recommended; starting with cultural changes
emphasizing commitment on all levels, forceful expression of policies, employee training and
incentive awards and conducting detailed engineering evaluations of the waste generating processes.
This would require a concerted effort from both Coast Guard Headquarters in Washington and
Support Center New York for policy changes, funding, implementation and technology transfer.
Additional recommendations include addressing storage areas and procurement practices. Central
warehousing could eliminate loss of paint from .unheated lockers. A centralized purchasing system
would eliminate duplication of supplies with short shelf life. Individual commands need to be more
cognizant of waste generating potential. Accountability would be conducive to better waste
minimization practices.
Technical Evaluation for Waste Minimization
Several technological operations and processes were identified as key areas for consideration
of waste minimization opportunities. These included the following areas with suggested
recommendations:
Maintenance of buoys used as navigational aids is one the Coast Guard's responsibilities.
Buoys require refurbishing every 4-6 years. Old and degraded paint is removed by blasting with steel
shot and several coats of durable paint are reapplied. Frequent painting is the means to preserve
the appearance and protect the integrity of equipment exposed to aggressive saltwater environment.
Currently items are moved into a large blasting room and paint, biological growth and rust
are removed by application of steel shot with a high pressure air gun. The steel shot is recycled ,
approximately five times, until it becomes too fragmented for use. Spent shot is collected, stored
and disposed as a hazardous waste because of low levels of lead. Estimated cost for purchase and
disposal of the shot approximates $38,000 annually.
Equipment is spray painted using a Binks Airless 1 spraygun and a high pressure air system.
Buoys are spray painted with several different paints, including an epoxy anti-foulant with a total of
approximately 5,300 gallons of various coatings used annually. It is estimated that the transfer
efficiency of paints using this equipment is only about 50%
Several options were examined and the following were considered most attractive for the
near future, both from an environmental impact and cost viewpoint.
• Low pressure spray guns to replace the airless guns. The high volume/low pressure spray
gun significantly reduces overspray from an estimated 50% with an airless gun to only 15% with the
HVLP gun. This immediately translates into a reduction in the amount of paint used and,
consequently into a comparable reduction in the amount of VOCs emitted to the atmosphere. In
addition, by reducing the overspray, the sludge buildup in the water curtain will be decreased or the
time between required cleanouts can be reduced. The cost of the new gun system, with its
compressor, is less than $1000 and retraining of operators is minimal, making this a most attractive
option. Closed system spray gun cleaners are now available at relatively low cost ($500). These
systems avoid discharge of solvent to the air while using a minimum amount of solvent. Estimated
payout for the conversion is only 0.5 months.
248
-------
• Replacement of steel shot with plastic media. The use of plastic shot as an alternate
blasting media is an emerging technology and is readily implementable at Governors Island. The
changeover from steel shot to plastic shot can be made with essentially no capital investment. Only
minor adjustments are needed. The changeover would markedly reduce the weight of the dust due
to lighter density material and increase the recycle capability from 5 for steel to 20 for plastic. Lead
from the older lead-based paints will remain a problem until all the buoys have been painted with
the new no-lead paint, however the plastic dust can be incinerated where the steel could not.
Factors of effective rust removal by the plastic media and coating durability on the plastic-media
cleaned buoy need to be considered. This option is highly cost-effective, with payout occurring in
only 3.4 months.
• On-site still for recovery of reusable solvent from spent thinners and waste paints. Waste
paint and solvent thinner generation on the site was recognized as a major source of waste and a
significant contributor to the disposal cost. The key in this area is to effectuate improved
implementation of management systems for the purchase, inventory and distribution of paint and
thinner on the island. Small scale stills were evaluated as a possible alternative to disposal.
Estimates were that from 50% to 90% of the volatile solvents can be recovered from paints and
contaminated solvents. This option, however, is not highly recommended at this time, in light of the
cost and anticipated changes in paint and solvent management.
RESEARCH, DEVELOPMENT AND DEMONSTRATION
No obvious research, development and demonstration projects were recognized during this
opportunity assessment and feasibility analysis of recommended options. This project focused on
management initiatives and applied technology changes.
VETERANS MEDICAL CENTER
CINCINNATI, OHIO
As part of the WREAFS program, pollution prevention opportunities were assessed at the
Department of Veterans Affairs' Cincinnati - Fort Thomas Medical Center (VA-Cin.). This report
serves as a case study for identifiying pollution prevention opportunities in a hospital setting and
focuses on ways of reducing the discarded medical supply wastestream,
FACILITY DESCRIPTION
The Veteran Affairs' Medical Center in Cincinnati (VA-Cin) is a government-owned, general
medical and surgical hospital offering four principal areas of service: medical, surgical, psychiatrical
and neurological. The facility maintains 415 authorized and 342 operating beds and is large in
comparison to other private and federal hospitals. The facility provides outpatient services for
approximately 500 individuals per day. In addition to the medical waste generated on-site, the
facility also manages wastes for an associated research facility, nursing home and home health care
services.
249
-------
AREAS OF WM OPPORTUNITY EVALUATIONS
VA-Cin segregates its waste to minimize the amount transported by the infectious waste
contractor because unit costs for infectious waste far exceed unit costs for disposal of general refuse.
Opportunities were evaluated through a "mass balance" approach, assuming that material entering
a system will be equal to material leaving a system, plus material accumulated. Material balances
allow for realizing losses that may have gone undetected if waste streams were characterized solely
upon disposal information. VA-Cin estimated that approximately 80% of hospital supplies are
disposable and preference for disposables includes cost, convenience, improved quality
assurance/quality control of manufacturing, constraints on space and staff for reprocessing and health
and safety assurance for sterile integrity. The popularity of disposables emerged about fifteen years
ago and the spread of acquired immune deficiency syndrome (AIDS) has escalated this practice
recently.
Infectious waste control is defined by the medical waste handling practices based on the
Center of Disease Control's (CDC) Universal Precautions of body substance isolation. The wastes
are categorized into five groups: general, chemotherapy, blood and body fluids, sharps and
radioactive. Disposal methods include infectious waste contract, toxic waste contract, sewer system,
animal crematorium and autoclaving on-site prior to offsite disposal.
The study revealed that the three largest consumers of disposable supplies in the hospital are
the Supply, Purchasing and Distribution department, the laboratory and the operating room. Eighty-
five to ninety percent of all disposables are attributed to these three areas. These areas were studied
individually and all disposable usage tracked. The two general types of disposable medical supplies
used are plastics and paper (non-woven) products.
Supply. Purchasing and Distribution
The Supply and Processing Department is a central group that distributes supplies to
designated patient wards and services, including the outpatient clinic, recovery room and nursing
services. Supplies ordered from here are termed "posted" and eighty percent of hospital's supplies
are posted or ordered through central supply.
LaboratoryServices
The Laboratory Services Department performs analyses on specimens taken from patients
throughout the Medical Center. It consists of four separate areas: hematology, clinical chemistry,
microbiology and histopathology. All laboratory wastes are currently placed in orange biohazard
bags, autoclaved and disposed in the general trash as non-infectious waste. There are three sizes
of bags: 1 gallon (small), 5 gallon (medium) and 30 gallon (large).
Hematology Laboratory: This group draws and analyzes blood samples from 50-60 patients
per day. Cloth gowns are currently worn by staff when blood is drawn and replaced with a second
cloth gown for work in the laboratory. All gowns are laundered for reuse. This area generates
approximately 2 large bags of autoclaved wastes per day.
Clinical Chemistry Laboratory: Blood sera and urine analyses are conducted in this
laboratory. This area generates 1.5 to 2 large bags of autoclaved waste per day.
250
-------
Microbiology Laboratory: This section generates the greatest amount of disposables by
weight and disposable products comprise 98% of all autoclaved waste. The principal disposable is
glass Petri plates pre-prepared with agar, a culture media. Approximately 3 large bags of autoclaved
wastes are generated daily.
Histopathology Laboratory: This area is responsible for analyzing tissue specimens and body
parts from surgery and the morgue. A limited number of disposable items are used here with no
more than one medium, 5 gallon bag of autoclaved waste generated per day.
Surgery Department
Surgery handles approximately 15 cases per day. The greatest volume of disposable medical
supplies used and disposed of in surgery are lap sponges. Other disposables from surgery include
procedure products that are found in operating room packs. Safety, quality assurance and product
availability are three major concerns providing the impetus for disposable, operating room packs.
Surgery also carefully segregates wastes as they are generated, however in order to increase
efficiency, the amount of segregation may be reduced in the future. Blood and body fluids are
brought to the storage room in the basement and then transported by the infectious waste hauler
to the final treatment and disposal site.
Other Areas of The Hospital
The Surgical Intensive Care Unit uses cloth gowns and launders them for reuse. Blood and
body fluid wastes are strictly segregated into 1-2 large bags per day. Waste generated in patient
rooms is segregated into three categories: (1) sharps, (2) blood and body fluids and (3) general
trash.
Patient Floors include such care as administering medication and changing dressings. Waste
is segregated into same three categories and amounts generated obviously vary with occupation rate
of beds.
The Medical Intensive Care Unit/Cardiac Care Unit normally operates at 100% occupancy.
The two reusable items employed here are cotton gowns and pressure bags.
Hemodialysis unit uses nearly all disposable products. Most are discarded in the blood and
body fluids receptacle.
The Outpatient Clinic provides services to approximately 500 patients per day. Services
provided are surgical procedures, medical exams, chemotherapy, dermatology, urology, plastic
surgery, orthopedics and ear, nose and throat. Ninety percent of the supplies used in the outpatient
clinic are disposable. Plastic-coated paper gowns are used by staff members administering
chemotherapy treatment. Reusable wovens that are commonly used include sheets, pillow cases,
towels and blankets.
The Incinerator is located on the ninth floor of the building. Sharps, pathological wastes and
expired pharmacy drugs are incinerated every Friday. The capacity is not great enough to
accommodate any additional waste, and consequently, the hospital is planning to build one with
251
-------
increased capacity to accommodate all the medical waste, this eliminating the need to contract with
waste haulers.
The Storage Area for blood and body fluid waste and cytotoxic wastes is located in the
basement. These wastes are packaged in large brown plastic garbage bags, transported to the
basement, placed in cardboard boxes lined with red biohazard bags. These are transported offsite
to be incinerated at a commercial treatment/disposal facility. The hospital is charged $0.03 per
pound of blood and body fluid waste transported for disposal.
WASTE MINIMIZATION OPTIONS AND RECOMMENDATIONS
Through a review of available literature, the site visit at VA-Cin and an understanding of the
limitations facing waste reduction in a hospital setting, recommendations are made pertaining to
product substitution, the reuse of disposables and recycling:
• Reuse of Disposables; As the cost of solid waste disposal (including incineration)
escalates, particularly the cost for medical wastes, the reintroduction of reusables may be warranted.
Hospital automation, specifically in the processing and sterilization of soiled linens, over the past few
years is enabling many institutions to reconsider the use of reusable surgical linens as a cost-effective
option to the disposal of paper products. The factors which must be considered when making the
decision to reuse a single-use product include possible contamination, increased liability, decreased
functional reliability, compromised patient safety and the associated costs. It must be determined
if the quality assurance program is compatible with reprocessing disposable items, and if not,
evaluate the economic feasibility to make it so. As an item becomes more critical and the potential
for infection increases, the likelihood that an item will be reused decreases. In the end, safety takes
precedence over economics of pollution prevention in a health care environment.
• Wovens versus Nonwovens: The use of wovens would decrease the volume and weight
of hospital waste significantly. Therefore, employing wovens throughout the hospital should be given
serious consideration and each of the reasons for choosing disposables re-evaluated. Health care
personnel often choose paper products to ensure the sterlity of an item even though wovens, when
laundered at sufficiently high temperatures and sterilized, present an equally sanitary product.
Reusable fabric can be treated and made water repellent, therefore resistant to blood and body fluid
penetration, and density of the treated fabric provides an effective barrier to bacteria. The
advantages of woven material include nonabrasiveness and allowance for freedom of movement. It
is more puncture resistant than .paper and allows for ease of maneuverability and examination.
When costs are integrated, the use of wovens may also represent a better use of hospital resources.
Although, in some cases, paper products will offer a superior basis for use in administering the best
and safest medical care, the universal use of paper products in any health care facility should be
avoided.
• Product Substitution: Plastic covers for pillows can be replaced by using vinyl/nylon
laminate covers. These covers would prolong the life of the pillows, decrease the risk of infection
and reduce waste by continuing the use of woven pillow covers. In some cases within the
laboratories, reprocessing of glassware may prove an economical alternative to plastic disposables.
252
-------
RESEARCH, DEVELOPMENT AND DEMONSTRATION
Although pollution prevention opportunities may appear limited at first, the implementation
of alternatives whenever possible will, in sum, achieve significant waste reduction. In addition,
greater opportunities may be unveiled with further research. Suggestions for research and
development possibilities in the health care industry are presented below:
• Costing: There may be a need to conduct cost studies for certain health care products in
cooperation with other Federal agencies, such as Veterans Affairs and Health and Human Services.
• Quality Assurance: There may be a need to consider working with trade associations and
other Federal agencies, such as the Food and Drug Administration in reviewing technical, legal and
policy impacts of reusing disposables. The ultimate goal would be a protocol for reuse.
• Development of Reprocessing Capacity: There may be potential for reestablishing the
viability of reprocessing, perhaps by stimulating the development of cooperative reprocessing service
centers in areas with a high density of health care facilities.
• Developing Reusable Market: The EPA and VA may want to work together in developing
procurement guidelines for the VA which will stimulate the production and distribution of reusables
and recyclables. This could lead to waste minimization technology transfer opportunities throughout
the health care community.
Officials expressed the belief that it was unlikely that hospitals would convert back to the use
of reusables on a wholesale basis due to concerns over worker health and safety and cost efficiency.
ON-GOING ASSESSMENTS
OPTICAL FABRICATION LABORATORY
FITZSIMMONS ARMY MEDICAL CENTER
DENVER, COLORADO
This project was established to develop waste minimization options for the principal hazardous
waste generating areas at the Fitzsimmons Army Medical Center, Denver, Colorado, Optical
Fabrication Laboratory (FAMC/OFL) installation. It was conducted in cooperation with the
Environmental Office of the Directorate of Engineering and Housing, which has a on-going waste
management program.
FACILITY DESCRIPTION
The Optical Fabrication Laboratory of the Fitzsimmons Army Medical Center, Aurora,
Colorado, is housed in the Charles W. Carter Optical Center. The OFL produces about 1,400 pairs
of spectacles per month, with 85 to 90% of the production involving fabrication of glass lenses. The
remaining 10 to 15% involves plastic lens fabrication.
253
-------
AREAS OF WM OPPORTUNITY EVALUATIONS
The areas involved in this project were the glass and plastic lens fabrication where uncut
lenses are received from optical suppliers and matched with eyeglass prescription orders. The lenses
are precoated with a polymer film of volatile solvents - methyl ethyl ketone (MEK), methanol and
ethanol. The precoated lenses are then blocked, ground to desired curvature, washed and deblocked.
The cleaned lenses are ground to fit frames, chemically hardened and placed in frames.
There are several waste streams produced by the lens production and of these the following
were selected for waste minimization or cost-savings options:
• Fine glass particulates from lens grinding are produced in amounts averaging 300 pounds
per day.
• Spent alkaline wastewater from glass lens deblocking and cleaning operations may contain
suspended alloy particles as well as possibly small amounts of dissolved lead salts.
WASTE MINIMIZATION OPTIONS AND RECOMMENDATIONS
Of the two waste streams studied, the nonhazardous glass fines were examined from a recycle
standpoint, eliminating a disposal cost. Several options were considered for the RCRA suspect
hazardous materials in the alkaline wastewaters.
• Fine glass particulates are possible feedstock for glass or ceramic tile manufacturing.
Transportation costs would limit marketable area as approximately 37.5 tons are produced
per year.
• Three potential options were recommended for the alkaline wastewaters: (1) substitution
of blocking alloy; (2) filtration of wastewater prior to disposal; and (3) removal of dissolved
lead. Of the three, substitution proved to be uneconomical. Installation of a filter on the
spent washwater effluent stream may be economical if sufficient alloy particulates can be
recovered to justify its operation. Onsite formulation of the alkaline washing solution was
explored, however exact ingredients and quantities are unknown at present.
RESEARCH, DEVELOPMENT AND DEMONSTRATION
Three specific needs were identified as a result of this study:
• Develop a milder glass cleaning solution to replace alkaline one.
• Assess the feasibility of developing another less costly alloy containing no toxic metals.
• Development and/or adaption of an aqueous cleaner for tool cleaning operations to
replace Stoddard solvent.
The experience and insight gained during this assessment should be of definite value for a
similar assessment proposed to be conducted at the U.S. Navy's plastic lens fabrication facility in
Yorktown, Virginia.
254
-------
SCOTT AIR FORCE BASE, ILLINOIS
This project will produce an assessment of three operations at Scott Air Force Base, Illinois.
These are circuit board manufacturing, the non-destructive wheel inspection and a painting
operation. The primary focus of the project is the non-destructive wheel inspection which possibly
involves a toxic penetrant used for detection of landing wheel fatigue, such as cracks or other
discontinuities that penetrate the surface of the metal.
EVALUATION OF EMULSION CLEANERS AT AIR FORCE PLANT
NUMBER 6
This project will evaluate the conformance of emulsion cleaners as replacement for
trichloroethylene (TCE) as degreasers in aluminum and steel preparation for manufacturing. This
process is used in production of military transport aircraft. The results can be transferred to similar
operations in DOD and/or DOE and other facilities.
WASTE REDUCTION FROM CHLORINATED & PETROLEUM-BASED
DECREASING OPERATIONS
This project will serve in formulating a model technology service program for DOD's
chlorinated solvents program. Auburn University will ascertain what is required to make state-of-the-
art solvent recycling technology available and minimize the risk to operators, liability and damage
to parts being cleaned.
OTHER RESEARCH AND DEVELOPMENT WITHIN THE FEDERAL COMMUNITY
Other Pollution Prevention Research Branch research projects with the Federal Community are:
• Wet to Dry System Evaluation in a Navy Paint Spray Booth
This project will evaluate the conversion of a Navy paint spray booth from a water curtain
particulate emission control technology to a dry filtration technology.
• Reclaiming Fiber from Newsprint
This project, funded under an interagency agreement with the United States Department of
Agriculture's Forest Products Laboratory in Madison, Wisconsin, is designed to investigate the
potential for newsprint reclamation through a dry fiberizing process.
• Composites from Recycled Plastics. Wood and Recycled Wood Fiber
This is a three-year interagency agreement between the Forest Products Laboratory of the
U.S. Department of Agriculture and EPA to investigate and develop wood/plastic composites.
255
-------
NEWLY PROPOSED PROJECTS
• Waste Minimization Assessments and Reviews Within The Federal Community
A waste minimization assessment is being conducted at the Department of the Treasury's
Bureau of Printing and Engraving in Washington, D.C., where the principal waste generating
activities result in metal and ink wastes.
Waste Minimization Assessments are planned with:
i/ The U.S. Navy and the City of San Diego to consider joint pollution prevention options.
S The Department of Interior's Bureau of Mines.
S The Department of Agriculture's Agriculture Research Service in Beltsville, Maryland.
S The U.S. Army's facility in Ft. Carson, Colorado.
S The Military Facility Model Community Pollution Prevention Demonstration Program within
the Chesapeake Bay.
* USCGS Ketchi'kan Pollution Prevention Project
This project will provide support to the Alaska Department of Environmental Conservation and
the United States Coast Guard Service to assess pollution prevention opportunities at the USCGS
facilities at Ketchikan, Alaska.
MATRIX OF INTERCHANGE
Table 1 describes the Pollution Prevention Research Branch's matrix of interchange with the
Federal Community. This represents completed, on-going and scheduled projects.
256
-------
POLLUTION PREVENTION RESEARCH MATRIX OF INTERCHANGE
IN THE FEDERAL COMMUNITY
(COMPLETED, ONGOING AND SCHEDULED PROJECTS)
DEPARTMENTS
AGRICULTURE
COMMERCE
DEFENSE: AIR FORCE
ARMY
NAVY
EDUCATION
ENERGY
HEALTH & HUMAN SERVICES
HOUSING & URBAN RENEWAL
INTERIOR
JUSTICE
LABOR
STATE
TRANSPORTATION
TREASURY
VETERANS AFFAIRS
ASSESSMENT
X
x
X
X
X
X
X
X
X
RD&D
X
X
X
X
TECH TRANS
X
X
X
X
X
X
X
X
X
SUMMARY/PROJECTIONS
Assessments have been initiated and have ongoing or completed research with seven of
fourteen of the Departments. We anticipate further cooperative projects with these federal facilities
and encourage the other seven Departments to seek cooperative participation in research projects
for pollution prevention in industrial processes. Each of these assessments can be used as reference
technologies in similar areas and each source reduction of a toxic waste will be a positive step toward
environmental protection.
257
-------
TOXIC SUBSTANCE REDUCTION FOR NARROW-WEB FLEXOGRAPHIC PRINTERS
by
Paul M. Randall
U.S. Environmental Protection Agency
Risk Reduction Engineering Laboratory
Cincinnati, Ohio 45268
Dr, Gary Miller and W, J. Tancig
Illinois Hazardous Waste Research
and Information Center
Champaign, Illinois 61820
and
Dr, Michael Plewa
Institute for Environmental Studies
University of Illinois
Urbana, Illinois 6I80I
ABSTRACT
This project is one of five undertaken as part of the Illinois WRITE (Waste Reduction Innovative
Technology Evaluation) Program, and has quantitatively evaluated, for one narrow-web flexographic printing
firm, the amount of waste reduction (both volume and toxicity) and the economic impact resulting from
modification of a traditional technology. Two main changes in the printing process were: substitutuion of
water based inks for solvent based inks and substitution of a nontoxic liquid cleaner (terpene based) for a
haiogenated solvent cleaner. The paper presents an in-plant evaluation and the impact of these changes
on environmental, health, cost, and other factors.
258
-------
INTRODUCTION
This project is one of five undertaken as part of the Illinois WRITE Program (Waste Reduction
innovative Technology Evaluation), and has quantitatively evaluated, for one narrow-web flexographic
printing firm, the amount of waste reduction (both volume and toxicity), and the economic impact resulting
from modification of a traditional technology. Two main changes in the printing process of this company
were considered for evaluation in this project:
1. Substitution of water-based inks for solvent-based inks.
2. Substitution of a non-toxic liquid cleaner (terpene based) for a halogenated solvent cleaner.
The report is based on work performed under terms of the Illinois/EPA WRITE Program.
Specifically, the project has been a joint effort of MPI Label Systems, University Park, Illinois, the Hazardous
Waste Research and Information Center, Illinois Dept. of Energy and Natural Resources. Champaign, Illinois,
and the U.S. EPA Office of Research and Development, Cincinnati, Ohio
All project testing was conducted in the printing plant of MPI Label Systems, Inc., University Park
(Monee), Illinois. The facilities are housed in a modern one-story, clear-span building of about 15,000 square
feet, half of which Is used for storage of supplies, and half for the actual printing operation and the
administrative offices.
MPI has been in operation at the current site slightly more than two years, having moved from
an older nearby location. The firm is one of eight separate corporate printing plants and has always relied
on narrow-web flexography to produce a wide variety of labels. Flexographic printing derives Its name from
the flexible, roll-mounted printing plates used, as opposed to the classical non-flexible metal printing plates.
Several of the press lines have the capacity to print up to eight colors per label. Each line is operated by
259
-------
a single individual who is responsible for ail steps of a complete label run.
During 1988 the parent organization mandated that all 'As plants eliminate, as quickly as practical,
every toxic material then in use. Further, all future operations were to avoid use of toxic and hazardous
materials, even barring storage of small quantities of such materials in each plant. Part of the motivation
for this mandate was concern about the exposure by their employees to alcohol fumes in the work area.
The first modification, and the most significant, was to change to water-based inks as a substitute
for the then common alcohol-based inks. A related change, though Instituted primarily for technological
improvement and not waste-related, was the substitution of newly developed plastic printing plates for the
older type rubber printing plates. Water-based inks did not produce satisfactory images with rubber printing
plates. Thus, the development of plastic printing plates was necessary before water-based inks could be
adopted. The plate making operation is not a part of this evaluation because at MPI the plates are not made
in house.
The second step involved elimination of the several halogenated solvents used to clean presses
at the end of a run. Before these changes, the solvent- and ink-soaked wipers resulting from cleaning had
to be disposed of as hazardous waste materials, an expensive handling and disposal operation. To make
this particular change, a variety of relatively new cleaning agents were examined. These agents ranged from
Industrial detergent cleaners to terpene-based mixtures. This step continues to be one of periodic
reevaluation as more satisfactory cleaners become available. Although IvlPl now does all its cleaning with
a dilute aqueous solution of detergent, it is prepared to look at any promising product.
PRINTING PROCESS BACKGROUND
The Printing industries of America (1991 )1 estimates there are about 57,000 printing, publishing
and related facilities within the United States. Of these about 40,000 are commercial printers. The remainder
Include newspaper and magazine publishers, photocopiers and in-house printers. The five most common
printing processes in order of their market share are lithography (also called offset), gravure, flexography,
260
-------
letterpress and screen. Presses are also categorized according to whether they print on individual sheets,
called sheet-fed, or print on a continuous roll, called web, of paper or other substrates.
The printing process begins by making an image, including words and designs, of what is to be
printed. For each color that is used, a different image is made. These images are made by various
processes on various types of metal, rubber or plastic plates. Common image-making processes are similar
to developing a photographic image. Recently developed plastic plates are made of photosensitive
polymers. The image to be printed is exposed onto the plate, and the exposed polymers harden. The
unexposed and softer areas of the plate are washed away with various solvents, leaving a raised image for
printing. Recently water-based plate developing systems have been developed for some segments of the
industry.
The next step in the printing process is to apply ink to paper or some other substrate. The
common printing processes listed above accomplish this step in different ways, however, they generally
function in much the same manner.
Die
Cutter
Waste )
I»prc»-)
sioa
Label*
Figure I. Single print-itucion tchnicic.
Figure 1 presents a simplified schematic of a single print station. The schematic shown in Figure
261
-------
1 ts drawn to represent a typical fiexographic arrangement. However, ft generally represents most other
printing processes. Ink from a reservoir is picked up by a roller in contact with the ink. That roller then
contacts another roller (sometimes several dozen rollers are involved) to develop an ink film of ideal
thickness for transfer to the printing plate. The base material on which the printing is to be imposed (usually
paper, fabric or plastic) is pressed against the printing plate by another roller to receive the plate's ink
Impression. Rollers are used so the entire process can be continuous (using either sheets or roils of paper),
not Intermittent as had been the case from Gutenberg's time (15th century) till the early 20th century when
rotary presses were developed.
The printing station shown in Figure 1 is limited to a single color ink. Multicolor printing uses
several printing stations in series, each applying a different ink color with controls to assure perfect
registration, all assembled to form a single print line. In Figure 2 a four-color flexographic press is shown.
After each inking station, the labels pass through a drying station which dries the ink in a few seconds.
Each station contains heaters maintaining a temperature of approximately 70" C, Depending on the type
of material receiving the plate's impression, and the customer's desire, the dry printed surface may be
sprayed with a coat of gloss varnish or plastic to protect it. At the end the completed labels are wound onto
a roll. The speed or rate at which paper is fed through the press can be varied, depending on the type and
amount of Ink used.
Inlc
area
\
\
Oryer \
\
«.
Ink
area
/
J
;
->
»
,.*
\
\
\
V
Dryer v
\
i
J
Ink
area
-
"N
^
\
\
\
V
Dryer \
\
t
Ink
area
/
/
/
J
~--Xr._
^^V ~"
* s.
V.
Dryer
FIGURE 2. Typical 4-color flexographic printing press.
262
-------
Within the flexographic printing industry, press types are separated into narrow-web (paper up
to 16" wide) and wide-web(over 16" wide). Narrow-web is typically used to print labels, packaging,
envelopes, and plastic bags. Wide-web presses are used for newspapers, wallpaper, soft-drink cartons and
similar surfaces. Most of the MPI presses (narrow web) use a web that is 9" wide.
Paper label stock for printing is made of two-ply material. The bottom ply consists of silicone-
treated paper. The top ply is the paper on which the labels are printed, and ft clings to the bottom ply by
virtue of an adhesive able to be peeled away from the silicone treated base stock. After printing and drying,
possibly receiving a gloss coat, each label is "parted" (cut) or readied for removal from the parent paper web
along each label's entire perimeter by a steel-die roller cutter. The steel-die, working to very close
tolerances, cuts through the single top layer of paper to separate the finished label from the surrounding
unprinted stock. Although the latter (stock waste) is peeled away automatically onto a separate roll for
disposal by shipment to a landfill, the entire label industry is presently seeking a recycler who can handle
this material with the adhesive on one side. The bottom ply, with adhering labels, is made up into rolls of
a specific label count for each customer, then packed and shipped. The eventual customer will remove the
finished labels for application either automatically or manually.
The most recent industry survey (Rexographic Technical Association(FTA), 1989)2 concluded that
over 4,000 U.S. printing plants utilize more than 22,000 narrow-web flexographic presses, employ about
150,000 individuals, and generate about $4.5 billion annually of product. Annual growth over the past
decade is estimated to be 3 - 5%.
Flexo inks were originally formulated to resemble the inks used in other printing processes.
Hence, they were usually compounded of colored metal compounds for pigments, a quick-drying oil vehicle
in which the coloring substance was dispersed, plus an organic solvent to control the ink's viscosity and
speed of drying. The organic solvent was frequently of the trichloroethane family because of Its rapid
evaporation rate. As the hazards of these solvents were finally recognized and accepted, substitute solvents
were sought as safe replacements. Aliphatic alcohols, and their derivatives, are now used as solvents in
263
-------
many current solvent-based inks. At the end of 1989 It was estimated that 90-95%3 of all flexographic label
printers had changed to water-based inks. Demand for the latter amounts to more than 300,000,000 pounds
of flexo ink per year, excluding the newspaper industry.
MPi's water-based inks contain organic chemical pigments, no heavy metals, an emulsifier, an
acrylic resin thickener similar to that found in water-based house paints, up to 5% of isopropyl
alcohol and SO - 65% water. Flexographic inks have a viscosity which approximates that of ordinary white
on or glycerine. By contrast, the inks used on offset presses possess a viscosity closely resembling warm
taffy. Although the waste reduction aspect of this new ink technology has not been completely and
quantitatively documented, the flexo industry's success with water-based inks has spurred other printing
firms, notably some using the rotogravure process, to change to water-based inks.
Because the cleaning agents used most frequently by printers were often composed of hazardous
organic solvents, the solvent-wetted wiping materials used as sponging pads also had to be classified as
hazardous waste and disposed of In hazardous waste landfills at relatively high cost. At MPI the fiber wipers,
In addition to the hazardous solvent cleaners, have now been abandoned in favor of fabric shop towels
which, when wetted with the current detergent cleaning solution, can be laundered and reused - an
additional economy.
WASTE REDUCTION IN PRINTING
Wastes are produced at each step in the printing process. Solid and liquid wastes generated in
making an image on a plate include damaged plates, developed film, photographic chemicals, silver (rf not
recovered), and the solutions used to develop the plates. Spent photoprocessing chemicals are generally
regarded as being biodegradable and are usually discharged to the sewer. Certain solutions may contain
mercury compounds which require special handling. Plate-making wastes can include acids and alkalis
which also may require special handling. Depending on the specific materials used, some solvents may also
be released to the air during image making and plate processing.
During the printing process, solvents in the inks (e.g., alcohols, aliphatics, ketones or xyienes)
264
-------
and cleaning solutions are released to the air. Most of the ink solvents are evaporated while drying the inks.
The amount of solvents released depends on the ink formulation and the type of printing process being
used. Solvent emissions from commercial gravure applications are of particular concern. Most inks used
in lithography dry by oxidative polymerization which produces little solvent emissions. There are also non-
heatset web inks which dry by absorption in the newsprint substrate. In lithography, fountain solutions are
used which contain 5-15% isopropyl alcohol (IPA), defoamers, fungicides, gum arable, phosphoric acid and
water. The alcohol evaporates along with the water. Fountain solutions have recently been developed that
use soaps and detergents instead of IPA, Point source control technologies (catalytic or thermal
incineration, carbon adsorption, or cooler/condenser systems) have been installed in some plants to capture
these solvent releases.4"7
A wide variety of cleaning solutions are used by printers. These cleaners can contain volatile
organic chemicals such as benzene, toluene, kerosene, naphtha, methanol, trichloroethane and methylene
chloride.
Waste inks are the primary liquid wastes generated in the printing process. Most inks can be
recycled such as by blending to make a black ink, either by the ink manufacturer or in-house. Waste inks
that contain organic solvents may be classified as hazardous wastes and must be properly incinerated or
landfiiled. Some nonhazardous inks are sent to the sanitary sewer. Small amounts of lubricating oils are
the other liquid waste that is generated from operation of the printing presses. The used oil can often be
recycled.5
Waste paper is the main solid waste generated In printing. In lithography, almost 98% of the total
waste generated is reported to be waste paper and paper scrap. This comes from rejected or off-quality
runs, scraps from the start and end of runs, and overruns. The paper consumed as waste during setup is
mostly a function of the press operator's experience, complexity of the label to be printed, the number of
ink colors required, and the size of the label. Operators try to use paper stock of such a width as to
minimize trim waste. Other solid waste produced includes empty ink containers and cleanup rags or wipes.
265
-------
Some printers dispose of their rags In the trash while others have the rags dry-cleaned for reuse. The
sludge produced In the dry cleaning process contains the materials removed from the rags (inks, cleaners,
oil. dirt and other contaminants) and this sludge may require disposal as a hazardous waste.5
Options for reducing waste generated by printers have recently been reviewed by the State of
California8 and the USEPA5. Methods for waste reduction in materials handling and storage, image
processing, plate making, printing, and finishing have been listed. Techniques which can be used to reduce
waste during printing include using less hazardous inks and cleaners plus generally being more careful
during setup and cleaning. Waste ink was reduced in one case by spraying a protective coat over the ink
In the fountain at the end of each day. As a result, waste ink was reduced by 5 pounds per day. Thus, less
waste ink would need to be disposed of and less new ink purchased. The total operating savings were
estimated at $3,375 per year for this technique8. Techniques for reducing waste paper include installing
break detectors and automatic splicers in web operations.
Changing to less hazardous inks is not always straightforward. Water-based inks generally
require more energy to dry than do inks with a high solvent content, though this is offset (can run 10% faster
than alcohol inks) by their quicker absorption on a paper base. Other reported disadvantages of water-
based Inks are a need for more frequent equipment cleaning, and a tendency to cause the paper to curt;
many are also low-gloss. Another alternative which can reduce solvent emissions is the use of UV inks.
These inks set or harden when exposed to UV light. Disadvantages of these inks include higher cost,
formation of ozone, hazards of UV to personnel, difficulty in recycling the printed paper, and high toxicity
of some of the chemicals in the inks. Eiectron-beam-dried inks are also available that contain no solvents,
but operator protection from X-rays created by the process is required, and the system often degrades the
paper5.
Three waste reduction case studies of printing plants were recently conducted for the state of
California5'8, At one facility the option of installing In-house recycling equipment was evaluated. It was
found for this plant that in-house recycle would pay back in 8 months. These case studies reviewed
266
-------
operations at two other plants and described practices already in place to reduce the amount of waste being
generated. In-plant measurements of the waste being generated were not taken in these studies. Instead,
the amount of materials used (ink, paper, fountain solution, etc.) was reported5'8.
A case study of switching from alcohol-based inks to water-based inks on low density
polyethylene film by a flexographic printer was presented by Makrauer9. He reported considerable difficulty
in making the conversion. Technical difficulties included pH control for the ink, a need to modify the drying
equipment, ink metering modifications and increased roller wear. All problems were gradually overcome
through improved Ink formulations, experimentation, and facing the rollers with more durable materials.
Makrauer also reported that cleaning water-based inks was more difficult. Benefits of water-based inks at
that time were compliance with environmental regulations, improved color control, greater coverage yield,
and improved working conditions due to reduced alcohol vapors. Quantitative measurements of emissions
and other wastes generated when using alcohol-based inks compared with water-based inks were not
reported. In fact, a quantitative evaluation of the benefits of using water-based inks in flexographic printing
has not been published.
It is worth noting, however, that though printing on plastic materials via flexography had some
problems early in the introduction of water-based inks, using the same technology for newspapers has been
considerably simpler. Many of the major newspapers have already transferred to this type of printing with
excellent results. In several recent installations the printing line has been designed so there are no ink
wastes. Excess ink and ink washings are collected in a holding tank and used to dilute new ink to the
proper viscosity. This closed-cycle system, of course, does not reduce any of the paper wastes. Although
none of these newspapers have published an economic comparison of the old versus the new system,
private discussions with plant managers confirm each plant is producing a better product at less cost.
PROJECT OBJECTIVES
This project had two main objectives: compare the volume and toxicity of any wastes
released to the air, liquid wastes and solid wastes generated during printing before and after switching to
267
-------
water-based Inks and a detergent cleaner; determine the cost impact on the company as a result of the
process changes.
Solvent loss by emission from the inks was estimated by materials balance, laboratory
measurement of alcohol evaporation from the two types of inks, and calculations based on the composition
of the inks. The methods used to compare emissions from the two ink types are summarized in Table 1.
Since MPI no longer uses alcohol-based Inks, and will not permit their use in the plant, it was not possible
to measure alcohol emissions by in-plant use of those inks. However, the volume of alcohols evaporated
from the two types of inks during a printing run can be calculated from known ink formulations if the total
amount of ink used is known. In-piant measurements of ink and cleaner usage were taken for two single-
color printing runs to obtain an estimate of variability.
Table 1. Methods Used for Estimation of Ink Emissions
TYPE OF INK
Water-based
Alcohol-based
MATERIALS
BALANCE
(In-plant)
X
-
EVAPORA-
TION RATE
LOSS
X
X
INK COMPOSITION
(Calculation)
X
X
For the materials balance method, the weight of ink used during each printing run, U, was
calculated by weighing the various items before and after the printing runs using the following formula:
U - (A + W) - (R 4- I + S) (1)
where, A = weight of ink in reservoir and weight of reservoir at beginning
W = weight of water and other materials added during run
R = weight of ink returned to reservoir and weight of reservoir at end of run
I = weight of ink retained in ink pan, plus gaskets, at end of run
S - weight of ink lost by spilling
Mass measurements were taken on an electronic balance (capacity: 12 kgs,, ± 0.1
gm) which was transported to the printing site. The weight of alcohol lost upon drying to constant weight
268
-------
in the laboratory was also determined gravimetrically. The percent solids of both ink types were determined
by drying known weights (approximately 100 gms in triplicate) of ink samples in an oven at 70° C, the drying
temperature maintained on each flexo press.
Laboratory measurements of evaporative loss of solvents for each ink were used to estimate the
percent volatiles. The weight of evaporative loss, E, was calculated using the total weight of ink used (U
from equation (1)) as follows;
(U x % voiatilized)/100 = E (2)
The weight of ink that was retained on the labels (Q) was calculated by:
Q = U - E (3)
The amount of ink on waste labels was then estimated by the proportion of total labels printed to good
labels sold as product, or not wasted.
These results were compared with the weight of volatiles in each ink as reported on the material
safety data sheets. This information is shown in Table 2. Both inks have similar amounts of total volatiles.
The alcohol-based ink contains six volatile components, four of them being alcohols. Ethyl alcohol and
isopropyl alcohol are present in the largest amounts. By comparison, the water-based ink contains four
volatile components. Most of the volatiles are water and isopropyl alcohol. Some of the water (about 24%)
is bound to the resins and does not evaporate upon drying. Both the solvent cleaner previously used and
the detergent cleaner contain over 97% volatiles.
The amount of Ink and other materials that were disposed of as liquid waste was determined
gravimetrically for the two printing runs. No liquid ink wastes were sent to the sanitary sewer prior to using
water-based inks. The solvent-based waste ink had to be disposed of as a hazardous waste and was thus
manifested. While the total amount of waste manifested in a year was available from company records, it
was not possible to determine the amount of liquid solvent-based ink that would have been generated from
printing runs similar to those we evaluated with the water-based inks. Company officials reported that in
their experience the amount of solid and liquid wastes generated are essentially the same for the two types
269
-------
Table 2. Composition of Inks and Cleaners Evaluated10"13
COMPOSITION
Material
Component
Percent by wt.
INKS
Alcohol-based Ink
(Manufacturers data)
Water-based ink
(Manufacturers data)
Methyl alcohol (v)
Isopropyl alcohol (v)
n-propyl alcohol (v)
Ethyl alcohol (v)
Ethyl acetate (v)
VM&P naphtha (v)
Resins (nonv)
Pigment (nonv)
Volatiles
Isopropyl alcohol (v)
Ammonia (v)
Dimethylethanolamine (v)
Acrylic resin (nonv)
Azo pigment (nonv)
Water (v)
Volatiles
4.7
10.6
6.5
21.4
4.2
6.6
Unknown
Unknown
54.5
5.0
1.0
1.0
20.0
8.0
65.0
56.5
CLEANERS
Solvent cleaner
(Manufacturers data)
Detergent cleaner
Toluene (v)
Acetone (v)
Isopropyl alcohol (v)
Diacetone alcohol (v)
Volatiles
Volatiles
54.5
20.0
20.0
5.5
99.+
97.8
(v) = volatile; (nonv) = nonvolatile. NOTE: both ink types contain plastic-based resins which
react and bind with some of the other materials present on drying. Hence, one cannot simply add the
volatile percentages to obtain total volatiles.
270
-------
of inks. The main difference is that liquid wastes from the water-based ink do not have to be disposed of
as a hazardous waste.
Currently, each printing line has a 50 gallon drum of water used to rinse off ink wastes. Each
drum is emptied every week into a commercial ink removal/filtering device called an Ink Splitter. This unit
absorbs the colored pigments on cellulose fibers, and the slightly grayish filtrate is run to the sewer - as
approved by the local treatment plant. The colored absorbent is acceptable in landfills as non-hazardous
material, primarily because none of the ink pigments contain metals of any type.
Ink wastes on the press rollers, pans and plate are removed by scrubbing with a brush and fabric
towel wetted with an aqueous detergent solution. This quickly removes the ink residues. The rollers, pans
and plates are then dried with another fabric towel. The towels are rinsed in the barrel of water at each
press, then sent to an industrial laundry service for cleaning. By appearance, a negligible amount of ink was
retained by these towels. The amount of detergent cieaner used was measured for each run (although this
depends largely on the press operator's general practice), but it was not possible to measure amount of
solvent cleaner previously used. Hence, the toxicity of the two cleaners was compared using the degree
of hazard system.
DEGREE OF HAZARD REDUCTION
Toxicity reduction evaluations on the ink and cleaner wastes were accomplished with the Degree
of Hazard14 scheme by calculating the equivalent toxic concentration (Ceq) as follows:
Ceq = A SUM (C,/B,T,) (4)
where,
SUM means the sum of the results of the calculation in parentheses for each component
of the waste stream.
Cj is the concentration of component i as a percent of the waste by weight.
T| is a measure of the toxicity of component i,
A is a constant equal to 300. It is used to allow entry of percent values for Cj, and to adjust
271
-------
the results so that a reference material, 100% copper sulfate, with an oral toxicity of 300 mg/kg, achieves
an equivalent toxicrty of 100,
B, Is a conversion factor used to convert toxicities (T;) to equivalent oral toxicities, Bj is
determined from Table 3. For carcinogens and mutagens, a TDgg oral rat is used when available.
Otherwise, carcinogens are assigned a T, of 0.1 mg/kg; and mutagens are assigned a Tj of 0.6 mg/kg.
Toxlcfties are converted to equivalent oral toxicities as specified in Table 3. The equivalent toxicity given
in this Table has the same lexicological response as referenced In the RCRA listing criteria15.
Table 3. Conversion Factors
Conversion factors for the equivalent oral toxicities (Bj) -
TOXICITY MEASURE
Oral - LDgg
Carclnogen/mutagen - LDgg
Aquatic - 48 or 96 hr - LCgg
Inhalation - LC^
Dermal - LD^
UNITS
mg/kg
mg/kg
ppm
mg/l
mg/kg
B,
1.00
1.00
5.00
25.00
0.25
Toxicity values are ranked by source according to the following priorities, with the best sources
listed first: oral rat; inhalation rat; dermal rabbit; or, aquatic toxicity and other mammalian toxicity values.
If there Is more than one value for the toxicrty from the best available source, the lowest (most
toxic) equivalent oral toxicrty value is used. If a carcinogen or mutagen is assigned a value for Tj In the
absence of a TDgg, Bf Is assigned a value of 1.
The toxic amount, M, is calculated as follows:
M = SCeq (5)
where, S is the maximum size of a waste stream produced in kg/month.
Ceq is the equivalent concentration from equation (4).
272
-------
The result of these calculations will be an estimate of the toxic amount (M) for each ink and
cleaner type generated during a printing run. This toxic amount takes into account the toxicity and amount
of each component of the inks and cleaner. The toxic amount, for M, can range from 0 to greater than
10,000. This toxic amount can be considered a relative toxicity of each ink and cleaner type. These relative
toxicrties can then be compared for the air, liquid and solid wastes produced while printing with the two
types of inks and cleaners. At MPI Label Systems, no waste was disposed of in wash water prior to the ink
t
substitution, so any determination of the amount and toxicity of waste disposed of in waste water will be
absolute.
The economic analysis of these various technology changes is based on comparison of the
factors shown in Table 4. As far as possible, monetary values are based on annual costs. This is the only
valid approach since no capital investment was required; hence, such terms as annual rate of return and
payback are not applicable. The factors listed in Table 4 were selected after a tour of the plant and
discussions with the plant manager.
Table 4. Summary of Cost Comparison Factors Evaluated
COST COMPARISON FACTORS
INKS
Print speed
Raw materials
Waste disposal and handling
CLEANERS
Disposal
Raw materials
OVERALL
Insurance liability
Inventory
273
-------
RESULTS OF ON-SITE TESTING
Two individual label runs were evaluated at widely separated times, with different size and different
colored labels, different operators and with different label totals.
1. Green labels. Approximately 3.25 X 13 Inches, printed with green ink (yellow plus blue) on
non-glossy white stock for a total run of approximately 55,000 labels.
2. Purple labels. Approximately 0.75 X 1.75 inches, printed with purple Ink on glossy white stock
for a total run of approximately 250,000 labels.
In each case the total weight of materials added (equation 1) and the weight of materials
remaining at the end of the run was measured. The difference was the weight of material that was assumed
to be either evaporated to the shop air, dried on the labels, or wasted.
During the printing of the green labels, an ink pump was used to increase the size of the ink
reservoir. Ink was continually recirculated between the ink pump and the ink pan. The weight of the ink
pump and Ink contained in it was determined at the beginning before any labels were printed (A in equation
1). During the course of that run water was added to the ink to adjust the color and viscosity on 9
occasions totaling 842.7 grams (W in equation 1). At the end of the run the rollers, plate, and pans were
cleaned and the ink drained back into the ink pump for future use. The ink pump with all materials added
during the printing was then weighed (R in equation 1). One spill (S in equation 1) occurred during this run.
The Ink pan and gaskets adjacent to the roller were weighed before and after the run to measure the amount
of Ink retained on them after they were scraped {I in equation 1). During the printing of the purple labels
nothing was added, and there was no loss due to spillage.
The total weight of ink used during the printing of these two labels is shown in Table 5, For the
two water-based inks the total amount used was calculated according to equation 1. To estimate the
amount of ink volatilized the % loss as determined by laboratory evaporation of each ink was used
according to equation 2. The laboratory evaporation results are shown in Table 6. The amount of solids
retained on the labels was calculated by the difference between the total ink used and the weight
evaporated.
274
-------
Table 5. Ink Used and Estimated Emissions
Total ink used,
(grams)
Ink solids retained
on labels (grams)
Weight evaporated
GREEN LABELS
Water-based Ink
1 ,458.9
609,8
849.1
Solvent Ink
1,175.0
609.8
565.2
PURPLE LABELS
Water-based Ink
399.2
152.1
247.1
Solvent Ink
293.1
152.1
141.0
Table 6. Weight loss data from laboratory evaporation at 70° C.
Material
Black water based
ink
Green water based
ink
Purple water
based ink
Black alcohol
based ink
I Detergent cleaner
Averages of triplicate runs
Initial wt.
12.17 gms
12.05
12.22
15.38
12.56
Dry wt.
5.303
5.04
4.66
7.99
0.278
Wt. loss
6.867
7,01
7.56
7.39
12.28
% loss
56.4
58.2
61.9
48.1
97,8
% LOSS, Std
Dev
0.42
0.21
0.20
0.60
0.03
To estimate emissions that would have resulted from using solvent based inks, it was assumed
that the same amount of solids would have been used for the printing of the labels as was used for the
water-based inks. Then the total amount of ink that would have been used and the weight evaporated was
calculated by using the percent loss factor determined in the laboratory (Table 6). Since less percentage
of solvent-based ink was lost to evaporation, more solids per gram of ink would be applied to the labels than
with the water-based ink. Thus, less total solvent-ink would be used and less total weight of component
would be lost via evaporation. According to the operators at MPI Label Systems, they estimate that about
275
-------
the same total amount of ink is required for a job for either type of ink. Thus, this analysis of emissions is
conservative for the solvent-based ink. It should be noted that laboratory evaporation loss results given in
Table 6 agree favorably with the total volatiles data given in the material safety data sheets (Table 2).
The next step is to estimate the weight of emissions of each specific component of the inks
studied. These estimates are presented in Table 7. These estimates were made using the percent
composition data from Table 2. For the water-based inks it was assumed that all of the alcohol, ammonia
and amine evaporated and that the remainder of the loss was water.
Table 7. Amount of Component Emissions Estimated for each run and Type of Ink (grams)
Component
Isopropyl alcohol
Ammonia
Dlmethylethanol-
amine
Water
Methyl alcohol
n-propyl alcohol
Ethyl alcohol
Ethyl acetate
VM&P naphtha
TOTAL
GREEN LABELS
Water-based Ink
72.9
14.6
14.6
747.0
-
-
-
-
-
849.1
Solvent Ink
110.9
-
-
-
49.2
68.0
224.0
44.0
69.1
565.2
PURPLE LABELS
Water-based Ink
19.96
3.99
3.99
219.16
-
-
-
-
-
247.1
Solvent Ink
27.67
-
-
-
12.27
16.97
55.87
10.97
17.23 •
141.0
Liquid wastes were generated only during cleanup at the end of the press runs. These wastes
were minimal and consisted of ink left in the pan and on the rollers, gaskets and plates at the end of the run
after scraping and detergent cleaner. For the green run, 116.6 grams of ink remained in the pan. A total
of 44.3 grams (about 44 milliliters) of cleaner was used. All of this was disposed of as waste water for a total
of 160.9 grams. During the cleanup of the purple labels only 56.4 grams of liquid waste was produced.
There was a more experienced operator for this run which resulted in less cleanup being required and less
wastage.
276
-------
During these two runs most of the ink retained on the fabric rags resulted from a spill that
occurred during printing of the green labels. A total of 24.6 grams of ink was cleaned up as a result. These
rags would have been sent to an industrial laundry for cleaning and reuse. Thus, this spilled ink would also
result in a liquid waste.
ECONOMIC EVALUATION
The approximate annual savings at MPI Label Systems as estimated by the plant manager is
summarized in Table 8. They have observed no significant difference in costs between the two types of inks
or cleaners being used. They can print on the same type of paper so there is also no difference in that
regard. Although for most labels and paper stock the press speed with water-based inks can be increased
about 10% over that with solvent-type inks, the economic impact of this slight increase in printing speed is
difficult to quantify. At least the rate of printing can be slightly increased as a result of adopting the water-
based inks. The company also reported that there is no difference to them in the cost of the printing plates.
The savings in waste disposal and handling results from the fact that the waste inks and cleaners
no longer have to be disposed of as a hazardous waste, but can be released to the sanitary sewer. The
other savings resulted from a reduced insurance rate when the company stopped using solvent-based inks
and cleaners The rationale was that the work environment was improved for the employees.
A significant off-setting factor is that the company decided to install a unit to treat its waste ink
prior to discharge to the sewer. This filtration unit removes most of the color. The capital cost of that unit
was about $18,000. The colored absorbent is acceptable at the local municipal landfill. Since this treatment
unit was not required as part of the change to water-based inks, its purchase and operating costs were not
included in this analysis.
277
-------
Table 8. Summary Economic Factors
COST COMPARISON FACTORS
INKS
Raw materials
Printing speed
Waste disposal and handling
CLEANERS
Raw materials
Disposal
OVERALL
Insurance liability
Inventory
Wiping materials
Total Annual Savings
SAVINGS WITH WATER-BASED INKS
None
Approximately 10% faster
Minimum annual savings =
SAVINGS WITH AQUEOUS
$10,000
CLEANER
None
Minimum annual savings =
$5,000
Approximately $500/yr.
None
Annually at least $1 ,000
At least $16,500.
DEGREE-QF-HAZARD ANALYSIS
The degree-of-hazard evaluations were conducted for three printing scenarios. The total mass of each
scenario was set at 600 g/SO,000 labels printed with a total usage of 1,000 gal or approximately 3,500 kg
of material each. The large volume of material was necessary in order to conduct a degree-of-hazard
analysis because of the RCRA Small Generator limit. The relevant analysis is the combined component
equivalent toxic concentration for each printing type run.
» Scenario Nfi 1 employed solvent-based ink with solvent-based cleaner.
» Scenario N° 2 employed solvent-based ink with detergent cleaner.
• Scenario Nfl 3 employed water-based ink with detergent cleaner.
278
-------
Table 9. Basis for Degree-of-Hazard Evaluation for Three Printing Scenarios
Component
Amount
(%) of
Ink or
Cleaner
Total (g)
per Sce-
nario
(600 g)
%of
Total
Scenario
Comments
Solvent-Based Ink with Detergent Cleaner
Black Ink = 400 g + 200 g Cleaner/50,000 Labels Printed
Polyamide Black
Methyl alcohol
Isopropyl alcohol
N-propyl alcohol
Ethyl alcohol
Ethyl acetate
VM&P naphtha
46.0
4.7
10.6
6.5
21.4
4.2
6.6
184.0
18.8
42.4
26.0
85.6
16.8
26.4
30.7
3.1
7.1
4.3
14.3
2.8
4.4
Considered as an innoc-
uous toxic hazard16. Mr.
Fishman, Sun Chemical
Co. 201-365-3479.
Solvent-Based Cleaner (200 g/50,000 Labels Printed)
Toluene
Acetone
Isopropyl alcohol
Diacetone alcohol
54.5
20.0
20.0
5.5
109.0
40.0
40.0
11.0
18.2
6.7
6.7
0.9
Solvent-Based Ink with Detergent Cleaner
Black Ink = 400 g + 200 g Cleaner/50,000 Labels Printed
Polyamide Black
Methyl alcohol
Isopropyl alcohol
Ethyl alcohol
N-propyl alcohol
Ethyl acetate
VM&P naphtha
46.0
4.7
10.6
21.4
6.5
4.2
6.6
184.0
18.8
42.4
85.6
26.0
16.8
26.4
30.7
3.1
7.1
14.3
4.3
2.8
4.4
Detergent Cleaner (200 g/50,000 Labels Printed)
279
-------
Component
Amount
(%)of
Ink or
Cleaner
Total (g)
per Sce-
nario
(600 g)
%of
Total
Scenario
Comments
Solvent-Based Ink with Detergent Cleaner
Black Ink = 400 g + 200 g Cleaner/50.000 Labels Printed
Detergent cleaner
100.0
200.0
33.3
99.4% water plus 0.6%
nonhazardous surfacta-
nts17 (Union Carbide
MP10) Mr. Dave Hostel-
ler, ArrowChem, Inc.
800-438-5883
Water-based Ink with Detergent Cleaner
Black Ink = 400 g + 200 g Cleaner/50,000 Labels Printed
Ammonia
Dlmethylethamolamlne
tsopropyl alcohol
Water
Acrylic resin
Azo pigments
1.0
1.0
5.0
65.0
20.0
8.0
4.0
4.0
20.0
260.0
80.0
32.0
0.7
0.7
3.3
43.3
13.3
3.3
LDgo <5 g/kg18, Mr.
Raleigh Turk 41 4-631-
2443.
LDso <5 g/kg18
Detergent Cleaner (200 g/50.000 Labels Printed)
Detergent cleaner
100.0
200.0
33.3
99.4% water plus 0.6%
nonhazardous surfacta-
nts (Union Carbide
MP10)17
EQUIVALENT TOXIC CONCENTRATION MEASUREMENTS
The equivalent toxic concentration values of the solvent-based ink with the solvent-based cleaners, the
solvent-base ink with water-based deaner, and the water-based ink with the water-based deaner are, 923,
121 and 347, respectively (Figure 3). The unexpected higher value for the water-based ink was due entirely
to the oral toxic'rty of the ammonia used in the ink formulation.
280
-------
This evaluation is for the bulk chemicals as
they are received and assumes that the exposure
is by ingestion. Thus, the results are most applica-
ble to the liquid wastes that are generated as a
result of printing
if there were no evaporative losses. If the ammonia
is removed due to evaporation, the equivalent
toxiclty of the water-based ink and detergent
cleaner would be about 20. The lowest concentra-
1000
-g 900
_w
§ 800
'l.i
o
700
v a 600
o °
I § 500
O.g 400
n
o
o
300
200
TOO
Solvent Ink SoJv«nt Ink Aqueoua Ink
Sol»«rit Q«or>»r Afluaoua CUorwr Aqueous Cleaner
... .. , ••«... Figure 3 The combined component
tion of ammoma m air at wh.ch any toxic effect to eqjivalent toxic concentration
. , , ,. „ .„, for the three printing scenarios.
humans has been observed is 20 parts per million-
19, The effects noted were with the nose and
changes in trachea or bronchii. At MPI Label Systems, with ammonia in the water-based inks present at
a concentration of 1 percent, and with the rapid changeover in the plant air, ammonia would be present in
the air far below any measurable toxic effect,
By contrast, components of the solvent-based cleaner are much more toxic than the detergent cleaner.
This can be seen by comparing the equivalent toxicity of the first scenario with the second in which the only
difference is in the cleaners evaluated.
CONCLUSIONS AND DISCUSSION
The change from solvent-based inks and cleaners to water-based ink and a detergent cleaner has
resulted in less waste being generated at MPI Label Systems. This paper presents an in-plant quantitative
measurement of the waste reduction benefits and trade-offs that have occurred. As with most changes in
processes, the switch in materials, including changing the type of plate used, required many adjustments
and fine-tuning. Also, in the past two years advances have been made in the water-ink formulation so they
are easier to clean, and the rate at which labels can be printed has increased. The cooperative approach
281
-------
between Industry and government developed in this project, and the evaluation methods described can be
used to evaluate the engineering and economic benefits of other similar chemical substitution projects.
The amount of solvent emissions to the air has been reduced per run by over 80 percent as a result of
this change according to the measurements taken at MPI Label Systems. In addition, the
components emitted are considerably less toxic to the press operators and the environment than was being
emitted from the solvent-based inks. Since MPI Label Systems uses approximately 1,800 pounds of ink per
year, and estimate can be made of the total weight of solvent emissions they currently release and what
would be released if they were still using the solvent-based inks. This comparison is shown in Table 10 for
MPI and the entire flexographic industry in the United States. For this estimate the laboratory evaporative
loss percent for black ink as determined in this study was used. For the entire industry it has been reported
that 300 million Ibs/yr of water-based inks are currently required. Thus, at MPI Label Systems solvent
emissions are estimated to have been reduced by over 800 Ibs. per year. For the entire industry, almost
150 million pounds per year of toxic solvents are no longer emitted. Less air toxics are also released from
the detergent cleaner than were released from the solvent cleaner previously used at MPI Label Systems.
Table 10. Comparative Solvent Loss to Air from Using Water-Based Inks
INK TYPE
Solvent Inks
Water-based Inks
Total Solvent Reduction
MPI LABEL SYSTEMS, INC,
972 Ibs/yr organic solvents
126 Ibs/yr
846 Ibs/yr
ENTIRE INDUSTRY
162,000,000 Ibs/yr solvents
15,000,000 Ibs/yr
Approx. 147,000,000 Ibs/yr
Another benefit found Is that hazardous wastes have been eliminated at MPI Label Systems. However,
aqueous wash liquids discharged to the sanitary sewer have increased from essentially none to
approximately 10,000 gallons per year.
Solid wastes generated in the form of wasted labels, wrap, trimmings and other paper has remained
282
-------
about the same. In total, MPI estimates an annual cost savings of at least $16,500 per year, as a result of
this change.
In addition to the above dollar advantages accruing from the use of water-based inks, and a
relatively harmless cleaner, it is the opinion of the plant manager and his shop superintendent that the
company is also realizing the following subjective benefits:
1. Water-based inks are easier to clean from pans, plates and rollers;
2. Waste inks is more easily disposed of;
3. Spilled ink is easier to dean up;
4. Waste going to a landfill is not classified as hazardous, giving MPI less long-term liability;
5. Expensive solvents are not required for cleanup; and
6. Employees are enjoying a cleaner, safer work environment.
An important lesson learned in this effort Is the importance of agreeing on all planned plant tests with
the company personnel to be involved. In our case we made it a point to closely monitor several label runs
to learn the various steps, solutions, etc., to be encountered before collecting any in-plant measurements.
However, by the time we were ready to take measurements, the plant had made changes in Its operations.
The operators prepared for a run the day before our arrival by completing preliminary tasks "to speed things
up." Unfortunately, these well-meaning preparations short-circuited some of our measurements. As better
materials appear, as less expensive procedures are noted - but, mostly, if quality can be maintained or
improved, and costs reduced, almost any process change will be considered.
During the planning stages of this project it was intended to quantitatively measure ink and cleaner
usage at every step of the printing process. After a preliminary run it became apparent that several of these
measurements would be very difficult to carry out. Two examples are worth noting. First, we
intended to determine by weighing the amount of dry ink actually deposited on labels (Q In equation 1).
This was to be accomplished by weighing approximately 1000 blank labels as they came from the press,
and 1000 printed labels. After numerous measurements, at least for the labels we measured, it became
283
-------
obvious that the amount of ink on anything but a very large number of labels Is negligible. It appears that
variations In the paper weight and perhaps amount of adhesive are much greater than the amount of Ink
applied. Second, we intended initially to measure the amount of ink wasted on the paper trimmed from
around each label. The trimmed top layer is peeled away from its paper backing, then collected as a roll
of waste. It was not possible to separate one layer of these trimmings from another.
The result of these two experiences was that we decided to simply weigh the ink reservoir before
and after a run and consider the difference due to all ink uses - labels, trim waste, spills, cleanup of the ink
pans, rollers and printing plates. With the exception of the ink lost during cleanup (and on a relatively short
run this will represent most of the ink used) the balance will have its solvents lost to the shop air.
It should be kept in mind that the scope of this evaluation was limited in several important aspects.
First, the Image and plate making steps were not included in this evaluation of waste produced. As
mentioned earlier, the plates used at the MPI Label Systems plant evaluated were produced by another
company. The wastes generated In formulating the inks and cleaners were not comparatively evaluated In
this study. It could be possible that the reduction in waste produced during the printing operation Is more
than offset by increased waste produced during ink or cleaner manufacturing. This seems unlikely since
the solvent components of both materials currently being used are much less than the solvent-based
materials. Finally, the Impact of using water-based inks on the recyclability of the product labels was not
evaluated. Since these labels are placed on many types of products, they will generally not be recycled
as waste paper.
MPI Label Systems and the entire flexographic printing industry has benefited economically, technically
and In the physical well-being of its employees by changing from solvent-based to water-based inks and
cleaners. The environment has also benefited. Additional benefits will be realized as the use of solvent-type
Industrial cleaners Is eliminated. Label customers are also benefiting from the change in technology with
better quality labels.
284
-------
ACKNOWLEDGMENT
This work was accomplished in large part because of generous cooperation by MPI Label Systems,
University Park, Illinois, Water Technologies, Inc., Iron Station, North Carolina, BASF Corp., Chicago, Illinois,
and the Flexographic Technical Association, Ronkonkoma, New York. The work was performed jointly by
the USEPA and the Hazardous Waste Research and Information Center, Illinois Dept. of Enery and Natural
Resources, Champaign, Illinois, under cooperative agreement CR-815829-01-0.
BIBLIOGRAPHY
1. Printing Industries of America (Natl HQ), Arlington, VA., Private communication, Jan. 22, 1991.
2. Third Flexo Plant Survey. Fiexographic Technical Association, Appleton, Wl, 1983.
3. Graphic Arts Technical Foundation, Pittsburgh, PA., Private communication. Dec. 1989
4, Duzinskas, Donald R., 1983. The systems approach to pressroom ventilation in solvent
recovery. W. F, Hall Printing Co., Gravure Research Institute, 36th Annual Meeting, Nov. 1983.
5. Guides to Pollution Prevention -The Commercial Printing Industry. U.S. EPA, Cincinnati, OH. EPA
document EPA/625/7-90/008, August 1990;
6. James, Christopher A., 1987. RACT Compliance: VOC Emission Reductions from Rhode island
Printing and Surface Coating Sources. Rhode Island Div. of Air and Hazardous Materials,
Providence, Rl. 1987.
7. Rosen, D. R., and Wool, M. R. Microprocessor Control of Rotogravure Airflows. Acurex
Corp., Mountain View, CA. August 1986.
8. Waste Audit Study - Commercial Printing Industry. California Dept. of Health Services,
Alternative Technology Section, Sacramento, CA. May 1986.
9. Makrauer, Geo. A., Innovations in Flexoqraphic Printing: Reducing VOC's with Water-based Inks
When Printing on High-slip Polyethylene Films. AMKO Plastics, Inc., Cincinnati, OH. 1987.
10. MSDS for Water-based Ink, Water Ink Technologies, inc., Iron Station, NC. May 1988.
285
-------
11. MSDS for Solvent-based Ink, BASF Corp., Clifton, NJ. December 1988.
12. Advance Process Supply Co., Chicago, !L Private communication. January 1991.
13. MSDS for Detergent Cleaner, Water Ink Technologies, Inc., Iron Station, NC. November 1990.
14. Plewa, Michael, Dowd, P., Ades, D., Wagner, E., Assigning a Degree of Hazardous Ranking to Illinois
Waste Streams. Hazardous Waste Research and Information Center, Champaign, IL HWRIC RR
013, November 1986,
15. Title 40 CFR, Chapter 1, Section 261.33.
16. Sun Chemical Co. Private communication, Mr. Fishman, January 1991.
17. Union Carbide Corp. Private communication, Mr. Dave Hosteller, January 1991.
18. S. C. Johnson Co. Private communication, Mr. Raleigh Turk, January 1991.
19. Registry of Toxic Effects of Chemical Substances. 1985-1986 Edition, Vol. 1. U.S. Dept. of Health
and Human Services,
286
-------
THE TREATABILITY OF URBAN STORMWATER TOXICANTS
Robert E. Pitt, P.E., Ph.D., Department of Civil Engineering,
University of Alabama at Birmingham, Birmingham, Alabama 35294
Patricia F, Barren, Law Engineering, Birmingham, Alabama 35244
Ali Ayyoubi, Department of Civil Engineering, University of Alabama at Birmingham,
Birmingham, Alabama 35294
Richard Field, P.E., Chief, Storm and Combined Sewer Pollution Control Program,
U.S. EPA, Edison, New Jersey 08837
ABSTRACT
This paper summarizes some of the information obtained during a research project sponsored and
directed by the EPA's Storm and Combined Sewer Research Program and conducted under a subcontract from
Foster-Wheeler/Enviresponse. Earlier research, that was reported at last year's Sixteenth Annual Hazardous Waste
Research Symposium, investigated typical toxicant concentrations in stormwater, the origins of the toxicants
found, and rain and land use factors that influenced the toxicant concentrations. The most recent research, which
is summarized in this paper, investigated the control of stormwater toxicants through conventional treatment unit
processes.
Twelve sheetflow samples were collected from the source areas that were found previously to generally
produce the most toxic runoff waters. These areas were automobile service areas (gas stations, car washes, oil
change and other automobile maintenance facilities), industrial parking and loading dock areas, and automobile
salvage yards. These samples were subjected to a variety of treatment processes. The bench scale treatability tests
included settling columns, sieving screens, membrane filters, aeration, photo-degradation, aeration and photo-
degradation combined, floatation, and alum addition. Toxicity changes were monitored using the Microtox
bioassay test. The Microtox test was extensively compared to conventional bioassay and chemical tests during
previous research phases.
The benefits of the treatment processes varied for the different samples. However, some of the treatment
processes consistently provided the greatest toxicity reductions. The most beneficial treatment tests included
settling for at least 24 hours (generally 40 to 90% reductions), screening through at least 40 micron screens (20 to
70% reductions), and aeration and/or photo-degradation for at least 24 hours (up to 80% reductions). The
floatation tests produced floating sample layers that generally increased in toxicity with time and lower sample
layers that generally decreased in toxicity with time. However, the benefits were quite small (less than 30%
reduction). Alum additions substantially reduced the turbidity of the samples, but the changes in toxicity were
highly irregular.
287
-------
The next project phases will include more extensive laboratory and field tests, using prototype treatment
designs based on these initial bench scale tests.
This paper has been reviewed in accordance with the U.S. Environmental Protection Agency's peer and
administrative review policies and approved for presentation and publication.
INTRODUCTION
BACKGROUND
Urban stormwater runoff has been identified as a major contributor to the degradation of many urban
streams and rivers (1 - 4). Organic and metallic toxicants are expected to be responsible for much of these
detrimental effects, and have been found in urban runoff discharges during many previous studies (5 - 8).
All U.S. cities having populations greater than 100,000 (which total about 15,000 square miles) (9) will
be required to participate in the EPA's stormwater permit program (Federal Register, November 16,1990). The
Nationwide Urban Runoff Program (NURP) monitored toxicant discharges from 28 cities (5). Based on this
monitoring, it is expected that these large cities are responsible for substantial toxicant discharges that directly
enter the nation's surface receiving waters from stormwater outfalls. This limited NURP data was collected mostly
from residential areas, with some commercial areas represented. More recent information indicates that industrial
stormwater discharges can have many times the concentrations of the toxicants as the areas represented in the
NURP data (10). In addition, base flows occurring in storm drains during dry weather that may be contaminated
by non-stormwater discharges (such as industrial waste cross-connections), can also significantly increase these
estimated loadings (11). The EPA sponsored research summarized in this paper was conducted to obtain much
needed information concerning the sources and potential control of these stormwater toxicants.
FIRST RESEARCH PHASE SUMMARY
The first phase of this research included the collection and analysis of about 150 urban stormwater runoff
and combined sewer overflow (SCSO) samples from a variety of source areas and under different rain conditions.
A number of combined sewer overflow and detention pond samples were also evaluated. This effort was
significantly greater than had been attempted previously for toxic pollutants in stormwater.
Samples were analyzed for many organic pollutants using two gas chromatographs, one with a mass
selective detector (GC/MSD) and another with an electron capture detector (GC/ECD), and for metals using a
graphite furnace equipped atomic absorption spectrophotometer (GFAA). All samples were further analyzed for
particle size distributions (from about 1 to 100 microns) and for toxicity using the Microtox (from Microbics)
toxicity screening technique. All samples were also filtered to determine the liquid/solid partition coefficients of
the pollutants and the relative toxicities of the filterable and nonfilterable portions of the samples. Overall, about
300 sample components (filterable and total portions of 150 samples) were analyzed to determine toxicant
concentrations in sheetflows and other SCSOs as part of the first phase of this project. The following paragraphs
briefly summarize the first project phase results (from ref. 8).
Most pH values were in a narrow range of 7.0 to 8.5 and the suspended solids concentrations were
generally less than 100 mg/L. The particle size ranges were usually narrow for any one sample, but the
distribution ranges developed using all samples from a single source area category were substantially greater.
Only a small fraction of the toxic organic pollutants analyzed were frequently detected. Thirteen organics
were detected in more than ten percent of all samples analyzed. The greatest detection frequencies were for 1,3-
dichlorobenzene and iluoranthene, which each had detection frequencies of 23 percent. The organics most
frequently found in these samples were similar to the organics most frequently detected at outfalls in prior studies
288
-------
(such as during the NURP study, ref. 5), The PAHs, especially fluoranthenes and pyrenes were the most,
commonly detected organic compounds.
The organic compounds analyzed had the greatest frequencies of detection in roof runoff, urban creeks,
and CSO samples. Vehicle service areas and parking areas had several of the observed maximum organic
compound concentrations observed. Most of the organics were associated with nonfilterable sample portions.
In contrast to the organics, the heavy metals were detected in almost all samples analyzed, including the
filtered samples. Roof runoff had the highest observed concentrations of zinc, probably due to galvanized roof
drainage components. Parking areas had the highest nickel concentrations, while vehicle service areas had the
highest concentrations of cadmium and lead. Urban creek samples had the highest copper concentrations, probably
due to illicit discharges.
About 15 percent of all of the unfiltered samples analyzed were considered highly toxic using the
Microtox screening procedure. The remaining samples were approximately evenly split between being moderately
toxic and not being toxic. The Microtox screening tests found that CSOs had the greatest percentage of samples
considered the most toxic, followed by samples obtained from parking and industrial storage areas. Runoff from
paved areas all had relatively low suspended solids concentrations and turbidities, especially compared to samples
obtained from unpaved areas.
Preliminary data evaluations indicated that variations in observed Microtox toxicities and organic
toxicant concentrations may be greater for different rains than for the different source areas sampled. As an
example, high concentrations of PAHs were mostly associated with long-antecedent dry-periods.
The literature review conducted during the first project phase found that many processes will affect the
potential transport and fate mechanisms of these pollutants. Sedimentation in the receiving water is the most
common fate mechanism because many of the pollutants investigated are mostly associated with paniculate
matter. Exceptions included zinc and 1,3-dichlorobenzene which were mostly associated with the filterable
sample portions. Paniculate removal can occur in many SCSO control facilities, including catchbasins, swirl
concentrators, screens, drainage systems, and detention ponds. These control facilities allow removal of the
accumulated polluted sediment for final disposal in an appropriate manner. Uncontrolled sedimentation will occur
in receiving waters, such as lakes, reservoirs, or large rivers. In these cases, the wide dispersal of the contaminated
sediment is difficult to remove and can cause significant detrimental effects. Biological or chemical degradation
of the toxicants in the sediments may occur, but is quite slow for many of the pollutants in the expected anaerobic
environments. Degradation of the soluble pollutants in the water column may also occur, especially when near the
surface in aerated waters. Volatilization (evaporation) is also a mechanism that may affect many of the detected
organic toxicants. Increased turbulence and oxygen supplies would encourage these processes that may
significantly reduce pollutant concentrations. Sorption of pollutants onto suspended solids and metal precipitation
increases the sedimentation potential of the pollutants and also encourages more efficient bonding of the
pollutants in soils, preventing their leaching to surrounding waters.
The second project phase, summarized below, examined the usefulness of a variety of treatment unit
processes that were expected to reduce the toxicities found in the initial project phase. Later project phases will
examine the treatability of SCSO toxicants in greater detail, especiaEy in terms of making and testing
modifications to existing treatment processes.
METHODOLOGY
SAMPLING EFFORT AND EXPERIMENTAL ERROR
The relative importance of different source areas (such as roofs, streets, parking areas, etc.) in
contributing toxicants was determined from the first project phase activities that examined 150 source area
289
-------
samples (8). These previous samples were collected from the most significant potential toxic pollutant source
areas In residential, commercial, and industrial land uses. The areas that received the most sampling attention
during both project phases were parking and storage areas in industrial and commercial areas. These areas had
been noted in previous studies to have the largest potential of discharging toxicants (10). Sheetflow samples were
collected during five Birmingham Alabama rains. Replicate samples taken from many of the same source areas,
but during different rains, enabled differences due to rain conditions versus site locations to be statistically
evaluated.
The second research phase included extensive analyses of 12 samples. Tablel lists these samples,
including their sampling dates and source area categories. These samples represent practically all of the rains that
have occurred in the Birmingham area since the second project phase field activities were able to begin in June,
1990. Information is also shown concerning the toxicities of the samples before treatment. These independent
replicates were used to measure the measurement errors associated with the Microtox procedure. The means,
standard deviations, and relative standard deviations (standard deviation divided by the mean, times 100) of the
replicate toxicity values are also shown on Table 1. The number of analyses refer to the total number of Microtox
analyses that were conducted for all of the treatability tests for each sample.
The initial toxicity values were plotted on normal-probability plots to indicate their probability
distributions. Almost all of the samples had initial toxicity values that were shown to be normally distributed.
Therefore, the relative standard deviation values shown on Table 1 can be used as an indication of the confidence
intervals of the Microtox measurements. The relative standard deviations ranged from 2.3 to 9.8 percent, with an
average value of 5.1 percent. Therefore, the 95 percent confidence interval for the Microtox procedure ranged
between 5 and 20 percent, and averaged about 10 percent (two times the relative standard deviation values include
95.4 percent of the values, if normally distributed). These confidence intervals are quite narrow for a bioassay test
and indicate the good repeatability of the procedure. One of the important factors of the Microtox test is the use of
a very large number of organisms (about one million) for each analysis, reducing erratic test responses that may be
caused by unusual individual organisms. In all cases, statistical tests were performed on the test results to indicate
the significance of the different treatability tests.
Figure 1 contains box plots of the initial toxicity values (18). These indicate the spread of toxicity values
that were represented by the samples. Two samples (B and D) were found to be highly toxic, while the remainder
were moderately toxic.
SAMPLING PROCEDURES
The sheetflow samples were collected using manual grab procedures. For deep sheetflows, samples were
collected directly into the sample bottles, or dipped using glass beakers. For shallow sheetflows, hand operated
pumps created a vacuum in the sample bottle which then drew the sample directly into the container through
Teflon tubes. About ten to twenty liters of each sample were collected for the treatability analyses. The samples
were all obtained from the Birmingham, Alabama area.
TOXICITY SCREENING TESTS
A number of previous studies have found high concentrations of toxic pollutants in storm water samples
(as summarized in ref. 4). Some urban stortnwater runoff studies attempted to use conventional 96-hr fathead
minnow fish bioassay toxicity tests (such as in ref.12), but very few fish died during the tests. However, in situ
taxonomic studies of urban runoff receiving waters found significant evidence of toxic effects from the long-term
exposure to these pollutants (such as reported in ref, 2, for the same stream as the negative fish bioassay tests).
More recent bioassay tests have used more sensitive organisms and have detected significant SCSO toxicities
(ref.13, from Syracuse, NY; ref. 14 from Birmingham, AL; ref. 15 from Waterbury, CN; and ref. 16 from San
Francisco Bay, CA).
290
-------
The objective of this toxicity monitoring task was to obtain toxicity measurements from a large number
of subsamples obtained during different stages of bench scale treatment tests. These tests were not used to
determine the absolute toxicities of the samples, but only to examine the toxicity differences between the sample
partitions from different treatment tests. To evaluate the different treatment options, it was necessary to use a
rapid screening method that only used small sample volumes . The toxicity testing procedure that was used
(Microtox from Microbics, Inc.) uses luminescent microorganisms to indicate relative toxicities of samples. A
series of special tests were made during the first project phase to compare the toxicities of about 20 selected
sheetflow and CSO samples to both the Microtox screening method and conventional bioassay methods.
The toxicity, as determined by the Microtox procedure, was expressed as three values, I „ (the
percentage light decrease after 10 minutes of exposure), I _ (the percentage light decrease after 35 minutes of
•5.3
exposure), and the EC_n. The EC_n is the sample dilution corresponding to a 50 percent light decrease after a 35
minute exposure. Therefore, only samples that have L . values greater than 50 were further tested to determine the
EC__ values. Higher values of I _ and L-, and lower fractions of EC,-, correspond to greater toxicities.
Microbics suggests that light decrease values greater than 60 percent correspond to "highly" toxic
samples, light decrease values between 20 and 60 percent correspond to "moderately" toxic samples, and light
decrease values less than 20 percent correspond to "not" toxic samples.
During the first project phase, a number of special tests were conducted that examined problems
associated with sample storage time, preservation, and sample containers. Teflon and glass were exclusively used
to reduce the effects of the containers on the sample toxicities and samples were all examined within 24 hours of
sample collection. Samples were also stored at 3° C to 5° C.
PARTICLE SIZE ANALYSES
Most SCSO physical treatment removal efficiencies significantly relate to the particle size distributions
and settling velocities of solids (17). Wet detention ponds, catchbasins, grass filters, street cleaning,
microscreening, filtration, and swirl concentrators are some of the pollutant control methods that require a
knowledge of particle size and/or settling characteristics. Additionally, the fates of many toxic pollutants in
receiving waters are also very sensitive to the physical characteristics of particles. It is not possible to correctly
design many of these physical treatment devices without knowing the specific particle size distributions and
settling velocities of the SCSOs.
A laser particle counter (SPC-510 from Spectrex Corp.) was used to analyze particle size distributions for
all of the samples during the different treatment phases. This instrument produces particle size distribution plots
for particle sizes ranging from 0.5 microns to more than 100 microns. Settling column tests were also concurrently
conducted during this research phase to determine the specific gravities and settling velocities of SCSO samples.
SOLIDS AND TURBIDITY ANALYSES
Nephelometric turbidity analyses were conducted for all subsamples during the treatability tests, using
EPA method 180.1. Gravimetric solids analyses were conducted on all settling column subsamples to calculate
settling rates and specific gravity. EPA method 160 was used for these solids analyses.
291
-------
TOEATABUJTY TESTS
Phase 2 included tests to examine the treatability of source area samples. As noted above, the samples
were all relatively toxic. This has allowed a wide range of laboratory partitioning and treatability analyses to be
conducted without having detection limit problems. The following tests were used for this study:
o Settling column (a 1.5 inch by 30 inch Teflon column)
o Floatation (a series of eight narrow neck glass 100 mL volumetric flasks)
o Screening and filtering (a series of eleven stainless steel sieves, from 20 to 106 microns, plus a 0.45
micron membrane filter)
o Photo-degradation (a two liter glass beaker with a 60 W "grow light" incandescent light placed 6 inches
above the water, stirred with a magnetic stirrer, temperature of water and evaporation rate also monitored)
o Aeration (the same beaker arrangement as above, without the grow light, but with filtered compressed
air keeping the test solution supersaturated and well mixed)
o Photo-degradation and aeration combined (the same beaker arrangement as above, with compressed air,
grow light, and stirrer)
o Chemical addition (a standard glass jar tests using alum and 800 mL samples)
o Undisturbed control sample (a sealed and covered glass jar at room temperature)
These bench scale tests were all designed to use small sample volumes because of the difficulty of
obtaining large sample volumes from many of the source areas that were to be examined.
Each test (except the filtration and chemical addition tests) was conducted over a period of time,
Subsamples were typically obtained for toxicity analyses at the following time intervals during the tests: 0,1,2,3,
6,12,24,48, and 72 hours. In addition, settling column samples were also obtained at many times within the first
hour: 1,3,5,10,15,25, and 40 minutes. The chemical addition tests were conducted using alum at several
concentrations in a standard jar test. In addition to the Microtox toxicity tests, most samples were analyzed for
turbidity and particle sizes. All settling column samples were also analyzed for gravimetric suspended solids
content to enable calculations of settling velocity to be made.
Future project phases will include pilot- and full-scale tests of various control and treatment practices.
Especially important in the future project phases will be the testing of modifications to conventional stormwater
and CSO treatment processes and the design and testing of combination treatment systems suitable for small
source areas (such as pavement at automobile service facilities, especially gas stations).
DATA OBSERVATIONS
The Microtox procedure alowed toxicity screening tests to be conducted on each sample partition during
the treatment tests. This efficient procedure enabled more than 900 toxicity tests (and turbidity and particle size
distribution tests) to be made.
Figures 2 through 10 are plots of the toxicity reductions observed during these tests. Each of these figures
contains the data for one of the treatment tests conducted, including the control test. Each figure contains three
plots, one contains the treatment responses for the automobile service facility samples (samples B, C, E, and H),
another for the industrial loading and parking area samples (samples D, F, G, I, J, and K), and the last one for the
automobile salvage yard samples (samples L and M). Even though the data are plotted into these three groups,
very few consistent differences are noted in the way the samples responded to the treatments. As expected, there
are greater apparent differences between the treatment methods than between the sample groupings. Statistical
tests that wiU be conducted during the current project phase will examine these groupings in detail.
292
-------
These initial data plots show the percentage reduction of toxicity, as measured by the Microtox
procedure. It is expected that some of the treatment processes will have varying effects, depending on the initial
toxicity values, which were shown to vary considerably. Other future statistical tests will therefore also examine
the effects of these treatment schemes on samples having different initial toxicities.
Tables 2 through 4 summarize results from the non parametric Wilcoxon signed ranks test (using
SYSTAT: The System for Statistics, Version 5, SYSTAT, Inc., Evanston, 111.) for different treatment
combinations. This statistical test indicates the two-sided probabilities that the sample groups are the same. A
probability of 0.05, or less, is used here to indicate significant differences in the data sets. As an example, Table 3
indicates that for sample D, the undisturbed control sample was significantly different (with probabilities of 0.02)
compared to all of the treatment tests.
The aeration test provided the most samples that had significant probabilities of being different from the
control condition. Settling, photo-degradation, and aeration and photo-degradation combined, were tied in
providing the next greatest number of samples that had significant probabilities of being different from the control
condition. The floatation test had many samples that had significant differences in toxicity of the top floating layer
compared to the control sample. However, the more important contrast between the middle sample layers (below
the top floating layer) and the control sample, which would indicate a reduction in toxicity of post-treated water,
had very few samples that were significantly different from the control sample.
The absolute magnitudes of toxicity reductions must also be considered. As an example, it may be
significant, but unimportant, if a treatment test provided many (and therefore consistent) samples having
significant differences compared to the control sample, if the toxicity reductions realized were very small.
CONCLUSIONS
As shown on Figures 2 through 10, good separation of toxicant responses were found during many of the
treatment tests. The most beneficial treatment tests included settling for at least 24 hours (providing generally 40
to 90% reductions), screening through at least 40 micron screens (20 to 70% reductions), and aeration and/or
photo-degradation for at least 24 hours (up to 80% reductions). Increased settling, aeration or photo-degradation
times, and screening through finer meshes, all resulted in greater toxicity reductions. The floatation tests produced
floating sample layers that generally increased in toxicity with time and lower sample layers that generally
decreased in toxicity with time. However, the benefits were quite small (less than 30% reduction). Alum additions
substantially reduced the turbidity of the samples, but the changes in toxicity were highly irregular. These results,
in conjunction with results from the first project phase, will enable us to modify treatment designs to optimize
toxicant removals from critical stormwater runoff source areas.
REFERENCES
1. Field, R., and Turkeltaub, R. Urban runoff receiving water impacts: program overview. Journal of
Environmental Engineering, 107:83-100,1981.
2. Pitt, R. and Bozeman, M. Sources of urban runoff pollution and its effects on an urban creek. EPA-600/2-
82-090, U.S. Environmental Protection Agency, Cincinnati, December 1982.
3. Pitt, R. and Bissonnette, P. Bellevue urban runoff program, summary report. U.S. Environmental Protection
Agency and the Storm and Surface Water Utility, BeUevue, Washington, 1984.
4. Field, R. and Pitt, R. Urban storm-induced discharge impacts: US Environmental Protection Agency research
program review. Water Science and Technology. Vol. 22, No. 10/11. pp. 1-7, 1990.
293
-------
5. EPA. Results of the nationwide urban runoff program. Water Planning Division, PB 84-185552, Washington,
D.C., December 1983.
6. Hoffman, E J., Mills, G.L, Latimer, J.S., and Quinn, J.G. Urban runoff as a source of polycyclic aromatic
hydrocarbons to coastal waters. Environment Science and Technology, 18:580-587,1984.
7. Fam, S., Stensirom, M.K., and Silverman, G. Hydrocarbons in urban runoff. Journal of Environmental
Engineering, 113:1032-1046,1987.
8. Pitt, R., and Barron, P. Assessment of urban and industrial stormwater runoff toxicity and the
evaluation/development of treatment for runoff toxicity abatement - phase 1. U.S. Environmental Protection
Agency, Office of Research and Development, EPA Contract #68-C9-0033. Edison, New Jersey, 1990.
9. Depf. of Commerce. Statistical Abstract of the United States, 1980. U.S. Bureau of the Census. 101st. edition.
Washington, D.C., 1980.
10. Pitt, R, and McLean, J. Toronto area watershed management strategy study: Humber River pilot watershed
project. Ontario Ministry of the Environment, Toronto, Ontario, 1986.
11. Pitt, R., Lalor, M., Miller, M. and Driscoll, G. Assessment of non-storm water discharges into separate storm
drainage networks - phase 1: development of methodology for a manual of practice. U.S. Environmental
Protection Agency, Office of Research and Development, EPA Contract #68-C9-0033. Edison, Edison, New
Jersey, 1990.
12. Pilt, R. Demonstration of nonpoint pollution abatement through improved street cleaning practices. EPA-
600/2-79-161, U.S. Environmental Protection Agency, Cincinnati, Ohio, August 1979.
13. Spiegel, S.J., Tifft, E.G., Murphy, C.B. and Oil, R.R. Evaluation of urban runoff and combined sewer
overflow mutagenicity. EPA-600/2-84-116. U.S. Environmental Protection Agency, Cincinnati, June 1984.
14. Mount, D.I., Steen, A.E. and Norberg-King, TJ. Validity of effluent and ambient toxicity for predicting
biological impact on Five Mile Creek, Birmingham, Alabama. EPA/600/8-85/015. U.S. Environmental
Protection Agency. Duluth, Minn., December 1985.
15. Mount, D.I,, Norberg-King, TJ. and Steen, A .E. Validity of effluent and ambient toxicity for predicting
biological impact, Naugatuck River, Waterbury, Connecticut. EPA/600/8-86/005. U.S. Environmental
Protection Agency. Duluth, Minn., May 1986.
16. Norberg-King, T.J., Mount,D.L, Amato, J.R. and Taraldsen,J.E. Application of the water quality based
approach on a regional scale: toxicity testing and identification of toxicants in effluents from the San
Francisco Bay Region. National Effluent Toxicity Assessment Center Technical Report 01-88. U.S.
Environmental Protection Agency. Duluth, Minn., May 1988.
17. Dalrymple, R.J., Hodd, S.L. and Morin, D.C. Physical and settling characteristics of particulates in storm and
sanitary wastewaters. EPA-670/2-7S-011. U.S. Environmental Protection Agency, Cincinnati, Ohio, 1975.
18. Tukey, J.W. Exploratory Data Analysis. Addison-Wesley. Reading, Mass. 1977.
294
-------
/ TABLE 1. SAMPLE DESCRIPTIONS
Automobile Service Area Samples:
Sample Date
Toxicity Number Standard Relative
(% light of Analyses Deviation Standard
reduction) Deviation
(percent)
B
C
E
H
7/10/90
7/21/90
8/19/90
10/17/90
78
34
43
50
28
42
74
88
7,6
2,9
1.3
1.5
9,8
8.5
3
3
Industrial Loading and Parking Area Samples:
Sample
D
F
G
I
J
K
Date
8/2/90
9/12/90
10/3/90
10/24/90
11/5/90
11/9/90
Toxicity
(% light
reduction)
67
31
53
55
49
28
Number
of Analyses
74
88
88
89
89
89
Standard
Deviation
2.1
1.5
3
1.9
1.1
2.2
Relative
Standard
Deviation
(percent)
3.1
4,9
5.7
3.4
2.3
8.1
Automobile Salvage Yard Samples:
Sample Date
L 11/28/90
M 12/3/90
Toxicity Number Standard Relative
(% light of Analyses Deviation Standard
reduction) Deviation
(percent)
26
54
89
89
1.4
1.8
5.5
3.4
minimum:
maximum:
mean:
st. dev.:
total:
26
78
47
16
1.1
7.6
2.4
2.3
9.8
5.1
927
295
-------
TABLE 2. TWO-SIDED PROBABILITIES COMPARING DIFFERENT TREATMENT
TESTS FOR AUTOMOBILE SERVICE AREA SAMPLES CO
Automobile Service Areas Samples:
Undisturbed versus; B C E H
settling n/a 0,25 0,02 0.41
aeration n/a 0.31 0.25 0.07
photo-degradation n/a 0.12 0.06 0.16
aeration and photo-degrad, n/a 0.35 0.24 0,06
floatation - top layer n/a n/a 0.74 0.02
floatation - middle layer n/a n/a 0,31 0.87
Aeration and Photo-degradation:
aeration vs. photo-degrad. 0.23 0.02 0.49 0.08
aeration vs. aeration and photo. n/a 0.03 0.99 0.14
photo vs. aeration and photo. n/a 0.25 0.14 0.02
Floatation:
top layer vs. middle layer n/a n/a 0.49 0,01
Settling versus:
aeration 0.46 0.02 0.02 0.45
photo-degradation 0.12 0.25 0.02 0.79
aeration and photo-degradation n/a 0,61 0,02 0.09
floatation - top layer n/a n/a 0.02 0,05
floatation - middle layer n/a r\/a 0.02 0.09
Aeration versus:
floatation - top layer n/a n/a 0.39 0,02
floatation - middle layer n/a n/a 0.21 0,02
Photo-Degradation versus:
floatation - top layer n/a n/a 0,18 0.02
floatation - middle layer n/a n/a 0.03 0.02
Aeration and Photo-Degradation versus:
floatation - top layer n/a n/a 0,49 0.02
floatation - middle layer n/a n/a 0.04 0.02
(1) Probabilities were calculated using the Wileoxon signed-rank test for paired data sets.
Comparisons having probabilities less than, or equal to, 0.05 are considered significantly
different.
296
-------
TABLE 3. TWO-SIDED PROBABILITIES COMPARING DIFFERENT TREATMENT
TESTS FOR INDUSTRIAL LOADING AND PARKING AREA SAMPLES CO
Industrial Loading and Parking Area Samples:
D F G I J K
Undisturbed versus:
settling 0,02 0.12 0.09 0.07 0.01 0.01
aeration 0.02 0.05 0.06 0.04 0.01 0.01
photo-degradation 0.02 0.04 0.03 0.07 0.01 0.01
aeration and photo-degrad. 0.02 0.05 0.03 0.09 0.01 0.01
floatation - top layer 0.02 0,05 0.13 0.01 0.03 0.21
floatation - middle layer 0.02 0.78 0.02 0,26 0.16 0.17
Aeration and Photo-degradation:
aeration vs. photo-degrad. 0.21 0.24 0.74 0.01 0.04 0.05
aeration vs. aeration and photo. 0.61 0.18 0,04 0.01 0.11 0.51
photo vs. aeration and photo. 0.21 0.16 0.25 0,79 0.74 0.12
Floatation:
top layer vs. middle layer 0.72 0.41 0.05 0.02 0.07 0.12
Settling versus:
aeration 0.18 0.33 0.61 0.48 0.41 , 0.02
photo-degradation 0.02 0.78 0.61 0.06 0.12 0.02
aeration and photo-degradation 0.03 0.67 0.75 0.05 0.12 0.03
floatation - top layer 0.14 0.05 0.13 0.04 0.01 0.02
floatation - middle layer 0.72 0,09 0.31 0,05 0.02 0,01
Aeration versus:
floatation - top layer 0.39 0.04 0.09 0,02 0.01 0,09
floatation - middle layer 0,12 0.09 0.18 0.03 0.01 0.01
Photo-Degradation versus:
floatation - top layer 0.04 0.04 0.04 0.02 0.01 0.09
floatation - middle layer 0.04 0.05 0.24 0.04 0.01 0.01
Aeration and Photo-Degradation versus:
floatation - top layer 0.03 0.04 0.06 0.02 0.01 0.05
floatation - middle layer 0.18 0.21 0.04 0.04 0.01 0.01
(1) Probabilities were calculated using the Wilcoxon signed-rank test for paired data sets.
Comparisons having probabilities less than, or equal to, 0.05 are considered significantly
different.
297
-------
TABLE 4 TWO-SIDED PROBABILITIES COMPARING DIFFERENT TREATMENT
TESTS FOR AUTOMOBILE SALVAGE YARD SAMPLES (1)
Automobile Salvage Yard Samples:
L M
Undisturbed versus:
settling 0.02 0,02
aeration 0.02 0.03
photo-degradation 0.02 0.16
aeration and photo-degrad. 0.02 0.09
floatation - top layer 0.01 0.09
floatation - middle layer 0.59 0.89
Aeration and Photo-degradation:
aeration vs. photo-degrad. 0.08 0.01
aeration vs. aeration and photo. 0.07 0.08
photo vs. aeration and photo. 0.99 0.14
Floatation:
top layer vs. middle layer 0.02 0.07
Settling versus;
aeration 0.02 0.12
photo-degradation 0.02 0.01
aeration and photo-degradation 0.02 0.02
floatation - top layer 0.02 0.02
floatation - middle layer 0.02 0.03
Aeration versus:
floatation - top layer 0.02 0.01
floatation - middle layer 0.02 0.01
Photo-Degradation versus:
floatation - top layer 0.02 0.03
floatation - middle layer 0.02 0.21
Aeration and Photo-Degradation versus:
floatation - top layer 0,02 0.01
floatation - middle layer 0.02 0.16
(1) Probabilities were calculated using the Wilcoxon signed-rank test for paired data sets.
Comparisons having probabilities less than, or equal to, 0.05 are considered significantly
different,
298
-------
»u
o 80
b
§' 70
H
o
i so
O.
'
fr-
LJU
0.
TOXTOXIC1TY
^ C
0 C
o
DC
i 30
^> w v-/
on
1 i
— T X75th percentile —
.. . -
-median value (50th percentile)
1 X25th percentile
T
T
-1- 5|C L
<=P f
e
i
— T *
L
^0
1 1 ! i 1 1
BCDEFGH I JKL'M
SAMPLES (SEE TABLE 1 FOR SAMPLING DATES AND LOCATIONS)
Figure 1, Box plot of initial sample Microtox toxicities (ret 18).
299
-------
AUTOMOBILE SERVICE AREA SAMPLES
too
80
60
40
20
0
-20
-40
-60
-80
-100
GO
O
o
Sample "H"
$/\ / Sample ^E", *>,.
0 6 12 18 24 30 36 «2 IB 54 60 66 72
Undisturbed (room temperature, dark, and sealed) (hours)
INDUSTRIAL LOADING/PARKING AREA SAMPLES
c
V
o
I.
«
0.
s«*
C
.0
o
3
XJ
u
Q£
2"
"3
'x
o
1—
X
o
o
o
if
IUU
80
60
40
20
0
-20
-40
-60
-80
— inn
i i i i . . i , , . .
• .
»
Somplg "0" - — *
V, '_~»' — -— r"TTTT7 -.-.,. Sample "1" "
-A^3/**? '"-.._ Sample "a" Sample "'1C „
1i Sample "G" •"
.••»• " r •
""** -
\
-
*• — .«_ Sample "F"
i i "7 i 1 1 1. c 1 r 1 .
0 61218 24 30 36 42 48 54 60 66 72
Undisturbed (room temperature, dark, and sealed) (hours)
AUTOMOBILE SALVAGE YARD SAMPLES
Figure 2. Undisturbed sample toxicity trends.
C
a>
o
w
&
a.
c
g
o
3
•a
(U
01
o
X
£
X
o
o
u
S
10U
80
60
40
20
0
-20
-40
-60
-80
_mn
1 1 1 1 1 1 1 1 ! 1 1
-
.
Sample "M"
.--*"
-~'
L
I
A Sample "L" ___— — — "•
\s* — ~~
-
1, , 1 1 1 ....!._ ,1 1 1 1 ! ..1
"o 6 12 18 24 30 36 42 48 54 6O 66 72
Undisturbed (room temperature, dark, and sealed) (hours)
-------
AUTOMOBILE SERVICE FACILITY SAMPLES
INDUSTRIAL LOADING/PARKING AREAS
c
u
o,
c
3
'o
X
o
X
o
o
o
2
C"
go - Sample "E" ..-••'" -j y
-•••' - - • " 1 as
80 - .----" -| -S
70 - / .,- --"'"Sample "C" ., j c
eo - .—/—'''' _. •••• ' 4 "•§
so - ,' / _- — — """Sample "H" J %
S-~"~"' >
'o
'x
o
l_
X
o
o
o
"J
80 , Sotnple 2£^--*y ~I^-~~ — 'Sample "J" --..-••' ^,.
70 - ,s"~ / jM.ias.~K' '.'.'•'••"""" ,---'"""" -
60 - / /--'~ ••"'' ,-''"
__ i,~-r—7 Somple "F" ...;-;' -'
40 ~ ,' ;'' Sample "0" / .-''""
30 Lv(/ / __-;,-'---'S0mple "C"
20 m^ ...--"'""•'
10 |l /••"
o J ;V
-10 - ,'/•
-20 ;/',,'
-30 f; 'v'
-40 \j
0 6 12 18 24 30 36 42 48 54 60 66 7
Settling Time (hours)
AUTOMOBILE SALVAGE YARD SAMPLES
90 - ^/Sample "L" ,.-
80 - / -
70 - / __._---•""
60 - f Sample "M" .,--'"""
50 /
40 J /
30 • S
20 f"
10 ;
0 •
-10 •
-20 -
-30 -
-40 •
_sn 1 — — .1 —i 1 — .1 1 1 1 1- 1 1
6 12 18 24 30 36 42 48 54 60
Settling Time (hours)
66 72
-------
co
o
ro
100
80
60
& 40
20
AUTOMOBILE SERVICE FACILIFC SAMPLES
\ Sample "H"
Sample "E"
0 10 20 30 40 50 60 70 80 90 100 110
Sieve Size (microns)
INDUSTRIAL LOADING/PARKING AREA SAMPLES
- 100
•, ^ --. _*"V._ Somple "0"
Sample "J" ^*"--l;^
0 10 20 30 40 50 60 70 80 90 100 110
Sieve Size (microns)
Figure 4. Sieve treatability test toxicity trends.
•a
-------
AUTOMOBILE SERVICE FACILITY SAMPLE:
o
GO
100
90
BO
70
60
50
40
30
20
10
0
-10
-20
-30
-40 K-
-50
Sample "E"
0 6 12 18 24 30 36 42 48 54 60
Aeration Period (hours)
66
INDUSTRIAL LOADING/PARKING AREA SAMPLES
100
90
80
70
60
50
•40
30
20
10
0
-10
-20
-30
-40
-50
Sample "I"
barnple j"
Scruple "C"...--
0 6 12 18 24 30 36 42 48 54 60 66 • 72
Aeration Period (hours)
AUTOMOBILE SALVAGE YARD SAMPLES
Figure 5. Aeration treatability test toxicity trends.
-------
OJ
o
•c*
t:
4)
v
O
"x
O
I—
X
j)
"o
u
ii
100
90
eo
70
60
50
40
30
20
10
0
-10
-20
-30
-40
-50
AUTOMOBILE SERVICE AREA SAMPLES
Sample "C"
•f^ ..-'"" Sample "B" ___ — - — - " I
**Z
1 1 1 f f 1 1 i i f
INDUSTRIAL LOADING/PARKING AREA SAMPLES
12 18 24 30 36 42 48 54
Photo-degradation Period (hours)
SO 66
6 12 18 24 30 36 42 48 34 60 66 72
Photo—degradation Period (hours)
AUTOMOBILE SALVAGE YARD SAMPLES
Figure 6. Photo-degradation treatability test toxicity trends.
12 18 24 30 36 42 48 54
Photo-degradation Period (hours)
-------
AUTOMOBILE SERVICE AREA SAMPLES
INDUSTRIAL LOADING/PARKING AREA SAMPLES
6 12 18 24 30 36 42 48 54 60 66
Aeration and Phota—degradation Combined (hours)
c
OJ
o
e*
a.
C
.9
tj
3
•o
u
rr
>*
'0
'x
o
I—
X
o
"o
u
3
I Uw
. 90
80
70
60
50
40
.30
20
10
0
-10
-20
-30
-40
— sn
1 1 1 1 1 1 1 1 1 1 1
-
Sample "J" -
^^•f-'" ~~~ Sample "C".-1'"
Sample •r.,--- Somole-Tr ..."
S" j>- ... '"" "'""— ~~ ' __---" '"„'" , - - " ""I
^>>^*-"~*"~" Sample "K" ^'. "-'-"-- *"""*"'
,1' ..,- — -' " ^^ *"* ,
i(\^r ^--'
Sample "F"rf
,''"
,--""'
»' • .
- /
'I i t i i ,,, i ) t i I ( , '
6 12 18 24 30 36 42 48 54 60 66
Aeration and Photo—degradation Combined (hours)
CO
o
CJl
AUTOMOBILE SALVAGE YARD SAMPLES
Figure 7, Aeration and photo-degradation combined
treatability test toxicity trends.
Sample "L"
Sample "M" .---""'
6 12 18 24 30 36 42 48 54 60 66 72
Aeration and Photo—degradation Combined (hours)
-------
AUTOMOBILE SERVICE AREA SAMPLES
INDUSTRIAL LOADING/PARKING AREA SAMPLES
o'
01
~
c
"
a*
c
o
1
]u
'x
X
0
O
w
,O
90
60
20
°
-60
-90
-,20
-150
1 ' ' ' *
^. Sample "H"
Sample "£"
i , , i ,,,,,,,
c
41
V
CL
c
0
u
•a
iti
"o
'x
o
t-
X
(3
w
O
2
90
60
30
0
-30
-60
-90
-120
1 1 — ' i -^i 1 1 1 1 1 1 1 —
-X^
Sonipio "0" jf
jS
S^
' /
ry . ^Sample "J"
i\— • '" SamDle ^" """'-.
- \ *'».,., Sample "K"
- *•--, Somple "P1
-
1 1 ' i i i 1 1 II i r
0 6 12 IB 24 30 36 42 48 54 60 06
Floatation period (floating top layer samples) (hours)
72
0 6 12 18 -24 30 36 42 48 54 60 66 . 12
Floatation period (floating top layer samples) (hours}
AUTOMOBILE SALVAGE YARD SAMPLES
Figure 8. Floatation treatability test toxicity trends
(top layer samples).
90 -
60 -
-120 -
-150
Sample "M"
Somple "L"
0 6 12 IB 24 30 36 42 48 54 60 66 72
Floatation period (floating top layer samples) (hours)
-------
AUTOMOBILE SERVICE AREA SAMPLES
'•5 -'80 h
6 12 18 24 30 36 42 48 54 60
floatation period (middle layer samples) (hours)
66 72
INDUSTRIAL LOADING/PARKING AREA SAMPLES
c
£
"x
o
t—
X
2
o
o
3
90
60
30
0
-30
-60
-90
-120
-150
-180
f — —
I Sample "D"
- I
. I Sample "I" S°TS!*."?.'..
AJ,-— ~~~~ _^^=^ , _____lsamEi£"J"
*,,V"-"-'""'-
'«*; Somple "K"
i
" \
\ **""
\ Sample "F"_ ,.-"'
\ „•""
\ -"'"
v **
*j. ' • -
J 1 1 1 1 1 1 i t i t
6 12 18 24- 30 36 42 48 54' 60 66 72
Floatation period (middle layer samples) (hours)
o
"xl
AUTOMOBILE SALVAGE YARD SAMPLES
Figure 9. Floatation treatability test toxteity trends
(middle layer samples).
90
60
3tr
0
-30
-60
-90
-120
-150
-180
Sample "M"
Sample "L"
6 12 18 24 30 36 42 48 54 60 66 72
Floatation period (middle layer samples) (hours)
-------
GO
o
CO
AUTOMOBILE SERVICE AREA SAMPLES
INDUSTRIAL LOADING/PARKING AREA SAMPLES
^
"£
V
g
s^.
C
O
u
•o
V
or:
^
»—
1
X
o
I
2
1OU
75
50
25
0
25
—
-50
-75
-100
-125
-ISO
1 1 1 r 1 i 1 1 1
.
.
Sample "H"
•*"""**» -''
.
-
.
Sample "E"
,,,,,,,,
c
3
V
a.
c
_o
"a
u
5,
^
X
o
t—
X
o
"3
u
3
100
75
SO
25
0
—25
-SO
-75
-100
-125
-150
1 I I 1 1 1 1 1 1—
/
/Soropte "D"
/
/
/ Sample "1"
/ S^*^ '•• Sample "C"
- Somple "K" s""**^££/^^ ^ "'•• ..-•' ""--. -
\ ,»f ,'*[ SompTe~"J" ""*"'
\ *'*^^***' '
* — -7**^ i
* \
\ Sample "P1
""*
i __i _ i _i i i i i ,
Alum dosage (mg alum/L sample)
0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18 0.20
Alum dosage (mg a!um/L sample)
AUTOMOBILE SALVAGE YARD SAMPLES
Figure 10. Alum addition treatability test toxicity trends.
c
a)
o
a*
a.
c
o
"8
"O
(E
~
X
o
X
o
"o
o
2
100
75
50
25
0
-25
-SO
-75
-100
-125
-150
0.
1 1 1 1 1 1 1 i — —I •
.
Sample "M" _,., ^.'-
"""'/- ^— "
Sample "L"
-
-
.
, , i , i i i i. 1
00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18 0.;
Alum dosage (mg alum/L sample)
-------
REPORT ON ENHANCING EFFECTS OF LOW FREQUENCY VIBRATION IN SOIL HASHING
by: Malvina Milkens
Releases Control Branch
Risk Reduction Engineering Laboratory
US Environmental Protection Agency
Edison, New Jersey 08837
Carl Gutterman
Foster Wheeler Development Corp.
Livingston, New Jersey 07039
Japan K. Basu
Foster Wheeler Enviresponse, Inc.
Livingston, New Jersey 07039
ABSTRACT
The project evaluated the use of low-frequency vibrations to enhance the
extraction of seven inorganic and three semivolatile organic contaminants from
synthetic soils. Acid and surfactant extraction solutions were used. The
contaminated soil and extractant solutions were stirred and vibrated at
different frequencies and amplitudes. Vibrations increased the removal of
inorganic contaminants, which originated in a compound that has a high degree
of insolubility in water. The results of analytical tests for the removal of
semivolatile organic contaminants were inconclusive.
This paper has been reviewed in accordance with the U.S. Environmental
Protection Agency's peer and administrative review policies and approved for
presentation and publication.
309
-------
INTRODUCTION
Under a U.S. EPA contract, Foster Wheeler Enviresponse, Inc. is evaluating
several technologies for cleaning soil contaminated with inorganic and organic
hazardous waste. One such evaluation washes contaminated soil in appropriate
extractants using mechanical mixing with and without vibration. Low-frequency
mechanical vibrations enhance chemical processes involving heat and mass
transfer. The experiments attempted to clean materials contaminated with
inorganic or semivolatile organic substances [1,2].
A literature search was conducted; a series of preliminary tests were then
performed with and without low-frequency vibrations. Results in both inorgan-
ic [arsenic, cadmium, chromium, copper, nickel, lead, and zinc] and semivolat-
ile organic [pentachlorophenol, anthracene, and bis(2-ethylhexyl)phthalate]
contaminants were examined.
The test setup and procedure, the test matrix, and the results will be
described; conclusions and recommendations for future work will also be
presented.
MATERIALS AND METHODS
Literature Search
The literature search used two databases -- NERAC™ and DIALOG™. The
initial goal was to find any research data available on soil washing assisted
by low-frequency vibration. Search of the two databases revealed only the
Harbauer method, for which only general information was available. No
quantitative data on this process are readily available.
The search in DIALOG™ was then broadened to include any information on
mass transfer enhancement by low-frequency mechanical vibrations. Papers in
this category discussed extraction of vegetable oils, metal recovery from
ores, and particle separation. These papers indicated that low-frequency
mechanical vibrations do indeed assist mass transfer for various processes.
310
-------
In most cases, the optimum vibration frequency was in the range of 25 to 60
hertz (Hz); amplitude was a few millimeters (mm). V.L. Demidov [3] reported
recovery of gold and silver from ores using low-frequency vibration mixing.
Vibration-Assisted Soil Cleaning Tests
The test equipment employed a mechanical shaker, a function generator, and
a power amplifier. (See Figure 1.) The function generator/amplifier system
provides vibratory response to the shaker table over a range of frequencies
and amplitudes for different wave forms. Tests were conducted at frequencies
ranging from 30 to 60 Hz and at double amplitudes from 1/32 to 1/8 inches (in)
[0.8 to 3.2 mm].
Soil and extractant were put in a stainless steel beaker, which was
fastened to the shaker table with a screw. An inverted plastic funnel was
press-fit into the beaker opening to prevent splashing. A mechanical mixer
was inserted through the funnel opening. This apparatus was used to duplicate
the base case soil cleaning [1].
In the beaker, three perforated, partial, annular plates (See Figure 2.)
facilitated transmission of vibratory energy to the soil-extractant mixture,
while maintaining a near-normal circulation pattern. The three plates were
spaced vertically to remain immersed in the solution during all tests.
In each test, the beaker was subjected to appropriate mixing and vibratory
action for 30 minutes. At the end of each test, the extracting solution was
decanted through a filter, followed by three 50-milliliter (ml) washings. The
filtrate was analyzed for the presence of inorganic (or semivolatile organic)
contaminants. For the inorganic runs, each test was interrupted every 5
minutes to adjust the pH to 1.0, by adding 6-molar nitric acid (6M HN03)
solution (a total of 20 ml added during each test).
311
-------
OJ
(—>
ro
j Mixer ;j;
%
.>!4i
S|
\\'
'r
tis
ii.
t :~
i -•
*
1 r-
,
:!*
Is
';; .
•; : •
tfi, ,
jiit « ;•;.
51- J i
M ' " . I :
Function
Generator
Stainless Steel
Beaker
Figure 1. Vibration-assisted soil washing setup.
-------
0.045" Thick
Stainless Steel
Plates (31
PLAN VIEW
/ 4"
\
z^
I/ £2
i\
FRONT
\
/
\|
/)
„
..^••^MM 1
1
L
ELEVATION
/ Sta
\
s
/
\
7
S
iniesa aleel
Beaker
1/2
/
\
1/2
/
\
1/2
/
Figure 2. Vibrating beaker configuration.
313
-------
Two separate sets of tests were performed: one analyzing the removal of
inorganics; the other, semivolatile organics. The test conditions were chosen
from the results of mixing tests conducted earlier [1,2].
For the removal of inorganics, a test from a previous study was chosen as
the base case [1, Test 35]. This test was based on the following parameters:
Soil
Reagent
PH
Time
Temperature
Liquid/soil
250 to 2000 micron fraction of SARM* III
Nitric acid (HN03)
Initial 1.0, final 1.3
30 minutes
75°F
15
This case was selected because the removal efficiencies for the contami-
nant metals [arsenic, cadmium, chromium, copper, nickel, lead, and zinc],
using mixing, ranged from 14.51 for lead to 76.0% for cadmium -- thus leaving
some room for improvement. The starting pH of 1.0 was desirable since this
level was found to be most effective [1]. The 30-minute time was considered
reasonable. The 74°F to 76°F temperature (ambient) eliminated the need for
any heating. A total of seven tests were performed using freshly wet-sieved
SARM III. All seven tests were performed in triplicate so that their preci-
sion could be assessed. The following conditions were maintained during all
tests for the removal of inorganic contaminants:
Soil
Reagent
Initial pH
Time
Temperature
Liquid/Soil
250 to 2000 micron fraction of SARM III (20 gm)
0.1M HN03 (300 gm)
1.0
30 minutes
Ambient
15
* Synthetic Analytical Reference Matrix
314
-------
The first test (in triplicate) was a control case. No mechanical vibra-
tions to the beaker were imparted; stirrer mixing was operated at 500 rpm.
The remainder of the test matrix repeated the first test with vibrations at
respective frequencies of 30, 45, and 60 Hz, Each frequency was tested at
double amplitudes of 1/32 and 1/16 in.
The tests for removal of semi volatile organics were performed in a similar
manner. Eight tests (each in triplicate) used freshly wet-sieved SARM I,
Test parameters were maintained as follows:
Soil 63 to 250 micron fraction of SARM I (20 gm)
Reagent 0.75% Achowet Surfactant (100 gm)
Time 30 minutes
Temperature Ambient
Liquid/Soil 5
The first test was the control test. No vibrations were employed; stirrer
mixing was operated at 500 rpm. The remaining seven tests added vibrations at
three different frequencies (30, 45, and 60 Hz) and varying amplitudes of
1/32, 1/16, and 1/8 in.
Technology Applications, Inc. of Cincinnati conducted all soil and
filtrate analyses for the presence of inorganic and semivolatile organic
contaminants. The inorganics analyses measured arsenic, cadmium, chromium,
copper, nickel, lead, and zinc. The semivolatile organics analyses quantified
the presence of pentachlorophenol, anthracene, and bis(2-ethylhexyl)phthalate.
The SARM III (250-2000 /an) and SARM I (63-250 urn) feed samples were
analyzed in triplicate to provide the base contaminant levels. All filtrates
from the seven triplicate inorganic tests and the eight triplicate semivola-
tile organic tests were analyzed to determine contaminant removal efficien-
cies.
315
-------
RESULTS AND DISCUSSION
.Inorganics
Feed Soil Analysis
Table 1 provides the analysis of the SARM III feed, performed in tripli'
cate. It also shows the arithmetic means (x).
TABLE 1. ANALYSIS OF SARM III (250-2000 urn fraction) FEED
Contaminants
Arsenic
Cadi urn
Chromium
Copper
Nickel
Lead
Zinc
1
183
432
86.5
1008
128
4725
3361
Replicates
2
165
402
102
1360
148
4416
4148
3
178
414
137
1771
178
5953
5927
X
175
416
109
1380
151
5031
4479
Replicates: Concentration of contaminants in soil (mg/kg) from
identical experiments.
x: Arithmetic mean of three values (mg/kg)
FiltrateAnalysis
Table 2 summarizes the results of the seven triplicate tests for removal
of inorganic contaminants. It also shows the arithmetic means'(x).
MaterialBalance
Table 3 summarizes the material balance data for six tests on removal of
inorganics. It gives "Percent Extracted" and "Percent Remaining on Soil"
316
-------
TABLE 2. SUMMARY OF EXTRACTION RESULTS -- INORGANIC CONTAMINANTS
Test
description
Stirring, no vibration
30 Hz, 1/16 in, da
(low- frequency
vibration)
30 Hz, 1/32 in, da •
(low- frequency
vibration)
45 Hz, 1/16 in, da
(low-frequency
vibration)
Contaminant
Arsenic
Cadmium
Chromium
Copper
Nickel
Lead
Zinc
Arsenic
Cadmium
Chromium
Copper
Nickel
Lead
Zinc
Arsenic
Cadmium
Chromium
Copper
Nickel
Lead
Zinc
Arsenic
Cadmium
Chromium
Copper
Nickel
Lead
Zinc
Test
replicates
81
57.8
72.8
84.5
82.6
107.7
60.1
71.6
2A
77.8
123.8
96.7
85.2
105.2
107.9
107.2
3A '
59.0
85.9
76.0
67.9
73.9
61.9
73.1
4A
70.8
87.1
92.7
84.9
89.8
94.0
90.2
B2
60.1
75.2
84.9
89.1
92.2
61.0
76.5
2B
62.4
99.9
89.3
78.9
90.5
74.2
87.1
3B
60.7
83.1
70.4
64.4
70.2
61.6
66.0
4B
64.9
87.4
77.0
69.3
78.4
87.4
75.4
B3
62.3
85.8
83.3
75.3
78.6
59.0
78.2
2C
62.6
86.2
83.8
77.3
83.1
80.7
82.6
3C
62.0
84.1
80.0
12.2
77.7
65.8
74.6
4C
64.3
86.0
77.3
71.7
77.3
87.8
77.6
X
60.1
77.9
84 ."2
82.3
92.8
60.0
75.4
67.6
103.3
89.9
80.5
92.9
87.6
92.3
60.6
84.4
75.5
68.2
73.9
63.1
71.2
66.7
86.8
82.3
75.3
81.8
89.7
81.1
(continued)
317
-------
TABLE 2. Continued
Test
description
45 Hz, 1/32 in, da
(low-frequency
vibration)
60 Hz, 1/16 in, da
(low-frequency
vibration)
60 Hz, 1/32 in, da
(low- frequency
vibration)
Replicates: Percentages
x: Arithmetic
da: Double amp!
Contaminant
Arsenic
Cadmium
Chromium
Copper
Nickel
Lead
Zinc
Arsenic
Cadmium
Chromium
Copper
Nickel
Lead
Zinc
Arsenic
Cadmium
Chromium
Copper
Nickel
Lead
Zinc
extracted
5A
66.0
86.7
80.0
73.8
80.0
81.1
79.0
6A
65.5
86.1
82.8
75.8
81.0
94.9
. 78.9
7A
63.2
86.4
80.5
75.0
81.4
83.0
82.0
from three identi
Test
replicates
5B
62.9
92.1
76.4
72.7
79.0
76.1
78.1
6B
63.8
86.2
82.3
77.6
80.0
85.9
75.5
7B
62.6
89.6
76.4
67.8
74.7
78.1
71.7
5C
61,5
82.3
77.3
73.1
77.0
74.5
75.2
6C
66.3
85.6
82.3
73.6
77.7
93.7
79.9
7C
60.3
82.3
80.0
71.6
77.7
91.7
77.6
X
63.5
88.7
77.9
73.2
78.7
77.2
77.4
65.2
86.0
82.5
75.7
79.6
91.5
78.1
62.0
86.1
79.0
71.5
77.9
84.3
77.1
cal experiments
mean of three values, %
itude
318
-------
TABLE 3. MATERIAL BALANCE DATA -- INORGANICS
Test
Bl
B3
2A
2C
7A
Contaminant
Arsenic
Cadmium
Chromium
Copper
Nickel
Lead
Zinc
Arsenic
Cadmium
Chromium
Copper
Nickel
Lead
Zinc
Arsenic
Cadmium
Chromium
Copper
Nickel
Lead
Zinc
Arsenic
Cadmium
Chromium
Copper
Nickel
Lead
Zinc
Arsenic
Cadmium
Chromium
Copper
Nickel
Lead
Zinc
Percent
extracted
57.8
72.8
84.5
82.6
107.7
60.1
71.6
62.3
85.8
83.3
75.3
78.6
59.0
78.2
77.8
123.8
96.7
85.2
105.2
107.9
107.2
62.6
86.2
83.8
77.3
83.1
80.7
82.6
63.2
86.4
80.5
75.0
81.4
83.0
82.0
Percent1
remaining
on soil
46.6
13.5
12.3
8.7
10.7
40.9
7.7
56.9
22.3
11.4
14.1
12.7
56.5
10.0
43.1
10.6
9.5
7.1
9.3
21.8
4.6
42.3
8.3
15.5
8.0
8.7
34.6
6.2
55.4
18.1
12.3
8.5
10.0
13.8
5.6
Percent
recovered*
104.4
86.3
96.8
91.3
118.4
101.0
79.3
119.2
108.1
94.7
89.4
91.3
115.5
88.2
120.9
123.4
106.2
92.3
114.5
129.7
111.8
104.9
94.5
99.3
85.3
41.8
115.3
88.8
118.6
104.5
92.8
83.5
91.4
96.8
87.6
(continued)
319
-------
TABLE 3. Continued
Test
78
Contaminant
Arsenic
Cadmium
Chromium
Copper
Nickel
Lead
Zinc
Percent
extracted
62.6
89.6
76.4
67.8
74.7
78.1
71.7
Percent
remaining
on soil
40.0
13.7
16.4
26.1
10.7
15.1
5.4
Percent
recovered*
102.6
103.3
92.8
93.9
85.4
93.2
77.1
* Amount of contaminant in feed soil
values for each inorganic contaminant. The sum of these two percentages
should ideally be 100; the calculated values are shown as "Percent Recovered."
As seen in Table 3, most "Percent Recovered" values fall between 80 and 120,
indicating good material balance.
Extraction^..Ef f_ici encv
Table 4 summarizes the average percent removal efficiencies for the seven
inorganics tests. Symbols indicate improvement (y) or lack of improvement
(n)» relative to the no-vibration control case.
Table 4 also lists the seven inorganic compounds in SARM III and their
solubility in cold water. Table 5 shows the percent increase in extraction
efficiency over the no-vibration control case.
320
-------
TABLE 4. AVERAGE PERCENTAGE EXTRACTION RESULTS -- INORGANIC CONTAMINANTS
Description
Double amplitude (in)
Test
Hetals
Arsenic
Cadmium
Chromium
Copper
Nickel
Lead
Zinc
(As)
(Cd)
(CD
(Cu)
(Ni)
(Pb)
(Zn)
Stirring,
no vibration
81,2,3
60
77
84
82
92
60
75
.1
.9
.2
.3
.8
.0
.4
30 HZ
1/16
2A.B.C
67.6
103.3
C9.9
80.5
92.9
87.6
92.3
45 Hz
60 Hz
1/32 1/16 1/32
3A,B,C 4A,B,C 5A,B,C
y 60.6
y 84.4
y 75.5
n 68.2
y 73.9
y 63.1
y 71.2
y 66.7
y 86.8
n 82.3
n- 75.3
n 81.8
y 89.7
n 81.1
y 63
y 88
n 77
n 73
n 78
y 77
y 77
.5 y
.7 y
.9 n
.2 n
.7 n
.2 y
.4 y
1/16
6A.B.C
65.2 y
86.0 y
82.5 n
75.7 n
79.6 n
91.5 y
78.1 y
1/32
7A,B,C
62.0 y
86.1 y
79.0 n
71.5 n
77.9 n
84.3 y
77.1 y
Metal source
Solubility*
1.2(0') ASjO.,
114(0') 3CdS04"8H20**
s er(N03)3'9H2Q**
24(0") CuSOySHjD**
243(0") Ni(N03)2"6H20»*
0.004(18") PbS04'PbO
0.004(18') ZnO
* Solubility in 100 parts cold water [degrees in Centigrade; s = soluble]
** Water-soluble compounds
y/n Indicates improvement (y) or lack of improvement (n) relative to "no-vibratfon" mode
-------
TABLE 5. PERCENT INCREASE IN INORGANICS EXTRACTION
EFFICIENCY OVER CONTROL CASE
Test
Metal
Arsenic
Cadmium
Chromium
Copper
Nickel
Lead
Zinc
Stirring,
no vibration
Bl, 2, 3*
-
-
-
-
-
-
30
1/16"
2A,B,C
12
32
7
-2
0
46
22
Hz
1/32"
3A,B,C
1
8
-10
-17
-20
5
-6
45
1/16"
4A,B,C
11
11
-2
-9
-12
50
8
Hz
1/32"
5A,B,C
6
14
-7
-11
-15
29
3
60
1/16"
6A,B,C
8
10
-2
-8
-14
53
4
Hz
1/32"
7A,B,C
3
11
-6
-13
-16
41
2
* Control Case
The following observations are based on the results presented in Tables 4
and 5:
o Vibration improved the extraction of arsenic, lead, cadmium, and zinc.
o The effect of vibration was more enhanced at 1/16 in (double amplitude)
than that at 1/32 in.
o In general, the enhancement due to vibration improved at lower frequen-
cies; 30 Hz at 1/16 in (double amplitude) resulted in maximum enhancement.
o Lead showed the maximum enhancement due to vibration.
o Test precision is well within the guideline value of 30% relative standard
deviation for all tests.
322
-------
Semivolatile Organics
Feed Soil Analysis
Table 6 provides the analysis of SARM I (63-250 fun) feed performed in
triplicate. It also shows the arithmetic means (x).
TABLE 6. ANALYSIS OF SARM I (63-250 urn fraction) FEED
Contaminant Re.pl icaies.
1 2
Pentachlorophenol 300 305 310 305
Anthracene 7895 8230 6860 7662
Bis(2-ethylhexyl)phthalate 801 775 850 809
Replicates: Concentration of contaminants in soil (mg/kg) from identical
experiments
x: Arithmetic mean of three values (mg/kg)
Filtrate Analysis
Table 7 presents the results of the eight triplicate tests for the removal
of semivolatile organics.
MaterialBalance
Very poor material balance data were obtained for all semivolatile organic
tests. The lack of material balance may be caused by the analytical tech-
niques employed. This makes suspect any evaluation of the degree of extrac-
tion enhancement due to vibration.
323
-------
TABLE 7. AVERAGE PERCENTAGE EXTRACTION RESULTS -- ORGANIC CONTAMINANTS
Stirring,
Description no vibration 30 Hz 45 Hz 60 Hz
Double amplitude (in) 0 1/8 1/16 1/32 1/16 1/32 1/16 1/32
Test 1A,B,C 2A,B,C 3A,B,C 4A,B,C 5A,B,C 6A.B.C 7A,B,C 8A,B,C
Compound
ro Pentachlorophenol . 28.2 13.0 21.6 27.1 37.5 33.5 17.9 26.2
% Increase over "Control" -54 n -23 n -4 n +33 y +19 y -37 n -7 n
Anthracene 0.15 0.043 0.10 0.35 0.29 0.34 0.29 0.34
Bis(2-ethylhexyl)phthalate 2.01 1.11 0.29 10.8 12.9 5.0 4.4 2.63
% Increase over "Control" -45 n -86 n +437 y +541 y +149 y +119 y +29 y
y/n Indicates improvement Cy) or lack of improvement (n) relative to "no-vibration" mode
-------
Extraction Efficiency
Table 7 summarizes the average percent removal efficiencies for the eight
semivolatile organic tests. It shows the percent increase in extraction
efficiency over the no-vibration control case for pentachlorophenol and bis(2-
ethylhexyl)phthalate. Such values for anthracene are eliminated due to very
small removal efficiencies.
CONCLUSIONS
Inorganics
>-
1. The limited tests of vibration-assisted soil washing show an improvement
in removal efficiency for basic lead sulfate (PbS04-PbO), As203 and ZnO
which is relatively insoluble in water. Water-soluble compounds were
extracted, to a large extent, by the mixing action alone; there was little
enhancement with vibration, with the exception of hydrated cadmium sulfate
(3CdS04«8H20) -- which is water-soluble -- showed extraction enhancement
due to vibration.
2. In general, vibration enhancement for inorganics improved at the larger of
the two amplitudes tested (1/16 in and 1/32 in), and at lower frequencies
(30 Hz).
3. The triplicate test data analysis shows good precision for the inorganics.
Good material balances were also obtained.
Semi volatile Organics
1. The semivolatile organics tests showed very little extraction of anthra-
cene with and without vibration. There was a complete lack of material
balance for all three semivolatile organics. The analyses of the tripli-
cate test results showed inadequate precision for most tests. Therefore,
no meaningful conclusions can be drawn from the results obtained for the
extraction of semivolatile organic contaminants.
325
-------
RECOMMENDATIONS
Further work should be performed to establish definitive quantitative
trends. It should include the following:
General
o Further tests on alternate extractants are necessary to draw sound conclu-
sions on the enhancement of soil washing by low-frequency vibration.
o Frequency and amplitude effects should be explored; a large range of
amplitudes should be investigated. This would necessitate the use of
improved vibration equipment that would allow testing at a wider range of
frequencies and amplitudes (such as a vibrating screw device).
o Tests with different time durations should be conducted to evaluate
optimum elapsed times.
o Further testing should include evaluation of vibration enhancement for
differing sizes of soil fractions. Varying liquid-to-solid ratios should
also be investigated.
o The use of alternative extractants and the effect of heating in the tests
should be considered.
Inorganics
o Further testing should incorporate better pH control so its influence can
be excluded from the results. This will facilitate more meaningful
comparisons and performance evaluations. It could involve two sets of
experiments. In one set, short-time (10-min) tests without any pH adjust-
ments can determine relative performance. In another, longer-term (30- to
60-min) tests can be conducted, but with more frequent pH adjustments so
that all tests have close pH control.
326
-------
Semi volatile Organics
o Additional surfactants should be tested.
ACKNOWLEDGEMENTS
This study was sponsored by the Releases Control Branch of the EPA Risk
Reduction Engineering Laboratory, under Contract No. 68-C9-0033 to Foster
Wheeler Enviresponse, Inc. and Foster Wheeler Development Corporation. This
paper has been reviewed in accordance with the U.S. Environmental Protection
Agency's peer and administrative review policy and approved for presentation
and publication.
The authors gratefully acknowledge the assistance provided by P. Steiner,
K. Ahluwalia, and E. Gonzalez of Foster Wheeler Development Corp. in con-
ducting the program. Additional thanks are due to M. Krawchuk who assembled
the test apparatus. Technology Applications, Inc. provided all analytical
services.
The support of the Foster Wheeler Enviresponse, Inc. staff is deeply
appreciated, especially the technical input of R. Gaire and the editing by
Marilyn Avery. RGB technical reviewers also assisted in editing this work.
327
-------
REFERENCES
1. "Laboratory Feasibility for the Removal of Lead and Other Heavy Metals
from Soils, Sludges, and Sediments at Superfund Sites", by T.K. Basu, C.
Gutterman, R. Raghavan. FWEI, 1989 (unpublished draft report to EPA RREL,
Edison).
2. "Ultrasonic Soil Washing Process Evaluation and Pilot-Plant Conceptual
Design", by E. T. Coles, B. Rubin, and R. Raghavan. FWEI, 1988 (unpub-
lished draft report to EPA RREL, Edison).
3. Recovery of Au and Ag from Ores and Products with Use of Vibration Mixing,
V.I. Demidov, T. Metally (USSR), Vol. 15, November 4, 1974,
328
-------
VACUUM-ASSISTED STEAM STRIPPING TQ REMOVE POLLUTANTS. FftOM CONTAMINATED
A LABORATORY STUDY
by: Arthur E. Lord, Jr., Leonard J. Sansone and
Robert M. Koerner
Geosynthetic Research Institute
Drexel University
Philadelphia, PA 19104
and
John E, Brugger
Risk Reduction Engineering Laboratory
U.S. Environmental Protection Agency
Edison, New Jersey 08837
ABSTRACT
Previous work on this project was involved with a laboratory study
of the viability of steam stripping a variety of organic chemicals from a
number of soil types with water permeabilities between 10~3 and 10"6
cm/s. The range of permeabilities was achieved by mixing appropriate
amounts of sand and silt. In this study the effects of a clay fraction
and organic fraction in the soil are investigated. Results are shown for
a predominately sand/silt soil but also with:
*
• varying amounts of clay (kaolinite and bentonite were used)
* varying amounts of organic material (a commercial top soil was
used)
The effect of a "delay time" (i.e., interruption time) in
decontamination procedure is studied. The initial rate of chemical
removal immediately upon resumption of treatment is compared with the
rate immediately before interruption.
A new analytical model for vacuum-assisted steam stripping using a
circular symmetry model is developed. Some relevant experiments were run
to obtain data which allow the model to be used more confidently
estimating decontamination times for a given spill scenario.
INTRODUCTION AND OVERVIEW
In the case of contaminated soils at Superfund (and other) sites, it
is important that the chemicals present be prevented from reaching the
329
-------
groundwater. In many locations fortunately a partially saturated or
vadose zone exists and acts as a temporary containment retarding the
downward movement of the pollutant. The possible remediation options
are:
• Excavation and off-site disposal.
* Excavation and on-site treatment.
• Insitu treatment (via a number of possible methods, e.g.,
biological, physical or chemical).
A number of these techniques (and others) have been reviewed in
other articles (1,2). These in-situ techniques have been discussed by
the authors (3) . A review of decontamination techniques is given in
reference 4.
The present study falls Into thein-situ treatment category wherein
the authors propose to have pipes inject steam into the soil beneath the
contaminated zone. Steam stripping of the chemical occurs and aided by a
vacuum at the ground surface, transports the contaminants to a collection
point where they can be properly treated. A unique aspect of the study
is the development of a geosynthetic cap assembly consisting Of
a high transmissivity geotextile and a flexible membrane liner
(geomembrane). The vacuum is applied to the underside of this liner and
the contaminated gas and/or liquid rooves beneath the liner in the
geotextile to the outlet ports. A schematic diagram of a proposed system
is given in Figure 1.
There have been a few steam stripping soil decontamination studies
reported in the literature (5,6,7,8). These works were field-oriented,
sites-of-opportunity projects with no attempt to look at the general
problem of the feasibility of steam stripping a wide variety of* chemicals
from a wide variety of soils.
The present work continues a long term study to determine the
ability of vacuum-assisted steam stripping to decontaminate general
organic chemical species from a variety of soil types.
PREVIOUS WORK ON THIS PROJECT
The work performed previously on this project has been reported in
detail elsewhere (3,9,10,11,12). Only a brief review of the results of
this work will be given here. More detail can be found in the cited
references. Among the tasks undertaken, in a wide variety of soils:
• Observations were made of the transient steam front movements in
two dimensional flow.
• The steam permeabilities were determined in conventional one-
dimensional flow.
• The efficiency of steam stripping kerosene and a number of
330
-------
individual organic chemicals from soils was determined. The
chemicals included dodecane, decane, octane, octanol and butanol.
The analytical methods included either volume separation of the
outflow material (for kerosene) or gas chromatography (GC) of
extractions from the soil (for the compounds). Work was done in
regard to determining the limits of confidence in the soil
extraction/gas chromatography analytical procedure.
* The effect of steam pressure and temperature on the steam
stripping capability was determined.
• Comparison of the efficiency of steam stripping versus the other
common in-situ techniques, i.e., air stripping, vacuum extraction
and heat alone.
* A steady state analytical model was developed where steam flows
upward to the collection cap from pipes embedded in the soil. Use
of this model allows calculation of the decontamination time for a
given spill. Due to certain objections raised concerning the
previous model and its use, a new model is presented here and
analytical results given.
* A small scale model of the geosynthetic cap was fabricated and
used to determine its feasibility as a cover assembly during steam
stripping.
It was felt that the results from the above topics indicated that the
vacuum-assisted steam stripping technique showed significant promise as a
soil decontamination method for a wide range of soil types and chemicals.
The authors deemed that some additional work would be necessary in
helping to evaluate the feasibility of using vacuum-assisted steam
stripping to remove organic chemicals from soil. These items were:
» The effect of a clay fraction in the soil on the chemica*! removal
efficiency.
« The effect of an organic fraction in the soil on the chemical
removal efficiency.
* The development of a better theory with which to determine the
time to decontaminate a given soil/chemical situation,
• The effect of a delay time in the decontamination processing upon
the relative rate of chemical removal.
The next Section describes the results of this work.
PRESENT WORK
EFFICIENCY OF ORGANIC CHEMICAL REMOVAL VIA VACUUM-ASSISTED STEAM
STRIPPING
It was decided in the interest of better use of the graduate
student's time that the soil extractions and gas chromatography work
would be handled by a firm specializing in this type of chemical
331
-------
analytical work. Fortunately we were able to obtain the services of
Technology Application, Inc. , Who do soil chemistry work the the U.S. EPA,
EMSL Analytical Support Laboratory in Cincinnati.
The matrix of chemicals and soil types which were used in vacuum-
assisted steam stripping studies are shown in Table 1. We chose to work
with dodecane because it is a very difficult chemical to remove due to
its low vapor pressure (boiling point = 216°C) . It thus presents an
extreme test for the vacuum-assisted, steam stripping decontamination
procedure. The benzene/ethyl benzene/toluene/xylene (BETX) was chosen
for it is a commonly-used simulant for gasoline. The soil extraction
procedure and gas chromatography approach were the method of analysis, as
was the case in earlier work described in references 12 and 13. The only
differences were that Soxhlet extraction was used here instead of
mechanical agitation of the mixed "extraction fluid and soil used earlier
(11,12), In the Soxhlet work here, methyl ene Chloride was used as the
extraction fluid. In our earlier work, either methylene chloride (11) or
ethyl ether (12) was used as the extraction fluid.
The vacuum-assisted steam stripping runs were performed with the
apparatus described earlier (3) and shown in Figure 2. The sample cells
were now made of brass instead of plexiglass, for the BETX was extremely
corrosive to plexiglass. The steam pressure was 5 psi gauge at the input
end in all cases, and full vacuum was applied at the other end. (The
vacuum amounted to about 10-inches of Hg for the sand and up to 20-
inches of Hg for the silt.)
The results are given in Table 1. It is seen that the level of
dodecane can be reduced much lower in sand than in the 50 sand/50 silt.
Of course the flow is smaller in the 50/50 soil than in the sand (about
6-times slower). It is felt that the results are quite favorable
considering that dodecane has an extremely high boiling point and the
steam stripping treatment times were not overly long.
The other materials involved in the other entries in Table 1 are
currently being worked on and results will be presented at the
Conference.
EFFICIENCY OF ORGANIC CHEMICAL REMOVAL VIA VACUUM EXTRACTION
Although vacuum extraction studies per se, were not part of our
initial work plan, we have found them over the course of our work to be
very productive experiments to ascertain the "binding ability" of organic
chemicals to particular soil types. Screening experiments (using a
variety of soils) can be performed much quicker with vacuum extraction
than with the much more time-consuming steam stripping measurements. Of
course our ultimate goal is to ascertain the feasibility of steam
stripping organic chemicals from a wide range of soils.
332
-------
TABLE 1. RESULTS OF VACUUM-ASSISTED STEAM STRIPPING EXPERIMENTS
Chemical
Dodecane
Dodecane
Dodecane
Dodecane
Dodecane
Initial
Concentration
(% and ppm)
5%
50,000 ppm
5%
50,000 ppm
5%
50,000 ppm
5%
50,000 ppm
5%
50,000 ppm
« Benzene 2% each
•Ethyl— 20,000 ppm
Benzene each
'Toluene
•Xylene (BETX)
BETX
BETX
BETX
BETX
2% each
20,000 ppm
each
2% each
20,000 ppm
each
2% each
20,000 ppm
each
2% each
20,000 ppm
each
Soil Time of
Type Treatment
(hours)
Beach 2
Sand
50 Sand 8
50 Silt
45 Sand 8
45 Silt
10 Kaolinite
48 Sand 8
48 Silt
4 Bentonite
45 Sand
45 Silt
10 organic
Beach 2
Sand
50 Sand 8
50 Silt
45 Sand 8
45 Silt
10 Kaolinite
48 Sand 8
48 Silt
4 Bentonite
45 Sand 8
45 Silt
10 Organic
Average Final
Concentration
(with Standard Amount
Deviation) Removed
(ppm) (%)
421 ± 53 99.2
(n = 5)
3826 ± 2070 92.4
(n - 5)
333
-------
Figure 3 shows a schematic diagram of the soil cell in the furnace.
The cell size is the same as for the steam stripping experiments. The
main thrust in these experiments was to ascertain how the rate of
kerosene removal was affected by the addition of a clay and an organic
fraction to the 50 sand/50 silt type soil. The experimental procedure is
the same as discussed in reference 13. Figure 4 shows the results of a
45 sand/45 silt/10 kaolinite soil. (The results of the 50/50 soil are
shown for reference purposes.) It is observed that the rate of kerosene
removal has been perceptively slowed {from the 50 sand/50 silt soil) by
the addition of the kaolinite fraction. Figure 5 indicates results for
the addition of 10% bentonite and 10% organic (i.e., top soil) with the
sand/silt mixture. It is surprising that the bentonite-doped-soil does
not have a slower removal rate than the kaolinite-doped-soil.
EFFECTS OF A DELAY TIME IN DECONTAMINATION PROCEDURE
There have been indications in the field that if vacuum extraction
is performed for a given period, and then stopped, the initial rate of
contaminant removal .upon resumption of treatment is much larger than the
final rate before stopping treatment. We have attempted to model this
"delay effect" in the laboratory in order-to gain insight into the
problem. The emphasis will be upon the effect of soil type on the "delay
effect". The procedure is the same as in the vacuum extraction studies
described in the previous section with the exception that the treatment
is stopped at about 50% kerosene removal. Then the sample is sealed
tightly and stored for a given "delay time" before the resumption of the
vacuum treatment.
Some results for the 50/50 soil is shown in Figure 6.
Results for 45/45/10 kaolinite clay are shown in Figure 7.
334
-------
ANALYTICAL MODEL FOR VACUUM-ASSISTED STEAM STRIPPING
In a previous work (3) a theoretical model was developed, with the
ultimate goal of determining the decontamination time for a given
chemical spill. The model used a point steam source buried benesth the
surface and collection over a wide area on the surface. The model could
be called a point steam source buried in a semi-infinite half space.
Unfortunately, much too high steam pressures (unrealistic values) were
used in the calculation. In the present work a model based on circular
symmetry will be used. This symmetry may be closer to that actually used
in the field. The model is shown in Figure 8. Electrical field mapping
experiments (13) have shown that the pressure contours and flow lines
will essentially conform to circular symmetry whenever eight, or more
sources (sinks) are employed symmetrically around the center sink
(source) . Both the dc electric field and the steady state pressure
profiles are governed by Laplace's Equation.
In steady state, with constant temperature1, the steam
pressure, P, will be governed by Laplace's Equation, which in the
cylindrically-symmetric case can be written:
(1)
r dr
where
r is the radial position as shown in Figure 9.
With the boundary conditions:
P = 0 at r = r0 (inner pipe radius)
P = Pd at r = rd (the outer pipe's distances from the center),
the solution of Equation <1) is:
r
(2)
We use Darcy's law in the following vector form
We have performed experiments on a small cylindrically-symmetric
steam flow device of dimension 10-inch diameter and 1 foot high, with
steam introduced in the center and sand as the soil. The temperature of
the entire mass of soil was 100°C, once steady state was achieved. The
same constancy of temperature was also achieved in earlier two-
dimensional work. (3,9)
335
-------
v - -k VP
where
V is the steam vector velocity
k is the permeability of the steam, and
V& is the pressure gradient vector
In the case of cylindrical symmetry, Equation (3) reduces to
v--,f .
dr
where V is the radial velocity. Using equation (2) in Equation (4) we
arrive at
V - -k
(5)
The volume flow rate, Q, through a lateral surface of length, 1, is (see
Figure 9)
Q = VA
f A is the lateral 'surface area, 2it
(6)
rl]
Q - -k
In
O)
In steady state, Q must be a constant as a function of r, as is reflected
in Equation (7), i.e., the volume flow rate does not depend on r.
The values of steam permeability, k, can be obtained most easily in
one dimensional flow experiments such as those described earlier
(3,10,11,12). Using the value of k determined earlier in our experiments
(actually k Pd •* kAP, is more readily accessible), the following results
for velocity flows, using Equation 4, are given in Table 2. Here we can
336
-------
TABLE 2. STEAM VELOCITIES AS A FUNCTION OF VARIOUS PARAMETERS
AP « 10 psi
r
r
0
1
2
3
4
5
d/ro
V(m/s)
Cm)
.5
.0
.0
.0
.0
,0
5m/0.2m = 25
3,9xlO~6 l/r{m)
7
3
1.
1
0.
0
.8xlO~6
. 9X10'6
95X1Q-6
-3xlO-6
98X10"6
.8xlO"6
3m/0.2m = 15 3m/0.4m =7.5 5m/0.5 = 10
4.6X10"6 l/r(m) 6.26xlO~6 l/r(m) 5.4X10"6 l/r(m)
9.2xlO~6 1.25X10"5 1
4.6X10'6 6.25X1Q-6 5
2.3X10"6 3.1X10'6 2
1.5X10'6 2.1X10"6 1
1.
1
. 1X10'5
.4X10-6
.7xlO~6
.8xlO~6
35X10'6
. IxlO"6
AP = 20 psi
r
r
0
1
2
3
4
5
d/ro
V(m/s)
(m)
.5
.0
.0
.0
.0
.0
5m/0.2m = 25
7 . 8xlO~6 1/r (m)
1.
7
3
2
1.
1.
56X10""5
.8xlO~6
. 9X10-6
. 6X1Q-6
95X10"6
56X10-6
3m/0.2m = 15 3m/0.4m =7,5 5m/0.5 = 10
A
9.2X10"6 l/r(ra) 1.25X10'6 l/r(m) 1.08X10""6 l/r(m)
1.84X10"5 2.5X10"5 2.
9.2X10-6 1.25X1Q-6 1
4.6X10-6 6.25X10"6 5
3.1X10-6 4.2X10"6 3
2
2.
16X10-5
. IxlO-6
. 4X10'6
.6X1Q-6
.7xlO-6
16X10-6
337
-------
use sand as the soil and allow AP (including the vacuum value at the
collector) to be either 10 psi or 20 psi, rd to vary between 3 meters and
5 meters, and r0 to run from 0,2 meters to 0.5 meters. It is seen that
over these ranges Of field-like parameters, the velocities vary from
0.78 X 10~6 m/s to 25 X 10~6 m/s. The value used in the laboratory
experiments (from which k was derived) was 83 X 10~6 m/s. It will be
difficult for logistical reasons, to lower this laboratory value by more
than about a factor of ten. Hence we may not be able to determine the
chemical removal rates in the laboratory at the needed low steam flow
values, more characteristic of the field situation. Nonetheless, we have
proceeded to perform experiments in organic chemical removal at various
flow rates to determine if there is a strong flow rate dependence.
Figure 10 shows the results of rate of removal of kerosene from sand at
various flow rates. Figure 11 shows the rate of removal of kerosene from
the 50/50 soil at various flow rates.
Before calculating the decontamination time for a given contaminated
soil using the model, it is imperative to decide upon the important
decontamination parameter. Several obvious choices present themselves.
Is it:
(a) the absolute volume of steam passed through the soil, i.e., the
velocity X time?
(b) the number of pore volumes of steam passed through the soil, or
(c) some other unknown parameter?
We must assume a critical parameter before proceeding with the
calculation. The data of Figure 11 and 12 favor (b) above. We have the
basic experimental data from earlier papers (3,10) with which *to
determine the decontamination time for a given kerosene spill in the
sand/silt soils that we have investigated.
The decontamination time using the critical parameter (a) above is
more straightforward than using (b). Assume we have a series of curves
such as those presented in Figures 11 and 12. The time for a given
percent removal, at a given velocity, can be read directly off the
appropriate figure. The velocity as a function of position (i.e., "r")
is given by Equation (5). Therefore we can determine the decontamination
as a function of time and radial position by Equation (5) and the
appropriate decontamination curve. The velocities are higher for smaller
r-values (goes as 1/r),
The calculation of decontamination time using critical parameter (b)
above is much more difficult than using concept (a). In the first place
the pore volume is not a unique function of r, because it depends on how
thick a layer, dr, you take for the calculation (see Figure 12). The
only possibility is to take the entire pore space of the spill volume and
hope that there are compensating factors between the high steam velocity
338
-------
and short steam dwell time at small "r" and the low velocity, and long
dwell time at larger "r". Of course, this can only be answered via
experimentation.
It is instructive, even though we don't have a complete set of
curves like Figures 10 and 11, to forge ahead and make some estimate of
decontamination times. The necessary data dPS ShOWfl in Figures 13 and 14,
which give the decontamination rates of kerosene and the flow rate of
steam in the various soils. As mentioned previously, the steam flow
velocities are 83 x 10~6 m/s in sand and 13 X 10~6 m/s in the 50/50 soil.
(These values are based on the volume of condensed steam) . By using
critical parameter (a) above and assuming the decontamination rate is
independent of velocity (this has certainly not been shown experimentally
as yet), the values of decontamination times can be read directly from
Figure 13 and are shown in Table 3. It must be strongly emphasized that
the decontamination times calculated here are certainly on the low side
due to the reasons stated in the footnotes to Table 3.
TABLE 3. TIMES FOR DECONTAMINATION USING METHODS (a)
(i.e., decontamination only depends on time that.the steam is passing)
Soil % Removal Decontamination Time
(minutes)
Sand
50 Sand/
50 Silt
60
80
100
60
80
100
10
40
100 (extrapolated value)
300
2800 (extrapolated -value)
5300 (extrapolated value)
The extrapolated values are certainly lower bounds on the decontamination
times due to:
* extrapolation of lab data
• the removal rate in the field will probably be smaller the smaller
the steam velocity
For critical parameter (b), i.e., the number of pore spaces passed,
we must first assume a model of the decontamination area. The "model
spill" described earlier (3) will again be utilized. Figure 15 shows the
"model spill". It is a kerosene spill in sand (or the 50/50 soil) of
pore volume = 40% and covers a circular area of 5 meter radius. The
saturation of the kerosene in the originally dry soil is uniform and has
a value of 25% (the same as in the lab experiments). The amount of pore
volume per meter of length in the spill is
339
-------
5t(5 ) (1) (0.40) = 31.4 m (8)
The volume flow rate per unit length in the circular model described
earlier is, from Equation (7):
fi. , -2Kk
The values of kAP can be obtained from the lab results for the two soils
{3,105. The results are:
sand: kAP - 1.76 X 10~ — (for AP = 10 psi)
s
50/50: kAP *= 0.27 x10~5 — (for AP = 15 psi)
s
Putting these values in Equation (7) with rd = 5 m and r0 = 0.5 m we
obtain:
2n {1.76 xio" )
ly In 10
sand
2n (1.76 X10~ )
(f)
2.3
~5 in *
4.8 XIO — (this is per meter of length)
s
2n (0.27 XIO
50/50
3
~5 m
- 0.75 XIO — (this is per meter of length)
s
For 100% removal in sand it takes (extrapolated value) about 1000 minutes
(see Figure 14). From Figure 14 it is seen that in 1000 minutes about
20,000 cm3 of steam (condensed) has flowed through the sample. The lab
sample has a volume of 456 cm3 and a pore volume of 178 cm3. Therefore
20,000/178 «= 112 pore volumes of steam (condensed) must flow through the
sample for 100% decontamination. The volume of 112 pore spaces in the
field spill is
112 (31.4) m3 - 3520 m3
Thus the decontamination time for .100% removal is
340
-------
volume of steam to be passed
decent. rate of steam volume passage
decont
.y
3520 m
sand
4.2 xl(f5 %-
s
= 7.3 Xl07s
=28 months
Results using the above approach are given in Table 4. Included in
the table are results for sand and the 50/50 soil at steam pressures of
both 5 psi and 10 psi and various r^ and r0 values.
TABLE 4(a). DECONTAMINATION TIMES IN MONTHS, FOR THE TWO SOILS AT THE
VARIOUS PARAMETERS
Sand 50/50 Soil
-------
TABLE 4(b). DECONTAMINATION TIMES IN MONTHS, FOR THE TWO SOILS AT THE
VARIOUS PARAMETERS
Sand
(AP = 15 psi)
removal amount
60%
rd « 5 m
r0 - 0
r0 - 0
r0 - 0
rd « 3 m
r0 - 0
r0 - 0
r0 - 0
r^ - 2 m
r0 - 0
r0 - 0
r0 - 0
.5 m
.25 m
.10 m
.5 m
.25 m
.10 m
.5 m
.25 m
.10 m
0
0
0
0
0
0
0
0
0
.28
.37
.48
.22
.30
.42
.17
.25
.37
50/50 Soil
(AP - 20 psi)
removal amount
80% 100% 60% 80% 100%
(extrapolated) (extrapolated)
0.7
0.9
1.2
0.56
0.75
1.1
0.42
0.63
0.92
14
18.2
23.8
10.9
15
20.7
8.4
12.6
18.3
14.2 87.5
18.6 114
24.3 148
11.1 68
15.3 94
21.1 130
8.7 53
12.9 79
18.6 114
190
248
323
148
204
281
115
172
248
SUMMARY AND CONCLUSIONS
To be presented at Conference when data acquisition is more
complete.
ACKNOWLEDGEMENTS
The Drexel authors would like to thank the Risk Reduction
Engineering Laboratory of the U.S. Environmental Protection Agency,
Edision, New Jersey for support through Cooperative Agreement Number CR
813022.
342
-------
• injeeiiop fipes
* Valves
Pip« Manifold Syslwm-
Vocoum CoNaclion
Unar
Injoclion Pipss
ta.im Dislillation
Gioundwafor Tablo
ELEVATION VIEW
Figure 1 - Schematic diagram of geosynthetic steam collection cap.
coo/ant
sfeam
two phase
-condensatf
• flask
Figure 2 - Schematic diagram of the experimental setup for vacuum-
assisted steam stripping studies in the laboratory.
343
-------
FUnNACE
VJ
-. VACUUM (air out)
/- "S- AIR (air in)
S//S//77.
SOIL
AND
KEROSENE
/v/xxx7?
1
\
:OPEN
X X x X x s/sssss/ss
Figure 3 - Schematic diagram of the experimental setup for vacuum
extraction studies in the laboratory.
Results of Vaccum Strip on 45-45-10 (Kaollnlle) 11-29-90 (Run 1)
too
1
cr
o
g
90-
00-
70
60-
50
40-
30-
20-
10
z
45-45-10 ton
50-50 soil
100 200 300 400 SOO 600 700 800
lime (mln)
Figure 4 - Results for vacuum stripping of kerosene from a 50 sand/50
silt soil and from a 45 sand/45 silt/10 kaolinite soil (soil
temperature = 100°C)
344
-------
Vacant! Slrip of Sand /Silt / Clay Fraction (1st. series)
IIX)
•Ml
100 2tm 300
500 600 700
Tinic (n«n)
Figure 5 - Results for vacuum stripping of kerosene from various
sand/silt/clay or sand/silt/organio mixtures (soil temperature
= 100°C)
HYPOTHETICAL
2
o
S
et
DELAY TIME-
[SAND
TIME »•
Figure € - Hypothetical delay_time data in sand.
(THIS IS NOT ACTUAL DATA!)
345
-------
5
I
HYPOTHETICAL
DELAY TIME-
TIME-
Figure 7 - Hypothetical delay time data in the 50/50 soil,
(THIS IS NOT ACTUAL DATA!)
STEAM FLOW
PIPES
Figure 8 - The circular-symmetry model for the analytical theory
calculations for soil decontamination times using vacuum-
assisted steam stripping.
346
-------
Figure 9 - Steam flow through a lateral surface area.
o
Q:
..«
HYPOTHETICAL
TIME
Figure 10 - Hypo t he tic a. X data for steam stripping removal efficiency in
sand'for various steam flow velocities.
{THIS IS NOT ACTUAL DATA!)
347
-------
I
I
Q:
HYPOTHETICAL
TIME-
Figure 11 - Hypothetical data for steam stripping removal efficiency in
the 50/50 soil for various velocities.
(THIS IS NOT ACTUAL DATA!)
Figure 12 - Diagram showing different volumes present in cylindrical
shells. (V - 2n rl dr)
348
-------
100
o
3
o
fr.
loo ^ (3) 2600 m/n.
100 X (S> 5300
- O/ IOO (Posiitlt Strom ffaetuft I
0 200 400 600 SOO 1000 1200 I4OO 1600 1800 2000 2200
TIME ( MINUTES )
Figure 13 - Kerosene removal efficiency in the various sand/silt
mixtures. (These data apply for only one steam flow velocity
for each soil).
o
>
4OOO
3000
2000
IOOO
100/0 / 75/25
I No O/IOO Corn — Fan,kit Sltam Fraelurt)
IOO . . 200 - 300 40O 500 600
TIME (MINUTES »
Figure 14 - Steam volume flow rate for the various sand/silt mixtures,
349
-------
confamfnaftd area
Figure 15 - Schematic diagram of assumed kerosene spill volume.
350
-------
REFERENCES
1. Kovalic, J. M. and Klucsik, J. F., "Loathing for Landfills Sets Stage
for Innovative Hazardous Waste Treatment Technology," Hazard. Mat.
and Waste Manag. 5,, 1987, pp. 17-18.
2. Cheremisinoff, P. N., "Update: Hazardous Waste Treatment," Pollut.
Eng. 11, Feb. 1987, pp. 42-49.
3. Lord, A. E., Jr., Koerner, R. M. and Murphy, V. P., "Laboratory
Studies of Vacuum-Assisted Steam Stripping of Organic Contaminants
from Soil," Proc. 14th Annual Conference on Land Disposal,Remedial
Action and Treatment ofHazardous Waste, Cincinnati, Ohio, April
1988, sponsored by the Risk Reduction Engineering Laboratory, U.S.
Environmental Protection Agency, Cincinnati, Ohio, pp. 65-92.
4. Handbook onIn Situ Treatment ofHazardousWaste-Contaminated Soils,
U.S. Environmental Protection Agency, Risk Reduction Engineering
Laboratory, Cincinnati, Ohio, 45268, Report Number EPA/540/2-90/022,
January 1990.
5. Hilberts, B. (1985), "In-Situ Steam Stripping," Assink, J". W. and Van
Den Brink, W. J., Eds., Cpotaminated Soil, Proc. of 1st Intern. TNO
Conf. on Contaminated Soil, Utrecht, The'Netherlands, Nov. 11-15,
1985, pp. 680-687.
6. Baker, R., Steinke, J., Manchak, F., Jr. and Ghassemi, M., Wln-Situ
Treatment for Site Remediation," Proc,Third Annual Conference on
Hazardous Waste Law an
-------
in Civil Engineering, Drexel University, Philadelphia, PA, June 1988.
11. Lord, A. E., Jr., Koerner, R. M., Hullings, D. E. and Brugger, J. E.,
"Laboratory Studies of Vacuum-Assisted Steam Stripping of Organic
Contaminants from Soil," Proc. 15th Annual Conf. on Land Disposal,
RemedialActionand Treatment of Hazardous Waste, Cincinnati, Ohio,
April 1989, sponsored by the Risk Reduction Engineering Laboratory,
U.S. Environmental Protection Agency, Cincinnati, Ohio.
12. Lord, A. E., Jr., Hullings, D. E., Koerner, R. M. and Brugger, J. E.,
"Vacuum Assisted Steam Stripping to Remove Pollutants from
Contaminated Soil, A Laboratory Study," Proc.16thAnnual Conf. on
Lantj Disposal. Remedial Action and Treatment of Hazardous Waste,.
Cincinnati, Ohio, April 1990, sponsored by the Risk Reduction
Engineerning Laboratory, U.S. Environmental Protection Agency,
Cincinnati, Ohio.
13. Hullings, D. E., Masters Thesis work at Drexel University, 1990.
352
-------
EFFECTIVENESS OF COMMERCIAL MICROBIAL PRODUCTS IN ENHANCING OIL DEGRADATION IN
PRINCE WILLIAM SOUND FIELD PLOTS
lAlbert D. Venosa, !john R. Haines, and 2David M. Allen
ly.S. Environmental Protection Agency
Risk Reduction Engineering Laboratory
Cincinnati, OH 45268
^University of Kentucky
Department of Statistics
Lexington, KY 40506
ABSTRACT
In the spring of 1990, previously reported laboratory experiments were
conducted on 10 commercial microbial products to test for enhanced biodegrada-
tion of weathered crude oil from the Exxon VaTdez oil spill. The laboratory
tests measured the rate and extent of oil degradation in closed flasks.
Weathered oil from the beaches in Alaska and seawater from Prince William
Sound were used in the tests. Two of the 10 products were found to provide
significantly greater alkane degradation than flasks supplemented with mineral
nutrients alone. These two products were selected for further testing on a
beach in Prince William Sound.
A randomized complete block experiment was designed to compare the effec-
tiveness of these two products in enhancing oil degradation compared to simple
fertilizer alone. Four small plots consisting of a no nutrient control, a
mineral nutrient plot, and two plots receiving mineral nutrients plus the two
products, were laid out on a contaminated beach. These four plots comprised a
"block" of treatments, and this block was replicated four times on the same
beach. The plots were positioned in random order within each block. Tripli-
cate samples of beach sediment were collected at four equally spaced time
intervals and analyzed for oil residue weight and alkane hydrocarbon profile
changes with time. The objective was to determine if either of the two com-
mercial microbiological products was able to enhance bioremediation of an oil-
contaminated beach in Prince William Sound to an extent greater than that
achievable by simple fertilizer application. Results indicated no significant
differences among the four treatments in the 27-day time period of the experi-
ment.
This paper has been reviewed in accordance with the U.S. Environmental
Protection Agency's peer and administrative review policies and approved for
presentation and publication.
353
-------
INTRODUCTION
BACKGROUND
The largest field bioremediation test ever attempted was conducted by
the U.S. Environmental Protection Agency and the Exxon Corporation on the
shorelines of Prince William Sound, Alaska, following the oil spill from the
supertanker Exxon Valdez in March, 1989 (1). In that study, investigators
concluded that application of nitrogen and phosphorus nutrients enhanced bio-
degradation of the crude oil. Furthermore, no adverse environmental effects
were observed as a result of the fertilizer application.
Other studies have met with mixed success. In February 1970, the tanker
Arrow spilled approximately 108,000 barrels of bunker C oil into Chedabucto
Bay, Nova Scotia. Rashid (2) reported 3.5 years later substantial losses of
ff-alkanes resulting largely from microbial degradation, aided by the wave-
driven mixing permitting infusion of dissolved oxygen and exogenous nutrients.
Cretney et a?. (3) reported that biodegradation accounted for almost complete
removal of n-alkanes during the first year after the tanker Irish Stardust ran
aground near Vancouver Island, B.C. In contrast, Colwell et a?. (4) observed
prolonged persistence of oil spilled from the tanker MetuTa in the Straits of
Magellan in 1978. They attributed the slow biodegradation rates not to the
cold temperatures typical of the area but to the limited concentration of
nitrogen and phosphorus available in the seawater, as well as restricted
accessibility to degradable compounds within aggregrated oils or tar balls.
Atlas et a7. (5), after experimentally contaminating a coastal Arctic ecosys-
tem with Prudhoe Bay crude oil, measured low rates of natural hydrocarbon
biodegradation. They found that temperature and availability of nutrients
limited the biodegradation rate. Several research teams investigated the fate
of the Amoco Cadiz crude oil spilled off the Brittany Coast in 1978 (6-11).
Microbial degradation was found to have played a crucial role in the weather-
ing of the oil contaminating the shoreline in the intertidal zone. There were
rapid changes in the n-alkane/isoprenoid hydrocarbon ratios within weeks
following the spill. The isoprenoid alkanes, /j-alkanes with carbon number C£7
to Csi, hopanes, alkylated dibenzothiophenes, and alkylated phenanthrenes were
the classes of compounds most resistant to biodegradation.
Another way to enhance bioremediation in the field is inoculation with
allochthonous microorganisms. Cultures and cultural products have been added
to different environments to stimulate biological removal of contaminants.
Some of the studies have demonstrated enhancement, while others have not (12).
In a recent study, Dott et aL (13) compared fuel oil degradation rates of
activated sludge microorganisms with nine different commercial bacterial cul-
tures in separate laboratory flasks. They found that the rate and extent of
o-alkane and total hydrocarbon degradation by the diverse populations in
activated sludge were significantly higher than any of the highly adapted
commercially available cultures, Lehtomaki and Niemela (14) found that addi-
tion of brewers' yeast to oil-contaminated soil enhanced oil removal 2- to
10-fold, This was most likely due to the supply of critical nutrients,
vitamins, or cofactors that were naturally deficient in the soil. Christian-
son and Spraker (15) reported a series of case histories of refinery wastewa-
354
-------
ter treatment plants using commercial cultures to overcome various specific
problems, such as foaming, toxic loads, low biomass, etc. Most success with
biodegradation enhancement by allochthonous microbial cultures has been
achieved when chemostats or fermentors were used to control conditions or
reduce competition from indigenous microflora (16) .
Venosa et a?, (17) recently conducted laboratory tests on 10 commercial
products for microbial degradation of weathered crude oil from the Exxon Val-
dez spill. The products were selected from a public solicitation by EPA and
review of proposals by a panel of experts convened by the National
Environmental Technology Applications Corporation (NETAC), a non-profit orga-
nization dedicated to the commercialization of environmental technologies.
Laboratory tests on the products were conducted to measure the rate and extent
of oil degradation in closed ecosystems. Weathered oil from the beaches in
Alaska and seawater from Prince William Sound were used in the tests. The
NETAC panel reviewed the results of the tests and agreed with the recommenda-
tion for further testing of two products that exceeded the performance of
inorganic nutrient addition. This paper presents the results of the field
testing of the two selected products. The objective was to determine if
commercial microbiological products were able to enhance bioremediation of an
oil-contaminated beach in Prince William Sound to an extent greater than that
achievable by simple fertilizer application. The two companies that partici-
pated in the testing were Sybron, Inc. and ERI-Waste Microbes, Inc.
MATERIALS AND METHODS
PLOT DESCRIPTION
ure I.
A schematic representation of the experimental layout is depicted in Fig-
1 A: No Nutrients H B: Mineral Nutrient]
B8 C: S)-lron * NutricnU F/3 D: ERI * Nutrients
BLOCK I BLOCK 2
BLOCK 3 BLOCK 4
B C A D
C B D A
D C B
I
C D A B
Figure 1.
Randomized Complete Block Design Showing Location of
Replicate Plots at Disk Island.
355
-------
The experiment was a randomized complete block design. Four beach segments
("blocks"), each 20 m wide (labeled 1 through 4} were staked out in the inter-
tidal zone at Disk Island (designated DI-67A), an island in Prince William
Sound located between Eleanor and Knight Islands. Within each block were 4
treatment plots, labeled A through D, 2 m wide by 5 m long (top-to-bottom).
The plots were separated from each other by a buffer zone measuring 3 m in
width. Plot A, color-coded blue on the shoreline, was the no-treatment con-
trol. Plot B, color-coded orange, was the nutrient-only treatment. Plot C,
color-coded green, was the plot receiving nutrients plus Sybron's product.
Plot D, color-coded red, was the plot receiving nutrients plus ERI's product.
The treatment plots within each block were randomly distributed according to
the following scheme: block 1, BCAD; block 2, CBDA; block 3, ADCB; and
block 4, CDAB.
Each of the 16 plots was subdivided horizontally into three equal seg-
ments 2 m wide by 1.67 m long, as shown schematically in Figure 2.
2m
8 m
Figure 2. Typical Plot Showing Location of Sampling Bags.
In each of the three segments, four bags, each made of fiberglass screening
material and containing approximately 750 to 1000 g of uniformly sized oily
gravel, were buried approximately 5 to 10 cm below the surface and covered
with mixed sand and gravel. The four bags corresponded to the four sampling
events that were planned for the experiment. A surveyor's ribbon was attached
to each bag for easy identification. The 12 samples within each block were
numbered 1 through 4 in the top third, 5 through 8 in the middle third, and 9
through 12 in the bottom third.
The bags had previously been filled with gravel that had first been
sieved through a 25 mm coarse screen to remove large stones and then a 4.75 mm
sieve to remove the small sand granules that compact the beach material. The
gravel was mixed manually by shovels and hoes in a large wooden container to
356
-------
achieve reasonable homogeneity with respect to oil contamination and rock
size. These bags served as samples to be taken on the appropriate sampling
days.
SAMPLING
On a given sampling day, triplicate samples from each plot within a block
were collected according to a random schedule. One sample was randomly taken
from each of the three identical sectors of each plot. Some of the gravel was
poured into 500 ml I-Chem jars, labeled, and placed in a cooler to be carried
back to Valdez for freezing and shipment via Federal Express to the analytical
chemistry laboratory located in Pittsburgh, PA. The rest of the gravel was
archived in aluminum foil and frozen. Thus, 48 samples were collected on each
of the four sampling days, giving a total of 192 samples for the entire
experiment.
SEDIMENT CHEMISTRY
The 48 samples were analyzed for oil residue weight by methylene chloride
extraction followed by evaporation to dryness and weighing on an analytical
balance. After weighing, each sample was reconstituted with methylene chlo-
ride, passed through a silica gel fractionation column, and analyzed for the
normal alkanes C12 through C34 plus the isoprenoid hydrocarbons pristane and
phytane by gas chromatography using a flame ionization detector. The ali-
phatic fraction was eluted from the silica gel column with hexane prior to GC
injection.
MICROBIOLOGY
Subsamples from the 8 plots of blocks 2 and 3 were analyzed for oil
degrading bacteria by standard plate count, using Bushnell-Haas medium
supplemented with Prudhoe Bay crude oil as the carbon source. Only one of the
three triplicates from those 8 plots was analyzed for microbial numbers. The
plates were incubated at 15°C for 21 days and the colonies counted.
NUTRIENTS
Within 8 of the 16 plots on the shoreline, a well was installed for
collecting nutrient samples. Two wells extending approximately 60 cm below
the surface were driven into each of the four blocks, one in the no-nutrient
control plot and one in the nutrient-treated plot. Subsurface water from
these 8 wells served as samples for nutrient analysis.
APPLICATION OF NUTRIENTS
The source of nitrogen was ammonium nitrate. Each 2 m x 5 m plot
received 200 g of N (20 g/m2). At 35% N, the amount of NH4N03 containing 200
g of N was 570 g or 1.25 Ib per plot. This amount, less approximately 40 g to
account for the N in the product containing the phosphate salt (see next
paragraph), was added to 6 gallons of seawater and the contents stirred until .
dissolved. A 2-gallon plastic sprinkling can was filled with the solution and
357
-------
the entire contents poured onto the top third of a plot earmarked for nutri-
ents, The sprinkling can was again filled and the contents poured onto the
middle third. The procedure was repeated for the bottom third.
The source of phosphorus was an Ortho product named "Upstart," which had
an N-P-K analysis of 3-10-3, At 101 PgOg, the amount of Upstart used was 450
g (1 Ib) per plot. This corresponded to a phosphorus loading of 20 g P per
plot (2 g P/m2). The 450 g of Upstart was added to the 6 gallons of seawater
above (after the NH4N03 had been dissolved) before applying to each plot.
Note that this product contained 3% N in the form of NHdNOs- The amount of N
in Upstart had already been accounted for in the above 530 g computation of
needs.
SCHEDULE
The entire experiment lasted only 27 days because severe Alaskan winter
weather precludes field activities beyond the month of August. Day 0 occurred
on Sunday, July 29, 1990. Nutrients and commercial products were applied on
days 0, 4, 8, 12, 16, 20, and 24. One extra application day, day 2, was used
for an additional commercial product application, as specified by the two
vendors. After nutrients and products had been delivered to the appropriate
plots, randomly assigned triplicate sampling bags were removed from the plots
for time 0 sediment chemistry and microbiology analysis. The other triplicate
sampling bags were collected on days 9, 18, and 27. Nutrient sampling took
place on days 1, 2, 3, 4, 17, 18, 19, and 20. This allowed determination of
nutrient concentrations throughout the four-day interval between applications
at two different times in the experiment.
RESULTS
PERSISTENCE OF NUTRIENTS
Figures 3-5 summarize the average changes in nutrient levels with time in
each block on Disk Island. Figure 3 shows the ammonia-N data, Figure 4 the
nitrate-N data, and Figure 5 the phosphate-P data.
No Xulrltnt
Control
Hlntrt!
NutritDti
0,80
n ftfl
•
1 I
13 - -
ft,
»
1134 1234 1 Z 3 4 1*34
rmt. BAYS
Figure 3. Changes in NH3-N in each block within 4 days after application.
358
-------
Ho Nutrient
Control
Hiaenl
Nulricnll
l.«J
0.80
O.«0
0.40
MO
Lr,
i I J
J
LnL
1234 1134 1234 1234
TIKE, Dm
Figure 4. Changes in N03-N in each block within 4 days after application,
He NulrltBl
Coitlol
Itaeril
Hutrkatt
0.20
lJ
1234
234 1234 1234
TUB, PAYS
Figure 5. Changes in PQ4-P in each block within 4 days after application.
Persistence of ammonia-N was the most erratic. In block 1 the levels of
in the nutrient-treated plot were measured at 4.0 and 1.1 mg/L one and
two days after application, respectively, and in block 2 the NHs-N was 1.7
mg/L one day after application. Little NH3-N was measured in any of the
control plots at any time except in block 4, where 0.1 mg/L was measured after
one day and almost 1.0 mg/L after four days. The source of the high NH3-N
spike in the control plot of the fourth block may have been caused by carry-
over of nutrients from the nutrient-treated plot onto the control plot. The
nutrient-treated plot had to be placed above the control plot (see Figure 1}
because of the presence of compacted peat on the extreme right end of the
359
-------
beach. There was a surface flow of water from a saltwater lagoon located
approximately 50 m above the test area that flowed across the nutrient-treated
plot onto the control plot. This stream was not noticed when the plots were
first laid out. Although this explains the higher levels of NHs-N measured
one day after application, it does not explain why such a high spike was
observed on the fourth day.
The nitrate and phosphate data indicate significant but decreasing levels
of nutrients in the nutrient-treated plots as time progressed to four days
after application (Figures 4 and 5). Again, high levels of N03-N and measur-
able levels of P04-P appeared in the control plot of the fourth block four
days after application.
CHANGES IN OIL DE6RADERS
Oil degrader counts in all plots of blocks 2 and 3 are shown in Figure 6.
Although the levels of oil degraders were high in each of the plots, there
were no significant changes or differences in any of the plots after 27 days
of field testing.
us
bl
CO
ta
(A
Ob
o
10'
10'
10'
10'
io4
10'
io1
io1
10'
0
No Nutrient
Control
Mineral
Nutrients
Sybron +
Nutrients
ERI +
Nutrients
10 15 20
TIME, DAYS
25 30
Figure 6. Changes in Oil Degrader Counts on All Plots of Blocks 2 and 3.
OIL RESIDUE WEIGHT
Changes in oil residue weight, averaged over all four blocks, are summa-
rized in Figure 7 as a function of time. The hatched bars are the mean
residue weights for each of the four treatments, and the error bars depict one
360
-------
standard deviation unit above and below the means. These error bars represent
the variation in oil residue weight among the four blocks and are indicative
of the overall experimental error.
EES1 No
Nutrients
Mln.rol t
Nutrientj
Sybron*
Nutrients
ERI 4
Nutrients
jj 8000
9 it
TIME, DAYS
Figure 7. Average Oil Residue Weights for All Treatments.
Visual inspection of the data from the plots treated with mineral nutrients
alone and mineral nutrients supplemented with Sybron's product indicates a
decrease in oil residue of approximately 33% at the end of the experimental
period compared with no net change in the no-nutrient control plot and a
slight increase in the ERI plot. When the data were subjected to analysis of
variance, however, there were no statistically significant differences among
any of the four treatments at the 5% significance level. This was true even
after the data were log transformed to stabilize the variance.
Note the broad error bars on Figure 7 at the day 0 sampling time compared
with the other three sampling times. Despite the effort to control the heter-
ogeneity of rock size and contamination by the sieving and mixing techniques
prior to start-up, there was still substantial variation in oil residue weight
from plot to plot and block to block at day 0. To ascertain the source of
this variation, a breakout of plot oil residue weights by block was conducted.
Results are shown in Figure 8.
361
-------
Ko
Nutrient*
MiD«r«I
NutrienU
Syfcron +
Nutrients
EEI +
Nutrient*
E-
ss
o
W
a;
llrtJUU
SOOO
6000
4000
2000
10000
BOOO
6000
4000
SOOO
n
-
-
- ;
BLOCK
I
$
1
1
X
y,
**'
V
E
t
i
>
•v
T
i
.*;
i
^
f
1
g
!
i
i.
\
*$
u
>
>
\
•> '
\
s :
1
i
j, :>; •
.[: ¥ '
;| -^
BLOCK
1
W
m.
]
i
11^
i
1
II
>x
/y
it
i
X,
.AAAAAXXX;
3
rl
s
• >>
:>
« ><
i
m
iy
•
x;
s
BLOCK 2
rii
<•
¥
1
aftifi
; ill
* !!«
X i|^
> 1^
BLOCK 4
nil
i
|
1
t
*sL.
ll
IXXXAX/MH
1
S
|i
IK
ill
j
i
18
TIME. DAYS
Figure 8. Oil Residue Weights for All Treatments Broken Out by Block.
Examination of these data reveals the differences in the distribution of oil
from plot to plot. The error bars shown on this figure are the standard
deviations of the triplicate samples within each plot and are indicative of
the sampling error. At day 0 the agreement of the triplicate samples averaged
within each plot (Figure 8) is better than the agreement of identical plots
averaged over blocks (Figure 7). This suggests that the cause of the varia-
tion among plots was consistent within each of the plots.
362
-------
TOTAL RESOLVABLE ALKANES
All samples were subjected to GC analysis to determine the changes in the
aliphatic profiles of the oil among the various treatments. The concentra-
tions of all the normal alkanes and the isoprenoid alkanes pristane and phy-
tane resolvable by GC/FID were summed together for each treatment, averaged
over all four blocks, and plotted as a function of time. The data with
associated error bars are shown on Figure 9.
No
Nutrients
Mineral I
Nutrients
Sybron + I
Nutrients
ERi +
Nutrients
StO
a
X
E-*
O
S-
9 18
TIME, DAYS
27
Figure 9. Average Total Resolvable Alkanes for All Treatments.
Except for the day 0 data, the error bars in Figure 9 are generally higher
than the corresponding residue weight error bars (Figure 7). Although a down-
ward trend in resolvable alkanes is perceptible after 27 days among all treat-
ments, the analysis of variance revealed no significant differences among the
treatments (p<0.05). This agrees with the findings of no significance among
treatments in the oil residue weight data.
Figure 10 was constructed to examine the behavior of the GC data among
the individual plots within each block. The error bars represent the sampling
error associated with the triplicate samples in each plot.
363
-------
tllBfel
ou
25
20
W)
^5 15
a 10
sz
^.
Si
^
-3 25
-J5
E-i
0
E- 20
^ 15
10
5
0
•
•
~ '
•
i ';
•
;
•
1
• '
' J
No
Nutrients
BLOCK
5r
£
'f
«
i F
' W-
ra^
! :^
k
>
BLOCK
T
1
1
Is
;rfl
>]
J
E
s
i;
M3 MiDeral S/bron + LXlJ EP.I 4
Nutrient* Nutrients Nutrient*
1
T
I
i Jm H
!
1 5
T BLOCK
':j i
> :
< 1
-------
N-NONADECANE
To determine if an individual alkane hydrocarbon exhibited biotransforma-
tion differences among treatments, the data from a representative normal
alkane, n-nonadecane (n-C19), were analyzed in several different ways. First,
if n-C19 were biotransformed at the same rate as the total resolvable alkane
hydrocarbons, one would expect to see no net change with time when the
n-C19/total alkane ratio was plotted. Figure 11 summarizes such a plot.
E23 NO
Nutrients
Miners! E
Nutrients
Sybron +i
Nutrients
m +
Nutrient*
0,10
0.00
» 18
TIME, DAYS
Figure 11. Average n-C19/Total Alkane Ratio for All Treatments.
No significant changes are evident among the treatments within 27 days.
Second, if n-C19 were biotransformed in the same way as the total resolv-
able alkanes, one would expect to observe the same behavior in the
n-C19/residue weight ratio as the total alkane/residue weight ratio. Figures
12 and 13 show that this is the case.
365
-------
He HulHcal
Coslrcl
Vlncnl
HitrUnti
SjbroB +
N«lri«Bti
E2Z3 in
9 18
TIME, DAYS
Figure 12. Average n-C19/Res1due Weight Ratio for All Treatments,
WM No
Nutrients
Mineral
Nutrients
Sybron +
Nutrients
ERI +
Nutrients
9 18
TIME, DAYS
Figure 13. Average Total Alkane/Residue Weight Ratio for All Treatments,
366
-------
Thus, whether one examines the total alkane hydrocarbon fraction of the oil or
a representative normal alkane within it, the results are the same: no
enhancement either by nutrients alone or by nutrients supplemented with com-
mercial inocula.
One important observation from Figure 13 is the magnitude of the total
alkane/residue weight ratio. The sum total alkane hydrocarbons resolvable by
6C/FID are less than 0.5% of the total oil residue weight 1.5 years after the
Exxon Valdez spill. This is the most likely explanation for the lack of
enhancement of bioremediation by either nutrient addition alone or nutrient
addition supplemented with commercial microbial cultures. Over 99.5% of the
oil remaining on Disk Island 1.5 years after the spill is not resolvable by
conventional gas chromatography. The compounds comprising this recalcitrant
fraction are likely the tars and asphaltines that will only slowly degrade
with time. The 27-day period of this investigation was much too short to
determine if enhancement of bioremediation is possible.
DISCUSSION
The conclusions reached in this field study were based on three sources
of information: nutrient persistence, microbiology, and sediment chemistry.
The nutrient data clearly demonstrated that nitrogen and phosphorus persisted
at measurably higher levels in the treated plots compared with the control
plots throughout the four days between applications. These measurements were
taken approximately 60 cm below the surface of the beach, suggesting that
nutrients were in constant contact with the subsurface sediment layers for
relatively long periods of time.
The microbiology data clearly demonstrated no net increase in oil
degrader populations in any of the plots after 27 days and no differences
among the four treatments at any time during the 27 day period. The oil
degrader populations were high to begin with and were maintained with or with-
out the presence of excess nutrients. Either the oil degraders were dormant
or, more likely, they were sufficiently able to sustain their activity with
the oligotrophic levels of nutrients present in the ambient environment.
Sediment chemistry revealed the most definitive information, because it
was the basis of the statistical analyses conducted. No significant differ-
ences were found among the four treatments at the 5% significance level either
from the standpoint of oil residue weight, total resolvable alkane
hydrocarbons, n-C19/total alkane ratio, n-c!9/residue weight ratio, or total
alkane/residue weight ratio. The experimental error from the residue weight
data was higher than the sampling error, which clearly points out the neces-
sity to replicate treatments when conducting field experiments.
An instructive piece of information obtained from this investigation was
the fact that most of the readily biodegradable compounds in the aliphatic
fraction of the contaminating oil has disappeared in the 1.5 years since the
spill took place off Bligh Reef in Prince William Sound. This is the most
likely explanation for the lack of any significant enhancement observed in the
short time period allotted for this study. Further evidence supporting this
conclusion derives from examining the n-alkane/isoprenoid alkane ratios.
367
-------
These ratios have been used in past literature to indicate extent of biodegra-
dation; the lower the ratio, the more extensive the biodegradation. The
average n-C17/pristane and n-C18/phytane ratios on day 0 for all the plots on
Disk Island were 0.18 and 0,27, respectively. This compares with approxi-
mately 1.5 to 1.8 for unweathered Prudhoe Bay crude oil. Thus, the remaining
oil present on Disk Island will likely degrade very slowly from now on because
of the recalcitrant nature of the substrate. If either nutrient application
or commercial inoculation can accelerate this rate, the time period must
extend significantly beyond the 27 days allotted for this study or the trial
must be conducted on beaches with fresher oil contamination.
REFERENCES
1. Pritchard, P. H., R. Araujo, J. R. Clark, L. D. Claxton, R. B. Coffin, C.
F. Costa, J. A. Glaser, J. R. Haines, D. T. Heggem, F. V. Kremer, S. C.
McCutcheon, 0. E. Rogers, and A. D. Venosa. 1990. Interim report. Oil
spill bioremediation project. U.S. Environmental Protection Agency,
Washington, D. C.
2. Rashid, H. A. 1974. Degradation of bunker C oil under different coastal
environments of Chedabueto Bay, Nova Scotia. Estuarine Coastal Mar. Sci.
2:137-144.
3. Cretney, W. I., C. S. Wong, D. R. Green, and C. A. Bawden. 1978.
Long-term fate of a heavy fuel oil in a spill-contaminated coastal bay.
J. Fish. Res. Board Can. 35:521-527.
4. Colwell, R. R., A. L. Mills, and J. D. Walker. 1978. Microbial ecology
studies of the Metula spill in the Straits of Magellan. J. Fish. Res.
Board Can. 35:573-580.
5. Atlas, R. M., A. Horowitz, and H. Busdosh. 1978. Prudhoe crude oil in
arctic marine ice, water, and sediment ecosystems: degradation and
interactions with microbial and benthic communities. J. Fish. Res. Board
Can. 35:585-590.
6. Atlas, R. H., P. D. Boehm, and J. A. Calder. 1981. Chemical and
biological weathering of oil from the Amoco Cadiz oil spillage in the
littoral zone. Estuarine Coastal Mar. Sci.
7. Atlas, R. M. and A. Bronner. 1980. Microbial hydrocarbon degradation
within intertidal zones impacted by the Amoco Cadiz oil spillage. In
Proceedings of the International Symposium on the Amoco Cadiz: Fates and
Effects of the Oil Spill. Centre Oceanologique de Bretagne, Brest,
France.
8. Boehm, P. D. and D. L. Fiest. 1980. Comparative weathering patterns of
hydrocarbons from the Amoco Cadiz oil spill observed at a variety of
coastal environments. ' In Proceedings of the International Symposium on
the Amoco Cadiz: Fates and Effects of the Oil Spill. Centre Oceanologi-
que de Bretagne, Brest, France.
368
-------
11.
12.
13.
14.
15.
16.
17.
Carder, 0. A. and P. D. Boehm. 1980. Year study of weathering processes
acting on the Amoco Cadiz oil spill. In Proceedings of the International
Symposium on the Amoco Cadiz: Fates and Effects of the Oil Spill.
Centre Oceanologique de Bretagne, Brest, France.
10. Vandermeulen, J. H. and R. W. Traxler. 1980. Hydrocarbon-utilizing
microbial activity in marsh, mudflat, and sandy sediments from north
of the International Symposium on the Amoco
of the Oil Spill. Centre Oceanologique de
Brittany. In Proceedings
Cadiz: Fates and Effects
Bretagne, Brest, France.
Ward, D., R. M. Atlas,
biodegradation and the
Ambio 9:277-283.
P. D. Boehm, and J. A. Calder. 1980. Microbial
chemical evolution of Amoco Cadiz oil pollutants.
Leahy, J. 6. and R. R. Colwell. 1990.
carbons in the environment. Microbiol
Microbial degradation of hydro-
Rev. 54:305-315.
Dott, W., D. Feidieker, P. Kampfer, H. Schleibinger, and S. Strechel.
1989. Comparison of autochthonous bacteria and commercially available
cultures with respect to their effectiveness in fuel oil degradation. J,
Ind. Microbiol. 4:365-374.
Lehtomaki, M. and S.
oil in soil. Ambio
Niemela. 1975.
4:126-129.
Improving microbial degradation of
Christiansen, J. A. and P. W. Spraker. 1982. Improving effluent quality
of petrochemical wastewaters with mutant bacterial cultures. Proc. Pur-
due Ind. Wastes Conf. 37:567-576.
Wong, A. D. And C. D. Goldsmith. 1988. The impact of a chemostat
discharge containing oil degrading bacteria on the biological kinetics of
a refinery activated sludge process. Water Sci. Technol. 20:131-136.
Venosa, A. D., J. R. Haines, W. Nisamaneepong, R. Govind, S. Pradhan, and
B. Siddique. 1990. Screening of commercial products for enhancement of
oil biodegradation in closed microcosms. Paper presented at the 17th
Annual Hazardous Waste Research Symposium, Risk Reduction Engineering
Laboratory, U.S. Environmental Protection Agency, Cincinnati, OH. April
9-11, 1991.
369
-------
BIODEGRADATION OF VOLATILE ORGANIC COMPOUNDS IN AEROBIC AND
ANAEROBIC BIOFILTERS
V. Utgikar, Y. Shan and R. Govind
Department of Chemical Engineering
University of Cincinnati
Cincinnati, OH 45221
ABSTRACT
Biodegradation is an attractive alternative for the treatment
of volatile organic compounds (VOCs) present in landfill
leachates. A stripper-biofilter system can be used to control the
air pollution due to these VOCs. Preliminary experimental results
on the stripping of volatiles from landfill leachate streams and
their biodegradation in aerobic and anaerobic biofilters are
presented in this paper. This paper has been reviewed in accordance
with the U.S. Environmental Protection Agency's peer and
administrative review policies and approved for presentation and
publication.
370
-------
INTRODUCTION
In recent years, emission of volatile organic compounds (VOCs)
has received increased attention from EPA, QSHA and other
government agencies due to the serious human health hazards these
compounds present as pollutants. The origins of these VOCs can be
from manufacturing process or wastewater treatment plants, where
the waste stream is stripped of the VOCs during aeration. Another
significant source of these pollutants is from landfill leachate.
Chemicals present in landfill solid waste leach into water
(precipitation run-off or groundwater) forming a pollutant landfill
leachate stream. This stream is treated in wastewater treatment
plants, emitting VOCs during aeration. Table 1 shows a summary of
leachate composition from landfill sites (1).
TABLE 1: COMPOSITION OF LANDFILL LEACHATES - VOLATILE ORGANICS
Compounds
Benzene and alkylated
benzenes
Toluene
Acetone
Higher Ketones (methyl
ethyl , methyl isobutyl
and methyl butyl)
Chloroethylenes (di~,
tri- and tetra-)
Chloroform
Methylene chloride
Chloroalkanes
Chlorobenzene
Concentration (/ug/L)
Range
1100-2600
4400-12300
14000-32000
11000-27000
1300-4300
1000-3100
8000-22000
10-3850
190-770
The conventional treatment methods for these gaseous pollutants are
adsorption on a solid, absorption in a solvent, incineration or
catalytic conversion. An alternative to these conventional
treatment methods is the biological destruction of the VOCs. This
method has the advantages of pollution destruction (as compared to
transfer to another medium) and lower operation and maintenance
costs. The proposed treatment scheme for the VOCs in the landfill
leachate stream is shown in Figure 1. The leachate stream is first
fed to a stripper where'the'VOCs are stripped from the liquid. The
liquid effluent from the stripper is sent to the wastewater
treatment plant. The gases coming out of stripper are fed to a
biofilter where the VOCs are biologically degraded. The contaminant
free gas can be discharged to atmosphere or recycled to the
stripper. The biofilter process is discussed in the next section.
371
-------
DESCRIPTION OF THE BIOFILTER
A biofilter consists of a packed column containing
biologically active mass. The support material can be of the
following four types: 1) nonbiodegradable inactive material, such
as glass or sand, which has no significant adsorption potential for
the organics; 2) biodegradable inactive material, such as peat,
with low adsorption potential for the organics, but has organic
matter; 3) nonbiodegradable active material, such as activated
carbon, which has high adsorption potential for organics; and, 4)
biodegradable active material, such as polymeric adsorbent, which
has adsorption potential for the organics and has biodegradable
organic groups.
The biologically active matter (biomass) can exist either as
a uniform biofilm on the support medium, or as a biomass particle
trapped in the void spaces between the support material. In the
case of a biofilm, the biomass is attached to the support material
with simultaneous diffusion and degradation of the organics. In the
case of a biomass particle, the organics degrade as they diffuse
through the active biomass.
There are several important differences between the biofilter
concept and conventional treatment technologies. The most important
differences are as follows:
1. Vertical stratification of the microorganisms, with
different predominant organisms existing at various levels of the
biofilter bed height. Through the process of natural selection,
microorganisms of a" certain type will dominate at a specific height
which maximizes their growth due to the existence of optimum
conditions, such as, concentration of organics, pH, temperature,
humidity, etc.
2. No breakthrough of the organic(s) due to continuous
degradation as compared to breakthrough in an activated carbon
system when its capacity is reached. Initially, the concentration
of the organic(s) in the support material will increase until a
steady state is established, when the rate of transport of the
organic(s) from the gas phase to the support material is balanced
by the rate of biodegradation of the organic (s) .
3. Higher rate of biodegradation than in activated sludge
systems due to the existence of an immobilized biofilm, which can
contain a significantly higher concentration of the microorganisms
than found in conventional activated sludge. Since the rate of
biodegradation is dependent on the concentration of the
microorganisms, a significantly higher concentration in the biofilm
will result in an increased rate of biodegradation.
4. Potential of using a variety of organisms, either under
aerobic or anaerobic conditions. Mixed cultures that have been
acclimated to specific organics can be used as easily as pure
cultures, which are capable of degrading certain organics only.
372
-------
Aerobic and anaerobic biofilters can be used sequentially to
degrade a mixture of organics containing components that are
recalcitrant under aerobic conditions.
5. Less potential for contamination of support material by
nonbiodegradable organics or high molecular weight contaminants,
which is likely in the case of completely mixed continuous systems,
such as activated sludge plants or fluidized bed reactors, handling
aqueous waste streams. For the biofilter, the organic contaminants
that are introduced through the gas phase would not have a high
molecular weight or be recalcitrant compounds that can accumulate
in the support material.
The above mentioned differences make the biofilter concept
unique when compared to conventional technologies for removing
organics from gaseous streams. In the following section, the
mechanism of biodegradation in the biofilter has been discussed
both qualitatively and modeled quantitatively.
MECHANISM OF BIODEGRADATION AND MODEL DEVELOPMENT
The microbial degradation of substrate is assumed to take
place through the following steps:
1. Diffusion of substrate through bulk gas phase to gas-
liquid interface;
2. Dissolution of substrate in the liquid phase (formed by
the nutrient solution);
3. Diffusion of substrate through the liquid film which
covers the biofilm on support;
4. Simultaneous diffusion with biodegradation of the
substrate in the biofilm; and
5. Adsorption of the substrate on the support not covered by
the biofilm.
Each of the above steps can be quantified using transport
theory and can be mathematically expressed using equations
containing the design and operating parameters of the biofilter,
properties of the organic(s), and biodegradation kinetics
variables.
The model assumptions are as follows:
1. The support is homogeneous and has a flat geometry.
2. Biodegradation takes place in the biofilm only.
3. The gas-liquid and biofilm-support interfaces are at
equilibrium.
373
-------
4. Biodegradation is first order with respect to the
substrate. This assumption is justified at low
concentrations of the substrate.
The material balance for the substrate in the biofilm yields
the following equation:
dt
where,
C = concentration of the substrate in the biofilm,
Dc = effective diffusivity of the substrate in the biofilm,
kt = first order biodegradation rate constant,
z SB dimension co-ordinate, and
t «= time.
The initial and boundary conditions are given by the following
equations:
C-0, t<:Q, all z (2)
Boundary condition at liquid-biofilm interface:
Boundary condition at biofilm-support interface:
Ue___ __^ cdzj (4)
where;
KL = overall gas-liquid mass transfer coefficient,
H = Henry's constant for the substrate,
PA = partial pressure of the substrate in the bulk gas phase,
C2 = concentration of the substrate at the liquid-biofilm
interface,
p? = support density,
C — substrate concentration in the support, and
L *= characteristic dimension of the support.
The first boundary condition (equation 3) represents the continuity
of the substrate flux at the liquid-biofilm interface. The second
boundary condition (equation 4) is a material balance for the
substrate in the support. In addition to these equations, the
374
-------
following equation can be written for biofilm growth:
X-j- (da) =[L*d* (yk^c-bX) dz (5)
where ;
y = bacterial yield,
b = bacterial decay coefficient,
X » biomass concentration, and
d^ = biofilm thickness.
The first term in parentheses after the integral sign in equation
5 gives the biomass yield due to biodegradation of substrate and
the second term describes the decay of the biomass.
At steady state the solution of equations 1-5 becomes
simplified. Steady state in the biofilm implies that the
concentration profiles of the substrates are independent of time.
There is no net growth of the biofilm and the thickness of the
biofilm attains a constant value. The biodegradation rate attains
a final value that does not change with time. The rate of transport
of the substrate through the biofilm equals the rate of
biodegradation. This results in a. constant, non-zero concentration
of the substrate in the biofilm-support material. The following
three equations describe the steady state solution?
) 2-l) (6)
d;)
where
bx
375
-------
z>.
and,
KT
(10)
C3 is the concentration of the substrate at the biofilm-support
interface. The dimensionless quantities are defined as follows:
c-- (12)
(13)
Equations 6, 7 and 8 can be solved simultaneously to obtain the
concentration profile of the substrate in the biofilm. The
degradation rate can then be obtained from the following equation:
Equation 14 yields the biodegradation rate for the substrate
on a single particle. This rate is utilized in the sizing of the
biofilter. The mass balance on a differential element of the
biofilter yields the following equation:
-G dyA-RAa A dZ (15)
where;
G ~ gas flow rate,
A = biofilter cross sectional area, and
yA = mole fraction of the substrate in the gas.
Substituting from equations 13 and 14, the following equation
is obtained for the height of the biofilter:
376
-------
HG (16)
(i-c2')yA
The kinetic parameters of the model can be evaluated from the
experimental data. These parameters can then be utilized in the
sizing and design of biofilter for a given load of volatile organic
compounds.
BACKGROUND
While biological treatment of aqueous waste streams has been an
established practice, using biofilters for treatment of air or
gaseous streams has not been extensively investigated. Pomeroy (2)
and Carlson and Leiser (3) have presented a similar approach for the
removal of the sewage related odors. The main mechanism for the
removal of odorous compounds, for example, hydrogen sulfide,
mercaptans, terpenes, amines, etc., was adsorption on support
material. There was no clear evidence of biodegradation. Smith et al.
(4) determined that soil beds are effective in removing sulfur
containing gases and can serve as sinks for hydrocarbons.
Hartenstein (5) has presented a range of operating conditions
for a biofilter. He found that an important operating parameter is
the moisture content of the filter bed, which plays an important
part in determining removal efficiency. Eitner (6) determined that
the most active microbes in a biofilter are the heterotrophic and
chemo-organotrophic groups. He also noted that Actinomvcesspp. are
able to exploit a wide variety of organics and are reported to kill
pathogens under certain conditions.
The distribution of microbes in the biofilter has been described
by Ottengraf and van den Oever (7), Eitner (6) and Kampbell (8). The
population density of the microbes is highest at the gas entrance at
the bottom of the biofilter, and these microbes preferentially
metabolize the more readily degradable influent compounds. The less
degradable compounds are assimilated in the upper portions of the
biofilter.
Eitner (6) presented data which indicated that significant
reduction of the hydrocarbon concentration was achieved in
approximately one week, and maximum removal rates were attained
within one month of the operation. Ottengraf and van den Oever (7)
investigated the survivability of the microbial flora in biofilters
which were not loaded and showed that biofilter operation can be
suspended for 14 days with only minimal loss in activity.
Biofilters have been traditionally used for controlling odors.
Eitner and Gethke (9) showed that odors were reduced by approximately
98% at sewage treatment facilities. Prokop and Bohn (10) reported
99.9% removal rates for odors from an animal rendering facility.
377
-------
Hartenstein (5) provided data which showed the percent removal
of several organics as shown in Table 2.
TABLE 2: PERCENT REMOVAL OF ORGANICS IN A BIOFILTER
[DATA OF HARTENSTEIN (5)]
Compound
Hydrogen sulfide
Dimethyldisulfide
Terpene
Organo- sulfur gases
Ethyl benzene
Tetrachloroethylene
Chlorobenzene
% Removal
99
91
98
95
92
86
69
Kampbell (8) measured removal rates between 95% and 99% for
propane, isobutane, n-butane and trichloroethylene. Ottengraf and
van den Oever (7) also measured high removal rates for toluene,
butanol, ethylacetate and butylacetate.
Don and Feenstra (11) presented data comparing several
alternative technologies for treatment of waste hydrocarbon gas
streams and showed that biofilter is the most cost effective
treatment method.
A biofilter system, consisting of prefabricated concrete parts
which form an aeration plate to give uniform air distribution and
drainage ducts, termed BIKOVENT system, was developed by Drs. Hans
Gethke and Detlef Eitner of Aachen, Germany. The BIKOVENT system
has been extensively used in Germany and Austria for odor control
and controlling volatile organic compounds (VOCs) in waste air
streams. The technology was recently introduced to the U.S. by
Biofiltration Inc., Gainesville, Florida.
This paper reports on the experimental study on the
biodegradation of volatile organic compounds present in landfill
leachates in aerobic and anaerobic biofilters. The most abundant
compounds in the leachate streams were targeted for study. A
stripping study was carried out on the selected compounds to
confirm Henry's law constant values.
EXPERIMENTAL STUDIES
STRIPPING OF LEACHATES
The characterization of vapor-liquid equilibria for landfill
leachates (with respect to the volatiles of interest) was necessary
in order to estimate the emission of the volatiles from the
pollutant stream. It is expected that the stripping of the
378
-------
volatiles will be governed by Henry's law for each compound. The
interaction between various compounds in the multicomponent mixture
can be neglected since the leachate streams form a dilute solution.
The aim of the stripping experiments is to obtain Henry's constants
for the compounds present in leachates. This was important to
ensure that leachate constituents were not complexing the organics
and thereby reducing Henry's law constant. The theoretical basis
for these experiments has been described by Mackay et al. (12).
The experimental set up is shown in Figure 2. The experiments
consist of bubbling the stripping agent (nitrogen gas) at a fixed
flow rate through a leachate solution containing the volatiles and
monitoring the concentration of the volatiles in this solution. The
experimental conditions and results for stripping of methylene
chloride, trichloroethylene and chloroform are shown in Table 3. It
can be seen that, within experimental errors, Henry's constant can
be used to describe the behavior of volatile organic compounds in
leachates in a stripping operation.
TABLE 3: HENRY'S CONSTANTS FOR STRIPPING OF LEACHATES
Compound
Methylene
Chloride
Trichloroethylene
Chloroform
Henry's constant (m3 atm/mol x 103)
Water
. 3.19
11.7
3.39
Leachate
1
2.75
14.0
5.3
Leachate
2
2.28
14.9
4.2
Leachate
3
2.0
12.5
4.34
from sites in
Experimental Conditions:
Volume of lig_uid: 1000 ml
Nitrogen flow rate: 100 ml/min
Time of run: 90 min
leachates 1 and 2 were two different leachates
Delaware and leachate 3 was from Cincinnati.
BIOFILTER OPERATION
The VOCs from the landfill leachate streams were treated in a
bench scale biofilter. The following three chemicals (substrates)
were targeted for this study at the concentration shown below:
Toluene: 520 ppm
Methylene Chloride: 180 ppm
Trichloroethylene: 25 ppm
These three compounds were present in large concentration in
the landfill leachate and were found to be present predominantly in
the gas phase during the stripping study. Hence, when a landfill
379
-------
leachate is stripped with gas, as shown in Figure I, these
chemicals are expected to be present in the gas stream which will
be treated in a biofilter. The compound concentrations were
selected to obtain preliminary experimental data. Further
experimentation will be conducted using various concentrations of
these compounds. A schematic of the bench scale biofilter apparatus
is shown in Figure 3. Both the aerobic and anaerobic modes of
degradation were investigated independently of each other. The
substrates were fed to the biofilter through the gas phase. The
requisite composition of the substrates in the gas phase was
achieved by making the synthetic gas mixtures in a cylinder and
subsequent blending with air or nitrogen. This was done to ensure
a uniform feed concentration to the biofilter. The biofilter was
packed with the support material. Nutrient solution with
constituents as specified in Table 4 or Table 5 was circulated
counter to the gas through the bed. The inlet and outlet gas
streams were analyzed for the above three chemicals.
The details of the biofilter dimensions and operating
conditions (for both aerobic and anaerobic operations) are shown
below:
Support medium: activated carbon
Support dimension: 4 mm
Biofilter diameter: 25 mm
Packed height: 600 mm (jacketed)
Gas flow rate: 150 ml/min
Gas superficial velocity: 0.533 cm/s
Liquid flow rate: 10 ml/min
Liquid superficial velocity: 0.033 cm/s
The biofilters contained active acclimated biomass. The
biomass acclimated to the above substrates was obtained in the
following manner: Biomass from a pilot scale activated sludge plant
treating hazardous waste was suspended in the bioreactor (column
100 mm dia., 700 mm height). The bioreactor was fed daily with the
three substrates. Nutrients necessary for the growth were also
added weekly. The aerobic bioreactor was aerated by air. The
anaerobic bioreactor was operated similarly (but without any
aeration) with seed biomass from the anaerobic digester. The
biomass in the anaerobic bioreactor was kept suspended by operating
a recirculation loop at a high flow rate.
The biomass from the bioreactor was transferred to the
biofilter by circulating the bioreactor suspension through the
biofilter. It was found that the biomass could be effectively
transferred from the bioreactor to the biofilter.
Specific details on the aerobic and anaerobic biofilters have
been discussed in the next two sections.
Aerobic Biofilter
The synthetic gas mixtures of the substrates were prepared in
air for aerobic operations. The composition of the aerobic nutrient
380
-------
solution is given in Table 4.
The inlet and outlet gas and liquid streams were analyzed for
the volatile organic compounds mentioned above using a gas
chromatograph (EPA standard method 602} .
TABLE 4:COMPOSITION OF NUTRIENT SOLUTION FOR AEROBIC BIOFILTER
Salt
KH?PO,
K^HPO^
Na?HPO4
NH4C1
MgSO4.7H?0
CaCl?
FeCl7.6H.,0
Trace Elements-
MnSO4.4H2O
H3B03
ZnS04.7H20
(NH4)6M07024
FeCl3.EDTA
Yeast Extract
Concentration
(mg/L)
85.0
217.5
266.4
25.0
22.5
27.5
0.25
0.0399
0.0572
0.0428
0.0347
0.10
0.15
AnaerobicBiofliter
For anaerobic operation, the synthetic mixture of the
compounds was prepared in nitrogen. Nitrogen was scrubbed in a
solution of sodium thiosulfite (with resazurin as an indicator) to
remove the traces of oxygen prior to being fed to the biofilter.
The biofilter was maintained at constant temperature (35 °C.) by
circulating hot water through the biof ilter jacket. The composition
of the anaerobic nutrient solution is shown in Table 5.
The inlet and outlet concentrations of the compounds in the
gas and liquid streams were monitored using a gas ehromatograph
(EPA standard method 602}.
381
-------
TABLE 5: COMPOSITION OF NUTRIENT SOLUTION FOR ANAEROBIC BIOFILTER
Salt
Ammonium Chloride
Magnesium Chloride
Potassium Chloride
Calcium Chloride
Ammonium Phosphate
Ferrous Chloride
Cobalt Chloride
Potassium Iodide
Manganese Chloride
Ammonium Vanadate
Zinc Chloride
Sodium Molybdate
Boric Acid,.., ,
Nickel Chloride
Cysteine
Sodium Bicarbonate
Concentration (mg/L)
1200
500
400
25
80
40
2.5
2.5
0.5
0.5
0.5
0.5
0.5
0.5
10
6000
RESULTS AND DISCUSSION
It was found that the amount of compound removed from the gas
phase that can be accounted for by the increase in the liquid phase
concentration of the compound was negligible for both aerobic and
anaerobic operations. This means that the compound removed in the
biofilter by the trickling nutrient flow was negligible as compared
to the feed rate of the compound. Removal efficiency of the
biofilter for a compound was defined as the amount of compound
removed from the gas phase expressed as a percentage of the amount
of that compound fed to the biofilter through the gas phase. The
removal efficiency can ^"'calculated simply by taking the ratio of
difference in the inlet and outlet concentrations of the compound
in the gas phase to the concentration of the compound at the inlet.
The steady state removal efficiencies of both the aerobic and
anaerobic biofilters for the three compounds are shown below in
Table 6.
382
-------
TABLE 6: REMOVAL EFFICIENCIES OF THE BIOFILTER
Compound
Toluene
Methylene Chloride
Trichloroethylene
Removal Efficiency, %
Aerobic Biofilter
80
25
40
Anaerobic Biofilter
(unsteady state)
60
20
40
It should be noted that these removal efficiencies are a
function of the nutrient concentration. Since the impact of the
nutrients on removal efficiency has not been fully studied, numbers
given in Table 6 are preliminary and subject to modification
pending further research. The average biodegradation rates obtained
in the biofilter for each of the compounds are shown in Table 7.
TABLE 7: AVERAGE BIODEGRADATION RATES
Compound
Toluene
Methylene Chloride
Trichloroethylene
Average Biodegradation Rate,
mg substrate/cm3 biofilter min
Aerobic Biofilter
8.7 x 10-*
8.7 x 10'5
3.0 x 10'5
Anaerobic Biofilter
(unsteady state)
6.5 x 10'*
6.96 x ID'5
3.0 x 10~5
It can be seen that the removal efficiency of the biofilters
for any of the compounds is less than 100%. However, efficiencies
approaching total removal of substrates can be obtained by
providing more residence time for the gas or increasing nutrient
concentration. This can be achieved by decreasing the gas flow rate
for the biofilter, or alternately, by increasing the height of the
biofilter.
It can also be seen that the removal efficiencies of the
aerobic and anaerobic biofilters differ for the three compounds.
Differing intrinsic kinetics of biodegradation and other factors
such as diffusional resistances and biomass concentration result in
different aerobic and anaerobic biofilter heights for a given
degree of removal for the substrate.
It was also found that trichloroethylene could be degraded
under aerobic conditions in the presence of a co-metabolite such as
toluene.
383
-------
CONCLUSIONS
1. The volatile organic compounds present in the leachate
streams can be stripped according to Henry's constant for these
compounds in water. This shows that leachate constituents do not
change Henry's law constant significantly.
2. The compounds can be removed from gas stream by aerobic as
well as anaerobic degradation in a biofilter.
ACKNOWLEDGEMENT
We are thankful to Professor M. Suidan, Civil and
Environmental Engineering, University of Cincinnati, and Mr. S. I.
Safferman and Mr. R.C. Brenner, U.S. EPA for their help in carrying
out this research.
NOMENCLATURE
A
b
C
C
6
H
KLa
L
Pt
PA
r
R.
X
y
yA
z
z
pp
Cross sectional area of the biofilter, cm2
Bacterial decay, sec"1
Substrate concentration in the biofilm, mol/cm3
concentration in the support, mol substrate/gm
concentration at the ligiaid-biofilm interface,
concentration at the biofilm-support interface,
of the substrate in the biofilm,
Substrate
support
Substrate
mol/cm3
Substrate
mol/cm3
Biofilm thickness, cm
Effective diffusivity
cm2/ sec
Gas flow rate, mol/sec
Henry' s constant, atm m3/mol
First order biodegradation rate constant, sec"1
Overall gas-liquid mass transfer coefficient, cm/sec
Overall volumetric gas-liguid mass transfer coefficient,
sec"1
Characteristic dimension of the support, cm
Total gas pressure, atm
Partial pressure of substrate A, atm
Biomass growth/decay ratio, dimensionless
Biodegradation rate of A, mol/cm2 sec
Volumetric degradation rate of A, mol/cm3 sec
Time, sec
Biomass concentration, gm/cin3
Bacterial yield, gm biomass/ mol substrate
Mol fraction of A in gas, dimensionless
Dimension co-ordinate, cm
Height of the biofilter, cm
Kinetic parameter, dimensionless
Mass transfer parameter, dimensionless
Support density, gm/cm3
384
-------
superscript
* Dimensionless quantity
subscripts
i At the inlet of the biofilter
o At the outlet of the biofilter
REFERENCES
1. Bramlett J. , Furman C., Johnson A., Nelson N., Ellis W. and Vick
W. Composition of Leachates from Actual Hazardous Waste Sites. EPA-
600/2-87/043, U.S. Environmental Protection Agency, 1987.
2. Pomeroy R.D. Biological treatment of odorous air. J. Wat.
Pollut. Contr. Fed. 54: 1541, 1982.
3. Carlson D.A. and Leiser C.P. Soil beds for the control of sewage
odors. J. Wat. Pollut. Contr. Fed. 38: 829, 1966.
4. Smith K.A., Bremmer J.A. and Tatabai M.A. Sorption of gaseous
atmospheric pollutants by soil. Soil Science, p.313, 1973.
5. Hartenstein H. Assessment and redesign of an existing
biofiltration system. M.S. Thesis, University of Florida, 1987.
6. Eitner D. Untersuchungen uber Einsatz und Leistungsfahigkeit von
Konpostfilteranlagen zur biologisciien Abluftreiningun ira Bereich
van Klaranlagen unter besonderer Berucksichtigung der Standzeit
(Investigations of the use and ability of compost filters for the
biological waste gas purification with special emphasis on the
operation time aspects). GWA, Band 71, TWTH Aachen, 1984.
7. Ottengraf S.P.P. and van den Oever H.C. Kinetics of organic
compound removal from waste gases with a biological filter.
Biotech. Bioeng. 25: 3089, 1983.
8, Kampbell D.H., Wilson J.T., Read H.W. and Stocksdale T.T.
Removal of volatile aliphatic hydrocarbons in a soil bioreactor. J.
Air Pollut. Contr. Assoc. 37: 1236, 1987.
9. Eitner D, and Gethke H.G. Design, construction and operation of
biofilters for odor control in sewage treatment plants. Paper
presented at 80th annual meeting of Air Pollution Control
Association, New York, New York, June 1987.
10. Prokop W.H. and Bohn H.L. Soil bed system for control of
rendering plant odors. J. Air Pollut. Contr. Assoc. 35: 1332, 1985.
11. Don J.A. and Feenstra L. Odor abatement through biofiltration.
385
-------
Paper presented at Symposium Louvain-La-Neuve, Belgium, April 1984.
12. Mackay D., Shiu W.¥. and Sutherland R.P. Determination of air-
water Henry's law constants for hydrophobia pollutants. Environ.
Sci. Technol. 13: 333, 1979.
Makeup
Water and-
Leachate
Air
or Nitrogen
Treated Air
or Nitrogen
Biofilter
Stripper
To Activated
Sludge Plant
Pump
Figure 1: Schematic of the Biofilter Treatment Scheme
386
-------
To flow meter
Flow
Controller
T
Sample tap
o
o
o
o
o
I
X
C
C
Water—
in -%
o(
1
XJ
i
— Water out
Stripping Vessel
Saturator
From Nitrogen
Cylinder
figure 2: Experimental Sat up for Stripping
To vent
To vent
Substrates
Bioreactor
bioniass
suspension
Flow i-l-i
Meter I I
— Sample
Point
water
'out
Support-
biomass a
complex"
pump
pump Nutrient
Solution Sample
Reservoir Point
Air
Figure 3: Schematic of Bioreactor-Biofilter System
387
-------
SCREENING OF COMMERCIAL BYPRODUCTS FOR ENHANCEMENT OF
OIL BIODEGRADATION IN CLOSED MICROCOSMS
by: Albert D. Venosa and John R. Haines
Risk Reduction Engineering Laboratory
Cincinnati, Ohio 45268
Wipawan Nisamaneepong
Engineering and Economics Research Institute
Reston, Virginia 22091
Rakesh Govind, Salil Pradhan, and Belal Siddique
University of Cincinnati
Cincinnati, Ohio 45221
ABSTRACT
Ten commercial products designed to enhance oil biodegradation were
tested in the laboratory. The products were used according to manufacturers'
directions. The performance of each product was assessed by measuring:
onset, rate, and extent of 02 uptake in a respirometer; disappearance of oil
components by GC and GC/MS; and growth of microorganisms on crude oil agar and
marine agar. Products selected for field testing were required to outperform
ordinary mineral nutrients in enhancing oil biodegradation. Two products, E
and G, were found to outperform mineral nutrients in 02 uptake and removal of
oil components. Both E and 6 produced rapid 02 uptake and a greater net 02
uptake than mineral nutrients alone. Similarly, alkane and aromatic
hydrocarbon reduction was more extensive than mineral nutrients alone. The
other products either matched performance with mineral nutrients or were much
poorer in the choice criteria.
388
-------
INTRODUCTION
Bacterial degradation of petroleum hydrocarbons has been known and
recognized for decades. The subject has been reviewed comprehensively in the
literature (1,2,4), the most recent one appearing this year (5). Vestal et
al. (6) reported that, although oil degraders comprise approximately 1% of the
total heterotrophic population in unpolluted waters, the oil degrader
population increases to as high as 10% in response to a spill. In 1989,
research conducted by the U.S. Environmental Protection Agency (EPA) in Prince
William Sound demonstrated that microbial communities on the contaminated
beaches were highly competent in their ability to degrade the Prudhoe Bay
crude that was spilled from the Exxon Valdez (7). The purpose of the latter
study was to determine if application of water soluble and oleophilic
nutrients could enhance the natural biodegradation rate.
After the EPA study showed that bioremediation of oil-polluted beaches
was enhanced by the addition of fertilizer, the question then arose whether
further enhancement was possible with the addition of microbial inocula
prepared from oil degrading populations not indigenous to Alaska. Seeding
experiments have been done in previous studies with mixed results (5). In a
recent study, Dott et al. (8) compared nine commercial mixed bacterial
cultures to activated sludge microorganisms for their ability to degrade fuel
oil in laboratory flasks. They found that fuel oil degradation by the
naturally occurring bacteria in activated sludge did not depend on, nor was it
enhanced by the application of highly adapted commercially available cultures.
Most success has been achieved when chemostats or fermenters are used to
control conditions or reduce competition from indigenous microflora (9).
In February, 1990, the U.S. Environmental Protection Agency issued a
public solicitation for proposals to the bioremediation industry on testing
the efficacy of commercial microbial products for enhancing degradation of
weathered Alaskan crude oil. The Agency commissioned the National
Environmental Technology Applications Corporation (NETAC), a non-profit
corporation dedicated to the commercialization of environmental technologies,
to convene a panel of experts to review the proposals and choose those that
offered the most promise for success in the field. Forty proposals were
submitted, and 11 were selected for the first phase of a two-tiered testing
protocol (only 10 were tested because one company did not participate). The
laboratory testing consisted of electrolytic respirometers set up to measure
oxygen uptake over time and shake flasks to measure oil degradation and
microbial growth. If one or more products were found effective, the second
tier would take place, consisting of small field plots on an actual
contaminated beach in Prince William Sound in the summer of 1990. This paper
discusses the first phase of testing, the laboratory batch flask and
respirometric evaluations.
The objective of the laboratory protocol was to determine if commercial
bioremediation products can enhance the biodegradation of weathered crude oil
to a degree significantly better than that achievable by simple fertilizer
application. Testing was conducted in a controlled and closed environment
designed to give quick results under ideal conditions. It was not meant to
389
-------
simulate the open environment of the oiled beaches of Prince William Sound,
where conditions are in a constant state of flux with respect to tidal cycles
and washout, temperature variation, climatic changes, freshwater/saltwater
interactions, etc. The organisms inside the respirometer vessels were in
continuous contact with the oil, seawater, and nutrients added initially, and
the seawater was not replenished every 12 hours as is the case in nature. The
test was merely a screening procedure that was designed to determine if there
was sufficient enhancement due to the commercial additives that would justify
proceeding to the next tier of testing. To proceed to the field phase, three
lines of evidence were used for decision-making: rapid onset and high rate of
oxygen uptake, substantial growth of oil degraders, and significant
degradation of the aliphatic and aromatic fractions of the weathered Prudhoe
Bay crude oil.
The 10 companies participating in the laboratory testing phase were (in
alphabetical order): Alpha Environmental, Bioversal, Elf Aquitaine,
ERI-Microbe Masters, Imbach, Hicrolife Technics, Polybac, Sybron, Waste
Microbes, and Woodward Clyde. Specific products and companies cannot be
identified to preserve confidentiality.
MATERIALS AND METHODS
ELECTROLYTIC RESPIROMETRY
The studies were conducted using four automated continuous oxygen-uptake
measuring Voith Sapromats (Model J-12). The instrument consists of a
temperature-controlled water bath containing measuring units; a recorder for
digital indication and direct plotting of the oxygen uptake velocity curves;
and a cooling unit for the conditioning and continuous recirculation of water
bath volume. The recorder displays a digital readout of oxygen uptake and
constructs a graph of the data for each measuring unit. The cooling unit
constantly recirculates water to maintain a uniform temperature in the water
bath. The measuring units are comprised of 12 reaction vessels each with a
carbon dioxide absorber mounted inside, 12 oxygen generators each connected to
its own reaction vessel by tubing, and 12 pressure indicators connected
electronically to the reaction vessels. The measuring units are
interconnected by tubing, forming an air-sealed system, so that the
atmospheric pressure fluctuations do not adversely affect the results.
Depletion of oxygen by microbiological activity creates a vacuum, which
is sensed by the pressure indicator. The oxygen generator is triggered to
produce just enough oxygen to counterbalance the negative pressure. The
current used to generate the oxygen is measured by the digital recorder, and
the data are converted directly into mg/L oxygen uptake. The C02 produced by
microbial activity is absorbed by soda lime. The nitrogen/oxygen ratio in the
gas phase above the sample is maintained throughout the experiment, and there
is no depletion of oxygen. The oxygen generators of the individual measuring
units are electrolytic cells that supply the required amount of oxygen by
electrolytic decomposition of copper sulfate/sulfuric acid solution.
390
-------
A recorder/plotter constructs an oxygen uptake graph as a function of
time and displays it on the computer screen while digitally saving the data on
disc. For frequent recording and storage of oxygen uptake data, the Sapromat
B-12 recorders are interfaced to an IBM-AT personal computer via the Metrobyte
interface system. A software package allows the collection of data at 15
minute intervals.
EXPERIMENTAL DESIGN
All commercial products were tested in duplicate at the concentration
recommended by the manufacturer. Each experimental respirometer flask was
charged with the following materials^in the order listed: weathered crude
oil, 250 mg; 250 mL seawater from Prince William Sound; and commercial
product at the concentration specified by the manufacturer. Seawater was
prepared as follows: 25 g of oiled rocks from a contaminated beach in Prince
William Sound was placed in a 4-L flask to which was added 2 L of seawater.
The mixture was shaken for approximately 30 minutes to wash off a microbial
inoculum from the rocks. The flask contents were allowed to settle, and 10ml
of supernatant was mixed in each respirometer vessel. The following table
presents the summarized experimental design showing all control and
experimental flasks.
TABLE 1. EXPERIMENTAL DESIGN FOR RESPIROMETRIC STUDIES.
Reaction Vessel Weathered Commercial Seawater TOTAL
Oil Product
TEST FLASKS:
TPn -f + + 20
Fl,2 ' + - + 2
CONTROL FLASKS:
CPn + + 20
CF1,2 - + 2
Cl-inoculum - - + 2
C2-no nutrients + + 2
TOTAL 48
TPn = duplicate commercial product flasks (n = 10)
Fl,2 = fertilizer flasks (mineral N and P nutrients)
CPn, CF1,2 = no-oil controls for products and fertilizer,respectively
Cl, C2 « inoculum and no-nutrient controls
391
-------
Flasks Fl and F2 represented simple inorganic fertilizer application and
contained the following ingredients (mg/L final concentration): KH2P04, 6.33;
K2HP04, 16.19; Na,HP04, 24.86; NH,C1, 38.5; MgSO,.?H20, 45; CaCl,.2H20, 55;
FeCl3«6H20, 2.5. the following additional trace elements were included in the
formulation (pg/L final concentration): MnS04-H20, 60.4; H,B03, 114.4;
ZnS04.7H20, 85.6; and (NH4)6Mor024«4H20, 69.4.
All respirometer flasks were incubated at 15"C in the dark and
continuously stirred at 300 rpm by magnetic stirrers. The first set of
control flasks (CPn, CF1,2) represented background oxygen uptake of the
product and seawater without oil. Results from these flasks were subtracted
from the appropriate test flasks to obtain the net oxygen uptake on the
weathered oil. The inoculum control represented the endogenous oxygen uptake
of the organisms from the v/ashed beach material and the seawater alone. The
no-nutrient control represented the oxygen uptake of the organisms from the
washed beach material and seawater on weathered oil without any external
source of nutrient addition (i.e., background nutrient levels from Prince
William Sound).
FLASK EXPERIMENTS
Shaker flasks duplicating the respirometer flasks were used to assess
the quantitative changes in oil composition by chromatographic separation of
the individual components. Although it was possible to remove samples from
the respirometer flasks, it was deemed more prudent not to disturb the
respirometric runs but instead have the shake flasks with proportionately
higher levels of oil, commercial products, etc., to facilitate sampling for
and precision/accuracy of the analytical chemistry. Table 2 summarizes the
shaker flask experimental design.
The test flasks corresponded exactly to the 22 test flasks listed in
Table 1 but with the following modifications: flask size, 250 ml; seawater,
100 ml; weathered oil and commercial products, 10 times the final
concentrations used in the respirometer flasks; and mineral nutrients, same
final concentration used in the respirometer flasks. The higher concentration
of weathered oil was used to improve the final sensitivity of the chemical
analyses.
In addition to the 22 test flasks, 18 supplemental flasks were set up.
These reactors represented 9 sterile product controls, which determined
whether the enhancement was due to the microorganisms or to the nutrients or
metabolites in the product, and 9 sterile background controls (i.e., sterile
oil and seawater, but non-sterile product) to evaluate the effect of
competition from naturally occurring organisms (one of the 10 products did not
receive these sterile treatments). Sterilization of materials was
accomplished by autoclaving at 12TC for 15 minutes.
SAMPLING
There were three sampling events for analytical chemistry and
microbiology: day 0, day 11, and day 20. These events were determined by the
392
-------
shape of the oxygen uptake curves from the respirometry experiments. Each
shaker flask was sacrificed at the indicated sampling time by mixing the
contents with methylene chloride and performing the extraction on the entire
mixture. Before sacrificing a flask, a small aliquot was removed for
determination of microbial density changes.
TABLE 2. EXPERIMENTAL DESIGN FOR THE SHAKER FLASK STUDIES.
Reaction Vessel Weathered Commercial Seawater TOTAL
Oil Product
TEST FLASKS:
TPn . + + -f 20a
SPn + sterile + 9a
TPnSb sterile + sterile 9a
Fl,2 + + 2a
CONTROL FLASKS:
CPn + + lOb
CF1 + Ib
Cl-inoculum - + Ib
C2-no nutrients + + 2a
TOTAL 54
TPn = duplicate commercial products (n = 10), non-sterile system
SPn = sterile products in non-sterile seawater/oil, non-duplicated
TPnSb = non-sterile products in sterile seawater/oil, non-duplicated
Fl,2 = fertilizer (mineral N and P nutrients) in non-sterile system
CPn, CF1 = no-oil controls for products and fertilizer, respectively
Cl, C2 = inoculum and no-nutrient controls
a - microbiological and chemical analysis
b = microbiological analysis only
393
-------
ANALYTICAL CHEMISTRY
The oil constituents were analyzed by measuring the aliphatic and
aromatic fractions of the methylene chloride extracts. The extracts were
concentrated and passed through a silica gel fractionation column to separate
the alkanes and the aromatics. The column was first eluted with hexane to
collect the alkane fraction and then a 1:1 mixture of hexane and benzene to
collect the aromatic fraction. Any polar compounds remaining in the extract
stayed bound to the silica gel column. Aliphatic fractions were measured by
gas chromatography using a flame ionization detector. The aromatic fractions
were characterized by gas chromatography/mass spectrometry (6S/MS).
NUTRIENT ANALYSIS
The nitrogen species NH3-N, N02--N, and N03--N were determined by U.S.
EPA Methods (10). The NH3-N method was No. 350.1 and the N02-N/N03-N method
was No. 353.1.
MICROBIOLOGICAL TESTING
Growth of oil degraders was measured by spread plates on oil agar
(Bushnell-Haas medium supplemented with Prudhoe Bay crude oil as the carbon
source) (3). Total heterotrophic bacterial numbers were estimated by spread
plate culture on Marine Agar 2216 (Difco). Plates were incubated at 15*C for
21 days prior to counting.
RESULTS
Nutrient Levels in Each Product. Product flasks requiring nutrient
addition, as specified by the product manufacturer, received the same level of
mineral nutrients as the fertilizer flasks. The ammonia-nitrogen concen-
trations measured in each product flask at day 0 are summarized in Table 3.
TABLE 3. NH3-N LEVELS IN EACH PRODUCT FLASK AT THE START OF THE EXPERIMENT.
PRODUCT
A
B
C
D
E
F
G
H
I
J
PR*
NH3-N
mg/L
8.0
2.1
1080.0
11.8
11.3
10.0
24.9
426.0
0.5
1.5
6.9
"W5
YES
NO
NO
YES
YES
YES
YES
NO
NO
NO
YES
*FR = mineral fertilizer
394
-------
STATISTICAL ANALYSIS OF ALKANE DEGRADATION DATA
The percent reductions of the resolved aliphatic constituents of the
weathered oil (n-C12 through n-C34 plus the branched-chain compounds pristane
and phytane) were computed at day,11 for each product flask and the results
compared to the percent reduction computed for the mineral nutrient flasks.
Table 4 summarizes the statistical differences observed using Tukey's
Studentized Range Test (11). The products are arranged in descending order of
significance. Only Products E and G gave significantly higher removals (p <
0.05) than inorganic fertilizer after 11 days. Six of the other products gave
results no different from mineral nutrients, while two actually gave
significantly lower removals. The latter results suggest that the products
may have been toxic to the biomass at the levels used in the closed flasks.
TABLE 4. TUKEY'S STUDENTIZED RANGE TEST FOR DETECTING DIFFERENCES IN MEAN
PERCENT REMOVAL OF ALKANES BY PRODUCTS IN 11 DAYS
PRODUCT % REMOVAL SIGNIFICANTLY DIFFERENT FROM
INORGANIC NUTRIENTS*
E
G
B
A
D
FR
C
J
H
F
I
94.5
93.6
87.9
75.9
74.2
68.4
67.8
59.9
49.5
33.3
27.9
YES
YES
NO
NO
NO
NO
NO
NO
NO
YES
YES
*Minimum Detection Difference = 21.3% at 5% Significance Level
TOTAL ALKANE REDUCTION
The total alkane degradation data from the product flasks and the
corresponding sterile controls at days 11 and 20 are summarized in Figure 1.
395
-------
-------
The products are arranged on the x-axis in the order determined by the .
statistical analysis (this same ordering has been made on all figures). At
day 11 (top half of Figure 1), better degradation was observed in every case
when the commercial products were first sterilized, suggesting that the
indigenous Alaskan populations were doing most if not all of the
bioremediation. In contrast, less degradation occurred in every case except
Product I when the background (seawater and oil) was first sterilized. This
suggests that, when left alone, the product organisms were less able to
degrade the alkane fraction than the indigenous organisms. In the non-sterile
treatments, enhancement was observed for Products E and 6 compared to mineral
nutrients, suggesting that the products exhibiting the enhancement were
providing metabolites or some other form of nutritional' benefit that was
lacking in the mineral nutrient flask. By day 20 (bottom half of Figure 1),
all products except Products F and I caught up, giving greater than 85%
reduction in the total alkane levels in the flasks. However, most of the
flasks containing oil and seawater that were first sterilized still
significantly lagged behind the non-sterile systems.
TOTAL AROMATICS REDUCTION
A summary of the total aromatics reduction data at day 11 and 20 is
presented in Figure 2. Differences are less clear among the products,
although Products C, F, H, and I gave total reductions considerably less than
mineral nutrients. By day 20, aromatic reduction by Product C was somewhat
closer to the others, while Products H, F, and I substantially lagged.
Excellent removal of aromatics was observed in all other flasks.
RESPIROMETRIC RESULTS
The net oxygen uptake curves (oxygen uptake in product flasks with oil
minus oxygen uptake in flasks without oil) for all 10 products (curves with
symbols) compared to mineral nutrients (curve with no symbols) are summarized
in Figure 3a and 3b. In Figure 3a the two products giving significantly
higher alkane degradation, E and 6, also exhibited higher net oxygen
consumption than mineral nutrients. The final plateau in total oxygen uptake
was slightly less than 500 mg/L for both Products E and 6 compared to about
340 mg/L for the mineral nutrient flasks. The acclimation lag period for
products E and 6 were approximately 2 and 4 days, respectively, compared to 5
days with mineral nutrients. Product A gave the highest maximum net uptake
(630 mg/L compared to 340 for mineral nutrients) but the lag period was almost
10 days. Products B and D exhibited 02 uptake characteristics no different
from the nutrient control.
In Figure 3b only products J and C gave higher overall net 0, consumption
than mineral nutrients, although product F exceeded the control after 27 days.
The lag period for both products J and C was only 1 day. The shape of the
product F curve was multi-phasic, suggesting the organisms were consuming
different substrates at different rates and at different times (diauxie).
Very little net oxygen consumption was observed with product I.
397
-------
DAY 11
DAY 20
B
H
A D FR C J
PRODUCT
Figure 2. Total aromatics reduction in the product flasks at days 11 and 20.
398
-------
(a)
en
E
.
rt
•*-•
o.
CM
O
J2
=3
3
O
(b)
FR
E
Q
B
A
D
20 SO 40
Time, days
TOO
O
1O
2O
3O
40
Time, days
Figure 3. Net ozjgen uptake cturei for products and mineral
nutrients: (B) products E.CJB4J) (b) prodacta C.J.H JJ.
399
-------
MICROBIOLOGICAL RESULTS
Virtually all changes in oil degrader densities occurred by day 11. The
populations levelled off in all flasks thereafter. Consequently, the growth
of oil degraders has been summarized for days 0 and 11 only, and the results
are presented in Figure 4.
Products E and G, which gave the best alkane degradation of all the
products (Table 4 and Figure 1) and displayed net oxygen uptake character-
istics superior to most (Figure 3a), also exhibited excellent growth of oil
degraders in 11 days. Products C, 0, and F yielded high levels of oil
degraders and good oxygen uptake curves, but alkane degradation was no better
than the populations growing in simple mineral nutrients. Oil degrader
populations actually declined in the product B flasks, and the increase in oil
degraders in the flasks containing products A, D, and I was minimal.
DISCUSSION
The objective was to determine whether a commercial bioremediation
product was able to effect weathered crude oil degradation better than natural
Prince William Sound organisms when stimulated with simple mineral nutrients.
Oil degradation chemistry, oxygen uptake in respirometer flasks, and microbial
density changes were used to decide which product(s) would proceed to field
testing.
Of all the products tested, the two that provided the most consistent
results in all three tests were products E and G. Both gave higher oxygen
uptake, greater growth in oil degraders, and superior alkane degradation than
mineral nutrients. Two of the products, C and J, showed good growth of oil
degrader populations and gave excellent net oxygen uptake curves but were no
better than indigenous populations stimulated with simple mineral nutrients.
Product F yielded the highest oil degrading population of all, yet its oxygen
uptake curve was no better than the mineral nutrient curve until after day 27,
and alkane degradation was relatively poor. Product A gave the best overall
net oxygen consumption, but the increase in oil degraders and the relative
alkane degradative capability were mediocre, as were the flasks containing
products B, D, H, and I.
The sterile controls revealed that the indigenous Alaskan oil-degrading
populations were performing most if not all of the biodegradative activity.
The organisms present in products E and G did not appear to contribute
significantly to such activity. This suggests that a co-metabolite or a
nutrient or some other unknown factor exists in these two products that
stimulates the indigenous microorganisms to degrade the crude oil constituents
at rates faster than is possible with simple nutrient addition. Further work
needs to be done to define the enhancement factor in these products.
Correlations have not as yet been made between weathered crude oil
degradation and oxygen uptake, nor have carbon balances been performed. Work
is being planned to measure carbonaceous metabolic end products, C02
production, and total biomass yield, and then to correlate this information
400
-------
DAY 0
DAY 11
—3
s
C/3
o
W
>_a
o
o
o
B
H
A D FR C J
PRODUCT
Figure 4. Growth of oil degraders in the product flasks.
401
-------
with the oxygen consumption data. If such correlations can be established,
then use of oxygen consumption data for estimating biodegradation efficacy as
part of a screening protocol will be made possible. The respirometric
technique requires much less effort than conventional shake flask studies
because data gathering is automated and computer!zed, and it is not necessary
to collect samples manually during the course of a biodegradation experiment.
All that is required, assuming the proper correlations have been established,
is the careful measurement of initial substrate and biomass values followed by
the measurement of the residual soluble product value at the plateau of the
uptake curve (12). From the analysis of this information, treatment decisions
can be facilitated.
CONCLUSIONS
Results from all three lines of evidence, i.e., respirometry,
microbiology, and oil chemistry, supported the decision to field test only
products E and G. It appears from all the available evidence that the
indigenous Alaskan microorganisms were primarily responsible for the
biodegradation in the closed flasks and respirometer vessels, and that any
enhancement provided by products E and G might have been due simply to
metabolites, nutrients, or co-substrates present fortuitously in the products.
Questions remain unanswered, and further research is being planned to increase
our knowledge base regarding oil spill bioremediation enhancement using
commercial inocula.
REFERENCES
1. Atlas, R. M. Hicrobial degradation of petroleum hydrocarbons: an
environmental perspective. Microbiol. Rev. 45: 180-209, 1981.
2. Cooney, J. J. The fate of petroleum pollutants in freshwater
ecosystems. In R. M. Atlas (ed.), Petroleum Microbiology, Macmillan
Publishing Co., New York, 1984.
3. Difco Products. 1985. Difco Manual, pp. 184-185. Tenth edition.
Difco Laboratories, Detroit, MI.
4. Floodgate, G. D. The fate of petroleum in marine ecosystems. In R. M.
Atlas (ed.), Petroleum Microbiology, Hacmillan Publishing Co., New York,
1984.
5. Leahy, J. G. and R. R. Colwell. Microbial degradation of hydrocarbons
in the environment. Microbiol. Rev. 54: 305-315, 1990.
6. Vestal, J. R., J. J. Cooney, S. Crow, and J. Berger, In R. M. Atlas
(ed.), Petroleum Microbiology, Macmillan Publishing Co., New York, 1984.
7. Pritchard, P. H., R. Araujo, J. R. Clark, L. D, Claxton, R. B. Coffin,
C. F. Costa, J. A. Glaser, J. R. Haines, D. T. Heggem, F. V. Kremer, S.
C. McCutcheon, J. E. Rogers, and A. D. Venosa. Interim report: oil
402
-------
spill bioremediation project. U.S. Environmental Protection Agency,
Office of Research and Development, Washington, D.C., 1990.
8. Dott, W., D. .Feidieker, P. Kampfer, H. Schleibinger, and S. Strechel.
Comparison of autochthonous bacteria and commercially available cultures
with respect to their effectiveness in fuel oil degradation. J. Ind.
Microbiol. 4: 365-374, 1989..
9. Wong, A. D. and C. D. Goldsmith. The impact of a chemostat discharge
containing oil degrading bacteria on the biological kinetics of a
refinery activated sludge process. Water Sci. Techno!. 20: 131-136,
1988.
10. U.S. Environmental Protection Agency. Methods for chemical analysis of
water and wastes. EPA 600/4-79-020, U.S. Environmental Protection
Agency, Washington, D.C., 1979.
11. Tukey, J. W. The problem of multiple comparisons. 396 pp. Princeton
University, Princeton, NO, 1953.
12. Grady, C. P., Jr., J. S. Dang, D. M, Harvey, A. Jobbagy, and X. -L.
Wang. Determination of biodegradation kinetics through use of
electrolytic respirometry. Water Science Techno!. 21: 957-968, 1989.
403
-------
ALTERNATING CURRENT ELECTROCOAGUIATION FOR
SUPERFUND SITE REMEDIATION
by: Dr. Clifton W. Farrell
Electro-Pure Systems, Inc.
Amherst, NY 14228-2298
ABSTRACT
A study is being conducted by Electro-Pure Systems, Inc. (EPS) under the
Emerging Technology portion of the U. S. Environmental Protection Agency's
(EPA's) Superfund Innovative Technology Evaluation (SITE) Program to study
alternating current electrocoagulation for Superfund site remediation.
Alternating current electrocoagulation has proven to be effective in
agglomerating and removing colloidal solids, metals and certain organic
contaminants from surrogate soils prepared from the U.S. EPA's Synthetic Soil
Matrix. Treatments under a wide range of operating conditions (power,
electrode configurations, retention time, frequency, mode of operation) have
enabled the optimum parameter settings to be established for multiple phase
separation. Electrocoagulation enables appreciably enhanced filtration and
dewatering rates to be realized for metals- and diesel fuel-spiked surrogate
soil slurries; such enhancements are prompted by growth in the mean particle
size of the clays and particulates from typically <10 microns to as much as
150 microns depending on the degree of electrocoagulation. Reduction in the
total suspended solids content of clays in all slurries in excess of 90% can
routinely be achieved. Bench-scale experiments of the metals-spiked surrogate
soils indicate that electrocoagulation preferentially concentrates soluble
metals into the sludge phase; excellent metals separation (Pb, Cr, Cu, Cd) can
be realized. Experiments on surrogate wastes spiked with volatile organics
suggest that this technology is not capable of effecting good volatile
extractions from the aqueous phase. Reductions in excess of 80% in the total
organic carbon (TOO) content of the diesel fuel-spiked surrogates can,
however, be achieved.
Alternating current electrocoagulation is effective in reducing the
volume of a potentially hazardous slurry by concentrating the clays and metals
into a readily dewaterable and filterable solid phase. Electrocoagulation
offers equivalent or slightly better treatment than chemical polymer addition
and has the added attributes of producing more readily filterable sludges
without introduction of soluble species. In terms of solids and metals
reductions the results achieved by electrocoagulation are far superior to
those achieved by alum addition,
This paper has been reviewed in accordance with the U. S. Environmental
Protection Agency's peer and administrative review policies and approved for
presentation and publication.
404
-------
INTRODUCTION
Electro-Pure Systems, Inc. (EPS) has completed the first year of
participation in the U. S. EPA's Emerging Technology portion of the SITE
Program. The primary objective of this project is demonstration of the
technical and economic viability of alternating current electrocoagulation
technology for use in Superfund site remediation. Alternating current
electrocoagulation offers a technologically simple mechanism to achieve phase
separation of liquid-liquid and solid-liquid slurries and emulsions, to remove
certain metals from solution and to destroy certain organic compounds. In the
first year of the program laboratory experiments conducted in bench-scale
electrocoagulation units, referred to as ACE Separators, were performed to
evaluate performance of the technology for treatment of surrogate waste
matrices containing metals and organic constituents that might be present at
Superfund sites. The second year of the program entails investigation of the
potential of a packed-bed catalyst version of the ACE Separator to oxidize
organic compounds within aqueous phases. Presented in this paper is an
overview of the technology, discussion of some preliminary results for the
Year One SITE program and for testing of industrial effluents and a summary of
the applications and benefits of ACE Separator usage.
TECHNOLOGY DESCRIPTION
BACKGROUND
Alternating current electrocoagulation (ACE Technology) was originally
developed as a treatment technology in the early 1980s to break stable aqueous
suspensions of clays and coal fines in the mining industry. The technology
was developed as a replacement for primary chemical coagulant addition to
simplify effluent treatment, realize cost savings, and facilitate recovery of
fine-grained products that would otherwise have been lost. The traditional
approach for treatment of such effluents entails addition of organic polymers
or inorganic salts to promote flocculation of fine particulates and colloidal-
sized oil droplets in aqueous suspensions. These flocculated materials are
then separated by sedimentation or filtration. Unfortunately, chemical
coagulant addition generates voluminous, hydrous sludges which are difficult
to de-water and slow to filter. As an alternative to chemical conditioning
and flocculation, ACE Technology agglomerates the particulates without adding
any soluble species to produce a sludge with a lower contained water content
and which will filter more rapidly. Another disadvantage of chemical
coagulation is the high susceptibility to filter shear of the particulates and
emulsion droplets entrained in the sweep floes. Through separation of the
hazardous components from an aqueous waste, the volume of potentially toxic
pollutants requiring special handling and disposal can be minimized. Waste
reduction goals may be accomplished by integrating this technology into a
variety of operations which generate contaminated water. Laboratory-scale
testing has also indicated the ACE Technology is capable of effecting removal
of soluble and insoluble metals and anions from aqueous streams.
405
-------
PRINCIPLES OF ACE TECHNOLOGY
Most suspended solid particles carry electrical charges on their
surfaces. Such charges commonly develop through preferential adsorption,
ionization and isomorphous replacement. These ions form a primary layer
which, in turn, attracts an oppositely charged secondary layer of ions. The
size of the particles and the strength of the ionic charge largely affect
whether the particles will aggregate and settle out of solution. When the
particles are sufficiently large, gravity will usually overcome the electrical
forces which tend to hold such particles in suspension. Smaller particles are
more susceptible to being suspended by electrical forces. For example, if a
particle has a weak ionic charge, the repulsive nature of the ionic double-
layer will prevent the particles approaching close enough for Van der Waals
forces to overcome the electrical repulsion. Thus, the particles are held in
suspension.
ACE Technology is based upon colloidal chemistry principles using
alternating current power and electrophoretic metal hydroxide coagulation.
Two basic mechanisms have been postulated for enhancing particle growth
characteristics as a result of electrocoagulation: electroflocculatiori,
whereby minute quantities of highly-charged cationic metal hydroxides are
released from the electrodes to facilitate flocculation, and electrostriction,
whereby the charges on colloidal particles are neutralized when subjected to
alternating current electrical field conditions. Electrostriction is believed
to reduce the main stabilizing force of the suspension and the production of
metal hydroxides assists in the flocculation and settling of the particles,
ACE Technology prompts agglomeration of charge neutralized particles in a
manner analogous to chemical polymerization resulting in an increase in the
mean particle size of the solids and thus, ease of filtration from the aqueous
phase. Each of these phenomena are discussed below.
Electroflocculation
The theory of electroflocculation or metal ion flocculation is very well
established. Iron and aluminum ions had been used for clarifying water as
early as 2000 B.C. Parekh et al(l) developed a coagulation model involving
use of metal hydroxides and fine particles. They reported that the optimum
coagulation of a metal ion-particle system takes place at the iso-electric
point of the metal hydroxide precipitate.
Electrof locculation causes an effect similar to that produced by the
addition of chemical coagulants such as aluminum or ferric sulfate. These
cationic salts destabilize colloidal suspensions by neutralizing negative
charges associated with these particles at neutral or alkaline pH. This, in
turn, enables the particles to come together in close enough proximity that
Van der Waals attractive forces prompt their aggregation and settling with the
neutral hydroxide floe. A report explaining electroflocculation chemistry was
prepared by James F. Grutsh(2) for the American Petroleum Institute Committee
on Environmental Control. Although the mechanism resembles chemical
coagulation in that cationic hydroxide species are produced from the energized
electrodes-, the characteristics of the electrocoagulated floe often differ
dramatically from those generated by chemical coagulation. They tend to be
less hydrous, more shear resistant, and are more readily filterable.
406
-------
Electrostriction
The theory of electrostriction is little known and is based primarily on
work done by Schwann.(3) Schwann examined the electrical properties of
colloidal suspensions in electrolytes subjected to alternating current
electric fields and discovered that colloidal suspensions exhibited variable,
and often extremely high, dielectric constants (E) in alternating fields. Not
only did the dielectric constant of the suspension decrease as the frequency
was increased, but the dielectric loss factor (E") went through a maximum, and
the conductivity (K) of the suspension increased dramatically (Figure 1). The
peak of the dielectric loss factor curve appears at the characteristics
frequency 'f * for the diagram, which is proportional to 1/R , where R is the
radius of spherical colloidal particles (for rod-shaped particles f o( 1/R ).
At the characteristic frequency, adsorbed ions, which give the colloidal
particle a surface charge, are least tightly held, and can move freely over
the particle surface in response to the electric field. For example, for
particles of 3 microns size f =60 Hertz has been found to be optimum.
Mixtures of particle sizes will broaden the 'f ' band considerably.
Theoretically, at a corresponding frequency the surface ion migration on a
given sized particle continually leads its restabilizing time forming a
continually shifting dipole (Figure 2). These polarized particles then can
theoretically agglomerate under the force of mutual electrostatic attraction,
rather than remain dispersed by the mutual electrostatic repulsion normal to
colloidal aqueous suspensions. The degree to which electrostriction functions
as a component of ACE Separator treatment is the subject of another ongoing
study at EPS.
Within the ACE Separator the particulates and metal ions will move
towards oppositely charged electrodes. Collisions are not only likely between
positive and negative particles moving in opposite directions, but also
between discharged particles of different sizes once the stabilizing charges
are eliminated. Consequently, loosely-held particle surface charges near the
characteristic frequency of the suspension could be neutralized and displaced
by oppositely charged electrophoretically mobile ions, eliminating or greatly
reducing the main stabilizing force of the suspension. Further collisions
between now uncharged and/or partially charged particles could then create
conglomerates which would eventually begin to move up or down in response to
the relative density of_the dispersing medium. The solution pH "will change
when highly mobile OH or H ions are used up to neutralize surface charges.
Co-precipitation is also known to reduce concentrations of other materials in
solution and suspension.
ACE SEPARATOR OPERATION
The basic process of the ACE Technology is illustrated in Figure 3. The
redistribution of charges and the onset of coagulation and flocculation occur
within the ACE Separator through exposure to an alternating electric field.
An ACE Separator contains electrodes in one of two design configurations: (1)
a series of vertically oriented, parallel plates or (2) a series of stacked,
cylindrical electrodes where a fluidized bed of metallic pellets is
maintained. Residence time in the ACE Separator is typically less than 60
seconds, after which the flow is transferred to a gravity separation tank or
to a filtration apparatus. Retention time in the ACE Separator may be reduced
407
-------
•
I
o
-------
Electrode Plates
Over Flow
Weir
Product Coagulation
and Separation
Control Panel
Alternating Current
Liquid
Raw Solution
or Slurry
Optional Air
For Turbulence
Figure 3. Schematic Diagram Of Ace Separator Unit
-------
in many applications by adding turbulence to agitate the solution as it passes
through the electric field. Turbulence is typically provided by bubbling air
from the base of the unit? treatment of the vent gas may be necessary when
volatile organics are involved due to air stripping effects. After
separation, each phase (water, oil and solid) is removed for reuse, recycle,
further treatment or disposal. The efficiency of electrostriction in removing
capillary or interparticle absorbed water, as well as adhesion water between
precipitated colloids, enhances the filtration characteristics of sludges.
Bench-scale tests and full-scale field applications performed by EPS as
another study have demonstrated a phenomenon referred to as residual
effectiveness whereby charge redistribution and coagulating forces appear to
remain effective for extended periods of time. This phenomenon is important
in that mixing and pumping of the solids or treated suspension can be
accommodated after coagulation, if so dictated by other system design
conditions, without losing the,phase separation effectiveness.
The parallel electrode ACE Separator operates on low voltage (20-40V) and
high current (50-500A) and for some applications, a frequency suited to the
characteristics of the waste stream. The fluidized bed ACE Separator in
contrast operates at a higher voltage (120-150V) and low amperage (0.5-5
amperes). ACE Separators of both designs operate at atmospheric pressure and
are vented to alleviate any problems associated with gas accumulation.
Solution characteristics such as particle size, conductivity, pH and chemical
constituent concentrations dictate operating parameters of the unit (for
example, electrode plate spacing, current density, frequency of power and
chemical pre-treatment, if required). Quantity and flow rate will affect
system sizing, retention time and mode of operation (recycle, batch,
continuous).
EXPERIMENTAL DESIGN
Experiments were conducted on two surrogate wastes prepared from the
EPA' s Synthetic Soil Matrix. Both were prepared to be stable aqueous
suspensions of silt, clay and top soil containing -1% suspended solids, with
or without spikes of toxic metals (Cd, Cr, Cu, Pb) or volatile organics (1,2-
dichloroethane, ethylbenzene, tetrachloroethylene, xylenes). The first
surrogate waste (Surrogate Waste A) contained solely the -40 mesh synthetic
soil matrix fines while the second (Surrogate Waste B) incorporated both the
fines fraction mixed with both 1.5% No. 2 diesel fuel and 1% of a strong
surfactant (Titan TX-100).
A preliminary series of experiments was undertaken to investigate the
effects of the five principal operating parameters of the ACE Separator:
electrode plate spacing (field strength), residence (or treatment) time,
applied current, current density and frequency of the applied current. These
studies indicated that the higher the electrical field strength, applied
current, retention time and current density (all of which correlate to higher
aluminum introduction into the solution), the more effective the phase
separation of the surrogate waste as judged by the clarity or suspended solids
loading (TSS) of the treated solution supernate. Electrocoagulation was found
to have a pronounced effect on the filtration rates of the slurries typically
410
-------
yielding increases of 60-9O%. The filter cake produced in all experiments was
generally quite compact, readily cracked and less voluminous than that
produced by comparative chemical treatments.
EXPERIMENTAL RESULTS
CLAY SUSPENSIONS (SURROGATE WASTE A)
Surrogate Waste A was spiked with four metal salts (Pb(NO ) , CuSO .5
HO, cdCl .2 1/2 HO, CrCl .6 HO) at concentrations of 10-50 mg/1, thoroughly
mixed and electrocoagulatea at operating conditions found to be optimum for
effecting separation of clays from the surrogate. While a majority of the
soluble metals - strongly adhere to the clays, electrocoagulation enables
agglomeration of the colloidal, metal-bearing clays and significant >90%
reductions in the soluble metals loadings (Table 1). The filtration time for
the treated surrogate decreased to.-50% of that required for the untreated
surrogate waste (3:02 minutes versus 6:00 minutes). Comparative chemical
coagulant addition experiments with alum (Al (SO ) ) and organic cationic
polyelectrolyte flqcculants (Drew Polymer 485) were also conducted on the
surrogate waste. Electrocoagulation yielded a faster filtration rate (2:15
minutes for 100 mis) than for either the untreated slurry (7:30 minutes) or
the alum-treated solution (3:53 minutes). Polymer treatment had the same
filtration rate. Filter cake volume expressed as a percentage of the pre-
filtered sludge volume seems to be a minimum for ACE Separator treatment (6%)
compared to alum and polymer treatment (13-15%). The volume data indicate
that the ACE Separator-treated solids cakes are more compact and easier to
dewater than those for coagulant-treated samples. Particle size analyses of
the treated and untreated slurries indicated that the mean size of the ACE
Separator-treated solids both in the supernate and filtrate (25.2 and 35.1
microns, respectively) increased by a factor of 3-4 over that in the original
slurry (9.1 microns). Larger particulate growth occurred as a result of
electrocoagulation than by either polymer or alum addition (15 and 10 microns,
respectively),
TABLE 1. SURROGATE WASTE A METALS-SPIKED•EXPERIMENTS
ANALYTICAL
PARAMETER
UNTREATED
SLURRY
. (mq/1)
ACE S E PARATOR TREATS D
SUPERNATE
(mq/1)
FILTRATE
(mq/1)
FILTER CAKE
(uq/q)
TOTAL SOLIDS' 17,000
.TOTAL SUSPENDED SOLIDS 22,000
TOTAL ORGANIC CARBON 220
TOTAL METALSs
CADMIUM 7.9
CHROMIUM 16
COPPER 44
LEAD 32
MOISTURE CONTENT N.D.
1,700
25
1.8
0.89
0.074
0.13
0.11
N.D.
1,700
3.0
26
2.2
0.018
0.053
1.7
N.D.
540,000
N.A.
N.D.
340
1,100
2,400
2,700
42.4
Spiking the surrogate waste with volatile organics at concentations of 5-
50 mg/1 was undertaken to examine the effect of electrocoagulation on
411
-------
volatiles extraction. Table 2 summarizes the results of one test.
Significant loss of volatiles to the vapor phase during electrocoagulation
appears to have occurred.
Comparison of the analytical data for supernate samples indicates
pronounced reductions achieved for electrocoagulation. Not documented,
TABLE 2. SURROGATE WASTE A VOLATILE ORGANICS-SPIKED EXPERIMENTS
ANALYTICAL UNTREATED
PARAMETER SLURRY
TOTAL SOLIDS (ug/g) 19,000
TOTAL SUSPENDED SOLIDS
(mg/1) 13,000
TOTAL ORGANIC CARBON
(mg/1) 150
TOTAL CARBON (mg/1) 465
ORGANIC CONTAMINANTS:
1,2-DICHLOROETHANE
(ug/1) 29,000
ETHYLBENZENE
(ug/1) 5,900
TETRACHLOROETHYLENE
(ug/1) 6,900
XYLENES (total)
(ug/1) 17,000
ACE SEPARATOR TREATED
SUPERNATE
1,400
41
2.4
7.8
14,000
1,700
1,300
6,300
FILTRATE
1,300
<2
3.2
27
2,500
590
390
1,900
FILTER CAKE
600,000
N.D.
N.D.
N.D.
1.8
<0.8
1.5
<5
however, was the loss of volatiles to the vapor phase as a result of stripping
by the compressed air introduced into the ACE Separator to create turbulent
conditions. Pronounced improvements in filtration rate for the ACE Separator-
treated slurries (-3:00 minutes per 100 mis) compared to the untreated stock
solution (>20:00 minutes per 100 mis) were again documented.
DIESEL FUEL CONTAMINATED SLURRY (SURROGATE WASTE B)
Electrocoagulation of the surrogate waste containing 1.5% diesel fuel
produced effective reductions in suspended solids (112 to 12 mg/1), total
carbon (230 to 110 mg/1) and total inorganic carbon (28 to 12 mg/1).
Electrocoagulation of Surrogate Waste B spiked with metals (Cu, Cd, Cr, Pb)
had respectable reductions (Table 3). For example, copper is reduced by 90-
94*, cadmium and chromium by 91-97% and lead by 86-89%. No appreciable change
in TOC loadings in the supernate resulted from treatment; the TSS was reduced
by approximately 90% from 222 to 19 mg/1. Comparative chemical coagulant
addition experiments were conducted for the diesel fuel contaminated slurry.
The following generalizations can be made for the treatment: alum and polymer
treatments generally require approximately 30% longer filtration times, ACE
Separator and polymer treatments reduce the TS and TSS loadings to an
equivalent degree, which is about 25% of the alum value, and better reductions
in soluble metal concentrations were usually achieved with polymer and
electrocoagulation treatment. Particle size data confirm the appreciable
enhancement in the clay fraction as a result of electrocoagulation (Table 4).
412
-------
The mean size of the ACE Separator—treated particulates both in the supernate
and filtrate (188 and 20 microns, respectively) has increased by a factor of
approximately 85 and 8, respectively, over that in the original slurry (2.2
microns).
TABLE 3. SURROGATE WASTE B METALS-SPIKED EXPERIMENTS
UNTREATED
ANALYTICAL SLURRY
PARAMETER (mq/1)
TOTAL SOLIDS 1,870
TOTAL SUSPENDED SOLIDS 222
TOTAL INORGANIC CARBON 15
TOTAL ORGANIC CARBON 130
TOTAL CARBON 150
TOTAL METALSt
CADMIUM 0.5
CHROMIUM 0.31
COPPER 0.30
LEAD 0.72
ACE SEPARATOR TREATED
SUPERNATE
(ma/1)
1,480
4.5
7.8
6.6
20
0.15
0.024
0.085
0.09
FILTRATE
(ma/1)
N.D.
N.D.
N.D.
N.D.
1,300
0.28
0.01
0.44
0.16
FILTER CAKE
(uq/a)
1,000,000
N.D.
N.D.
N.D.
N.D.
289
721
650
3,700
TABLE 4. DIESEL FUEL CONTAMINATED SURROGATE PARTICULATES SIZE DATA
TREATMENT METHOD
D(50%) SUPERNATE(microns)
Untreated slurry
Polymer 425
Alum
ACE Separator
21.97
7.91
9.32
188.6
Analytical results for electrocoagulation of organic volatiles-spiked
Surrogate Waste B slurries are similar to those for the clay suspension tests:
marked improvement in filtration time (10:30 versus 31:00 minutes for 100 mis)
and sizeable reductions in supernate TOC (91%), TSS (76-93%) and TC (86%)
loadings. While electrocoagulation did appear to reduce the concentrations of
the volatile organics in the supernate phase (Table 5), filtration of the
reslurried supernate shears the solids to release many of the volatiles back
into the aqueous phase.
With respect to the removal of the volatile organics, the concentrations
of all four spiked species in the treated supernates are all lower than in the
treated slurry. In contrast, the concentrations of these spikes in the
filtrates were nearly all higher than in the untreated slurries.
SUMMARY OF SURROGATE TREATMENT RESULTS
ACE Separator treatment is effective in removing particulates from the
soil suspensions (TS, TSS), in increasing their mean particle size and,
consequently, in improving their filtration properties (speed of filtration).
The technology has proven effective in reducing the metal concentrations in
such slurries both by removing the clays to which metal ions have adsorbed and
by coprecipitating soluble species. -Reducing the loadings of TC, TIC, and TOC
413
-------
attributable to fine particulates can be readily achieved; volatile organics
are removed either by air stripping or by flocculation of the clays to which
the organics have adhered. For volatile organics removal electrocoagulation
i» not a preferred treatment method.
TABLE S. SURROGATE WASTE B VOLATILE ORGANICS-SPIKSD EXPERIMENTS
ANALYTICAL UNTREATED ACE SEPARATOR TREATED
PARAMETER SLURRY SUPERNATE FILTRATE FILTER CAKE
TOTAL SOLIDS (ug/g) 10,900 867 N.D. 462,000
TOTAL SUSPENDED SOLIDS
(mg/1) 5,050 47.5 97.3 N.A.
TOTAL ORGANIC CARBON
(mg/13 14,500 86.4 540 N.D.
TOTAL CARBON (mg/1) 11,700 220 4,450 N.D.
ORGANIC CONTAMINANTS:
1,2-DICHLOROETHANE
(ug/1) 4,900 1,500 3,800 110
ETHYLBENZENE
(ug/1) 3,500 77 16,000 190
TETRACHLOROETHYLENE
(ug/1) 3,900 120 15,000 170
XYLENES (total)
(ug/1) 13,000 130 62,000 680
COMPLEMENTARY STUDIES
Several laboratory-scale amenability tests of industrial effluents have
indicated that ACE Separator treatment is effective in removing certain
metals, phosphate, fluoride, suspended solids and BOD from process waters.
PHOSPHATE REMOVAL
Three experiments on synthetic laboratory solutions and actual industrial
wastewaters have confirmed the feasibility of phosphate removal by
electrocoagulation. None of these experiments were optimized and so the
extent to which phosphate could be reduced remains unknown. Treatment of
wastewater from a commercial laundry reduced the phosphate concentration from
45 mg/1 to 5.4 mg/1 after minimal treatment (1.25 L, 0:45 min, 0.36 KW).
Electrocoagulation of process water from a Florida phosphate mining operation
reduced the PO level by 91% from 160 to 14 mg/1, once again with brief
treatment (1.5 L, 0:10 min, 3.3 KW) . Finally, treatment of a dilute
phosphoric acid solution resulted in a pronounced conductivity decrease and pH
increase suggesting phosphate removal by means of aluminum phosphate
precipitation. Electrocoagulation should constitute an efficient method to
remove orthophosphates from aqueous media; reduction of pyrophosphates and
polyphosphates from wastewater may not, however, be as easily achieved.
414
-------
COD/BOD REDUCTION
Treatment of production vat wash waters from a manufacturer of imitation
dairy pro.ducts containing 0.1-1.0% total solids reduced the BOD loading from
3,530 mg/1 to 343 mg/1, a 90% reduction. The filtration time for a 100 ml
sample decreased from 15 minutes for an untreated sample to 0:50 minutes after
ACE Separator treatment (7-8 amperes, 4 minute retention time), a 92%
reduction.
SYNTHETIC POLLUTANT METAL SOLUTIONS
Four synthetic solutions containing nickel, chromium (III), zinc and
copper were prepared for electrocoagulation under a variety of pH, power and
operating parameters. ACE separator treatment at pHs typical of those for
process wastewaters can routinely lead to >90% reduction in loadings of these
four metals (Table 6).
TABLE 6. SYNTHETIC EFFLUENT STUDY METALS REDUCTIONS
EFFLUENT pH
(units)
COPPER
6
8
10
CHROMIUM
6
8
10
NICKEL
6
8
10
ZINC
6
8
10
APPLIED
POWER
(KVA)
2.0
0.7
2.4
2.6
2.3
1.4
2.0
0.7
2.5
2.2
2.4
0.2
UNTREATED
FILTRATE
CONCENTRATION
(ma/I)
127
110
17
89
30
7.5
127
110
17
138
1.5
6.6
ACE SEPARATOR
FILTRATE
CONCENTRATION
(mq/11
3.4
0.85
1.1
11
2.4
0.82
3.4
0.85
1.1
30
0.55
0.56
METAL
REDUCTION
97
99
93
88
92
89
97
66
94
78
64
92
REFERENCES
1. Parekh, B. and Apian, F. Flocculation of fine particulates and metal
ions. Paper presented at the 1988 Annual Meeting of the American
Institute of Mining Engineers, Phoenix, Arizona.
2. Grutsch, J. The Chemistry and Chemicals of Coagulation and
Flocculation. In; Report-of the American Petroleum Institute
Committee on Refinery Environmental Control. 168 pp
3. Schwann, -H. On the low frequency dialectric dispersion of colloidal
particulates in electrolyte solutions. J. Phy. Chem 66:2626, 1962
415
-------
ENERGY ELECTRON BEAM IRRADIATION: AN EMERGING
TECHNOLOGY FOR THE REMOVAL OF HAZARDOUS ORGANIC CHEMICALS
FROM WATER AND SLUDGE - AN INTRODUCTION
William J. Cooper1
Thomas D.Waite2
Charles N. Kurucz3
Michael G. Nickelsen1
David A. Meacham1
1Drinking Water Research Center
Florida International University
Miami, FL 33199
Department of Civil and Architectural Engineering
3Department of Management Science and Industrial Engineering
University of Miami
Coral Gables, FL 33124
ABSTRACT
When high energy electrons impact an aqueous solution reactive transient species are
formed. The three transient species of most interest are the aqueous electron, e"aq, the
hydrogen radical, H-, and the hydroxyl radical, OH-. This paper describes preliminary
research conducted at the Electron Beam Research Facility (EBRF) in Miami, FL to
determine the removal efficiency for four organic chemicals, chloroform, trichloroethylene,
tetrachloroethylene, benzene and toluene, using this innovative treatment process. The
accelerator is a 1.5 MeV, 50 mA insulated core transformer type. The effect of water
quality on the destruction of the organic compounds is examined using potable water,
secondary wastewater and a secondary anaerobically digested sludge (2-5 % solids). Batch
experiments can be conducted using 6,000 gallon tank trucks. The nominal treatment flow
rate is 120 gallons min'1. The absorbed dose can be varied from 0 to 800 krads.
The work described in this paper was not funded by the U.S. Environmental
Protection Agency and therefore the contents do not necessarily reflect the views of the
Agency and no official endorsement should be inferred.
416
-------
INTRODUCTION
The ultimate disposal of toxic and hazardous organic chemicals is emerging as a
priority in the search for new and innovative treatment technologies. By ultimate disposal
we refer to the mineralization of the solutes of concern. Historically, for waste streams and
even for site remediation, treatment process efficiency focused on the removal of the solute
of interest. Little or no concern was voiced once the parent compound was "out of sight."
In some cases the removal of the solute or waste product was considered complete by
merely removing it from the manufacturing or remediation site to a landfill or by using deep
well injection. These options are less attractive when considering the long term
environmental effects and potential liability to the owner of the "disposed" waste. The "out
of sight-out of mind" mentality is being called into question and it is probably safe to say will
disappear in the future.
An extension of this approach is the use of carbon and aeration stripping, where the
chemical(s) of interest are transferred to another media. In the case of carbon the solutes
are concentrated and then are disposed of during the regeneration. If the carbon is not
regenerated it then must be disposed of either in a landfill or by incineration. Aeration
stripping for the removal of volatile chemicals, the cheapest alternative when using
extremely naive and simplistic economic analyses, at worst transfers the problem to the air
and at best transfers it to carbon or another adsorbent.
We feel that the more realistic solution to the problem of the disposal of toxic and
hazardous organic waste chemicals will be treatment processes that result in, or facilitate,
the mineralization of the chemicals. Probably the best known process to achieve this is the
use of ozone, O3, most often in the presence of various catalysts for its decomposition, e.g.
ultraviolet (UV) light and/or hydrogen peroxide, H2O2. Other chemical/physical processes
that are receiving attention are supercritical oxidation and wet oxidation. Bioremediation
can also be considered an ultimate disposal process. Incineration of wastes has certain
demonstrated advantages, but also a high potential for the formation of reaction by-products
that may be as bad or worst than the starting materials. Thus, this technology is now under
critical review when suggested as the process of choice.
This paper describes an innovative treatment process for the ultimate disposal of
toxic and hazardous organic chemicals in aqueous solutions. The underlying chemistry is
reviewed to acquaint the reader with the process and the technology of using high energy
electron irradiation. Experimental results are presented for five compounds of interest in
site remediation.
ELECTRON BEAM TECHNOLOGY
Electron beams have been in commercial use since the 1950s. Early applications
involved the cross-linking of polyethylene film and wire insulation. The number of
applications has since grown to include sterilization of medical supplies, rubber
vulcanization, disinfection of wastewater, food preservation, curing of coatings, etc. Today
417
-------
there are several hundred electron processing systems installed for industrial applications
in over 25 countries.
Polymerization of cable insulation and cross-linking of plastic film still account for
the bulk of the applications. More than half of the total installed world capacity of 15 MW
of electron beam power is devoted to these applications while less than 1 MW is used for
sterilization of medical products. Only a small amount of the installed capacity is used for
biological disinfection and detoxification (1). The reasons for the relatively slow growth in
the latter applications can only be partially explained by need and economic considerations,
and yet these applications hold the potential for the most social good (2).
Electron beam processing involves exposing the material to be irradiated to a stream
of high energy (fast) electrons. These electrons interact with the material in less than 10"12
seconds to produce electrons of lower and lower energy. Eventually a large number of slow
electrons with energies less than 50 eV is produced, and these electrons interact with
molecules to produce excited states of these molecules, positive ions and electrons.
Eventually the electrons slow to thermal energies and get trapped. In materials of low
dielectric constant most electrons do not escape the pull of the positive ions formed when
they were produced. The electrons are attracted back to the positive ions causing a chemical
reaction. This is termed direct radiolysis. In high dielectric materials such as water and
aqueous solutions, most electrons escape the pull thus leaving both the positive ions and
electrons free to react with the water or waste components in it. This is referred to as
indirect radiolysis. The ratio of direct to indirect radiolysis in wastewater is approximately
the weight fractions of waste to water (3). The radiation chemistry of aqueous solutions is
presented in more detail in another section of this paper.
ELECTRON BEAM RESEARCH FACILITY
The Electron Beam Research Facility (EBRF) is located at the Miami-Dade Central
District Wastewater Treatment Plant located on Virginia Key, Miami, Florida. Schematic
diagrams of the EBRF are shown in Figures 1 and 2. The facility consists of a horizontal
1.5 MeV electron accelerator. The accelerator is an insulated-core transformer (ICT) type,
capable of delivering up to 60 mA beam current. However, we usually use 50 mA as the
upper limit in beam current. Varying the beam current changes the absorbed dose in a
linear fashion, allowing for experimentation at doses from 0 to 800 krads. The electron
beam is scanned at 200 Hz to give a coverage of 48" wide and 2" high.
Influent streams at the EBRF are presented to the scanned beam in a falling stream
approximately 48" wide and at the design flow of 120 gpm is 0.15" thick. Since the
maximum penetration in water is approximately 0.29" for 1.5 MeV electrons, some electrons
pass through the stream and thus not all of the beam energy is transferred to the water.
With the addition of overscanning the waste stream to insure that the edges of the stream
are irradiated, more energy is lost with the result that the efficiency of energy transfer is
approximately 60 - 85% (see section on electron utilization efficiency). Thus when the
418
-------
Figure 1. Schematic diagram of electron beam research facility, ground view.
'4 s x. n
. &» / X
Figure 2. Schematic diagram of electron beam research facility, top view.
419
-------
electron beam is operating at 50 mA (75 kW) the waste stream is receiving an average dose
of approximately 650 krads. Total power consumption, including pumps, chillers and other
auxiliary equipment is about 120 kW.
Because the system is being used for research, and water quality is one of the main
experimental variables, three influent streams are directly connected to the facility. These
three influent streams are potable water, chlorinated secondary wastewater and secondary
anaerobically digested sludge that is 2-5% in solids. Batch experiments can be run at the
facility utilizing a 6000 gallon tank truck connected to the influent pump. Experiments have
been conducted using raw wastewater collected and transported in the tank trucks.
The EBRF is instrumented with resistance temperature devices (RTDs) to obtain
direct estimates of absorbed dose. Five RTDs are mounted in the influent (2 sensors) and
effluent (3 sensors) streams immediately before and after irradiation. The RTDs are
connected via an interface to a computer which continuously reads and records temperatures
and the absorbed dose is estimated by converting the observed temperature differences to
the energy transferred to the water. The average absorbed dose (DAV) in pure water is
calculated using the equation:
DAV = Kfc-tj) [1]
where, ^ and t^ are the before and after irradiation water temperature of the flowing stream
in °C, respectively; and K is the constant of proportionality:
K = 418 krads ''C1, [2]
Therefore, an increase 1°C in water temperature is equivalent to a dose of 418 krads in pure
water.
During experiments to determine the removal efficiency of parent compounds and
to collect samples to determine reaction by-products, samples are taken prior to and after
irradiation. These samples are obtained in the control room from continuously running
sample streams.
AQUEOUS CHEMISTRY OF HIGH-ENERGY ELECTRONS
The purpose of this section is to provide a brief overview of aqueous-based radiation
chemistry. This brief introduction should assist the reader in understanding the application
of high energy electron irradiation to the treatment of toxic and hazardous organic wastes
in water and sludge.
The studies reported in the literature relating to radiation chemistry have been
conducted in pure aqueous solutions. The extrapolation of pure water data to natural
waters is complicated by the presence of inorganic and organic matter (primarily humic
substances) found in natural waters. These compounds may interact with the transient
reactive species formed during irradiation and lead to side reactions not observed in pure
water. An example is the reaction of hydroxyl radical with carbonate ions, similar to that
observed in processes utilizing O3.
420
-------
High-energy electron irradiation of pure water results in the formation of
electronically excited states and/or free radicals along the path of the electron. 10"9 sec
after the electron has passed through a solution the products that are present are shown in
Equation [13] (4-8);
H20 -V\A- [2-7] OH- + [2.6] e-aq + [0.6] H-
+ [0.7] H2O2 +[2.6] H3O+ + [0.45] H2 [3]
Unlike photochemical reactions where one photon of light initiates one (molecular)
reaction, a high energy electron is capable of initiating several thousand reactions as it
dissipates its energy. The efficiency of conversion of a high energy electron, ionizing
radiation, to a chemical process is defined as G (shown in brackets in Equation 13). G is
the number of radicals, excited states or other products, formed or lost in a system
absorbing 100 eV of energy. Of the products formed in Equation [13], the most reactive are
the oxidizing, hydroxyl radical (OH-), and the reducing, aqueous electron (e"aq) and
hydrogen radical (H-). Thus, the chemistry of primary interest in the high energy electron
irradiation process is that of these three species.
CONCENTRATION OF REACTIVE SPECIES
One aspect of the research on high energy electron irradiation of toxic organic
chemicals is to develop an understanding of the underlying chemistry. This understanding
would be helpful in predicting, a priori, the success of the process for various applications.
It is possible, using absorbed dose and G values (equation 13), to determine the
approximate concentration of the transient reactive species in irradiated aqueous solutions.
For a chemical with a G = 1, and a dose of 1 Mrad the concentration of the reactive
species is 1.04 x 10'3 mole L'1 or 1.04 mM (16).
Thus, for a G value of 2.7, e.g., OH- from equation 1, and an absorbed dose of 1
Mrad, there are 2.81 mmol of OH- formed. Typical concentrations of these transient
reactive species and H2O2, at several doses, are summarized in Table I.
TABLE L ESTIMATED CONCENTRATION OF TRANSIENT REACTIVE SPECIES
AT SEVERAL DOSES USING HIGH ENERGY ELECTRON IRRADIATION
DOSE (krads) e' H- OH H2O2
100
500
1000
0.27
1.4
2.7
mM
0.06
0.3
0.6
0.28
1.4
2.8
0.07
0.4
0.7
421
-------
RADICAL SCAVENGER EFFECTS
An important consideration in extending laboratory data to natural waters is the
effect of naturally occurring radical scavengers. In this section we will present an overview
of some of the scavengers for which rate data exist. A limitation is the lack of rate
constants for inorganic constituents commonly found in natural waters.
OXYGEN
Both e~ag and H- rapidly reduce O, to form O2" (pIL, = 4.8) with second order rate
constants of 1.9 x 1010 and 2.1 x 1010 M'^s"1, respectively (12). Using a dissolved oxygen
concentration of 3.7 mg L"1 (0.12 mM). At a dose of 100 krads this O2 concentration would
remove approximately 35% of the two reactive species. However, at 800 krads only 5%
would be removed.
BICARBONATE/CARBONATE ION
A common OH- scavenger in natural waters is alkalinity. Alkalinity, the measure
of the total carbonate concentration is further complicated by the equilibrium that exist in
natural waters.
H2CO3 + H2O -----> H3O+ + HC03- [4]
HCO3- + H2O > H3O+ + CO32- [5]
Based on the pH and equilibrium calculations the carbonate/bicarbonate ion
distribution is quite different in the waters used for study at the EBRF. For example, the
secondary wastewater has a pH = 7.0 whereas the potable water is approximately 9.0. The
relative effects of these ions on radical scavenging can be calculated from the equations:
OH- + HCOa1' ' H2O + CO/'- [6]
OH- + CO32' OH' + CO/'- [7]
and, the second order reaction rate constants. The second order rate constants are 8.5 x 106
M"Vl and 3.9 x 108 M'V1, respectively (9).
More specifically, the two water sources at the EBRF were used to calculate the
relative effects of alkalinity on OH- concentration. The alkalinity of the secondary
wastewater, pH = 6.78, and the potable water, pH = 8.64, in one set of experiments was
1900 and 360 n M, respectively. Although the alkalinity of the secondary wastewater was 5-
fold higher than the potable water, the OH- scavenging is approximately 2.5 times higher
in the potable water based on bicarbonate/carbonate equilibria. This relationship was
determined by summing the product of the concentration of the carbonate and bicarbonate
ions with OH- reaction rate constants in each water.
422
-------
The reaction of the carbonate radical ion, CO3"-, is not known for many solutes, and
therefore this process is considered as a loss of reactive species with no removal of solute.
It is possible that the removal efficiency would be effected by the presence of this radical,
and further studies are necessary to determine the effects.
NITRATE ION
When comparing the removal of solutes in potable water and secondary wastewater,
the presence of higher concentrations of nitrate ion (NO/') in the secondary wastewater
may effected solute removal efficiency by acting as an e~(ag) scavenger. By efficiently
scavenging e"^ the effective concentration of the OH* might be increased (minimizing
recombination of the e'^ and OH-, thereby increasing the effective OH- concentration)
and removal enhanced. No information is available on the reactions of the nitrate radical
(NO3-) with the solutes in this study.
DISSOLVED ORGANIC CARBON
Another common component in natural waters is the ill-define fraction referred to
as dissolved organic carbon, DOC. The difference in DOC concentration in secondary
wastewater (6-fold higher than the potable water we used for our studies) may have also
reduced removal efficiency by acting as an OH- radical scavenger. There are no data on
the reactions of either e" or H- with DOC and only indirect evidence of the reaction with
OH.
METHANOL
An experimental artifact for much of the data obtained at the EBRF results from the
need to use methanol as a carrier of the organic solutes of interest. Based on the ratio of
the injection pump flow rate to the aqueous stream flow rate, the concentration of methanol
for all experiments was approximately 3.3 mM. Methanol reacts with OH-, and to a lesser
extent with H-, in aqueous solution with second order reaction rate constants of 9.7 x 108
and 2.6 x 106 M~*s , respectively (9). Therefore, under the experimental conditions
employed, the removal efficiency of the solutes is probably underestimated relative to
aqueous solution of the solutes in the absence of methanol, due to OH- scavenging.
EXPERIMENTAL RESULTS FOR THE REMOVAL OF TOXIC AND HAZARDOUS
ORGANIC CHEMICALS IN AQUEOUS SOLUTION
We have conducted numerous experiments on organic chemicals that may be of
interest in: water treatment, trihalomethanes; groundwater contamination, halogenated
ethanes and ethenes; leaking underground storage tanks, benzene and substituted benzenes;
as well as other organic chemicals now regulated as hazardous wastes. Results for selected
compounds will be reviewed below. The six compounds are representative of the major
groups mentioned above.
423
-------
TRfflALOMETHANES
One of the initial chemical groups studied at our facility was the halogenated
methanes. Others have reported studies using electron and gamma irradiation of aqueous
solutions of chloroform (10,11). We have studied CHC13 and CHBr3 and observed removal
efficiencies of CHC13 of approximately 85 - 99.9% in secondary and raw wastewater, and
potable water, respectively.
The removal of CHC13 is effected by water quality. That is, in potable water we
observed >99% removal at 800 krads and initial concentrations of between 75 - 750 ng L"1
(Figure 3). Similar initial concentrations in secondary and raw wastewater resulted in the
removal of approximately 85 - 90% (Figures 4 and 5). To achieve higher removal
efficiencies either higher doses would be required or, the water would have to be
recirculated a second time.
The removal efficiency of CHBr3 in raw wastewater and potable water is not affected
by the change in water quality and is independent of concentration in the range of 100 -
1500 ug Lr1 and above 100 krads is > 99.99% (16).
The reaction by-products have not been studied in detail as yet. However, in studies
conducted at low solute concentrations, none of the halogenated reaction by-products shown
in the mechanism equations 45 - 60 have been observed. The liquid-liquid extraction
method used for the quantification of the CHCl^ would also have determined the presence
of the chlorinated ethanes at detection limits of 0.01 \iUl. Research is now underway to
determine the reaction by-products, aldehydes and carboxylic acids.
HALOGENATED ETHENES
Another group of organic chemicals that have been studied at our treatment facility
are the halogenated solvents. The compounds most commonly found are trichloroethylene
(TCE) and tetrachloroethylene (PCE) (12). We have shown that the removal efficiency of
both compounds.in potable water, is >99.9%, at doses of approximately 800 krads. The
removal of TCE and PCE from secondary wastewater is 90 - 95% at 800 krads. Thus, as
we observed for CHC13, water quality does affect removal efficiency of these two
compounds.
Figures 6 and 7 show results of the removal of these two compounds at several doses
in potable water and secondary wastewater. No studies have been conductd to examine the
variables of solute concentration. However, we expect that results similar to those obtained
for CHC13 will be obtained when the experiments are completed. Although no studies have
been conducted to examine the reaction by-products, the analyses of TCE and PCE involve
liquid-liquid extraction, and no halogenated reaction products have been observed during
the chromatographic determinations.
424
-------
1.00
Figure 3.
D.7S -
K
O
g o.so
x
u
I
s
0.29 -
0.00
INFLUENT
EFFLUENT
• too
•75
-50
•28
100 200 300 400 500 800
ABSORBED DOSE (kradi)
700
BOO
* INFLUENT
* EFFLUENT - 900
•000
•sooi'
«h*
-400
•300
•200
• 100
0
100 200 300 400 SOO 600
ABSORBED DOSE (krod«)
700
BOO
Removal of chloroform at several irradiation doses in potable water at two
initial concentrations (error bars indicate one standard deviation from mean,
where no error bars are seen the error is within the data point).
425
-------
1.00
0.78 -
K
o
0.30
x
u
0.23 -
0.00
* INFLUENT
* EFFLUENT
100
200
—r_
300
—I—
400
BOO
ABSORBED DOSE (krodt)
• I""
eoo
-100
•78
-BO
-28
700
BOO
100
200
300 400 800
ABSORBED DOSE (krod»)
eoo
700
BOO
Figure 4. Removal of chloroform at several irradiation doses in secondary wastewater
at two initial concentrations (error bars indicate one standard deviation from
mean, where no error bars are seen the error is within the data point).
426
-------
1.00
• INFLUENT
• EFFLUENT
0.00
• 100
•78
-80
•28
100 200 300 400 500 800 700 800
ABSORBED DOSE (krodi)
7-
8 -
a a-
K
s
O
o
I
* INFLUENT
» EFFLUENT •
•00
800
•700
•800
100
—,—
200
•400
-300
•200
-100
300 400 800 600
ABSORBED DOSE (krad.)
700
800
Figure 5, Removal of chloroform at several irradiation doses in raw wastewater at two
initial concentrations (error bars indicate one standard deviation from mean,
where no error bars are seen the error is within the data point).
427
-------
POTABLE WATER
1500
INFLUENT
• EFFLUENT '
••150
••100
.-50
100 200 300 400 500 600 700 BOO
ABSORBED DOSE (krads)
SECONDARY WASTEWATER
1500
1-150
••100
o e INFLUENT
• . EFFLUENT t
.-50
100 200 300 400 500 600 700 600
ABSORBED DOSE (krads)
Figure 6. Removal of TCE at several irradiation doses in potable and secondary
wastewater (error bars indicate one standard deviation from mean, where no
error bars are seen the error is within the data point).
428
-------
POTABLE WATER
1000
750--
500 n
250--
150
100
-50
100 200 300 400 500 600 700 800
ABSORBED DOSE (krads)
SECONDARY WASTEWATER
1000
750-•
500-
250-
•150
'100
o—o INFLUENT
,—. EFFLUENT
100 200 300 400 500 600 700 800
ABSORBED DOSE (krads)
Figure 7.
Removal of PCE at several irradiation doses in potable and secondary
wastewater (error bars indicate one standard deviation from mean, where no
error bars are seen the error is within the data point).
429
-------
BENZENE AND SUBSTITUTED AROMATIC COMPOUNDS
A third group of compounds which we have studied are benzene and substituted
benzenes (13, 14).
Figure 8 shows the effective removal of benzene and toluene from potable water.
Three experimental factors affected removal efficiency: water quality, solute concentration,
and dose. Both compounds were removed to below detection limits in potable water when
added at the lower initial concentration. However, in the secondary wastewater, at the
lower solute concentration, 90 - 96% removal was observed for all four solutes at 787 krads.
Concentration effects were also examined by irradiating mixtures with solute
concentrations approximately 20 times higher in concentration. For example at 0.96 \i M (75
jig L"1) removal of benzene to below detection limits (0.01 jig L"1) was observed in potable
water while at 17.5 M. M (1370 jig L"1) the benzene was reduced by 93%, at an absorbed dose
of 787 krads. Similar results were observed for toluene in both potable water and secondary
wastewater.
The G value for solute removal at a given dose, Gd, is defined by the disappearance
of the solute in aqueous solutions, and is determined experimentally using the following
equation:
([org]d)(6.02 x 1023) [8]
(d)(6.24 x 1017)
where, [org]d is the change in organic solute concentration in mol L"1 at a given dose, d is
the dose in krads, 6.24 x 1017 is the constant to convert krads to molecules L"1, and 6.02 x
1023 is Avogadro's number.
The Gd values at all doses for potable water were very similar to those observed in
the secondary wastewater at the low solute concentration (Table n). At the high solute
concentration, the Gd values in potable water were higher when compared to secondary
wastewater, only at the lowest dose. At the two higher doses the Gd values were very
similar in both waters, with the exception of benzene where the Gd value appeared to be
higher in potable water. At the higher solute concentrations the observed Gd values were
an order of magnitude higher than those observed for low solute concentrations. This
observation is consistent with removal being first order in solute concentration.
A plot of In [solute] against dose for both compounds was linear in potable water and
secondary wastewater. Table VII summarizes the observed dose constants and half-dose
(cli/a) for benzene and toluene in potable water and secondary wastewater at two solute
concentrations.
430
-------
iZOU-
1000^
*S 750-
yj
z
LjJ
Z 50°"
LiJ
m
250-
0-
(
(A) 0 0 INFLUENT
• • EFFLUENT
3 100 200 300 400 500 600 700
ABSORBED DOSE (krads)
*•
,
-O
•
-•
8C
-90
-BO
-70
•50
40
•30
•20
-10
•0
)0
O O INFLUENT
« • EFFLUENT '
0
0 100 200 300 400 500 600 700 800
ABSORBED DOSE (krads)
Figure 8. Removal of benzene (a) and toluene (b), at various absorbed radiation doses,
from potable water.
431
-------
TABLE H. SUMMARY OF SOLUTE REMOVEC AND Gd FOR POTABLE WATER
AND SECONDARY WASTEWATER AT TWO SOLUTE CONCENTRATIONS
Potable Low Solute Concentration High Solute Concentration
Water Dose8 [OS]db Gd Dose" [OS]db Gd
(103) (uM) (103)
Benzene
155
469
797
0.74
0.92
>0.9
4.6
1.9
_
144
469
794
8.6
14
16
57.8
28.1
19.0
Toluene 155 0.88 5.5 144 10 68.0
469 1.1 2.3 469 13 27.1
797 >1.2 - 794 14 17.0
Secondary
Wastewater
Benzene 144 0.73 4.9 144 5.1 33.9
465 0.88 1.8 469 9.0 18.7
779 0.94 1.2 794 12 14.7
Toluene 144 0.78 5.2 144 7.8 51.8
465 1.2 2.5 469 13 27.3
779 1.3 1.6 794 14 17.5
"Units = krads.
b[OS],| = organic solute concentration removed at an absorbed dose d.
The dose required to remove half of the initial benzene and toluene concentration
in potable water was approximately 1.5-fold lower than the dose required to remove a
similar concentration in secondary wastewater, at low solute concentration. The dose
constant for benzene in secondary wastewater was higher when compared to potable water,
while the dose constant for toluene was similar in both waters. The significance of the
differences in d1/2 is not known at this time.
For the two aromatic compounds studied the OH- was the radical most responsible
for the removal of the individual solute. In the case of toluene the H- could account for
up to 16% of the removal. This may be significant in explaining the increased removal of
toluene relative to benzene. In a natural water the observed removal efficiencies will vary
because of the presence of natural radical scavengers and thus it is difficult to predict a
priori the removal efficiency of the various organic solutes.
432
-------
TABLE III. SUMMARY OF OBSERVED PSEUDO-FIRST-ORDER HIGH-ENERGY
ELECTRON IRRADIATION DOSE CONSTANTS
Potable
Water
Benzene
Toluene
Initial
Concentration
(u.M)
0.934
17.0
1.15
16.4
Observed jcei
(kracT1)
8,40 x 10'3
3.05 x 10'3
6.09 x 10-3
2.22 x 10'3
dl/2
(krad)
82.5
227
114
312
r2
0.99
0.99
0.99
0.86
Secondaiy
Wastewater
Benzene
Toluene
1.04
16.2
1.35
17.1
2.63 x 10-3
1.60 x 10"3
3.75 x 10'3
2.18 x 10'3
264
433
185
318
0.86
0.99
0.95
0.92
Several reaction by-products have been identified in irradiated solutions of benzene
and toluene. Our results are consistent with reported findings showing that phenols are an
initial product in the decomposition of the compounds studied. Figure 3 shows the results
of total phenol analysis, for the irradiation of an aqueous mixture of all compounds, at
various doses in secondary wastewater. These results support the reported studies in that
at low doses the effluent phenol concentration increases and then falls below the influent
(or background) concentration at high doses. In addition to phenol, more highly oxidized
species have also been observed. Glyoxal, at sub-jiM concentrations, was identified
qualitatively in this study. A control experiment in which only secondary wastewater (no
added organic solutes) was irradiated showed no formation of glyoxal. Several other
aldehydes were observed but the structures of these compounds have not yet been
determined.
CONCLUSIONS
In conclusion, we have shown that high energy electron beam irradiation is effective
in removing organic chemicals from aqueous streams. The examples shown are typical of
organic chemicals found at Superfund sites. The results indicate that this process is an
ultimate treatment process for the removal of toxic and hazardous organic chemicals from
aqueous solutions.
Additional studies are in progress to irradiate aqueous solutions of the solutes, in
similar matrices, in the absence of methanol. These results will provide better estimates of
the removal efficiencies likely to be encountered in contaminated groundwaters, and it will
enable a better characterization of reaction by-products.
433
-------
In summary, several of the features of this process that result in its unique ability as
a treatment process are that solutions that contain up to 10% solids (sludges) can be
irradiated and organic compounds destroyed. And, secondly the presence of both e"(aq) and
OH- in aqueous solution, at similar steady state concentrations, is unique to this process
and distinguishes it from other advanced oxidation processes (15). A third feature is the
presence of significant concentrations of H- that may react with solutes of interest and
increase the efficiency of removal of some toxic organic compounds.
TECHNOLOGY TREATMENT COST ESTIMATES
The cost of treatment using the electron beam technology depends on many factors
such as the dose required to obtain the desired detoxification, the volume of waste to be
treated, the size of the treatment facility, etc. For example, the cost of permanent facilities
such as the one in Miami will range from $600K to S2.500K for 300 KeV to 3 MeV systems.
Assuming the $1,750K Miami facility at 120 gpm and a total absorbed dose of approximately
650 krads is adequate for a specific treatment application, treatment costs would be
approximately $8 per 1000 gallons if capital costs are amortized at 10% over 20 years and
the facility is operated approximately 8,000 hours per year. Transportable units will
undoubtably be developed in the near future. The costs of operating such units will probably
be higher due to transportation costs and increased maintenance requirements.
ACKNOWLEDGEMENTS
All of the experimental results reported in this paper were obtained with supported
by the National Science Foundation (NSF), Grant Number CES-8714640. More recently
a Cooperative Agreement No. CR-816815-01-0 and a Grant No. R-816932-01-0, from the
US Environmental Protection Agency have been recieved and no experimental data have
been obtained as yet. The cooperation of the Miami-Dade Water and Sewer Authority was
essential in the completion of the work.
REFERENCES
1. Morganstern, K.H. Radiation Processing To-day and To-morrow. Conf. on Electro-
technologies in Industry, Montreal, Canada, 1982.
2. McKeown, J. Electron Accelerators - A New Approach. Radiat. Phys. Chem. 22; 419-430,
1983.
3. Singh, Ajit, Norman H. Sagert, Joseph Borsa, Harwant Singh and Graham S. Bennett. The
Use of High-Energy Radiation for the Treatment of Wastewater: A Review. Proceedings
of the 8th Symposium on Wastewater Treatment, Montreal, 1985.
4. Pikaev, A.K. Pulse Radiolysis of Water and Aqueous Solutions. Indiana Uni. Press.
Bloomington, IN. 295 pp., 1967.
434
-------
5. Bielski, B.H.J. and J,M. Gebicke. Species in Irradiated Oxygenated Water. Adv. Radiation
Chemistry, 2:177-279, 1970.
6. Draganic, I.G. and Z.D. Draganic. The Radiation Chemistry of Water. Academic Press,
N.Y., N.Y. 242pp. 1971.
7. Allen, A.O. The Radiation Chemistry of Water and Aqueous Solutions, van Nostrand-
Reinhold. Princeton, NJ. 1961.
8. Bensasson, R.U., EJ. Land and T.G. Trascott. Flash Photolysis and Pulse Radiolysis.
Contributions to the Chemistry of Biology and Medicine. Pergamon Press, N.Y. 229 pp.
1983.
9. Buxton, George V., Clive L. Greenstock, W. Phillip Helman and Alberta B. Ross. Critical
Review of Rate Constants for Reactions of Hydrated Electrons, Hydrogen Atoms and
Hydroxyl Radicals (OH/-O") in Aqueous Solution. Journal of Physical and Chemical
Reference Data, J7: 513-886, 1988.
10. Rezansoff, B.J., K.J. McCallum and RJ. Woods. Radiolysis of Aqueous Chloroform
Solutions. Canadian J. Chem. 4§:271-276 (1970).
11. Dickson, L.W., V.J. Lopata, A. Toft-Hall, W. Kremers and A. Singh, Radiolytic Removal
of Trihalomethanes from Water. Proceedings from the 6th Symp. on Radiation Chemistry,
1986, pp!73-182.
12. Gehringer, P., E. Proksch, W. Szinovatz and H. Eschweiler. Decomposition of
Trichloroethylene and Tetrachloroethylene in Drinking Water by a Combined
Radiation/Ozone Treatment. Water Res. 23: 645-646, 1988a.
13. Schested, KL, H. Corfitzen, H.C. Christensen, and E.J. Hart. Rates of Reaction of O', OH,
and H with Methylated Benzenes in Aqueous Solution. Optical Spectra of Radicals. J.
Phys. Chem., 79: 310-315.
14. Hashimoto, S., T. Miyata, M. Washino and W. Kawakaml A Liquid Chromatography Study
on the Radiolysis of Phenol in Aqueous Solution. Environ. Sci. Technol. JJ:71-75, 1979.
15. Cooper, W.J.; Niekelsen, M.G.; Waite, T.D.; Kurucz, C.N. High-Energy Electron Beam
Irradiation: An Advanced Oxidation Process for the Treatment of Aqueous Based Organic
Hazardous Wastes. Symposium on Advanced Oxidation Processes for the Treatment of
Contaminated Water and Air, Toronto, Canada, 1990.
435
-------
SOIL BARRIER ALTERNATIVES
by: Walter E. Grube, Jr.
Soil Scientist
Risk Reduction Engineering Laboratory
U. S. Environmental Protection Agency
Cincinnati, Ohio 45268
ABSTRACT
Manufactured clay products, cementitious materials, and soil
modification processes are being marketed as,landfill liners, cover systems
for closed waste sites, soil berms built as secondary containment structures,
and other waste management structures required to restrict liquid migration.
This paper describes currently available materials and processes.
Differences in properties, construction, installation, and hydraulic
performance are presented. The advantages and disadvantages of alternatives
to several feet of compacted soil are discussed. Limited published
information on alternative barriers is critiqued. Conclusions regarding the
merits of use of alternative materials and processes in various residuals
management applications are presented.
This paper has been reviewed in accordance with the U. S. Environmental
Protection Agency's peer and administrative review policies and approved for
presentation and publication.
INTRODUCTION
A 3-ft(0.9 m) thick layer of low-permeability, compacted soil is a
required component of secondary liners for hazardous waste landfills and
surface impoundments regulated under the Hazardous and Solid Waste Amendments
(HSWA) to the Resource Conservation and Recovery Act (RCRA) [EPA, 1985]. The
recommended designs for cover systems over RCRA hazardous waste landfills and
closed surface impoundments include a 60-cm thick layer of low-permeability,
compacted soil [EPA, 1989]. Minimum design requirements for liner and cover
systems for non-hazardous waste landfills vary from state to state, but many
include a layer of low-permeability compacted soil.
436
-------
Waste disposal facility owners and operators seeking RCRA permits, and
responsible parties seeking designs for closure of remedial action sites are
requesting approval of commercial materials and soil treatment processes as
alternatives to several feet thickness of compacted soil. Alternative barrier
materials possess both advantages and disadvantages in each unique
application. This report summarizes the Agency's current state of knowledge
and expresses technical concerns regarding the structural and hydraulic
performance of available or proposed alternative barrier materials and soil
modification processes.
EQUIVALENT OR ALTERNATIVE ?
The main function of low-permeability, compacted soil is either to
restrict infiltration of water into buried waste (in cover systems) or to
limit seepage of leachate from the waste (in liner systems). Other objectives
may include enhancement of the efficiency of an overlying drainage system,
enhancement of the effectiveness of an overlying geomembrane, adsorption and
attenuation of leachate, restriction of gas migration, and others. In the
case of a cover system, compacted soil must also be able to withstand
subsidence or differential settlement, and must be repairable if damaged by
freezing, desiccation, or biologic intruders. For liner systems, the liner
must be able to withstand chemical degradation from the liquids to be
contained. In addition, low permeability compacted soil must have adequate
shear strength to support itself on slopes and to support the weight of
overlying materials or equipment.
Fundamental compositional and structural differences between compacted,
low-permeability soil and alternative materials create inevitable differences
in hydraulic properties, solute attenuation capacity, time of travel of
chemical compounds, structural strength, desiccation resistance, freeze/thaw
resistance, reaction to settlement, ease of repair, and useful life. Table 1
presents a qualitative list of factors differentiating between compacted soil
and clay-blanket alternative barrier materials.
An alternative barrier material, in order to be fully equivalent to a
compacted soil layer, must serve the same functions as compacted soil. Due to
inherent differences in the composition and construction of compacted soil and
alternative materials, the two categories of barrier structures can never be
"equivalent" in all possible respects. For example, compacted soil is usually
from 2 to 5 ft (0.6 to 1.5 m) thick, whereas the alternative barriers are all
typically a fraction of an inch to possibly a few inches (a few mm to a few
cm) in total thickness. Due to differences in thickness, an alternative
barrier is bound to be more vulnerable to puncture or other damage than a much
thicker layer of compacted soil.
Materials such as cements, grouts, and asphalts which are applied as
viscous liquids in layers one to four inches ( 2.5 to 10 centimeters) in
thickness must maintain their integrity after curing. Shrinkage cracks which
develop with time must not be allowed. The problem of quality control, for
example assuring consistent thickness of the applied material, has not been
addressed in use of these
437
-------
TABLE 1. COMPARISON OF DIFFERENCES IN ALTERNATIVE BARRIER MATERIALS
Compacted Soil
Thick (2 - 5 ft, or 0.6 - 1.5 m)
Field Constructed
Difficult to build correctly
Impossible to puncture
Constructed with heavy equipment
Usually requires test fill at each site
Site-specific data on soils needed
High leachate attenuation capacity
Relatively long containment time
Large thickness takes up space
Cost is highly variable
Soil has low tensile strength
Can desiccate and crack
Difficult to repair
Vulnerable to freeze/thaw damage
Performance is highly dependent upon
quality of construction
Slow construction
Clay Blanket Alternative
Thin (< 10 mm)
Manufactured
Easy to build (Unroll & Place)
Possible to damage or puncture
Light construction equip.used
Repeated field testing not needed
Manufactured product; data available
Low leachate attenuation capacity
Shorter containment time
Little space is required
More predictable cost
Higher tensile strength
Can't crack until wetted
Not difficult to repair
Probably less vulnerable to
freeze/thaw damage
Hydraulic properties are less
sensitive to construction
variables
Much faster construction
materials in waste management structures. Soil particle binders, such as
numerous types of organic polymers, must be proven stable over the expected
lifetime of the waste management facility. Although many of these materials,
such as polyacrylamides and urethanes, have proven applicability to
agricultural soil sealing, their long-term structural and hydraulic
performance in waste containment or infiltration prohibition at hazardous
waste sites has not been clearly demonstrated. Barrier materials created by
binding mineral particles together are unlikely to possess contaminant
sorption properties found in compacted soils.
Materials installed as discreet panels of impervious material suffer
from lack of clear demonstration of seam integrity. Mechanical overlapping
appears to be adequate with some materials, primarily bentonite blankets
installed in cover systems using the "shingle" approach on sloped areas.
Joint compounds and installed integrity proposed for rigid panels such as
fiberglass, compressed concrete, or other materials must be demonstrated by
objective studies before general acceptance can be recognized.
When the potential use of an alternative barrier is evaluated for a
specific project, the critical functions of the barrier should be identified.
"Equivalency" should be evaluated on the basis of the critical parameters and
not necessarily upon all potential areas of comparison. Further, it must be
remembered that all liner materials possess inherent advantages and
438
-------
disadvantages—no single type of liner material can be considered optimum for
all applications. The site-specific design function of a waste containment
liner, a precipitation infiltration barrier, or a groundwater control
structure, must the basis upon which alternative barriers are compared.
AVAILABLE MATERIALS AND PROCESSES
Currently, the marketplace is dominated by thin, manufactured, low-
permeability blanket-like products containing a thin layer of bentonite clay.
Other products such as spray-on concretes, grout formulas, asphalt coatings,
polyacrylamide or urethane liquid formulations, resins, epoxies, latexes, and
prefabricated panels have been proposed and advertised.
Clay-blanket types of materials include at least four products from
different manufacturers. These are supplied as rolls and installed in a
manner similar to that used for geomembranes. Seams joining adjacent sheets
are achieved by overlapping by amount specified by the supplier for each
product.
One of the alternative barrier materials is Bentomat(R), which consists
of 1 lb/ft2 (4.9 kg/m2) of bentonite sandwiched between two geotextiles that
are needlepunched together. Hydraulic conductivity of small specimens
permeated in the laboratory with water was found to vary with effective
confining stress but is generally in the range of 10"9 to 10"8 cm/s for
confining stresses of 8 to 12 psi (55 to 82 kPa). Practically no data are
available on hydraulic properties of overlapped seams or hydraulic properties
when the material is permeated with liquids other than water.
Claymax(R) consists of 1 lb/ft2 (4.9 kg/m2) of bentonite sandwiched
between and glued to two geotextiles. Hydraulic conductivity of small
laboratory specimens was found to vary from approximately 1 x 10'8 cm/s at an
effective confining stress of 2 psi (14 Kpa) to 3 x 10" cm/s at an effective
confining stress of 30 psi (207 Kpa). Hydraulic conductivity to chemicals was
found to vary with the chemical and to be very sensitive to whether or not the
bentonite was prehydrated with water prior to introduction of the chemical.
Under carefully-controlled test conditions, overlapped seams were found to
self-seal, provided the minimum recommended overlap width (6 in. or 150 mm)
was provided.
Gundseal consists of 1 lb/ft2 (4.9 kg/m2) of bentonite glued to a
20-mil (0.5 mm) high density polyethylene (HOPE) sheet. Practically no data
are available on the hydraulic conductivity of the bentonite or the shear
strength of the material. Under carefully-controlled test conditions,
overlapped seams were found to self-seal, even with as little as 1.5 in.
(38 mm) of overlap.
Bentofix is similar to Bentomat in that bentonite is sandwiched between
two geotextiles that are needlepunched together. The hydraulic conductivity
of small samples of the material is reported to be approximately 1 x 10"9
cm/s (testing conditions not reported). Information on other characteristics
439
-------
of the material could not be located.
Gunite, shotcrete, Portland cement grouts, and similar materials have
been proposed as low-permeability waste containment barriers. These are
claimed to possess high density and strength, high chemical resistance, and
low permeability. The desirable barrier properties are based on this
industry/s experience with wall coatings, reservoirs, and tanks.
Soil stabilizing reagents used in civil engineering have been proposed
by many suppliers as compacted soil substitutes. These include liquid
compounds used as dust palliatives, binders applied to stabilize sandy soils,
erosion control chemicals, and reagents applied to control water loss from
irrigated farmland in arid regions.
Hydraulic and structural performance of various other alternative
materials have only been sparsely documented in'published literature. Flyash-
bentonite-soil mixtures show promise in terms of providing low hydraulic
conductivity and high strength. Super-absorbent gebtextiles, such as
Fibersorb(R), have been proposed. Sprayed-on geomembranes, applied to a
bentonitic blanket material, have been manufactured. Custom-made bentonite
composites with geomembranes or geotextiles have been produced to meet
specific customer specifications.
ADVANTAGES AND DISADVANTAGES OF ALTERNATIVES
Alternative barriers have been claimed to possess several economic and
technical advantages over compacted soil:
+ Installation proceeds rapidly and with relative simplicity;
+ A more predictable (than with compacted soil) end-product results where
quality of a compacted soil has low assurance;
+ Cost may be as much as one-tenth that of compacted soil;
+ Much less volume is required, providing 1) more landfill ing space
available, 2) fewer truckloads of materials needed for construction, )3
less settlement of underlying wastes because alternative materials may
weigh less than thick soil;
•f Lighter construction equipment may be used, resulting in less stress on
underlying geosynthetic components;
-I- Retesting of material may be unnecessary after an alternative material
or process is initially thoroughly characterized;
-j- Unique self-repair or contaminant sorption characteristics may be
beneficial where bentonite or similar components are part of the
barrier.
440
-------
Disadvantages include several technical deficiencies In materials and the
unknown quantity of future performance due to lack of credible testing data:
A general sparsity of objective and independent performance data;
Limited field performance experience and data, especially long-term;
Vulnerability to damage during construction, due to thin physical nature
of materials;
Vulnerability of sodium bentonite to adverse chemical reactions with
leachate constituents;
Unknown tolerance to settlement of underlying waste deposits;
Unknown effects of cyclic wetting, drying, freezing, and/or thawing upon
bulk shrinkage;
Incomplete characterization of hydraulic and structural performance of
overlapped seams under field conditions;
Potential for instability when installed on slopes common in landfill
structures;
Unknown performance when overlain by a confining geomembrane which
intensifies temperature differentials.
CONCERNS
Several major structural and hydraulic concerns require clear answers
before many of these alternative materials can be proven as long-term
effective pollution control barriers. These include installation quality
assurance measures, slope stability, seam integrity, stability to settlement
or subsidence, and resistance to climatic impacts.
Although vendors of some alternative materials have developed clear
guidance regarding installation, such as permissible weather conditions,
traffic protection, and seam construction and testing, others have minimal or
absent guidance in these areas.
Stability of alternative barriers on slopes was the most important
factor requiring sound design and performance data concluded from a Workshop
conducted by the U.S.EPA [Daniel and Estornell, 1990]. The low friction angle
where wet bentonite is stable prevents use on steeper slopes that landfill
liner and cover designers need to provide maximum waste capacities. Attempts
to increase the stable slope of bentonite blanket materials include
needlepunching between the confining geofabrics, and texturing the exterior
geofabric surfaces. Slope stability of thin soils stabilized with polymers,
and prefabricated panels other than commonly used geomembranes has not been
documented in available literature.
441
-------
Changes in hydraulic performance after installation has been a
significant unanswered question. Clay blankets are installed dry, with any
hydration and subsequent sealing against liquid flow occurring as seepage
develops and hydrates the bentonite into a low-permeability mass. Hydration
of a field-installed unit by seepage of a contaminant liquid has not been
examined; although numerous laboratory test data show a variety of results.
The available laboratory data emphasize the need for compatibility testing
similar to that required for compacted soil and geomembrane materials. A
related hydraulic concern is the potential for migration of clay particles
from the blanket, with possible clogging of underlying drainage management
systems.
Seam integrity of alternative barrier products has been tested in
laboratory studies (Daniel and Estornell, 1990). Results are varied, and in
some cases conflict with manufacturers' claims. Dye seepage studies clearly
show solute migration through seam overlaps. Standardized testing procedures
are not yet available through standards organizations, so customers do not
have uniform acceptability guidance.
Barriers composed of both clay and geofabric components may possess
adequate strength and flexibility to tolerate moderate amounts of settlement
or subsidence of underlying material. This is a significant factor in closure
and cover design for landfills which may contain bio-decomposable materials.
The tensile strength of these products is readily measured by laboratory
techniques, but the actual stresses seen after field installation has been
poorly documented. Design guidance by manufacturers to provide installation
which will provide continuing conformance and hydraulic integrity with a
settling foundation is not available.
Liners and cover systems for waste management structures in dry climates
suffer from a greater paucity of performance data than those in humid regions.
Data from computer modelling suggests that desiccation of soil compacted wet
of optimum clearly is a long-term problem in dry climates. Unpublished
observations show that temperature cycling during normal seasons in humid
climates also causes moisture migration from compacted soils. Since
alternative barrier materials are composed of much thinner materials, it may
be expected that changes in structural and hydraulic characteristics may be
much more drastic with climatic changes. This effect has not been documented
for field installations of any of the proposed or installed alternative
barrier materials. Wetting, drying, heating, or freezing of only parts of a
barrier installation are also of concern. Such events may clearly require
different designs or materials for north and south facing slopes of cover
systems.
The resistance of all of the alternative barrier materials to biologic
intruders—plant roots and burrowing animals---has not been demonstrated.
These and other concerns exemplify the young stage of objective data gathering
from which alternative barriers suffer.
442
-------
VIABLE APPLICATIONS OF ALTERNATIVE MATERIALS
An early reported use of clay-blanket alternative barrier material was
as a backup for geomembranes in a double-composite liner system for a
hazardous waste landfill designed to meet EPA's Subtitle C regulations
(Schubert, 1987).
Alternative materials can serve as economic yet viable barriers when
incorporated into temporary caps for some type of RCRA or CERCLA sites where
settlement is expected which would damage a final cover. Although the
alternative barrier may be easy to repair, it is possible that unanticipated
practical problems may arise.
Alternative barriers may be less vulnerable to damage from desiccation
after installation because they are installed dry. Giving stronger
consideration to bentonite-rich alternatives in arid regions may be
problematic when short-duration infiltration events cause sudden hydration
followed by extended desiccation. Bulk shrinkage upon extended drying has not
been investigated in large-area liner installations.
Landfill covers on relatively flat surfaces, free from expected
subsidence, are among the least controversial applications. A well-designed
and operative lateral drainage layer overlying the barrier increases
confidence that the alternative construction will keep infiltration out of
underlying wastes. Monitoring devices, such as large collection lysimeters,
increase assurance that the barrier structures are operating according to
design.
CONCLUSIONS
Numerous alternative materials and installation processes are in the
marketplace to replace several feet of compacted soil. While clear economic
incentives exist to consider most of these, trustworthy performance data are
sparse, and need to be individually interpreted.
Slope stability, shear strength, and interfacial friction represent the
data gaps most needed to be filled in order for many alternative materials to
be installed with a high level of confidence in their ultimate performance.
Hydraulic properties are readily tested with adapted laboratory
procedures, but questions remain about field performance and seam integrity.
The capacity and value of solute attenuation in bentonite-rich alternative
barriers has not been defined.
Environmental (climatic) impacts are nearly all undocumented. These
include freeze/thaw resistance, desiccation resistance, effects-of settlement,
claimed self-healing capabilities, and impacts of rough bedding materials.
443
-------
The useful life of these materials is unknown because of their
relatively recent appearance in the marketplace. Some materials possess both
mineral and synthetic polymer components within one product, leading to claims
about the best features of each. Whether such compositions lead to
synergistic environmental barrier properties, or whether these may ultimately
be antagonistic is unknown. Science and the user industry have a strong need
for all experiences with alternative barrier installations to be objectively
reported.
Economic incentives, installation simplicity, structural consistency of
manufactured materials, and utility in geometrically complex waste management
structures point to a high potential value of alternative barrier materials.
With the completion of research studies currently underway, and further
publication of case study data by the user industry, both regulatory agencies
and waste management engineers will have a higher degree of confidence in
alternative barrier performance.
REFERENCES
Daniel, D. E. and P. M. Estornell. 1990. Compilation of Information on
Alternative Barriers for Liner and Cover Systems, EPA # R0944. RREL, U.S.EPA,
Cincinnati, Ohio 45268. 81pp.
Schubert, W. R. 1987. Bentonite matting in composite lining systems, jjn
Geotechnical Practice for Waste Disposal '87, edited by R. D. Woods.
published by ASCE, New York, pp 784-796.
U.S.EPA. 1985. Draft Minimum Technology Guidance on Double Liner Systems for
Landfills and Surface Impoundments -- Design, Construction, and Operation,
EPA/530-SW-014. Office of Solid Waste and Emergency Response, Washington, DC
20460. 70pp.
U.S.EPA. 1989. Technical Guidance Document: Final Covers on Hazardous Waste
Landfills and Surface Impoundments, EPA/530-SW-89-047. Office of Solid Waste
and Emergency Response, Washington, DC 20460. 39pp.
444
-------
WASTE MINIMIZATION ASSESSMENT CENTERS
COST SAVINGS RECOMMENDED AND IMPLEMENTED IN
TWELVE MANUFACTURING PLANTS
by: F. William Kirsch and Gwen P, Looby
Industrial Technology and Energy Management
UNIVERSITY CITY SCIENCE CENTER
3624 Market Street
Philadelphia, PA 19104
ABSTRACT
The Waste Minimization Assessment Center (WMAC) program was begun in 1988 under agreement
with the Risk Reduction Engineering Laboratory of the U.S. Environmental Protection Agency. University
City Science Center (Philadelphia, PA) manages the program through which Waste Minimization Assessment
Centers assist small and medium-size manufacturers who want to minimize their formation of waste but who
lack the in-house expertise to do so. Currently three WMACs at Colorado State University, the University
of Louisville, and the University of Tennessee participate in the program.
Each WMAC is staffed by engineering faculty and students who have considerable direct experience
with process operations in manufacturing plants and who also have the knowledge-and skills needed to
minimize hazardous waste generation. The waste minimization assessments, which are conducted at no
out-of-pocket cost to the client, require several site-visits for each client served. The WMAC staff locate the
sources of waste in each plant and identify the current disposal or treatment methods and their associated
costs. They then identify and analyze a variety of ways to reduce or eliminate the waste. Specific measures
to achieve that goal are recommended and the essential supporting technological and economic information
is developed. Finally, a confidential report which details the WMAC's findings and recommendations
including cost savings, implementation costs, and payback times is prepared for each client manufacturer.
This presentation will discuss the cost savings recommended and implemented in the 12
manufacturing plants served during the first program period of this project by the Colorado State University
and University of Tennessee WMACs.
This paper has been reviewed in accordance with the U.S. Environmental Protection Agency's peer
and administrative review policies and approved for presentation and publication.
445
-------
WASTE MINIMIZATION ASSESSMENT CENTERS
COST SAVINGS RECOMMENDED AND IMPLEMENTED
IN TWELVE MANUFACTURING PLANTS
INTRODUCTION
The amount of industrial waste generated by industrial plants has become an increasingly costly
problem for manufacturers and an additional stress on the environment. One solution to the problem of
industrial waste is to reduce or eliminate the waste at its source.
To take the first practical, effective steps toward that objective, many manufacturers need help which
is not available within their plants. However, that help can cost money, and it may not be as truly impartial
and disinterested as the manufacturer would like it to be. Unless some practical, effective actions are taken,
the problem can only get worse.
University City Science Center (Philadelphia, Pennsylvania) has begun a pilot project to assist small
and medium-size manufacturers who want to minimize their formation of waste but who lack the in-house
expertise to do so. Under agreement with the Risk Reduction Engineering Laboratory of the U.S.
Environmental Protection Agency, the Science Center's Industrial Technology and Energy Management
(ITEM) division in 1988 established two waste minimization assessment centers (WMACs) at Colorado State
University in Fort Collins and at the University of Tennessee in Knoxville. A third was recently initiated at
the University of Louisville.
Assessments are carried out by teams of university engineering faculty assisted by students, who
seek out interested manufacturers and arrange visits to their plants to gather detailed information on waste
generation and manufacturing operations. Then the university assessment team analyzes the data and other
information obtained, re-visits the plant for verification and further insight, develops specific quantified
recommendations (waste minimization opportunitiesorWMOs), and calculates theircost-effectiveness before
reporting them to the manufacturers. The WMACs that performed the assessments discussed here are
locatad at Colorado State University and the University of Tennessee.
The potential benefits of the pilot project include minimization of the amount of waste generated by
manufacturers, reduced waste treatment and disposal costs for participating plants, valuable experience for
graduate and undergraduate students who participate in the program, and a cleaner environment without
more regulations and higher costs for manufacturers.
This publication presents a detailed account of the recommendations offered to and implemented
by the first 12 manufacturers served by the WMACs at Tennessee and Colorado State. To obtain this
information, professional staff members from these WMACs contacted the manufacturers to learn the
outcome of every recommendation that had been offered -- specifically whether it was implemented, when,
and what the plant's experience had been with costs, savings, and waste reduction.
An analysis of the financial benefits of implementation is offered for all 12 manufacturers as a group
and for the federal government, which can derive revenue by taxing the manufacturers' cost savings as
Incremental income.
446
-------
SUMMARY OF PLANTS AND ASSESSMENT RESULTS
According to Table 1 these twelve plants were spending a total of $1.05 million/year on waste
management before the WMACs assessed their operations and identified waste minimization opportunities
(WMQs). In aggregate, the net cost savings recommended amount to almost $1.3 million/year. Including
savings in raw material costs enables the total cost saving to be larger than the cost of waste management.
The cost-saving approach taken with these results is generally conservative, because the WMOs
address only the avoidance of raw materials costs and the reduction of present and future costs associated
with waste treatment and disposal. Not claimed are the savings related to: possibly stricter emission
standards, any liability incurred from waste management practices, and costs arising from employe health
and safety problems.
There are some common characteristics found in these plants' operations, the wastes which they
generate, and the WMOs recommended to reduce the quantity and the cost of waste management. We have
utilized these commonalities to place the results from the 12 plants into these five descriptive (but somewhat
arbitrary) categories of cost-saving WMOs:
I. Substitution of Plant Operation or Material
II. Changes in Technique or Control Method
HI. Reduction of Solvent Use
IV. Reduction of Liquid Volumes
V. Reduction and Treatment of Sludge
The shares of cost savings attributable to each of these categories are shown in this tabulation:
Category Net Savings Value Share
($/yr minimum) (%)
I. 570,299 46.2
II. 410,443 33.2
III. 187,510 15.2
IV. 62,429 5.0
V. 4,804 0.4
Total 1,235,485 100.0
This total is slightly smaller than that shown in Table 1 because some options have been deleted here to
give a more conservative count. If two or more optional recommendations were given to a manufacturer,
the smaller savings options were excluded from this total.
It is obvious that substitution of a plant operation or a material (Category I) accounts for about 46%
of the total savings found. When that share is combined with changes in technique or method of process
control (Category II), almost 80% of the savings can be accounted for. This statistic clearly indicates the
kind of WMO to be sought in trying to reduce the costs of waste to small and medium-size manufacturers.
It is more significant because of the wide variety of plants included among the 12 plants served by the
WMACs.
Reduction of solvent use (Category III) is responsible for about 15% of the net cost savings found
among the twelve plants. This category includes not only solvent recovery and recycle but also taking
measures such as better control of solvent use to prevent its loss.
447
-------
IMPLEMENTATION OF WMACs' RECOMMENDATIONS
Implementation results for each of the 12 manufacturers served by a WMAC during the 1988-89
program period are summarized in Tables 2A and 2B. Collectively they show:
• Forty-five of the 87 waste minimization opportunities (WMOs) recommended have been
implemented or definitely will be within two years (not later than 1991). That represents a
52% implementation rate for WMOs related to waste management or raw material savings.
(If a WMO produced savings in waste management and raw material costs, the WMO was
treated as two recommendations, either or both of which could be implemented,)
• Total implemented savings for the manufacturers are $685,855/year, which represents 49%
of all the cost savings recommended by the WMACs.
• When examined separately, WMOs which produce waste management savings are being
implemented at a higher rate (66.5%) than those which produce raw material savings
(38.8%).
Implementation statistics can be affected dramatically by afew recommendations. Table 2A reveals
that decisions by Tennessee manufacturers not to implement four WMOs had a relatively large impact upon
overall results. If they had been implemented, raw material savings would have roughly doubled to
$706,000/year, which then would have raised the implementation rate to more than 80% for this savings
category.
The implementation rates found among the 12 manufacturers are consistent with the previously
reported results of m-person interviews with representatives of these plants. Eleven of the 12 said that the
WMAC's assessment had led to waste reduction and dollar savings or was expected to. There is a variety
of reasons given in Tables 2A and 2B for not implementing specific WMOs. All together, 26 WMOs were
rejected, and the most common reason given is that the WMO involves an unacceptable operating change
to the plant (9 times). In only three instances was the recommendation considered impractical, but in six
instances the WMO was considered unnecessary because the plant had, in the meanwhile, changed its
operating practices.
The twelve plants do not offer a large enough sample to draw broad conclusions about industries
(SIC codes) and the types of recommendations being implemented. As more plants are served and their
implementation data become available, conclusions of that kind will be more justifiable than they are at the
present time,
FINANCIAL ANALYSES OF WMOs IMPLEMENTED
Profitability and rate of return are the kinds of financial analyses applied to the implemented waste
minimization opportunities reported by the WMACs from the first twelve assessments made of small and
medium-size manufacturing plants.
The internal rate of return (IRR) is calculated from the following standard equation:
O « CF0 + {CF,/(1+i» + {CF2/(Ui)2} + ...
in which CF = cash flow
CFgubscrlpl - the year in which the cash flow occurs
i = IRR
448
-------
It is based upon a series of cash flows (usually annual) and, by inserting the cash flow data, the user is able
to calculate a value for], which is the internal rate of return. The initial investment, represented by CF0, is
negative because it represents an outflow of money. Earnings, on the other hand, are positive, and that
mathematical convention is observed when the individual cash flows are inserted into this equation. In
effect, 1RR is the rate of return at which the sum of discounted sequential cash flows equals the initial
investment.
Another way of interpreting IRR is to say that it is the value by which a series of cash flows is to be
discounted, in the manner prescribed by this equation, so that their algebraic sum is equal to the initial
investment. Obviously, when IRR Is large, the initial investment is considered to be more profitable.
Profitability is reported here as a profitability index, which is derived by first calculating the net
present value of the same series of annual cash flows discounted at a specified rate (usually 10, 15, or
20%), instead of the IRR. This equation can be expressed in the following manner:
NPV . CF0 + {CF,/(1+DR)} + {GF2/(1-fDR)2} + ... + {CFB/(1+DR)ft}
in which NPV = net present value
(^subscript = cash flow for a specific year
DR » discount rate specified
The profitability index {also known as leverage ratio) is the ratio of the net present value, calculated
according to this equation from a series of cash flows, to total capital investment needed to implement the
WMOs.
It is clear from Tables 2A and 28 that WMOs are implemented in various years rather than all at
once. The net cash flow for a given year is calculated under the following conditions:
• All costs are capitalized.
» All investments are depreciated linearly over 5 years to a net value of zero.
» All funds needed for implementation are borrowed at a specific rate, and loans are amortized
in equal annual payments over 5 years.
« Manufacturers' savings are taxed as incremental income at a rate of 25%.
» Savings from a specific capital investment do not begin until the year after the investment
Is first made. (Thus, the initial net cash flow is always negative and it diminishes
manufacturers' tax liability at the 25% rate.)
« Implementation costs and the savings they produce increase at a constant rate, which must
be chosen for each case.
As expected, lower borrowing rates are beneficial to the federal government's returns from its
investment in the WMAC program and to manufacturers' profitability as a result of implementing waste
minimization opportunities recommended by the WMACs (Table 3). However, investments in WMOs are still
very attractive even at borrowing rates of 15%. For example, at a 15% rate manufacturers as a group earn
$2.18 within five years for every dollar invested in a recommended waste minimization opportunity (10%
discount rate assumed on cash flows). If the borrowing rate is 9%, this earning figure goes up to $2.81 . The
federal government's financial returns are $2.85 at a manufacturers' 15% borrowing rate and $3.08 at a 9%
449
-------
rata. Internal rate of return follows a similar direction, but very large values of IRR (434 to 660%) are difficult
to interpret as measurements of profitability.
All but one of the financial analyses reported here were made with the assumption that
implementations costs and the savings in waste management and raw materials which they produce increase
at 6% per year. For example, an initial capital investment made in 1990 would then cost 6% more than it
did when its size was calculated in 1989, and the savings realized would also be 6% larger. The one
exception was calculated at a 3% rate, and the effects are shown in Table 4. It is clear that WMO
investments made with larger rates of increase in costs become even more profitable if savings in waste
management and raw materials increase at the same rate.
Risk is always a deterrent to action, and manufactures naturally hesitate to make capital investments
until they are confident about the outcome. The timing of implementation expenditures used in this report
is whatever the manufacturers stated during the WMACs' collection of implementation data. On a
hypothetical basis, a financial analysis was made by assuming that the manufacturers' second- and third-
year implementations were accelerated by a year. The results in Table 5 reveal even better profitability for
the federal government and the manufacturers for accelerated investments.
The policy implications of these two sets of calculations are these:
» A faster rate of cost increases does not harm profitability and, in fact, enhances profitability
if savings increase at the same rate.
« Accelerating the rate of expenditures to implement WMOs makes the investments even more
profitable, and incentives which accelerate the rate can be justified by the improved financial
returns.
The financial analyses of implemented WMACs' recommendations demonstrate rather convincingly
that investments in waste reduction are good for the manufacturers, for the federal government, and
eventually for the environment.
450
-------
TABLE 1. SUMMARY OF PLANTS SERVED BY HHAC PROGRAM
Plant
No.
1
2
3
4
5
6
7.
SIC
• Code
2759
2851
3411
3443
3471
3479
3585
Principal Product
Commercial Printing
Paints, Varnishes, and
Allied Products
Metal Cans
Fabricated Plate Work
Plating. Polishing,
Anodizing Parts
Coating, Engraving, and
Al 1 ied Services
Ai r- Condi tioning.
Cost of
Waste Mgmt.
($/yr)
53.540
88,860
249,850
99,510
14.910
35.250
147.940
Net Cost
Savings
C$/yr)*
145,750
22,110
133.060
337,870
9.438
234.880
217,135
Air-Handling,
Refrigeration Equipment
8 3672 Printed Circuit Boards 86.848 42,225
9 3672 Printed Circuit Boards 32.609 14,080
10 3743 Railroad Equipment 142.970 42,427
11 3823 Industrial Measuring 35,940 12.510
Instruments
12 3993 Signs and Advertising 61.210 62.210
Specialties
Totals 1.049.437 1,273.695
Means 87.453 106.141
*This figure includes savings in raw materials costs as well as waste management costs.
451
-------
ro
TABLE 2A
SUKHARr OF IHPLEHENTEO AND RECOMHENOEO WASTE HIHIH1ZATIOH OPPORTUNITIES
1988-89 UKAC PROGRAM
UNIVERSITY OF TEHHESSEE
Report
No.
42-1
42-2
SIC
Code Brief HMO Description
3479 Reduce peak solvent
concentration.
Reduce primer surface
area.
Reduce primer partial
pressure.
Minimize product
rejection.
Use physical stripping.
3993 Electrostatic spray
system.
Improve spray equipment.
Re-train spray
personnel .
Minimize residual paint.
Install screen cleaning
booth.
Use template for letter
fixation.
Use adhesive tape for
letters.
Use mechanical fixation.
Haste Mgt. Haste Mgt.
Savings Savings
Impl. Recommended Implemented
Date ($/yr> C$/yr)
1988 0
0
1991 0
8,247
10.410
18.657 OCOX)
14,022
1990 4.674 4,674
1989 1,870 1.870
4.816
• o --
1990 0
0
1989 0
25.382 6.544(26%)
Raw Material
Savings
Recommended
!$/yr)
4,995
11.238
2.997
200.711
46.241
266,182
11,566
3.855
1.542
6,"/67
10,824
1.980
1.980
5.260
43,774
Raw Material
Savings
Implemented
($/yr)
4.995
2.997
•^^
7.992C3X)
3,855
1,542
1.980
5.260
12,63?(29X)
Impl. Cost
($)
12.400
3,000
2.900
28.000
150,000
196,300
4,400
6.000
3.000
700
22,880
195
100
1.500
38.775
Reason for Not
Implementing.
Unacceptable operating
change.
Rework process now
done offslte.
Rework process now
done off site.
Unacceptable operating
change.
Unacceptable operating
change.
Impractical .
Impractical .
-------
TABLE ZA. conrd.
UNIVERSITY OF TENNESSEE
Report
No.
42-3
42-4
SIC
Code Brief WHO Description
3743 Reduce generation of
paint chips.
Electrostatic spray
system.
Re-train spray
personnel .
Minimize overspray.
3443 Cover degreaser tank.
Reduce oil pick-up.
Install ultrasonic
cleaner.
Minimize salt carry-
over.
Use vacuum brazing.
Reduce paint carry-over.
Electrostatic paint
Discontinue painting.
Waste Mgt.
Savings
Smpl . Recommended
Date ($/yr)
1990 14.334
5.431
1990 1.843
1989 J.S43
23.451
0
930
1991 3,075
6.000
1990 22.800
1989 2,996
5.414
•--• 6.743
Waste Mgt.
Savings
Implemented
($/yr)
14.334
1.843
1.S43
18.020(77*)
3.075
22.800
2.996
„- -_-.-.
Raw Material
Savings
Recommended
($/yr)
10.651
7,647
2.977
300
21.575
17.182
79
17.376
16,128
180,643
1.351
5.789
52.980
Raw Material
Savings
Implemented
(t/yr)
10,651
2.977
300
13.928(65%)
17.376
180.643
1.351
. .
--,;-
Imp! . Cost
<$)
13.500
58,320
3.500
0
75.320
220
290
50.000
43.800
720.640
2,790
13,200
28,440
Reason for Not
Implementing,
Too expensive
initially.
Unacceptable
change.
Impractical .
Switching to
system.
operating
a new
Installing new system.
Unacceptable
operating
change.
47.958
28.871(60*)
291.528
199.370(681)
859.380
-------
TABLE 2A. cont'd.
UNIVERSITY OF TE*«£SS£E
Report SIC
Ho. Code Brief HHO Description
42-5 3585 Alternate fastening.
Eliminate solvent
adhesives.
Hodify adhesive usage.
Re- train spray personnel
Minimize paint mist
loss.
.p..
en Reduce oil loss.
-t*
42-6 2759 Recover solvent.
Ayto. mixing.
Uaste Hgt. Waste Hgt.
Savings Savings
Itnpl . Recommended Implemented
Date ($/yr3 ($/yr)
1989 39.520 39,520
1989 27.750 27.750
28,42?
3.619
18.094
1990 0 _••••
117,410 67.270(57*3
0
32. 765 ---•
32.765 Q(05U
Raw Katerial
Savings
Recommended
($/yr>
29,330
-2,060
-1,854
5,191
26.B20
56.250
113.677
70,338
43,170
113.508
Raw Material
Savings
Implemented Impl. Cost
C$/yr3 {$)
29.330 6.400
-2.060 31,740
5,100
3,500
2.100
56.250 6,900
83, 520(73X3 55,740
63,760
27,880
0(013 91.640
Reason for Hot
Implementing.
Solvent adhesive no
longer used.
Plant now uses
prepainted parts.
Plant now uses
prepainted parts.
Too expensive
initially.
Switching to new
adhesive.
-------
TABLE 28
SUHHARY OF IMPLEMENTED AHO RECOHHENDED HASTE MINIMIZATION OPPORTUNITIES
1988-89 HMAC PROGRAM
Report SIC
No Code Brief HMO Description
56-1 3471 Recycle HjO.
Dewater sludge.
Reduce rinse volume.
4s.
01 56-2 3672 Reuse MEMTEK effluent.
01
Install filter.
Reduce H20 usage.
Segregate acid soap
waste.
Dewater Sn stripper.
Use more Dl H20.
Recycle CuS04.
Install drip bars.
COLORADO
Waste Mgt.
Savings
Impl . Recommended
Date ($/yr)
3.625
1989 2.914
1989 1.039
7.578
1,554
1990 1.540
1989 5.483
19B9 23,470
3,920
7,090
1989 280
£
STATE UNIVERSITY
Waste Mgt.
Savings
Implemented
(S/yr)
2,914
1.039
3.953(521)
1.540
5.483
23,470
....
400
11^1
Raw Material • Raw Material
Savings Savings
Recommended Implemented
$/yr} ($/yr)
0
0
0
0
1.966
150 150
357 35
0
0
0
o —
460 ----
Impl . Cost
($)
4.500
15.000*
JLfi
19,510
22.000
810
200*
300*
4.500
9.800
0
1_.200
Reason for Not
Implementing
Reduced H?0
consumption
Hater qual 1ty
problems
Unacceptabl e
operating
change.
Too expensive
initially.
Unacceptabl e
operating
change.
43,337
30.893(71?)
2.933
507(172S)
38.810
-------
TABLE 2B, confd.
COLORADO STATE UNIVERSITY
Report SIC Imp! .
Ho Code Brief MHO Description Date
56-1 3471 Recycle H20.
56-3 3672 Reuse effluent. 1991
Reduce HZ0 usage, 1989
Dewater spent reagents.
Recycle ,Cu waste. 1989
Reuse rinse HZ0. 1991
-P» Install drip bars. . 1991
en
en
56-4 2851 Pipe-cleaning system. 1991
Recover solvent. 1990
Eliminate Hg additive. 1990
56-5 3823 Use Cr-free coating.
Segregate waste oil.
Increase drain time. 1989
56'6 3411 use non hazardous 1989
wash.
Haste Hgt.
Savings
Recommended
(S/yr)
3,625
1.550
0
1,790
1.090
0
0
4.430
3.100
2.846
5.580
11.526
5,480
6.820
I
12.300
mAf\n
, SUU
Haste Hgt. Raw Haterlal
Savings Savings
Implemented Recommended
($/yr) $/yr)
0
1.550 5,680
2.670
0
1.090 0
1,270
•--- 370
2,640(605!) 9.990
3.100 8,010
2,846 2.574
5 ,580 ' £
11.526(100*) 10.584
0
0
---- 210
OC») . 210
m* nn r i f\fi9\ n
, suu \ iUU-*M «
Raw Haterfal
Savings
Implemented Imp!. Cost
<$/yr) ($)
4,100
5.680 13.000
2,670 ZSO
4.500
0
1,270 650
370 480
9.990(1001!) 18.880
8,010 1.600
2.574 17.000*
0
10.584(1001! 18.600
0
2.500
210 0
210(100%) 2,500
Reason for Not
Implementing
Reduced H20
consumption
Space
limitation
Unacceptable
operating
change.
Segregated
using
different
method.
•Manufacturer's estimate
-------
TABLE 3
EFFECTS OF BORROWING RATE ON PROFITABILITY FROM
IMPLEMENTING HASTE MINIMIZATION OPPORTUNITIES
Borrowing Rate Federal Government
LR10* LR15 IRR
9% 3.08 2.33 54.8
12% 2.96 2.23 52.4
15% 2.85 2.12 50.1
Manufacturers
LR10* LR1&
2.31 1.88
2.24 1.83
2.18 1.77
IRR
660
533
434
CJ1
--J
(LR
15'
IRR
Leverage ratio for five-year cash flows discounted at the stated per cent (10 or 15) to the initial
time period and compared to the program investment by the federal government and the capital investment
by the manufacturers.
Internal rate of return.
-------
Growth Rates
TABLE 4
EFFECTS OF GROWTH IN COSTS OH PROFITABILITY FROH
IHPLEHEHTIHG HASTE HIHIHIZATIOH OPPORTUNITIES
Returns to Federal Government
Returns to Manufacturers
WM & RH* Implementation
Costs. % Costs. %
6 6
3 3
1 D **
LKIO
2.96
1.81
LR.s
2.23
1.34
IRR
52.4
44.0
i p **
LK10
2.24
1.93
"As
1.83
1.57
IRR
533
509
(Jl
CO
Returns are calculated at a 12% borrowing rate after taxes at 25% have been levied against manufacturers* cost savings.
* WM =
RH =
** LR10URiq) =
IRR =
15
Waste Management
Raw Material
Leverage ratio for five-year cash flows discounted at the stated per cent (10 or 15) to the initial time
period and compared to the program investment by the federal government and the capital investment by the
manufacturers.
Internal rate of return.
-------
TABLE 5
ACCELERATED VERSUS NORHAL IHPLEMENTATION
45,
ui
IO
FINANCIAL CRITERION
IHPLEMENTATION SCHEDULE
Normal *
Accel crated
Leverage Ratio
10* discount
151 discount
Internal Rate of Return, IRR
federal Govt.
2.96
2.23
52.4
Manufacturers
2.24
1.83
533
Federal Govt.
3.41
2.60
55.1
Manufacturers
2.50
2.04
171
Normal implementation occurs at the actual rate stated by manufacturers between 1989 and 1991.
Accelerated implementation occurs at a hypothetical rate between 1989 and 1990; that is, during two years rather than
three.
Conditions of implementation:
All costs are capitalized.
All investments are depreciated linearly over 5 years to a net value of zero.
All implementation funds are borrowed at a 121 rate and loans are amortized in equal annual payments over 5
years.
Manufacturers* savings are taxed as incremental income at a rate of 251.
Implementation costs and the savings produced increase at 6%/year.
-------
EVALUATING FINAL COVERS FOR HAZARDOUS WASTE LANDFILLS
USING A RULE-BASED KNOWLEDGE SYSTEM
by: James T. Decker
Computer Sciences Corporation
Applied Technology Division
Cincinnati, Ohio 45224
Lewis A. Rossman
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
ABSTRACT
This paper examines the use of a rule-based knowledge system for the
evaluation of final covers used to close hazardous waste landfills.
Following a brief discussion of final cover design and associated
performance standards, a rule-based expert system which has been developed
to interpret these standards is described. The goal structure, control
structure, and other basic data structures used to implement the system are
discussed. Specific examples of rules are presented within the context of
discussion of the relationships between rules and control processes.
This paper has been reviewed in accordance with the U.S.
Environmental Protection Agency's peer and administrative review policies
and approved for presentation and publication.
460
-------
FINAL COVER DESIGN
In a typical landfill used for hazardous waste disposal, pits (cells)
are excavated in native soil or rock of low permeability to depths of 15 to
50 feet with the base of the cell above the water table. The practice of
multi-cell construction often makes use of groups of cells separated by
berms, liners and covers. A landfill under separate cover may range in
size from 1 to several dozen acres. Several such landfills, collectively
enclosing hundreds of acres, may comprise a single facility. Cells are
commonly lined with single or multiple soil and synthetic barriers of low
permeability to water (10~7 to 10*8 cm/sec), and are equipped with leachate
collection and monitoring systems.
A variety of wastes may be accepted at a hazardous waste landfill.
Wastes may be placed in fills as bulk material, be treated with
solidification agents, or be placed in drums prior to burial. Current
regulations prohibit the placement of free liquids in landfills, although
liquids inevitably enter the unit from precipitation and run-on that occurs
during filling. Older landfills often contain wastes with relatively high
liquid volumes.
Covers (or caps) are an essential feature of landfill closure. These
structures serve to minimize threats to the public health and the
environment which are posed by the potential or actual release of hazardous
materials from the facility. Covers prevent rainwater from infiltrating
into the unit and thus minimize the likelihood of leachate formation. A
cover will also prevent gases and leachate from exiting the top of the
fill.
A typical landfill cover consists of a multi-layer design. The top
layer consists of a vegetative or armored surface underlain by several feet
of topsoil. Its purpose is to minimize erosion and promote surface runoff
from the cover. Below this lies a permeable soil drainage layer for
removal of water which infiltrates through the top layer. Either natural
soil or geosynthetic materials are used in the drainage layer. The bottom
of the cover is a low-permeability layer designed to limit liquid
infiltration into the underlying wastes. The recommended low-permeability
layer design consists of two components: a flexible membrane liner (FML) of
at least 20-mil thickness, and a compacted clay layer at least 60-cm thick
with a permeability less than 10 cm/sec.
Covers may include several additional layers. Filter layers, made
either from natural soil or geosynthetic material, can be placed between
soil layers of differing grain size distributions. Filters prevent piping
of soil between the layers and help to maintain layer integrity over time.
A biotic barrier may also be included in the structure to prevent plant
roots and burrowing animals from disturbing the drainage and low-
permeability layers. A barrier layer typically consists of either tightly
packed cobbles or a thin polymeric material releasing a herbicide at a
controlled rate. Lastly, a gas vent layer is often included to collect and
transport landfill gases to a collection point where they can be safely
461
-------
treated or disposed. Perforated vertical collector pipes penetrating to
the bottom of the landfill may be used as vents in conjunction with the gas
collection layer.
Under the Resource Conservation and Recovery Act (RCRA) of 1976,
performance standards for hazardous waste landfill covers have been
published in the Code of Federal Regulations (40 CFR 264.310). These
regulations require that the final cover be designed and constructed to:
1) accommodate settling and subsidence in order to maintain the
structural integrity of the cover,
2) minimize erosion or abrasion of the cover,
3} promote drainage of surface and infiltrated water from the
cover,
4) minimize migration of fluids through the landfill to ground and
surface waters,
5) have a permeability no greater than that of the landfill's
bottom liner system or of the natural soils present, and
6) function with minimum maintenance over the post-closure period.
Regulatory agencies responsible for reviewing applications for
permits to operate or close hazardous waste landfills must ensure that
final cover designs submitted by the owner/operator will satisfy these
criteria. Since regulatory standards are based on cover performance and
not design characteristics, application of these standards at the time of
design often poses special problems.
Construction of a cover system which will last for a number of years
with minimal maintenance requirements depends on the use of sound
engineering practices derived from particular knowledge which has been
accumulated over time. Much of the knowledge needed for cover evaluation
also involves concepts that have been refined through practical design
experience. Properties of buried wastes, the materials and methods used to
construct the cover, and the effects of interactions between various cover
materials and the waste or its by-products, must be evaluated.
An expert system named F-Cover, which encodes guidelines for cover
design, is currently being developed. An initial version of this system
has been released for evaluation and testing. It is anticipated that F-
Cover will be particularly useful in assisting the permit review process of
evaluating landfill cover performance standards.
KNOWLEDGE REPRESENTATION
In operation, the F-Cover system first requests that the user
provides a layer design sequence for the proposed cover. The user also
462
-------
provides information about materials to be used, layer thicknesses,
permeabilities, and other basic data. The system uses this data and
internal information to generate questions for the user. Responses in turn
are used to determine other properties and characteristics of the cover
design. As required information is supplied through this interview
process, the system incrementally assembles conclusions by 'reasoning'
about the data. Eventually, the system determines that sufficient
information has been provided to determine final conclusions regarding the
adequacy of the design. A text report is output for review after the
interview and all evaluation processes have been completed.
The knowledge sources used to develop F-Cover include published
design manuals, proposed EPA guidelines, and experts from the EPA and
private consulting firms. Given the way in which cover performance
standards are stated, it is natural to express design knowledge and
conditions through which regulatory objectives are satisfied in the form of
rules.
Indeed, it is possible to express the basic operation of the cover
review process in a single rule: Abstracted from the system, this rule can
be stated as follows:
IF subsidence potential has been accommodated
AND erosion potential has been minimized
AND liquid migration is minimized
AND the drainage system is adequate
AND the cover meets permeability regulations
AND maintenance is minimized
THEN the cover design is approved
ELSE the cover design is (or may be) insufficient.
This rule corresponds to the top-level of the goal tree diagram shown
in Figure 1. Also, the conditions expressed in this rule's premise
correspond directly to the six performance standards listed in the previous
section. (See the Appendix for a more detailed explanation of the rule
syntax used in this chapter.)
Although the concept expressed by this rule is represented in a
slightly different manner in the F-Cover system, the underlying logic of
the system is essentially the same. In F-Cover, conditions expressed in
the top-level rule premises are represented by data objects (variables),
and each such variable is evaluated to determine whether or not the cover
design is approved. Additional rules are used to evaluate each condition.
Proceeding in a logical order through the knowledge domain, toward more
specific design parameters, the third premise in the main rule can be
considered as an example. This premise, "minimization of erosion
potential" is expanded through a rule such as the following:
IF no erosion deficiencies are found
THEN erosion potential is minimized
463
-------
ELSE erosion potential is not minimized
OR erosion potential is unknown.
In other words, if the rule base has been evaluated for all erosion-
related deficiencies and none were found, then erosion potential has been
minimized. Conversely, if a substantive erosion deficiency or certain
information about erosion potential is missing, it is concluded that either
erosion potential cannot be assessed from the information provided, or a
potential erosion problem exists. The first conclusion is reached if
calculations of the rate of erosion from the cover have not been supplied
and no other deficiencies are found. The second conclusion is satisfied
when any condition determines that an unacceptable rate of erosion is
likely (i.e., at least one substantive deficiency is detected). Through
similar processes, all higher-level goals are assigned values as a function
of lower-level processing.
COVER APPROVAL
I I I I I I
Subsidence Eros Con Water Drainage Permeability Maintenance
Aecofnmodax 1 on Minimization Migration Adequacy Criteria Minimized
Minimization Satisfied
Letera I Liquid
Leachate I rrf I I trat t on
Containment Minimization
Minimal No Gross Layer Layer Layer Drainage
Layer Mechanical Structural Infiltration Interactions Adequacy
Configuration Failure Soundness Minimized OK
Figure 1. Goal tree for the Cover system
Lower-level system rules are usually dependent on specific inputs,
and higher-level nodes or subgoals are determined by low-level conclusions.
In order to fix the values of the higher-level subgoals, the logic process
is switched to a "bottom-up" mode to search for specific characteristics of
the design which may cause an erosion design deficiency, or otherwise
determine that information is insufficient. Two examples of specific
deficiencies for the erosion goal are that the cover has a vegetative layer
thickness less than 60 cm and that calculations of soil erosion rate due to
precipitation runoff result in a value greater than 2 tons/acre/year. Rule
expressions for these concepts are straightforward:
464
-------
Table 1. Explanation of F-COVER's goals
Goal
Explanation
Subsidence
Accommodation
Erosion
Minimization
Water Migration
Minimization
Lateral Leachate
Containment
Liquid
Infiltration
Minimization
Minimal Layer
Configuration
No Gross Mech-
anical Failure
Layer Struct-
ural Soundness
Layer Infiltration
Minimization
Layer Interactions
OK
Settling of waste after closure accounted for; funds available to repair any
damage to cover.
Vegetative layer at least 60 cm deep and Its rate of soil loss is under
2 tons/acre/year.
Release of leachate to ground and surface waters Is minimized.
Leachate will not escape through sides or top of cover.
Amount of rainfall infiltrating the cover and causing the creation of
leachate is minimized.
All recommended cover layers are present (Top, Drainage, and Low-
Permeability layers)
No gross mechanical failure of cover due to subsidence, biotic
intrusions, soil creep, waste gases, leakage through penetration.
No piping of soil into adjacent layers; no clogging of drainage layer;
FML resistant to heat and UV attack; soil liner below frostline.
Drainage layer of suitable thickness and permeability; FML of sufficient
thickness and free of imperfections; no leakage through FML penetrations;
soil liner thickness, permeability, plasticity and organic content are OK.
No loss of fines from vegetative to filter or drainage layer; FML chemically
compatible with leachate; layer above FML doesn't creep; suitable blanket
and bedding layers provided for FML; soil liner compatible with leachate.
Drainage Adequacy Runoff capacity is adequate; no tendency for ponding of liquids.
Permeability
Criteria Satisfied
Maintenance
Minimization
Permeability of landfill's subsoil (or bottom liner) is no less than that
of the cover.
Efforts and resources needed to maintain cover over the post-closure period
are minimized.
465
-------
IF top layer is a Vegetative/Soil layer
AND layer thickness is < 60 cm
THEN a substantive erosion deficiency exists.
IF calculations of erosion are provided
AND calculations do not indicate an erosion rate less than 2
tons/acre/year
THEN a substantive erosion deficiency exists.
Figure 1 shows the high-level goal tree that was developed in this
manner for COVER'S rule base. It shows how the original landfill cover
performance standards have been abstracted into subgoals. Each of these is
briefly described in Table 1. The leaf nodes in the tree represent
subgoals. These nodes are evaluated by rules looking for specific design
or information deficiencies such as described above. A variety of input
values and value types from specific rules are accepted at this level.
These values are passed to subgoal rules for interpretation. If no
deficiencies are found at a subgoal node, then a subgoal variable is
assigned a value of 'YES'. Otherwise, the subgoal variable is assigned a
value of either 'NO' or '??'; the last of these indicating that information
is insufficient to determine a conclusion. All subgoals in the tree must
be satisfied for a cover design to be approved. The control rules needed
to accomplish this task are discussed next.
CONTROL STRUCTURE
It may appear sufficient to combine the goal-subgoal hierarchy and
lower-level rules with a 'backward chaining' control strategy (see
Appendix), and to thereby determine whether or not the cover should be
approved. There are no problems with this strategy when the system does
not detect design deficiencies, inasmuch as all performance categories are
eventually searched and evaluated in order to determine a value for the
top-level goal. However, if any deficiency is detected, and a subgoal node
value is set as a function of this fact, then other deficiencies may not be
detected because the system will not continue to search for information
specifically related to the node. Such failure to detect a deficiency can
occur either at a level contributing to a particular node being evaluated,
or with respect to other nodes which have not yet been evaluated. Because
the purpose of design review is to identify all potential cover
deficiencies - not just those which are sufficient to determine a subgoal
value - a simple backward-chaining evaluation strategy is inadequate.
Rather, for a complete deficiency report, the system must explicitly
find values for all goals listed in Figure 1, regardless of the value
determined at any particular node. Moreover, it is also necessary for
potential deficiencies to be evaluated to the extent that all contributing
factors for each individual subgoal node are identified.
In F-Cover, the required exhaustive evaluation control strategy is
implemented through various control processes. At the top-level, a single
control function is used. This control function forces evaluation of
466
-------
processes to evaluate each subgoal. The subgoals evaluated are:
Minimal layer configuration
Subsidence accommodation
Erosion minimization
Lateral leachate containment
No gross mechanical failure
Layer structural soundness
Layer infiltration minimization
Layer interactions OK
Drainage adequacy
Infiltration minimization
Water migration minimization
Permeability criteria satisfied
Maintenance minimization
The meaning of each subgoal is provided in Table 1. Because each
item in the table represents both a process and a variable in the system,
all corresponding processes can be executed through the control function.
As a result, when the system is asked to evaluate the goal "final approval"
through application of the top-level rule, values have already been
determined by the control function for all categories of performance
standards and their attendant design deficiency topics. The control
function insures that all node evaluations are completed, regardless of the
value subsequently assigned to "final approval".
The evaluation of subgoals is conducted by the application of various
rules and results in the assignment of a value to each subgoal node. To
indicate the result of each evaluation, subgoal node variables are set to a
single value of: 'YES', 'NO', or in some cases, '??'. As previously
described, the '??' value indicates that information is insufficient to
determine either a YES or NO value. For some subgoals the '??' value is
not allowed, and the final node value must be either 'YES' or 'NO'.
Lastly, additional control rules and procedures are used to build an
internal representation of the design, generate queries in response to
which the user provides design information and other data, generate the
text which is output in a deficiency report, and provide context-based
'help' information for the user. Many of the procedures governing these
functions are relatively complex in structure. Because these procedures
mostly serve to support the evaluation logic, they are not reviewed in
detail here. In particular, procedures which access help information or
generate user queries are not discussed.
The backbone of the F-Cover system is captured by the three-level
relationship of the high-level goal, to component subgoal rules, to sets of
lower-level analytical rules which determine the values of the subgoal
nodes and control output. Understanding the last part of this
relationship, the connections between subgoals and lower-level rules,
requires discussion of system data structures in three additional areas.
These are: 1) the representation of layers and layer-specific data, 2) the
467
-------
design deficiency rules, and 3) the rules which generate the deficiency
report. Examples of each of these representation classes are discussed in
the following three sections.
LAYER PARAMETER RULES
Many design deficiency rules refer to properties of a specific layer
in the cover. Other deficiency analysis rules evaluate relationships
between adjacent layer groups. For both kinds of rules, it is useful to
have methods (data structures) which convey information about the
positioning of layers with respect to adjacent layers, and which store
descriptive parameter values in a concise fashion. These needs are
satisfied through variables which store the layer sequence and design
parameters associated with each layer of the cover as lists. List elements
are ordered to correspond with cover layers and parameters, from top to
bottom. For example, the type of the second cover layer {usually a biotic
barrier, filter, or drainage layer) is stored in a "layer list" variable in
the second position. Similarly, layer thickness is stored in a "thickness
list" variable in the second position, its permeability in a "permeability
list" in the second position, and so on. These lists are referenced
whenever it is necessary to select information about layers as a function
of position in the cover design, or to evaluate inter-layer relationships.
Some layers, such as the vegetative top layer and the low-
permeability layer, may however, be composed of more than one distinct
component. The vegetative layer, for example, can include both topsoil and
fill soil components. Similarly the low-permeability layer usually
contains both an FML and compacted soil. For these layers, distinct
components are represented by sublists within the "layer list" variable.
Corresponding values of thickness, permeability and other key
characteristics are similarly represented in corresponding sublists. A
possible layer sequence of four layers might, for example, be represented
1n list form as:
([Vegetative, top soil, fill soil], [filter],
[drainage], [low-permeability, FML, compacted-soil])
There are four main list elements [in brackets] for the four layers.
Each main element is itself a list; the first item of which identifies the
layer and remaining elements, if any, identify components. If a layer does
not have distinct components, the only sublist element is the layer name
itself.
Specialized procedures are used to identify layer structure and basic
design values at the start of F-COVER's consultation. First a procedure is
executed to build a set of lists representing the design sequence of layers
and layer components. Then list items are interpreted through rules to
determine questions required to identify key layer parameters. Finally
responses to these questions are used to build other lists representing the
required parameter values (e.g., thickness, permeability, etc.).
468
-------
DESIGN DEFICIENCY RULES
Lower-level rules are used for identifying design deficiencies in
each of the subgoal categories listed in Table 1. Compared to the
straightforward way in which subgoal nodes are combined to determine the
high-level goal value, deficiency analysis rules are more interdependent
and less conveniently categorized. Some analytical rules provide input to
a number of subgoal nodes. Others provide information to only one subgoal.
Many low-level rules are analyzed only when very particular preconditions
are satisfied. Also, various linkage structures exist between
deficiency-analysis rules.
Some deficiency rules are 'vertically' linked so that the result of
one rule serves as an input for another more general analysis rule. In
this way, a set of independent rules at one level of specificity are
combined in the premises of another more general deficiency. In other
cases, deficiency rules at the same level of specificity are related
through common antecedent conditions or conclusions.
As a result of the complexity with which lower-level rules are
organized, these design deficiency rules are more difficult to classify in
terms of content than the higher-level subgoals. Some deficiency rules
will be applied to more than one layer in the cover. Others will apply to
only one layer, but will be interpreted in terms of their effect on the
cover as a whole. Still other rules are applied to cover characteristics
that concern relationships between layers, as mentioned above.
Despite these complexities, deficiency rules can usually be
classified into one of three groups: (1) rules applying to the cover as a
whole, (2) rules referencing specific layers, and (3) rules applicable to
relationships between layer characteristics. Examples of rules for each of
these groups follow.
RULES APPLYING TO THE COVER AS A WHOLE
is
An example of a set of rules that apply to the cover as a whole
provided by the processes which determine if lateral containment of
infiltrated water is satisfactory. Abstracted from the system's code, some
rules in this category are:
IF there is a potential bathtub effect
OR potentially, there are perched water pockets in the waste
AND also, the leachate collection system has not been analyzed or
otherwise found to be adequate
THEN Leachate Movement Containment is NO.
IF there is a bottom composite liner below the waste
AND there is no FML in the Low-Permeability layer
THEN there is a potential bathtub effect.
469
-------
IF there is no bottom composite liner below the waste
AND there is no FML in the Low-Permeability layer
AND the natural subsoil below the waste is less permeable than the
least permeable cover element
THEN there is a potential bathtub effect.
IF the waste contains sheets of material that are less permeable
than the least permeable cover element
AND these sheets are either directly below the cover or extend
across the entire unit
THEN potentially, there are perched water pockets in the waste.
IF the waste contains sheets of material that are less permeable
than the least permeable cover element
AND clays were used for cover when the unit was operational
OR there is an existing cover installed during a previous
partial closure.
THEN potentially, there are perched water pockets in the waste.
The conclusions of the last three of these four rules establish the
values needed in the premises of the first rule. The first rule by itself,
determines the value of the subgoal "leachate movement containment". In
terms of effect, the last three rules reference the structure of the low-
permeability layer. Because the conclusions of these rules are not
particular to an individual layer, the rules can nevertheless be regarded
as applying to the cover as a whole.
RULES REFERENCING SPECIFIC LAYERS
Rules applied either to the top cover layer or to another singular
layer such as the drainage, low-permeability, or gas vent layer, are
members of this group. Examples of layer-specific rules are provided by
the rules that determine whether or not erosion of the cover has been
minimized. These rules evaluate characteristics of the top layer in the
design. Four of these rules are:
IF the top layer in the design is a Vegetative layer
AND the thickness of the layer is < 60 cm
THEN Erosion Criteria Satisfied is NO.
IF the top layer in the design is a Vegetative layer
AND the thickness of the topsoil in this layer is < 15 cm
THEN Erosion Criteria Satisfied is NO.
IF the top layer in the design is a Surface Armor layer
AND the thickness of the layer is < 60 cm
THEN Erosion Criteria Satisfied is NO.
IF the applicant has provided calculations of the rate of erosion
from the top layer
AND these calculations indicate that the rate of erosion is more
470
-------
than 2 tons/acre/year
THEN Erosion Criteria Satisfied is NO.
Other layer-specific rule examples are found in the group of rules
used to evaluate the adequacy of the gas vent layer:
IF there is a Gas Vent layer in the cover design
AND it is a soil type layer
AND the USCS soil type is neither GP or SP
THEN the soil type is inappropriate and the Gas Vent layer may not
be structurally sound and the cover may not be structurally
sound.
IF both a Gas Vent layer and a Drainage layer are included in the
cover
AND at least one of these layers is composed of a geosynthetic
material
AND the transmissivity of the Gas Vent layer is less than the
transmissivity of the Drainage layer
THEN the cover may not be structurally sound.
IF both a Gas Vent layer and a Drainage layer are included in the
cover design
AND both of these layers are soil-type layers
AND the permeability of the Gas Vent layer is greater than the
permeability of the Drainage layer
THEN the cover may not be structurally sound.
Although these rules determine deficiencies particular to the gas
vent layer, the conclusions are generalized to the cover as a whole because
the goal "the cover may not be structurally sound" appears in each of the
three rules.
The software processes that interpret overall structural soundness
are, in practice, more complex than is implied by these examples. Much of
the added complexity stems from techniques which are used to preserve
information as conclusions are passed up the goal analysis tree. In
addition to the conclusions shown in the preceding examples, many lower-
level conclusions also generate a unique expression or value that becomes
an element of a list variable. Lists constructed by these events are
available for interpretation by subgoal nodes.
For example, each of the three rules shown for analysis of the gas-
vent layer also trigger processes passing information to the "cover
structural soundness" node. This subgoal node interprets the component
expression values so that if any one of the contributing rules is fired,
the source rule is identifiable. These unique identifiers are eventually
used in the report process. Hence, even though each of these three rules
can have the same higher-level result (to set the value of the subgoal node
to "NO"), lower-level contributing conditions can be identified when the
471
-------
report is generated. Similar processes are used in other rule groups for
other subgoal nodes.
RULES TESTING RELATIONSHIPS BETWEEN INDIVIDUAL LAYERS
Examples of layer-relationship rules are those used to evaluate the
potential for soil piping, or the gradual soil movement from one layer to
another:
IF the Filter layer to be analyzed is preceded in the layer
sequence (top-to-bottom) by a Vegetative/Soil top layer or a
Surface Armor/Soil Top layer
AND the Filter layer is a soil-type layer
AND the ratio of grain size diameters (in mm) for the soil in the
filter layer at the 15% sieve level, over the soil in the top
layer at the 85% sieve level, (D15[filter] / D85[top]), is
greater than 5
THEN the Filter layer may be subject to soil piping.
IF there is a Drainage layer in the design
AND this Drainage layer is constructed of geosynthetic material
AND the preceding (above) layer is not also constructed of
geosynthetic material
THEN the Drainage layer may be subject to soil piping.
IF there is a Drainage layer in the design
AND it is immediately preceded by a Filter layer
AND potential filter piping has been determined for the preceding
1 ayer
THEN Drainage layer integrity is threatened by the potential for
piping at the Filter layer.
In the first piping rule, D15 and D85 respectively represent the soil
grain diameters for soil particle size distributions at
the 15th and 85th percentiles.
Like the gas vent layer rules discussed above, the results of these
piping analysis rules also may supply information to the subgoal node
"layer structural soundness". Unlike the gas vent rule structure however,
piping rule results are chained through other rules; and it is these
intermediary rules that contribute the list elements interpreted by the
higher-level subgoal.
CONCLUSION AND DEFICIENCY REPORTING
The conclusion of a consultation is reached when all rules required
to evaluate subgoals have been tested, and all subgoals have consequently,
been assigned values. Usually, there are many rules in the rule base which
are neither tested or executed. In general, rules appropriate to the
specified design are tested, and lower-level rules which are inappropriate
472
-------
to the design are not considered. Of course, a rule only executes when
premises are satisfied, and many rules which are tested will nevertheless,
fail to be satisfied.
When all appropriate rules have been tested, the system writes a text
report of results and recommendations. The report is stored in a file
which is available for viewing, printing, or editing. The structure of F-
COVER's deficiency report follows, for the most part, the goal tree of
Figure 1 in a top-down rather than a bottom-up fashion. First the result -
of the major performance goal is reported; and if the goal has 'failed'
(i.e., one or more deficiencies were detected), then subgoal deficiencies
are listed. Under each subgoal heading, all specific deficiency findings
are presented. If there are no subgoal deficiencies (i.e., all subgoal
tests were satisfied), then only heading information and a statement that
the main goal was satisfied are included in the report.
Text for the output report is generated in various ways. Some text is
recovered from an external file, and other text is generated directly by
the program. List variable inputs to the subgoals are decoded to determine
report content and order. Report writing rules are also explicitly
triggered by the operations of deficiency analysis rules or subgoal rules.
Although text is output only after completion of a consultation, some
text is retrieved or generated during analysis, and other text is retrieved
afterwards. In either case, report text pieces are stored in a set of
program variables. These variables are combined and ordered as a function
of both logical content and the sequence of tested subgoal nodes. Finally,
all stored text is written to a designated file.
These functions are illustrated by the following report rule which
generates text, reads additional text from an external file, and stores the
combination of this information in a program variable:
IF the topsoil-type of the top layer is not one of the USCS class
values: MH ML CL OH OL SM or SC
THEN the goal-list gets 'soil class unacceptable'
AND the report gets 'The soil class specified,',[topsoil-type], 'is
unacceptable.'
AND the report gets READ(report text file,'soil class text').
The premise checks the Unified Soil Classification System (USCS)
topsoil class provided by the user [topsoil-type] against a list of
acceptable values. If the topsoil-type is not in the list (MH, ML, CL, OH,
OL, SM, or SC), a phrase representing this result ('soil class
unacceptable') is stored in a report variable (goal-list) which is
available to the subgoal node, and appropriate report text is created. The
report text is a combination of a sentence ('The soil class specified,'
[topsoil-type], 'is unacceptable.'), and text which is retrieved from an
external file via the READ function. READ obtains text labeled 'soil class
text' from a report text file. The first sentence of the report text
specifies the USCS classification of the topsoil entered by the user
473
-------
(topsoil-type). Combined text is stored in the variable named 'report' in
this example.
Although the processes that create, generate, and arrange text are
sometimes complicated, the procedure which writes the output file is
straightforward. This process, in pseudo-code, is as follows:
WRITE (,)
WRITE (,)
WRITE (,)
WRITE (,)
WRITE (,)
CLOSE ().
The WRITE operation is the converse of the READ function. It writes
the value of the second parameter to the file identified by the first
parameter. The first WRITE statement outputs heading information. The
second WRITE issues a statement corresponding to the state of the top goal
node (i.e., were any deficiencies detected?). All remaining WRITE
functions output text for various subgoals and accompanying deficiency
analysis text. If a particular subgoal node state is YES (meaning no
deficiencies were found), then the value of a particular report text
variable may be null or empty. If a final report variable is empty, the
corresponding WRITE function does nothing.
FINAL COMMENTS
Although F-COVER was created with a development tool which is
primarily a backward chaining system, the directions of rule linkage
processes in F-COVER are mixed. Both backward and forward rule linking
processes are used. The relationship between subgoals and deficiency
analysis rules is mostly driven through backward chaining. In order to
determine a subgoal value, some subset of deficiency rules must be tested.
A type of forward chaining is used however, for parts of the deficiency
analysis where it is necessary to ensure execution of some subset of
analysis rules regardless of the values of corresponding subgoal variables.
Also, forward chaining techniques are used to ensure that all subgoals are
evaluated during parts of the report writing process.
As well as forward and backward chaining structures, other methods
are used to control rule processing. Repetitive loop-like processing is
used at various stages of the analysis to ensure that rule tests are
efficiently applied to a sub-sequence of layers or layer parameters. Given
the list structures that store key values, loop processes control
'traversal' evaluation of selected lists or list subsets. In other words,
a rule process is conditionally applied to individual list elements through
a control operation. As a result, the number of explicit rules needed is
reduced, and the system's organizational coherence is increased.
474
-------
Lastly, other techniques are used as parts of the user interface and
to tie the report generation rule base to analytic and goal level rules.
For the report process, such methods are used to arrange findings in terms
of increasing levels of specificity. This ordering is more natural than
what is obtained when findings are displayed or written in the order
determined by discovery.
Rules for deficiency analysis, rules to determine subgoal and goal
values, and rules for controlling report generation, are the essential
elements of the F-COVER system. Rule-based structures were found to
provide a relatively flexible means for representing declarative knowledge
within the domain and for controlling communications with the system user.
These rules were combined with other procedures to enhance the user
interface and increase the efficiency of rule application and report
generation processes.
APPENDIX - RULE SYNTAX
Rules are expressions relating how variables, also called goals, can
be given particular values depending on the values of other variables.
Initially, all variables are assigned values of UNKNOWN, indicating that no
evaluation has occurred. Variables normally accept any number of values in
the form of a list. But a variable's definition can be narrowed, so that
it will accept only a single value or, perhaps, only a single numeric value
or character value. Rules are expressed in a format such as:
IF
AND/OR
» »
THEN
AND
ELSE
Clause in the rule's premise usually have the form:
where relations are either common logical comparison operators (e.g., AND,
OR), or comparison operators which operate on lists instead of single
values. Common numeric operators are: equals (=}, not equal (<>}, less
than (<}, and greater than (>). Other list-type relations, such as "is a
member of" or "is the first element of", also may appear in premise
clauses. If the clauses in the IF-portion of the rule are satisfied, then
the actions in the THEN-portion occur. Otherwise, the ELSE-actions occur.
Some common actions are:
Assign a variable a value or set of values,
Reset a variable value to UNKNOWN,
475
-------
Find the value of some variable,
Display text elements on the screen.,
Display text and request a value for a variable,
Write text to a file,
Read text from a file, and
Execute a procedure.
System rules are linked together through either backward or forward
chaining mechanisms. The difference between these two methods refers to
the ordering by which rule evaluations occur.
Backward chaining in F-Cover occurs as function of the rule
structure. If a rule references a variable in its premise for which a
value has not been determined, then a process is initiated to identify the
unknown value. Other rules which can conclude the needed value are also
tested. Satisfied rules are fired or executed. This process can be
continued so that additional rules are tested to establish the premises of
the rule or rules preceding in the chain. In short, backward chaining is
the process of 'reasoning' from conclusions to premises, or from goals to
the conditions which determine goal values.
Forward chaining is the process of reasoning from premises to
conclusions. In the F-Cover system forward chaining is accomplished by
explicitly forcing a rule evaluation process to occur and thereby
establishing a rule's consequent or action component.
REFERENCES
1. Technical Guidance Document: Final covers on hazardous waste
landfills and surface impoundments. EPA/530-SW-89-047, U.S.
Environmental Protection Agency, Cincinnati, Ohio, 1989. 39 pp.
2. Rossman, L. and Decker, 0. A rule-based system for evaluating final
covers for hazardous waste landfills. In: R. Allen (ed.), Expert
Systems in Civil Engineering: Knowledge Representation. American
Society of Civil Engineers, New York, New York, 1991 (in press).
476
-------
EVALUATION OF ASBESTOS RELEASE FROM BUILDING DEMOLITION
FOLLOWING THE OCTOBER 1989 CALIFORNIA EARTHQUAKE
by: Roger C. Wilmoth, Bruce A. Hollett, and Patrick J. Clark
Risk Reduction Engineering Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
ABSTRACT
Following the severe earthquake along the San Andreas fault on
October 17, 1989, the EPA Risk Reduction Engineering Laboratory was
requested to assist EPA Region IX and the Monterey Bay Unified Air Pollution
Control District in evaluation of fugitive emissions from building
demolition and subsequent disposal of the demolition waste. Buildings
damaged during the earthquake were considered to be structurally unsafe,
preventing any access to ascertain asbestos content. The presumption was
made that the buildings contained asbestos because similar undamaged
adjacent buildings were surveyed and asbestos was found.
Demolition of damaged buildings was monitored at the Santa Cruz Pacific
Garden Mall and Downtown Watsonville. Disposal operations for demolition
wastes were also monitored at the Santa Cruz Municipal Landfill.
Visible emissions were observed during the demolition process. These
were most apparent during structural destruction of the buildings and were
generally absent during handling of the debris. Because of copious wetting
during the process, no visible emissions were observed. There were measured
asbestos levels above background without visible emissions during the
handling of debris.
477
-------
Operators of bulldozers involved in landfill ing of the asbestos-
containing debris were exposed to personal breathing zone concentrations
above background level.
This paper has been reviewed in accordance with the U.S. Environmental
Protection Agency's peer and administrative review policies and approved for
presentation and publication.
INTRODUCTION
The October 17, 1989 San Andreas fault earthquake epicenter was only
about 10 miles from Watsonville, CA, 20 miles from Santa Cruz and 80 miles
from San Francisco (Figure 1). Numerous buildings were structurally damaged
and were scheduled for emergency demolition. Severely damaged buildings
were unsafe for re-entry particularly when aftershocks could be expected;
therefore, it was not possible to inspect these structures for asbestos.
Because inspection of similar undamaged adjacent buildings revealed that
they did contain asbestos, the presumption was made that the buildings being
demolished also contained asbestos. Since the buildings were condemned and
no access was permitted, asbestos removal prior to demolition could not be
accomplished. There was, therefore, a concern about emissions from the
demolition sites and the landfill where the debris was deposited.
The EPA Risk Reduction Engineering Laboratory (RREL) provided requested
assistance to the EPA Region IX office and the Monterey Bay Unified Air
Pollution Control District in monitoring these activities. RREL conducted
air monitoring for asbestos during demolition of one building in downtown
Watsonville, three buildings at the Santa Cruz Pacific Garden Mall, and at
the Santa Cruz Municipal landfill.
STUDY DESIGN AND METHODS
These observational studies were made as demolition work progressed.
Typically, there were multiple activities ongoing including demolition,
debris loading, and debris transportation. A few bulk samples were taken
from debris to determine the presence of asbestos. Air sampling locations
were selected to provide upwind (background ambient conditions), downwind
(migration of emissions), and near activity (source emission) observations.
Portable power generators were used and sampling locations were selected to
be accessible to generator power and out of harm's way to the extent
possible.
SAMPLING STRATEGY
Watsonville
The upwind samples were taken from a parking lot 300 meters from the
Canada Shoe Building demolition site. The downwind samples were taken from
a parking lot adjacent to the building at a distance of 70 meters. The
478
-------
nearby samples were taken across the street from the building and adjacent
to a debris removal activity from a previously demolished building (see
Figure 2).
Pacific Garden Mall
The upwind samples were taken on the roof of a parking garage on the
perimeter of the Pacific Garden Mall. The downwind samples were taken on
the roof of a building directly across the street from the three buildings
being demolished. The nearby samples were taken out a second story window
of a building adjacent to the three buildings being demolished. Another
nearby sample site was at ground level behind this building and adjacent to
the street access of the debris transport truck route (see Figure 3,
Sampling Locations in Santa Cruz).
Municipal Landfill
The upwind samples were collected at the landfill office. The downwind
samples were taken at the top of the hill where dumping occurred. Nearby
samples were taken at the bottom of the debris hill. Samples were also
collected with pumps mounted on corner posts inside the cab of the bulldozer
which compacted the debris (Figure 4). In addition, the bulldozer operator
wore a personal breathing zone (PBZ) sampler.
SAMPLE METHODOLOGY
Area air samples were collected in a open-faced, three-piece cassette
on 25 mm diameter, 0.45 urn pore-size, mixed cellulose ester membrane (MCE)
filter with a 5 im pore-size MCE backup diffusing filter and cellulose
support pad. Electric sample pumps were operated at about 10 liters per
minute (1pm) from portable gasoline powered generators. Filters were hung
about 5 feet above the ground and facing downward at a 45° angle. The PBZ
sampler and bulldozer cab samples were collected with a battery powered,
flow controlled, personal sampling pump at 1.7 1pm. A precision rotameter
was used to check pump calibration before and after sampling.
ANALYTICAL METHOD
Samples were analyzed by the RREL Transmission Electron Microscope
(TEM) laboratory. The filters were analyzed in accordance with the
non-mandatory TEM method in the Asbestos Hazards Emergency Response Act
(AHERA). The EPA provisional method counting rules were used.
STATISTICAL ANALYSIS METHOD
Airborne asbestos concentrations were compared using a one-way analysis
of variance (ANOVA) along with the Tukey mean comparison procedure. A
constant of 0.002 equal to counting one fiber was added to all counts to
analyze zero values by the ANOVA, i.e., In x +0.002.
479
-------
QUALITY ASSURANCE
Blanks were analyzed to assure quality of sampling media, field
handling procedures, and laboratory handling procedures.
One hundred blanks (5% from the 2000 lot purchase) were analyzed and no
structures were detected. Eight closed and fifteen open field blanks were
collected at the three sites and no asbestos structures were detected.
Laboratory blanks were processed with each preparation of samples and no
structures were detected.
RESULTS AND DISCUSSION
As a general observation, there were visible emissions during the
demolition of buildings even though water was applied by fire hoses for dust
suppression. The primary emission occurred when masonry walls were
demolished. A cloud of visible particulate drifted downwind from the
demolition sites during this phase of demolition. Once the walls of the
structures were collapsed, the fire hoses were much more effective in
controlling visible emissions. In general, the TEM analysis showed a
statistically significant increase in asbestos levels downwind of the
demolition activities.
WATSONVILLE
Figure 5 presents the arithmetic mean airborne asbestos concentrations
at the three sampling locations during the demolition of the building in
Watsonville.
The differences in mean airborne asbestos concentrations between the
samples taken at Main Street during bulldozing and at the parking lot during
demolition as compared to the background site are statistically significant
(p - 0.002).
PACIFIC GARDEN MALL
Figure 6 illustrates the arithmetic mean airborne asbestos
concentrations at the four sampling locations at the Pacific Garden Mall on
the three days of sampling.
On two of the three days, there was no statistically significant
difference in the downwind concentrations as compared to the background
asbestos levels.
There were significant differences in the mean concentrations measured
on November 4th (p = 0.0386). Specifically, the mean airborne asbestos
concentration on the second floor of the Rittenhouse Building (0.022 s/cm3),
adjacent to the demolition site, was significantly higher than the mean
concentration in the Rittenhouse parking lot (0.005 s/cm3) and significantly
greater than background (0.0006 s/cm).
480
-------
SANTA CRUZ MUNICIPAL LANDFILL
Figure 7 illustrates the arithmetic mean airborne asbestos
concentrations measured at the Santa Cruz Municipal Landfill during
bulldozing operations. There were no statistically significant differences
between upwind and downwind asbestos concentrations observed; however, the
samples collected in the personal breathing zone of the bulldozer operator
showed a mean concentration (0.06 s/cm3) that was significantly greater than
the mean background concentration (p = 0.03). Periodically, the operator
walked through and handled the debris. Hence, this activity may have
contributed to the elevated personal breathing zone concentrations.
CONCLUSIONS
Visible emissions were observed during the structural collapse of
buildings and were generally not apparent when firehoses were used to wet
the debris during loading operations; however, asbestos levels during the
debris handling were elevated above background (i.e., statistically
significant) even though there were no visible emissions.
These limited data support the premise in NESHAPS (Proposed Rules,
January 10, 1989; 40CFR Part 61, page 925) that the absence of visible
emission is not sufficient evidence to assume no fugitive particulate
emission.
Operators of bulldozers involved in landfill ing of asbestos-containing
debris can be exposed to personal breathing zone concentrations above
asbestos fiber background level.
481
-------
V*4.
20 Miles
N
Watsonville
v<..
Moss Landing
Earthquake Epicenter,
October 17, 1989
San Andreas Fault
Monterey
Figure 1. Geographical location of Earthquake.
482
-------
BACKGROUND SAMPLES
A
N
WELLS
FARGO
BANK
L FOX
ITHEATRE
1
* * *
J
UJ
UJ
CC
2
Q.
*
•
s
PARKING kS
LOT s 300 meters
s
MAIN STREET^'
DIRECTION
/
///^//^/
//CANADA'X
y SHOES
/
i—
UJ
UJ
CO
CO
CC
U-
RODRIGUEZ STREET
" /^~
— 74
0 Demolition s'rte
D Previously demolished site
• Sampling location
Figure 2. Location of sampling sites during the demolition of a building in Watsonville, California,
483
-------
WIND
DIRECTION
Demolition sites.
Denotes sampling locations.
Figure 3. Location of sampling sites at the Pacific Garden Mall in Santa Cruz, California.
484
-------
4
N
EQUIPMENT
BUILDING
SAMPLE
LOCATIONS
Figure 4, Location of sampling sites during the landfilling of demolition debris
at the Santa Cruz Municipal Landfill.
485
-------
Average Asbestos Concentration, s/cu cm
0.06-1
0.05-
0.04-
0.03-
0.02-
0.01 -
Background Main Street
.. °urlf!9
demolition
Main Street Parking lot
. gjjrlng . during
bullaozrng demolition
Figure 5. Average asbestos concentrations at Watsonsville.
Average Concentration, s/cm3
11/03/89
11/04/89
11/06/89
!• Background Kmi Rittenhouse parking tot
i I Rittenhouse second floor I I Oddfellows roof
Figure 6. Average airborne asbestos concentrations during
building demolition at the Pacific Garden Mall in
Santa Cruz, California.
486
-------
Average concentration, s/cm3
0.07 i
0.06-
0.05-
0.04-
0.03-
0.02 -
0.01 -
Background
Upwind Downwind
(Adjacent to debris)
Bulldozer cab
Personal
breathing
zone
Figure 7. Average airborne asbestos concentrations during
the landfilling of demolition debris at the Santa
Cruz municipal landfill.
487
-------
ASSESSING THE RISK OF REMEDIAL ALTERNATIVES AT SUPERFUND SITES
IMPLICATIONS FOR TECHNOLOGY DEMONSTRATORS
by: Patricia Lafornara
Releases Control Branch
Edison, NJ 08837-3679
ABSTRACT
Although the basics of risk assessment are familiar to those who must
administer the cleanup of Superfund sites, frequently technology developers
have little knowledge of information needs in this area. This paper
describes the users and use of risk assessment at sites in order to improve
technology demonstrators' understanding of these needs and make them aware
of potential data requests.
This paper has been reviewed in accordance with the U.S.
Environmental Protection Agency's peer and administrative review
policies and approved for presentation and publication.
INTRODUCTION
The importance of risk assessment in the selection of remedy process
at Superfund sites has many implications for technology demonstrators.
These implications will be related through a description of the use of risk
assessment in the process, a discussion of the users of the risk
information, and the information that technology developers can provide.
RISK ASSESSMENT IN THE SUPERFUND PROCESS
The introduction of a remedial technology to a Superfund site follows
the formal selection of remedy process. Information is gathered during the
site Remedial Investigation and Feasibility Study (RI/FS). Based on
information from the RI/FS process, the remedial project manager (RPM)
makes a selection of a remedy that is ultimately documented in the Record
of Decision (ROD).
It should be pointed out that the formal process of remedy selection
leading to a ROD is not the only way remedial actions are employed in
Superfund. "Interim actions" are taken at remedial sites when such an
488
-------
action is helpful in controlling contamination. These actions are taken
when time is critical, and may occur before the RI/FS process can be
completed. An example is the pumping of ground water to contain
contaminant plume migration.
For a ROD, the remedy is selected using nine criteria established in
the National Oil Spill and Hazardous Waste Contingency Plan (NCR). These
are: the threshold criteria -- (1) overall protection of human health and
the environment, (2) compliance with applicable or relevant and appropriate
requirements (ARARs); the balancing criteria -- (3) long-term effectiveness
and permanence, (4) reduction of toxicity, mobility, and volume through
treatment, (5) short-term effectiveness, (6) implementability, (7) cost;
and, the modifying criteria -- (8) state acceptance, and (9) community
acceptance. Except for the cost criterion, these criteria can involve
information from the risk assessment process,
A baseline risk assessment is part of the RI/FS process to determine
the risk if no action were taken. This assessment calculates risk to human
health and the environment through the assessments of hazard and exposure.
For human health, calculations are made for cancer and noncancer endpoints.
For environmental effects, other endpoints may be used. These methods are
detailed in the Risk Assessment Guidance for Superfund, Volume I -- Human
Health Evaluation Manual, Part A (1), and Volume II -- Environmental
Evaluation Manual (2). These documents discuss the current site risks, but
not the development of remediation goals or assessment of risks associated
with remedial actions. These topics are being addressed in guidance now
under development as Parts B and C, respectively. Some of the topics to be
included in Part C will be discussed here.
For emergency responses, especially in light of new land disposal
restrictions, various remedial technologies are now receiving more
consideration for controlling pollution at sites along with the traditional
removals. Risk assessment can assist in evaluating remedial alternatives.
Risk assessment may have another role in the five-year review. These
reviews are required where hazardous materials have been left on site as
part of the remedy. How these reviews will be carried out is under
development at this time.
THE USERS OF RISK INFORMATION
Before discussing the content of the risk information that technology
developers could provide, it is helpful to know who the audience is for
this information. The RPM has been mentioned. However, as the remedy is
selected, it has to be approved by both the regional management up to
Administrator and accepted by the public. The importance of community
acceptance is not to be underestimated. In most cases, the community's
list of concerns is topped by health effects.
For the RPM, the regional risk assessor or toxicologist is a resource
person who evaluates and interprets site risk information. They may be
called upon as expert witnesses to defend EPA actions in cases involving
489
-------
potentially responsible parties (PRPs). (Recently, a decision was made
that EPA would take over the assessment of risk at PRP sites.) The risk
assessors are committed to ensuring that the NCP-specified risk levels are
met at sites. The risk assessor may have a major role in assessing the
attainment of cleanup goals at remediated sites at delisting or five-year
review. In addition, risk assessors help the RPM find ways to offer the
community maximum protection during remedy implementation. Risk assessors
can help the community understand the balance between a choice of remedies
that have some probability of failure weighed against those with long-term
risk.
Besides risk assessors in regional offices, there are technical
experts in the specialized areas of ecological effects and air pathways.
These specialists help the RPM to interpret site information, or help to
advise what data are needed to assess impacts. There are ecological or
biological technical assistance groups (known as BTAGs) set up in most
regions. These groups have members with field experience from agencies
like the National Oceanic and Atmospheric Administration, U.S. Fish and
Wildlife Service, and state departments of the environment or natural
resources. They review site documents and provide advice on current and
predicted ecological effects at sites. Air/Superfund coordinators assist
with issues relating to air emissions from sites before and during remedy
implementation.
An additional, expanding audience for risk assessment information on
technologies is the technology transfer community. In order to facilitate
acceptance of new technologies, information on the associated risks is
becoming increasingly important.
TYPES OF AND RATIONALE FOR RISK INFORMATION
The Criterion of Long-Term Effectiveness and Permanence
The type of information needed to assess this criterion is either
extremely easy or extremely difficult to ascertain. There are two
components, both reduction of risk and probability of remedy failure.
Remedial technologies that result in complete destruction or separation of
contaminants from site media are effective and permanent, and result in
complete reduction of risk. These technologies incorporate chemical
processes, extraction, separation, bioremediation, etc. Their
effectiveness is contingent upon the extent to which the contaminant is
removed, provided that no new contaminants are added or created in the
process.
The health-based target concentration for remediation is based upon
the results of a risk assessment, or actually running the risk assessment
"backwards". The remediation goal is calculated based on a site-specific
cancer incidence risk of one-in-a-million to one-in-ten thousand or a non-
cancer hazard index of one. The individual contaminants are considered in
these calculations and the risk from several contaminants or pathways may
be summed. Reasonable maximum exposure is typically figured for an
individual as if the site were to be residential. The exposure point
490
-------
concentration is then derived from exposure pathways and toxicity data to
achieve a protective level. The Superfund program use EPA's IRIS data base
and also publishes toxicity data quarterly in the Health Effects Assessment
Summary Tables (HEAST) (3), available from the Environmental Criteria and
Assessment Office (ECAQ) in Cincinnati. These numbers can and do change
based on research results. In some cases, what was thought to be an
acceptable concentration could be found to be unprotective in the future.
Some remedial technologies have a finite reliable lifetime. Not only
is there the risk to consider, but also the probability that failure will
occur. Capping and containment technologies are examples. Choosing
between this type of remedy and one that completely removes the contaminant
cannot be done through simple comparisons of calculated risk. Where
available, engineering estimates of the lifetime of the remedy (including
probability of failure), in addition to the expected fate of the
contamination upon failure, is helpful in assessing the long-term risks
involved.
This long-term effectiveness and permanence information is
particularly valuable in light of the need for community acceptance. A
current example is in Region 7, where a Missouri community is steadfast
against the incineration of dioxin wastes, even though the Agency considers
incineration to be the only acceptable treatment alternative (4). The
community is apparently not accepting the Agency's assessment that
permanent remedial strategies are more desirable than others with a lesser
implementation risk.
Many remedies are complicated by edging the middle ground. Some
technologies involve intermedia transfer of contaminants or the production
of new contaminants through reaction and transformation. These types of
transfers and the production of new contaminants are under new scrutiny in
long-term effectiveness assessments. Usually the assessment is
qualitative, based on the toxicity of the contaminants, the medium
involved, and the characteristics of further treatment.
An example of contaminant transfer is in soil washing, where
contamination is transferred from the feed soil to the wash water. This
may result in a long-term problem depending upon the amount of residual
contamination or nature of the wash water treatment. In addition, the soil
fines separated from the wash water may require further treatment.
In incineration, contaminants can be transformed into other
contaminants altogether during quenching of hot gases, cooling of the ash,
or in the treated medium. All products and waste streams must be handled
further. For example, the ash then must be treated or stored, possibly
becoming a long-term problem. Also, other changes, like the lowering of
soil pH and destruction of organic matter, may have occurred in the treated
medium. In all technologies where contaminants are removed to other
locations, there is increasing scrutiny of the risks associated with both
the transport and redisposal of treated media and treatment products.
491
-------
Setting and Attainment of Remediation Goals
Data on the criterion of long-term effectiveness and permanence are
gathered at a point of compliance. The measure of attainment of the
remediation goal must be expressed in a sampling and analysis strategy. In
all long-term effectiveness assessments, it is important that the
remediation goals be worked out jointly with a risk assessor. The risk
assessor provides exposure pathway analysis. For example, the RPM and risk
assessor may need to establish the depth of soil treatment to protect
residents from excessive exposure. Goals for transformation products or
reagent residues could be set, as these would not be covered by preliminary
remediation goals that only addressed site contaminants and pathways in the
baseline risk assessment.
The technology demonstrator must consider how the treated medium
should be sampled to assure protective levels. This would include the
sampling frequency and any strategies such as compositing. Also, it may
help to make an assessment of the applicability of measurements like
destruction efficiency and process operating parameters in lieu of residual
concentration measurements. In addition, any waste streams must be tested
if they are to require storage or will need to meet standards like land
disposal restrictions.
Reliable data are being required from the treatability study to full
scale. Superfund has recently released a document that relates the quality
of data with its potential usefulness in risk assessment applications. It
is entitled Guidance for Data Useability in Risk Assessment (5), released
in October 1990.
Although strategies for compliance with remediation goals are common
for most RODs, the five-year review process may require more formal plans
to be available at the time of remedy selection. It is possible that
technologies that will necessitate the five-year review will be more
eagerly received if such a strategy is already in hand.
The Criterion of Short-Term Effectiveness
This criterion, covering what happens during remediation, is the most
frequently identified risk assessment concern for remedies. The exposed
receptors can be the community, site workers, or the environment. Although
the most common exposure pathway is through the air, the remedy may affect
multiple pathways. Partial or total habitat destruction during
implementation is a potential ecological effect.
Although remedy failure was discussed above, it is also a concern
here. Remedy failure as it relates to the short-term effectiveness
criterion deals with failure during implementation, which could include
trucking accidents involving hazardous materials, reagent spills, or other
mishaps.
Some common short-term effects are well-documented. Excavation
emissions are the most studied (6). In some cases, steps can be taken to
492
-------
reduce exposure that involve only minor changes in procedures. For
example, to avoid contaminated dust exposure in a residential area,
monitoring wind direction and spraying dry soil with water can be employed.
In some cases, like the Wide Beach site in New York, other factors can be
optimized. At that site, excavation began only after summer residents had
left. The entire operation was winterized and lighted for 24-hour
operation to permit the work to finish before the residents returned. For
the protection of permanent residents, air monitors were installed at the
construction fence line.
Other impacts are not as easily mitigated. In situ soil treatments,
for example, can lead to migration or transfer of contaminants or treatment
chemicals to uncontaminated media. This type of problem is largely
regarded as an engineering concern, addressed by the criterion of
implementability. However, assurances that impacts can be controlled are
now required before implementation in order to adequately address the
short-term effectiveness and community acceptance criteria. An exposure
analysis of all new pathways with all contaminants involved is needed. On-
site monitoring is required for any air releases, and may be required to
check other media as well.
For ecological effects, the BTAGs can help plan ecologically sound
remedy implementation. The environment of each site is unique. For
example, excavation of soils and sediments, or even bringing in excavating
equipment, can totally destroy habitats that are not easily replaced.
Region 3's Bioassessment Group was able to suggest an alternative way of
implementing a remedy to save a valuable hardwood wetland at a site.
Instead of the heavy equipment planned for use on site, wheelbarrows and
shovels were substituted, and the workers were shown the least disruptive
route to the contamination.
Summaryof Risk Information Needs From Technology Demonstrators
The following is presented as a quick checklist for information needs
to support eventual full scale implementation at sites. It is divided into
two parts to support the risk assessors concerns for both long-term
(chronic, seven years to a lifetime) and short-term (subchronic, two weeks
to seven years; and acute, less than two weeks) exposure. These divisions
correspond, usually, to risk following implementation and during
implementation, respectively.
Information for long-term assessments--
1) list of residual contaminants including transformation/reaction products
in the treated medium and contaminants transferred to other media, with
concentrations
2) long-term treatment/disposal concerns that include the probability of
remedy failure at some point in the future
3) receptor population/exposure pathways if different from the baseline
risk assessment
4) input on sampling and analysis plan for attainment of remediation goals
and a five-year review
493
-------
Information for short-term assessments--
1} contaminant release points during treatment, from excavation through
final disposition of residuals
2) estimates of duration and concentration of these releases
3) strategies for monitoring releases
4) transformation/reaction product concentrations in all streams
5) transportation information for equipment and hazardous materials
6} site preparation/restoration impacts
7) accident/failure probabilities and impacts
CONCLUSIONS
The following is a "short list" of guiding principles for technology
developers to promote environmentally safe and sound remediation of sites.
The developer should attempt to:
o Control contaminant fate (including the fate of introduced
contaminants). It is important to know what happens within and outside
of the technology black box.
o Make the remedy low in risk and low in probability of failure.
o Have a plan for long-term assessment of attainment of goals.
o Consider ecological issues, not just human health impacts.
o Seek community acceptance and address residents' concerns.
In general, the most desirable remedies are those that have little
potential for releases of contaminants at any time during the remedial
action. Remedies that do not disturb sensitive environments, or that
either destroy contaminants completely or permanently transform them into
non-toxic compounds are ideal. If the remedy does this with non-toxic
reagents, so much the better.
Most remedies, of course, will not meet the desired conditions. The
technology developer, however, must consider how to ameliorate the
resulting impacts. If the technology is to be part of a treatment train,
all of the issues must be addressed for all of the cars. Assessing the
risks before and after controls may soon become the responsibility of the
developer, not the user. As the public and risk assessors become aware of
the risks associated with certain technologies, replacement technologies
will be sought accompanied by assurances from the developer that risks are
low.
REFERENCES
1. Office of Emergency and Remedial Response. Risk Assessment Guidance
for Superfund Volume I -- Human Health Evaluation Manual (Part A).
EPA/540/1-89/002. U.S. Environmental Protection Agency, Washington,
D.C., December 1989.
2. Office of Emergency and Remedial Response. Risk Assessment Guidance
for Superfund Volume II -- Environmental Evaluation Manual. EPA/540/
494
-------
1-89/001. U.S. Environmental Protection Agency, Washington, D.C.,
March 1989.
3. Environmental Criteria and Assessment Office, Health Effects
Assessment Summary Tables. OERR 9200.6-303 (90-3). U.S. Environmental
Protection Agency, Cincinnati, Ohio, July 1990.
4. HMCRI Focus. Article, September 1990, p. 2.
5. Office of Emergency and Remedial Response. Data Useability in Risk
Assessment. EPA/540/G-90/008. U.S. Environmental Protection Agency,
Washington, D.C., October 1990.
6. Office of Air Quality Planning and Standards. Air/Superfund National
Technical Guidance Study Series -- Development of Example Procedures
for Evaluating the Air Impacts of Soil Excavation Associated with
Superfund Remedial Actions. EPA 450/4-90-014. U.S. Environmental
Protection Agency, Research Triangle Park, North Carolina, July 1990.
495
-------
CLOSURE OF A DIOXIN-CONTAMINATED SUPERFUND SITE
Alan Sherman and James P. Stumbar
Foster Wheeler Enviresponse, Inc.
Edison, NJ 08837
Joyce H. Perdek, Frank J. Freestone
USEPA Risk Reduction Engineering Laboratory
Edison, NJ 08837
ABSTRACT
The report summarizes the closure portion of the field demonstration
activities of the U.S. Environmental Protection Agency's (EPA) Mobile
Incineration System (MIS) at the Denney Farm site in southwest Missouri.
Sponsored by the EPA's Office of Research and Development, Office of Solid
Waste and Emergency Response, and Region VII, the field demonstration,
which began in October 1985, was completed with the certified clean
closure of the site in June 1989.
Over a four-year period, the EPA Mobile Incineration System, operating
at the Denney Farm site, treated more than 12.5 million pounds of
dioxin-contaminated wastes from eight southwestern Missouri sites. At the
conclusion of operations, the site soils, equipment, and buildings were
decontaminated following approved closure plans. The closure operation
and the closure certification process, detailed in this report, represent
the culmination of the project.
DISCLAIMER
The information in this document has been funded by the U.S. Environmental
Protection Agency under Contract No. 68-03-3255 under the sponsorship of
the Office of Research and Development. The document has been reviewed in
accordance with U.S. Environmental Protection Agency policy and approved
for publication. Mention of trade names or commercial products does not
constitute endorsement or recommendation for use.
496
-------
INTRODUCTION - CLOSURE PLANS AND CLEANUP CRITERIA
Under the sponsorship of EPA's Office of Research and Development, the
EPA Mobile Incineration System (MIS) was designed and constructed to
demonstrate high-temperature incineration of hazardous wastes [1, 2].
Shakedown of the unit took place at the EPA Edison, New Jersey facility.
After a successful demonstration of the unit's ability to incinerate PCBs
and other chlorinated organic liquids, the MIS was modified to handle
solids and brought to the Denney Farm site in southwest Missouri. The
full-scale field demonstration of the unit was sponsored by EPA's Office
of Research and Development, Office of Solid Waste and Emergency Response,
and Region VII. The field demonstration began in October 1985 and
culminated with the certified clean closure of the site in June 1989.
This report details the closure operation and closure certification
process.
After the MIS completed incineration of the dioxin-contaminated
materials from eight area sites, the unit was closed by executing approved
closure plans [3, 4]. Because incineration activities and site activities
were conducted by different parties and were covered under two separate
permits, two separate approved closure plans were required by the
regulatory authorities, one for the MIS and one for the site. Closure of
the MIS followed the plan presented in the 1987 final permit [3]. The
procedures in this plan were updated, however, to include greater detail.
These detailed procedures were included in the final closure plan for the
Denney Farm site [4]. In addition to this information, the site closure
plan included sampling procedures for soil, buildings and equipment, scope
of the decontamination work, and action levels to trigger decontamination
work.
Closure of the MIS and Denney Farm site, which took place from January
24, to June 30, 1989, involved decontamination of the MIS, Denney Farm
soil, and supporting equipment and buildings. The cleanup criteria for
the MIS and the site, included in the final site closure plan [4], are
presented in Table 1. The Agency for Toxic Substances and Disease
Registry provided health advisories used to develop these criteria, and
the criteria were approved by the Missouri Department of Natural
Resources.
The closure activities at the Denney Farm site provided a benchmark
example for clean closure operations at other incinerator sites. The
closure activities at Denney Farm were all encompassing and included
decontamination of the site soil, the hot-zone buildings, equipment used
at the site, and the incinerator itself with all of its ancillary
systems. Following state- and federally-approved closure plans, the
closure operation demonstrated that such activities could be completed
successfully, and in a timely and efficient manner.
CLOSURE OPERATION
Closure activities included excavation and incineration of dioxin-
497
-------
contaminated soil, decontamination or incineration of contaminated debris,
decontamination of buildings and equipment, and disassembly of buildings
and the MIS. Closure activities are summarized below.
Almost 3,000,000 Ib of dioxin-contaminated soil and debris were
incinerated. The debris included: the loading dock, HEPA filter and
associated ducting, wooden pallets, metal drums, and plastic sheeting used
to protect excavated areas. The four buildings in the hot zone, which
were constructed of sheet metal walls, wooden supports, and concrete
foundations, were decontaminated. The following equipment was
decontaminated: the four rented pieces of soil moving equipment used in
the hot zone, including a backhoe, a forklift, and two small front-end
loaders (Bobcats1*); and the MIS feed system, consisting of three feed
conveyors, the shredder, weigh chute, weigh scale, ram, ram trough, and
Hapman* conveyor. The rubber conveyor belts were incinerated as part of
the decontamination process.
Equipment located outside the hot zone was cleaned and wipe tested to
ensure that it was not contaminated. Wipe testing of the MIS building
located outside the hot zone showed that the walls were not contaminated.
However, the floor beneath the feed system had to be decontaminated. The
17 pieces of rental equipment located in the safe zone, consisting of two
air compressors, eight trailers serving as offices and crew quarters, four
storage trailers for spare parts and other materials, and three tanks for
wastewater storage, were cleaned and wipe tested. The MIS equipment was
also cleaned and wipe tested. This equipment included the three trailers
on which the major components of the MIS were mounted, the air pollution
monitoring trailer, HEPA trailer, electric generators, Monarch1* CPI
separator, cyclone, WEP, air dryer, water softener, four wastewater
storage tanks, process water storage tank, caustic storage tank, three
discharge water holding tanks, and several pumps.
Final MIS closure activities consisted of dismantling the unit and
preparing its components for shipment and storage at the EPA facility in
Edison, New Jersey. Final Denney Farm closure activities included
disassembly of the buildings and final grading of the site. The site was
covered with a minimum of one foot of soil and seeded. Closure of the MIS
and site were certified by an independent professional engineer registered
in the State of Missouri.
SOIL, BUILDING AND EQUIPMENT DECONTAMINATION
The extent of contamination was determined through preclosure sampling
and analysis conducted in November and December 1988. Soil samples,
samples from various parts of the buildings, and samples from the
equipment were taken and analyzed to determine the extent of contamination
resulting from activities at the site. The samples were analyzed using
approved test methods for solid waste (SW-846) for dioxins/furans, PCBs,
organics, and metals. The results of the analyses were compared with the
action levels for the site presented in Table 1.
498
-------
TABLE 1. ACTION LEVELS OR CLEANUP CRITERIA
FOR DENNEY FARM CLOSURE
Contaminant Surface Soils Subsurface Soils
(0 to 3 in. (>3 in. depth)
depth)
Buildings(a)
(Wood Framework/Sheet
Metal Siding/Foundations)
2.3,7.8-TCDD
ppb ' <10 ppb
+ 12-in soiI cover
or maximum 4-ft
excavation + 4-ft
soil cover
10 ng/m2 (1 pg/cm2)
or >10 ng/m + Sealant
PCBs
Volatile/
Semivolatile
Heavy Metals
<2 ppm <10 ppm
+ 12-in soil cover
<50 ppm <50 ppm
non-E.P. Toxic non-E.P. Toxic
100 ng/cm
or >1 ug/cm + Sealant
no significant
contamination
no significant
contamination
(a) Foundations had an additional criterion. Core samples were taken and the dust from
the samples was analyzed and had to show concentrations of 2,3,7,8-TCDD <10 ppb.
499
-------
Soil Sampling and Decontamination
Sampling was conducted prior to the start of site closure activities
to determine the extent of contamination of the site soil. This sampling
and analysis was primarily for dioxin using a statistical procedure to
guarantee that the dioxin levels reported were within 95% confidence
limits [5]. This means that there is a 5% chance that the actual
concentration exceeds the maximum concentration obtained from the
statistical analysis of the analytical results. In addition to employing
this statistical sampling and analysis method, grab samples were taken and
analyzed for dioxin and the other constituents listed in Table 1. The
results of this sampling are presented in Tables 2 and 3. The sample
locations are depicted on Figure 1.
The statistical dioxin sampling and analysis method is described
below. A site is divided into sections. Each section is divided into 50
equal areas, and three aliquots are taken from each area. The aliquots
taken correspond to locations 2, 4, and 5 as shown in Figure 2. The 50
aliquots from location 2 are composited to form one sample. The same is
done for the aliquots from location 4 and location 5. Therefore, a total
of three composite samples are analyzed. The results of the three
composite samples are combined statistically to arrive at the 95%
confidence limit for the section (see footnote a, Table 2).
The preclosure sampling at Denney Farm for dioxin was executed by
dividing the site into 16 sections of approximately 5000-ft^ areas
consisting of 6 sections inside the hot zone and 10 sections outside the
hot zone as shown in Figure 1. Dioxin contamination was found in the hot
zone as expected and was also found in the 10 outside sections. At the
beginning of closure activities in January 1989, the 10 outside sections
were divided into 2400-ft^ sections to further delineate the
dioxin-contaminated area (Figure 3). In addition, another 16 sections
outside the hot zone were added during the closure process to determine
the extent of the contamination. The final allocation of areas with their
levels of contamination is shown in Figure 3. The areas with dioxin
concentrations greater than the action level were designated as
contaminated areas and were targeted for remediation.
Contaminated soil was scraped in layers of at least three inches and
incinerated. After each scraping, the underlying soil was sampled and
analyzed for dioxin. Dioxin was selected as the indicator chemical for
the soil decontamination/excavation process due to its prevalence at the
site and the stringent action level for it. Scraping and incineration
continued until sampling and analysis showed that the contamination of the
remaining soil was below the action level for dioxin. Once an area was
below the dioxin action level, it was sampled and analyzed for the other
constituents listed on Table 1. Decontaminated areas were covered
temporarily with plastic sheeting to prevent the spread of contamination
from contaminated areas until the excavation was completed and the entire
site was covered with clean soil in accordance with the cleanup criteria.
The dioxin concentrations remaining after excavation are indicated in
Figure 4.
500
-------
x 2.94 ppB
\
L
:
= =
x-
;
n
j
1 — '
u
i r
r
h
i
1
J
L
6
4
C
C
c
c
n
U
59
D
r
D
n
—\
I
1
P
1
spb
3m Pi
a
1 /
CB
1
CD
Q
O
227 p
16 pp
O
C
PC
m
4
D
PCB
_
/ ppo
Q ppm
7
1'
1
MDR x
>VIUD C
7.75
1.6
16.8 ppB
190 ppm
PCB
.69 ppb
1.7 ppb
1.6 ppb
D.3 ppb
1.43 ppB
2.49 ppb
5
PPb
ppm PC!
PCB
f
EWI
16
C
5
3
0
-
6
P
]
\
3
1
l
1
:
3 PPO '
pm PCB
\
V
1
«_J
_.
1
L
\
\
\
\
\
-K-K-K
DSB
MDB-1
MDB-2
EWB
12.7 ppo
• Hot Zone Fence
Drum Storage Building
Hazardous Waste Storage Building
Hazardous Waste Storage Building
Equipment Washdown Building
O*— Not Detected
o Grab Sample Location (See Table 62)
Data = 2,3,7,8-TCDD concentrations
PCB = PCB concentrations
Figure l Oenney Farm Site Preclosure Sampling Results
November and December 1988
501
-------
TABLE 2. INITIAL PRECLOSURE SOIL SAMPLING RESULTS [November 1988]
Section No.
Sampling Date
2,3,7,8-TCDD(a)
(ppb)
Chromium(b)
(ppm)
PCBs(c)
(ppm)
Perimeter Sections(d)
1 SW
2 SE
3 NE
4 NW
5 Ash
Storage
11-7-88
11-7-88
11-7-88
11-7-88
11-7-88
69
14
33
10
2.94
16
11
19
13
200
160
ND
ND
ND
ND
Hot Zone Sections
6 SW
7 NW
8 SW-NE
9 Trench
10 NE
11 SE
11-7-88
11-7-88
11-7-88
11-7-88
11-8-88
11-8-88
127.44
659.00
227.37
16.77
7.75
6.68
30
4.7
8.5
13
7.8
15
240(e)
40(f)
16
190
1.6
ND
(a) All 2,3,7,8-TCDD concentrations are at the 95% Upper Confidence
Limit (UCL):
Ci = X + 2.92 S/1.731
where:
£.j = Maximum concentration of contamination at 95% UCL.
X - Mean concentration of three composite samples.
S = Standard Deviation of three composite samples.
(b) Chromium is total chromium, not EP toxicity chromium. The levels
detected were considered to be safe.
(c) PCBs expressed as Aroclor-1260.
(d) SW=southwest, NW=northwest, SE=southeast, NE=northeast.
(e) Also contained 200 ppm tetrachlorobenzene and 30 ppm
hexachlorophene.
(f) Aroclor-1016.
Note: Other organics were detected at low levels (<10 ppm).
Action levels: See Table 1.
502
-------
TABLE 3. FOLLOW-UP PRECLOSURE SOIL SAMPLING
RESULTS [December 1988]
Sample Type TCDD Concentration (ppb)
I. 95% UCL Soil Samples (0 to 2 in.)
1. Section 12 14.69
2. Section 13 11.61
3. Section 14 0.31
4. Section 15 1.43
5. Section 16 2.49
II. Depth Samples in Section l(a)
1. 0 to 3 in. 10.49
2. 3 to 6 in. 3.00
3. 6 to 9 in. 3.06
4. 9 to 12 in. ND(b)
III. Depth Samples in Section 6(a)
1. 0 to 3 in. 119.66
2. 3 to 6 in. 3.21
3. 6 to 9 in. 3.51
IV. Depth Samples in Section 8(a)
1. 0 to 3 in. 10.13
2. 3 to 6 in. 2.01
3. 6 to 9 in. 0.67
4. 9 to 12 in. 1.25
V. Grab Samples in Section 7 (0 to 2 in.)
(10 aliquots each)
1. Around waste oil containment 11.23
2. Underneath HEPA duct 8.23
3. Between MDB-1 and DSB 3798
VI. Wooded Area South of Trailers (0 to 2 in.)
(5 to 7 aliquots)
1. Soils south of the site ND(b)
2. Runoff behind trailer #5 12.74
(a) Grab samples
(b) ND = not detected
503
-------
50 SOIL SAMPLES COMPOSITED
TO FORM COMPOSITE 2
Figure 2 Systematic Sampling Design for Obtaining
the 3 Composite Samples.
504
-------
0,84 ppb
FACIL-ITY LAYOUT
MISSOURI FIELD DEHONSTRATIOH
OF U.S. E?A
MOBILE INCINERATION SYSTEM
AT OENNEY FARM SITE
Data = 2,3,7,8-TCDD concentrations.
U = Not detected at 0.300 ppb detection limit.
Figure 3 Final Allocation of Areas With Their
Levels of Qioxin Contamination.
505
-------
I \
nti. EMMPI.
IP
3.94 «JB ! :!2 ; o.?s BPO | ^B i°;|j
FACILITY LAYOUT
MISSOURI F'ELO DEMONSTRATION
OF U.S. EPA
MOBILE INCINERATION SYSTEM
AT OENHEY FARM SITE
Site covered by i minimum of 1 foot of clean soil.
Data » 2,3,7,8-TCDD concentrations.
U - Not detected at 0.300 ppb detection level.
Figure 4 Dioxin Levels at Denney Farm Remaining
After Excavation and Site Closure.
506
-------
Buildings and Equipment Sampling and Decontamination
Wipe sampling of the sheet metal walls, concrete foundations, and wood
framework of the buildings was conducted in order to determine the level
and extent of contamination in the buildings. The metal, wood, and
concrete were each wipe tested separately. Coring samples of the
foundations were also taken. The methods used are described below.
Wipe samples were taken using 3-in. by 3-in. sterile gauze pads that
were soaked with isooctane. The sampling procedure consisted of wiping
several areas over a designated portion of the building with the gauze so
that the total area wiped was 2500 cm2. The gauze was placed in an 8 oz
jar. This procedure was used to collect samples for metals,
organics/PCBs, and 2,3,7,8-TCDD analyses.
Concrete floor dust samples were collected from random locations
throughout the buildings by using an impact drill with 1-in. carbide
tipped bits to obtain dust from the entire depth of concrete. The dust
was composited into three 8-oz jars for metals, organics/PCBs, and
2,3,7,8-TCDD analyses.
«
Buildings were decontaminated by scrubbing with brushes using a
detergent solution and rinsing with high pressure water or steam
cleaning. This was preceded by scraping when necessary. The
decontamination process was repeated until sampling and analyses showed
that residual contamination was below the action levels.
All building materials were decontaminated to the action levels
indicated on Table 1 except for some of the wood. The contaminated wood
was incinerated. The buildings were disassembled and removed from the
site. The foundations remained and were covered with a minimum of one
foot of clean soil.
All equipment was decontaminated by scrubbing with brushes using
detergent solution, and rinsing with high pressure water or steam
cleaning. The process was repeated until sampling and analyses showed
that the residual contamination was below the action levels. It should be
noted that although the MIS closure plan [3] stipulated an equipment
cleanup level of 10 pg/cnr for dioxin, the equipment was cleaned to meet
the more stringent cleanup level of 1 pg/cnr indicated in the site
closure plan [6].
Wastewater generated during the decontamination process was passed
through a filter train consisting of 50 micron and 20 micron paper
filters, a sand filter, and two activated carbon filters. The filters
were mounted in series. The spent filter materials were incinerated
periodically. All water was discharged in accordance with the NPDES
permit for the site.
507
-------
Soine contaminated material was generated after the incinerator was
shut down for dismantling. This material included paper filters, carbon
filters, sludge from wastewater tanks, and floor sweepings. This material
was put into marked and sealed drums and sent to a permitted storage
facility.
DISASSEMBLY AND TRANSPORT OF INCINERATION SYSTEM
The disassembly and transport strategy for the MIS components depended
upon whether the equipment was rented, trailer-mounted or mounted in other
ways. Generally, the decontaminated components of the MIS were
disassembled only to the degree that all elements could be mounted on
over-the-road equipment and transported back to the EPA Edison, New
Jersey, facility. The non-rented equipment required 21 trailers for
transport. The disassembly and transportation strategies are discussed
below.
The rental equipment, listed above in the closure summary, was
returned to the appropriate vendors after being restored to its "as
supplied" condition.
•
Many of the major components consisting of the kiln, secondary
combustion chamber (SCC), MX scrubber, and flue gas analyzers were mounted
on seven trailers. Disassembly of the trailer-mounted units was conducted
as follows: The kiln and SCC were deslagged. The equipment was cleaned
and painted. The front one-third of the SCC was removed and placed on
another trailer to bring the SCC trailer within highway weight
restrictions. All service piping and three-phase cables were
disconnected, plugged, capped, and tagged. All instrumentation was
protected and sealed. The trailers were inspected, serviced, and prepared
for road operation.
The remaining equipment, consisting of the wet electrostatic
precipitator (WEP), cyclone, CPI separator, pumps, piping, hoses,
interconnecting electrical cable, storage tanks, etc., were loaded on flat
bed trailers after cleaning and disassembly.
The WEP and cyclone required special handling. The electrode
assemblies were removed and packed in separate crates, and the WEP was
filled with styrofoam packing. The support frame of the WEP was modified
to allow shipment of the unit within its frame. Cross-beams to which the
platforms were attached were replaced with shorter beams to bring the unit
within the permissible shipping clearances. The WEP was loaded on a flat
bed trailer and shipped in a horizontal position. The cyclone also was
shipped completely within its support frame. Preparation of the cyclone
for shipment consisted of removing the spring supports and replacing them
with rigid links. Shims were then welded between the unit and its
vertical guides.
The process water system and electrical system were disassembled and
inspected. All steel and plastic piping was scrapped because of reduced
508
-------
wall thickness or internal scale buildup. Only rubber hoses and
trailer-mounted piping for the process water system were salvaged. All
electrical cables were coiled and placed on trailers for shipping.
All remaining equipment was disassembled, placed on pallets, crated,
as required, and loaded onto flat bed trailers for shipment.
LESSONS LEARNED . , . .
Several important lessons were learned from this closure experience
that could make future closure operations of this nature more efficient.
These lessons are discussed below.
The hot zone at Denney Farm was separated from the surrounding area by
a fence, and the buildings holding the hazardous wastes were designed for
spill control. However, some materials handling occurred outside the
buildings creating the potential for contamination of hot zone soils.
Contamination in the hot zone, which was at a higher elevation than the
rest of the site, created the potential for contamination to migrate
outside the hot zone during periods of heavy rainfall. In fact, the
spread of contamination outside the hot zone boundary was discovered
during closure. A narrow plume of contaminants migrated out of the hot
zone and was carried along by stormwater runoff into the adjacent wooded
area. The discovery of this plume necessitated additional sampling and
analysis work during closure to define the area of contamination, as well
as additional remediation work to correct the condition.
Contamination outside of the buildings can be minimized by confining
feed handling to inside the buildings. The use of conveyors to transport
the contaminated soil between buildings did reduce handling outside the
buildings after they were installed in March, 1988. A dike at the fence
line would have been effective in minimizing the spread of contamination
from the hot zone to the outside.
Conducting routine soil sampling in both the hot zone and safe zone
during the operational phase of the field demonstration would have been
useful to minimize contamination of the site. Having such a program in
place would have allowed earlier discovery of the soil contamination.
This in turn would have allowed site personnel to remediate the situation
and prevent the spread of contamination beyond the hot zone, thus reducing
closure costs.
It would have been useful to have had routine wipe testing of the
equipment and buildings during the operational phase of the MIS program.
If this had been done, potential sources of contamination around the MIS
and in the personnel decontamination area could have been identified and
remediated quickly, and additional contamination of these areas could have
been prevented. This in turn would have helped simplify final closure
operations.
509
-------
REFERENCES
1. Mortensen, H., A. Sherman, etal. Destruction of Dioxin-Contaminated
Solids and Liquids by Mobile Incinerations. EPA-6QO/2-87-033,
Hazardous Waste Engineering Research Laboratory, Cincinnati, Ohio,
1987.
2. Stumbar, J., et al. Project Summary, EPA Mobile Incineration System
Modifications, Testing and Operations, February 1986 to June 1989,
USEPA, Risk Reduction Engineering Laboratory, Releases Control Branch,
Edison, New Jersey, 1990.
3. Sherman, A., et al. Final Permit Application USEPA Mobile
Incineration System at the James Denney Farm Site, McDowell,
Missouri. USEPA, Hazardous Waste Engineering Research Laboratory,
Releases Control Branch, Edison, New Jersey, 1987.
4. Wagoner, David A., Closure Plan for the Denney Farm Site, EPA ID
Number MOT300010980 USEPA Region VII, Kansas City, KS, January 31,
1989.
5. Exner, Jurgen H., etal. A Sampling Strategy for Remedial Actions at
Hazardous Waste Sites: Clean-up of Soil Contaminated by
Tetrachlorodibenzo-p-Dioxin Hazardous Waste & Hazardous Materials.
Vol. 2, No. 4, 1985.
6. USEPA Releases Control Branch, USEPA Technology Transfer
Presentation/Exhibit, The Mobile Incineration System. Edison, NJ,
March 29, 1989.
510
-------
THERMAL DESORPTION ATTAINABLE
REMEDIATION LEVELS
Paul R. de Percin
Superfund Technology Demonstration Division
Risk Reduction Engineering Laboratory
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
ABSTRACT
Thermal desorption is a physical separation process that
uses indirect heating to vaporize organic contaminants which can
then be removed and recovered. Initially, low temperature was
used as a descriptor for these systems to avoid being associated
with combustion processes, e.g., incinerators, plasma torches,
etc. This can be misleading since thermal desorbers can operate
between 200° - 1000° F.
In the past few years there has been an expanded use in the
application of thermal desorbers to contaminated soils and
sediments. The main reason for this growth is the wide range of
organics that can be effectively removed from the soils or
sediments. Many Records of Decision have identified thermal
desorption has the treatment of choice, and Feasibility Studies
frequently include thermal desorption in the list of applicable
treatment technologies.
This paper describes the thermal desorption technology (TD)
and presents data on the treatment performance and effectiveness
for organic compounds considered difficult to treat.
This paper has been reviewed in accordance with the US
Environmental Protection Agency's peer and administrative review
policies and approved for presentation and publication.
511
-------
INTRODUCTION
Thermal desorption (TD) is being used or has been proposed
as the most appropriate treatment technology for many superfund
sites. TD treatment technology is recommended and used because
of (1) the wide range of organic contaminants effectively
treated, (2) availability and mobility of commercial systems, and
(3) the public acceptance of the treatment approach. Thermal
desorption is applicable to many organic wastes and generally not
used for treating inorganics and metals. Commercial systems are
now in operation remediating Superfund sites, and more are under
construction and being designed. The public has shown a
preference for this technology over incineration because as a
separation process there is less likelihood of creating dioxins
and other oxidation products. In response to this increased
interest and use, the USEPA is now preparing an Engineering
Bulletin on thermal desorption. The bulletins will provide a 5
to 10 page summary of the TD technology, the available test data,
and a comprehensive reference list.
Because of the wide application of TD processes there are
ongoing studies examining TD effectiveness on Superfund wastes
under the SARA (Superfund Amendments and Reauthorization Act)
Land Ban program. Recently available data and other treatment
data make thermal desorption an excellent candidate for the Best
Demonstrated Treatment Technology (BOAT) designation.
It is always a good approach to perform a treatability study
to definitely determine if the selected technology can
effectively remove/destroy/immobilize the hazardous constituent.
The USEPA is now preparing an information bulletin on possible
treatability procedures for thermal desorption processes. Bench-
scale procedures would simply use a muffle furnace, exposing the
sample to a set temperature for a certain period of time.
There are several Superfund Innovative Technology Evaluation
(SITE) program field demonstrations scheduled for thermal
desorption processes this year (1991). At the Wide Beach
superfund site near Brant, New York the "Taciuk" thermal
desorption process of SoilTech, Inc. (Canonie Engineers) is being
used to remediate PCB contaminated soils. The USEPA proposes to
perform a field evaluation on this process in March 1991. The
"X*TRAX" process of Chemical Waste Management, Inc. will be used
at the Re-solve Superfund site near North Dartmouth,
Massachusetts and a USEPA SITE Demonstration is tentatively
scheduled for June 1991. In July a SITE demonstration of the
DAVE (Desorption and Vaporizaton Extraction) process of Recycling
Sciences, Inc. (RSI) is scheduled to take place. These SITE
field demonstrations are being designed to answer some of the
final, but most difficult, questions about thermal desorption;
(1) are there products of incomplete combustion, e.g., dioxins,
or (2) what are the impacts of high organic concentrations.
512
-------
PROCESS DESCRIPTION
Thermal desorption is any number of processes that use
either indirect or direct heat exchange to vaporize organic
contaminants from soil, sludge or sediments. These are physical
separation processes and are not designed to provide high levels
of organic destruction (e.g., 99%). Because of the high
temperatures of a few systems some localized oxidation and/or
pyrolysis may occur. System performance is typically measured by
comparison of untreated waste contaminant levels with those of
the processed waste.
Figure 1 is a general schematic of a thermal desorption
process. Waste material must be excavated and screened to remove
oversized objects (e.g., rocks > 1.5 inches) before being
conveyed to the desorber (step 1). There are generally four
designs for the desorber (step 2); an indirectly fired rotary
dryer, a single (or set of) internally heated screw auger(s), a
vertical mixed bed, and a series of externally heated
distillation chambers. In the desorber the waste is heated
(200°F to 1000°F) causing organics and water to vaporize. The
gaseous organics and water are moved out of the system to the
collection and control equipment (step 3), sometimes with an
inert gas (e.g., nitrogen). Volatiles may then be burned in an
afterburner, collected on activated carbon, or recovered in
condensation equipment.
The key variable controlling the effectiveness of the
thermal desorption process is the final temperature of the waste.
This temperature is mainly dependent on the residence time and
the heat transfer method.
Operation of a thermal desorber typically creates up to six
process residual streams; treated waste, oversized media
rejects, condensed contaminants and water, particulate control
dust, clean offgas, and spent carbon (if used). The treated
waste may be suitable for disposal on site. Water is needed to
control the dusting of the treated waste and the condensed water
is frequently used after treatment. The concentrated organic
liquids are stored for further treatment or recovery. Collected
particulate can be recycled through the desorber.
APPLICABILITY AND LIMITATIONS
Thermal desorption has been proven effective in treating
contaminated soils, sludges and filter cakes. Chemical
contaminants for which bench-scale through full-scale treatment
data exist include volatile organic compounds (VOCs) and
semivolatile organic compounds (SVOCs), and even higher boiling
point compounds such as PCBs and dioxins.
513
-------
This technology is not effective in separating inorganics
from contaminated media. Some metals (e.g., mercury) may be
volatilized by the thermal desorption process as the contaminated
media is heated. Normally the temperature of the waste achieved
by the process does not oxidize the metals .
The waste must contain at least 20 percent solids, and
sometimes at least 30 percent solids, to be fed into the
desorber. Higher solids content is preferred because of the cost
of evaporating, collecting and treating the water.
although VOCs and SVOCs are the primary target of the
thermal desorption technology, the total organic loading in the
feed is limited by some systems to 10 percent or less. There is
evidence that polymers (e.g., phenolic tars) may foul or plug
some of the systems.
TREATMENT PERFORMANCE
To support some of the above claims on treatment
effectiveness, data from four data sets are discussed;
manufactured gas plants, two PCB contaminated soils, and dioxin
contaminated soils. It is felt that these data will provide
support for considering thermal desorption for future
remediations (References 1,2,3 and 4).
HWR&IC Manufactured Gas Plants Treatment Study1
The Hazardous Waste Research & Information Center (HWR&IC)
and the Gas Research Institute (GRI) performed a study of the use
of thermal desorption of contamination from manufactured gas
plants. Wastes at these sites were coal tar contaminated soils
with between 400 and 2000 ppm total PAHs (Table 1). Eleven
pilot-scale tests were performed with the IT Corporation desorber
on three wastes using test conditions from 572°F to 752°F (300°C
to 400°C), and 5 to 9 minutes residence times. Approximately 30
to 60 kilograms/hour of soil was used for the pilot-scale tests.
Table 2 presents the pilot-scale test data with temperature and
time variables. Both temperature and residence time have a major
impact on the treatment results. Most commercially available
treatment systems can control both parameters and thus adjust the
operating conditions to meet the remediation requirements. One
very important piece of information from this study was the
comparability of the bench-scale and pilot-scale data. These
data should support the use of an inexpensive bench-scale test to
be used as an initial "proof-of-process" test.
514
-------
Chemical Waste Management X*TRAX Treatment of PCBs
Chemical Waste Management (CWM) evaluated the treatment
effectiveness of their pilot-scale (about 5 tons/day) X*TRAX
thermal desorber. Of particular interest was the treatment
effectiveness on heavy organics including PCBs. Tables 3 and 4
show PCS treatment data for the pilot-scale desorber. Good
percentage removal was achieved (>99 %). While the PCB residual
levels (8 to 19 ppm) shown may not be acceptable to the
regulatory agencies, the residence time and temperature of the
process can be increased to improve the removal efficiency.
Table 5 shows some residual levels achieved for the more
frequently seen contaminants.
RSI Treatability Tests3
Between July 1984 and April 1985 the pilot-scale version
(10 tons/day) of the RSI's DAVE process treated PCB contaminated
sediments from Waukegan Harbor, IL and the Hudson River, NY.
Table 6 presents the treatability data for these two wastes and
associated process data. The PCB concentration in the clean
soils averaged 1.89 ppm (1.11 ppm standard deviation), below the
2 ppm informal guideline goal. In some cases low removal
efficiencies occurred, primarily because of the low starting
concentrations.
NCBC Dioxin Treatment Study4
In June 1985, Herbicide orange contaminated soil was treated
using the pilot-scale IT Corporation thermal desorber. Of
specific concern was the dioxin (2,3,7,8-TCDD), and the 2,4-D and
2,4,5-T pesticides. Table 7 shows the operating conditions and
treatment results for the dioxins, furans, 2,4,5-T and 2,4-D.
The project goal of less than one part per billion (
-------
treating PCS contaminated soils. By the end of this year there
will be data from field demonstrations of PCB contaminated soils.
It is projected that thermal desorption can remove VOCs, SVOCs
and TPHs (total pretroleum hydrocarbons) from soils, sludges and
sediments to the part per billion range in most cases.
REFERENCES
1) Helsel, R,, E. Alperin, A. Groen of IT Corporation,
"Engineering-Scale Demonstration of Thermal Desorption
Technology for Manufactured Gas Plant Site Soils" for the
Hazardous Waste Research & Information Center, Savoy,
Illinois, 61874, Report # HWRIC RR-038, November 1989.
2} Swanstrom, C., C. Palmer of Chemical Waste Management, Inc.,
"X*TRAX Transportable Thermal Separator for Solids
Contaminated With Organics", Presented at the Air and Waste
Management Association - International Symposium on
Hazardous Waste Treatment: Treatment of Contaminated Soils,
Cincinnati, Ohio, February 5-8, 1990.
3) Personal Communication with Laurel Staley, USEPA, Risk
Reduction Engineering Laboratory, Demonstration Section,
Cincinnati, Ohio, January 22, 1991.
4) Helsel, R.W., R.W. Thomas of IT Corporation for EG&G Idaho,
"Thermal Desorption/Ultraviolet Photolysis Process
Technology Research, Test, and Evaluation Performed at the
Naval Construction Battalion Center, Gulfport, MS., for the
USAF Installation Restoration Program, Volume I",
Engineering & Services Laboratory, Air Force Engineering &
Services Center, Tyndall Air Force Base, Florida, 32403,
Report | ESL-TR-87-28, December 1987.
516
-------
Figure 1
Schematic Diagram of Low Temperature Thermal DesorpHon
»—
r
G*t TfM&mnt
Syesm
(3)
Con«ntr«t»d Conttnin*nts
Table 1: HWR&IC Untreated Soil Concentrations
PAH Concentrations (ppm)
Compound Soil A Soil B Soil C
Acenaphthene
Anthracene
Benzo(a)anthracene
Benzo(b)fluoranthene
Benzo(k)fluoranthene
Benzo(a)pyrene
Benzo(g,h,i)perylene
Chrysene
Dibenzo(a,h)anthracene
Fluoranthene
Fluorene
Indeno(1,2,3-cd)pyrene
Naphthalene
Phenanthrene
Pyrene
Total Quantified 2,107 1,999 366
8 ND - not detected
210
ND
ND
46
39
ND
18
ND
ne ND
260
170
ne 14
680
410
260
390
190
55
34
18
40
18
ND
ND
230
230
18
66
490
220
ND*
ND
'15
34
25
32
110
19
11
28
ND
37
ND
ND
55
517
-------
Table 2: HWR&IC Pilot-Scale Test Results3
Initial
Soil Concentration 300/5b 300/9 400/5 400/9
A 2107 85.4 140.6 10.78/8.94C 0.97
B 1999 69.4 22.0 7.31 0.50
C 366 ndd 79.8 nd 3.41
* PAHs in parts per million (ppm)
Nominal soil temperature in C/Time at temperature in minutes
c duplicate tests
nd - no data
Table 3: CWM X*TRAX Pilot-Scale Test Data
Sandy Soil with PCBs
Feed Product
Compound (ppm) (ppm) % Removal
PCBs 1,480 8.7 99.4
1,2,4-Trichlorobenzene 2.9 ND >99.9
Di-n-Butylphthalate 1.0 0.24 76.0
Bis(2-Ethylhexyl)phthalate 9.1 0.18 98.0
Table 4: CWM X*TRAX Pilot-Scale Test Data
Clay, Silt & Gravel with PCBs
Feed Product
Compound (ppm) (ppm) % Removal
TPH If400 34 97.6
PCBs (Arochlor 1254) 2,800 19 99.3
1,2, 4-Trichlorobenzene 6.8 ND >98.0
Di-N-Butylphthalate 6.9 0.18 97.4
Bis(2-Ethylhexyl)phthalate 4.7 ND >97.2
518
-------
Table 5: CWM X*TRAX Pilot-Scale Test Data
Surrogate Feed
Feed Product
Compound (ppm) (ppm) % Removal
Methyl Ethyl Ketone 100.9 < 0.1 >99.90
Tetrachloroethylene 91.0 0.015 99.98
Chlorobenzene 61.8 0.0065 99.98
Xylene 56.4 0.0028 99.99
1,4 Dichlorobenzene 78.4 0.0014 99.99
1,2 Dichlorobenzene 537.0 0.0741 99.99
Hexachlorobenzene 79.2 0.300 99.62
Table 6; RSI Pilot-Scale Test Results at
Waukegan Harbor and Hudson River
Run Temperature °F Feed Product
Date Gas Soil (ppm) (PPm)
7-19 1000 325 44 1.6
8-27 1400 325 109 1.0
8-29 1400 300 53 1.5
8-30 1500 350 37 3.2
9-04 1600 350 31 0.9
9-10 1450 400 38 1.5
9-11 1500 400 28 4.0
9-12 1600 300 27 1.4
3-12 1500 325 12.8 0.5
3-15 1450 275 12.9 0.5
3-26 1200 270 8.6 1.3
3-27 1500 275 8.6 1.3
4-18 1400 275 206 , 0.8
519
-------
Table 7: NBCB Dioxin Treatment Study
Test Runs
Parameter
Feed Rate
Ib/hr
Residence Time
minutes
Soil Temperature
Soil Processed
Ib
Total TCDD (ppb)
feed
treated
Total TCDF (ppb)
feed
treated
2,4,5-T (ppm)
feed
treated
2,4-D (ppm)
feed
treated
Rl
31
40
1040
150
__
— -
—
__
93
<0.2
0.37
<0.2
R2
31
40
1040
299
274
0.23
12.2
0.23
15
0.11
0.81
0.079
R3
55
19
1040
359
239
0.11
10.2
0.11
37
<0.2
0.58
<0.2
R4
97
10.5
1040
680
268
0.61
12.7
0.61
200
<0.2
17.0
<0.2
R5
44
J
24
860
220
235
0.75
"•
10.8
0.75
80
<0.2
1.4
<0.2
Note: (—) this data unreadable in source report.
520
-------
EFFECTIVENESS OF THE STABILIZATION/SOLIDIFICATION PROCESS IN
CONTAINING METALS FROM RCRA ELECTROPLATING WASTES
Ronald J. Turner
U.S. Environmental Protection Agency
Risk Reduction Engineering Laboratory
Cincinnati, Ohio 45268
Abstract: Wastewater treatment sludges from electroplating operations are subject to
the land disposal restrictions of the Resource Conservation and Recovery Act (RCRA).
The final rule (August 17, 1988) requires stabilization with cement or equivalent binders
for RGRA waste code F006. Samples from two electroplaters were stabilized with ce-
ment, lime/fly ash and kiln dust pozzoian materials. The Toxicity Characteristic Leach-
ing Procedure was performed on the 28-day cured specimens of the three binders.
The metals and total cyanide leachate data are presented.
Keywords: Stabilization, solidification, evaluation, metals, cyanide
INTRODUCTION
The U.S. EPA's Risk Reduction Engineering Laboratory conducted a study of
the stabilization/solidification (S/S) technology for metals binding to support the
development of treatment standards for electroplating wastes. This paper presents
the results of the S/S evaluations.
Stabilization/solidification involves the mixing of a waste (F006) with a binder
material to enhance the physical and chemical properties of the waste. The binder is
typically a cement, pozzoian, or thermoplastic. Stabilization produces a chemical re-
action that, in most cases, converts inorganic material to its least soluble and most
environmentally inert form. Solidification improves the handling and physical
521
-------
characteristics of the material. The two terms are commonly used together, as both
technologies are instrumental in immobilizing metals.
WASTE DESCRIPTION
The term F006 is defined as wastewater treatment sludges from electroplating
operations. Generation of F006 is depicted in Figure 1. Table 1 presents the stan-
dards promulgated for F006 on August 17, 1988.(1)
The F006 samples were obtained from a manufacturer of decorative hardware
plated with brass, copper, zinc, nickel, and chrome (Facility A) and from a facility with
barrel and rack lines dedicated to nickel-chromium, nickel-copper, cadmium, and
chromium plating (Facility B). Both facilities use alkaline chlorination, chemical precipi-
tation, and sludge dewatering for their electroplating wastes. The composition of Fa-
cility A's F006 was claimed to be confidential business information (CBI). The compo-
sition of Facility B's F006 is given in Table 2.
STABILIZATION/SOLIDIFICATION PROCESS
The three binding agents used for the F006 S/S tests were Type 1 Portland
cement, a 1:1 mixture of lime/fly ash, and kiln dust.0 Analyses of the binders indicat-
ed the presence of chromium, lead, and nickel (Table 3). During bench-scale testing,
different binder-to-waste ratios were used to stabilize/solidify the waste samples. (All
S/S processes increase the volume of the final product for disposal). A screening test
was performed to determine the appropriate amounts of water for hydration and the
binder/waste ratios. The final binder/waste ratios used for the tests were 1.9 (cement),
522
-------
PERIODIC
RINSE WATER
DISCHARGE
CHEMICAL
PRECIPITATION
POTW
F006
F006 - WASTEWATER TREATMENT SLUDGES
RJ07 - SPENT CYANIDE PLATING BATH SOLUTIONS
POOS - PLATING BATH SLUDGES
F009 - SPENT STRIPPING AND CLEANING BATH SOLUTIONS
Figure 1. Hypothetical Cyanide-Bearing Electroplating Process.
523
-------
TABLE 1. TREATMENT STANDARDS, METAL-FINISHING
SUBCATEGORY FOR F006 (NONWASTEWATERS)
Constituent
Cyanides (total)
Cyanides (amenable)
Cadmium
Chromium
Lead
Nickel
Silver
Maximum for any single
Total composition, mg/kg
590
30
a
a
a
a
a
grab sample
TCLP, mg/L
a
a
0.066
5.2
0.51
0.32
0.072
8 Not applicable
Source - Reference 1
TABLE 2. SUMMARY OF FACILITY B F006 METAL ANALYSES
(mg/kg)
Arsenic
Barium
Cadmium
Chromium (T)
Chromium (+6)
Copper
Iron
Lead
Nickel
Zinc
0.62
6.9
39,900
4,840
24
7,900
12,450
10
11,650
4,680
average of 6 tests
TABLE 3. BINDER TCLP ANALYSES
Cadmium
Average Cement <0.1
Average Kiln Dust <0.1
Average Lime/Fly Ash <0.1
Chromium
314
58
26
Lead
4
38
7
Nickel
1.5
2.6
1
average of 3 tests
Source - Reference 3
524
-------
0,4/0.4 (lime/fly ash), and 0.7 (kiln dust). The selected ratios contained the lowest
waste/binder ratios that exhibited a minimum strength of 50 psi.
After mixing, the 2-in.3 test specimens were prepared and cured for 28 days at
23°C and 98 percent relative humidity. One set of samples from the Facility A waste
was cured for just 24 hours to note any differences in teachability versus a 28-day
cure. Facility B samples were tested after 28-day curing. Figure 2 presents a flow
diagram of the S/S process used by the U.S. Army Corps of Engineers Waterways
Experimental Station, Vicksburg, MS, to prepare the test samples for the EPA. A
pass/fail eompressive test value of 50 psi was used to select specimens for chemical
analyses and leaching tests. All extractions were conducted in accordance with Test
Methods for Evaluating Solid Wastes: Physical/Chemical Methods, SW-846, third
edition.'4'
METALS LEACHABILITY RESULTS
Tables 4 and 5 present data for the Facility A tests and filter cake TCLP results
at 24-hour and 28-day cures for the regulated metals. The TCLP were performed in
triplicate for each binder, resulting in nine extractions. (The F006 total waste composi-
tion is claimed confidential by the generator and is not given.) The TCLP extracts for
the F006-cement binder contained about 1 mg/L copper and 0.2 mg/L zinc for both
the 24-hour and 28-day samples. The TCLP extracts for the F006-Iime/fly ash and kiln
dust binders had concentrations of these two metals approximately an order of mag-
nitude higher. The nickel and chromium analyses showed low concentrations in the
extracts of all three F006-binder combinations.
525
-------
WASTE TO BE
STABILIZED/ -
SOLIDIFIED
en
ro
en
INITIAL
SCREEN TESTING
WATER BINDER
1 I
WATER-TO-WASTE
AND
' BINDER-TO-WASTE
RATIO SaECTON
UCS
TESTING
WATER BINDER
1 1
BATCH
PREPARATION
CURING
DETERMINATION OF
COMPRESSIVE
STRENGTH AT 7,14,
21, AND 28 DAYS
TCLP
TESTING
TOXICtTY
CHARACTERISTIC ANALYSIS OF
LFACHING -M. ANALYblb Or
PROCEDURE (AFTER Lfc/tt,w\lb
28-DAY CURE)
Figure 2. Flowchart for stabilization/solidification tests.
-------
TABLE 4. TCLP RESULTS OF UNTREATED F006 AND 24-HOUR
F006 BINDER TESTS - FACILITY A
(mg/L)
Arsenic
Barium
Cadmi urn
Chromium
Chromium
Copper
Iron
Lead
Nickel
Zinc
Untreated Cement
0.006 0.002
<0.002 0.99
0.029 <0.004
(T) 0.055 0.15
(+6) 0.01 <0.25
135 0.96
<0.02 0.06
<0.001 0.002
26.8 <0.04
244 0.17
Source - Reference 2
Lime/fly
ash
0.002
0.74
<0.004
0.02
0.03
7.08
0.06
0.02
-------
TABLE 6. TCLP RESULTS OF UNTREATED F006 AND
28-DAY BINDER TESTS - FACILITY B
Metal
Arsenic
Barium
Cadmium
Chromium (T)
Chromium (+6)
Copper
Iron
Lead
Nickel
Zinc
Untreated
0.002
0.39
1298
0.66
0.01
15
<0.01
<0.001
255
88
Cement
<0.002
0.37
0.28
0.44
0.04
0.26
1.0
<0.001
<0.03
0.09
Lime/fly
ash
<0.002
0.32
0.08
0.28
0.34
0.21
0.14
<0.001
<0.03
0.09
Kiln dust
<0.002
0.29
64.2
0.07
0.06
0.49
<0.01
<0.001
8.78
1.3
528
-------
All silver results were less than the method detection limit and are not reported here.
Table 6 presents results of the analyses of Facility B's filter cake leachate and
the TCLP extracts of the 28-day cured samples. With the kiln dust binder, extracts
averaged about 64 mg/L cadmium, 0.06 mg/L hexavalent chromium, and 8.8 mg/L
nickel. The cement binder extracts showed 0.28 mg/L cadmium, 0.04 hexavalent
chromium, and nickel below the detection level (0.03 mg/L). The lime/fly ash binder
extracts showed 0.08 mg/L cadmium and 0.34 mg/L hexavalent chromium; no nickel
was detected in these extracts.
TOTAL CYANIDE RESULTS
The total cyanide, CN(T), represents all cyanide species in the sample, includ-
ing metal complexes but excluding cobalt, gold, and some platinum group metals.
The CN(T) is determined by SW-846 Methods 9010 or 9012 (reflux mineral acid distilla-
tion).
Table 7 summarizes the cyanide data for F006 from both facilities. The average
CN(T) in the untreated filter cake samples was approximately 722 mg/kg for Facility A
and about 2400 mg/kg for Facility B. As a special test, the CN(T) was determined for
the two unstabilized (untreated) wastes (before and after leaching) and for Facility A
stabilized samples and the TCLP leachates. (Note: Total cyanide is regulated by con-
tent after stabilization, not by the TCLP tests.) Apparently, about half of the cyanide
remained in the unstabilized F006 residue after leaching with TCLP extraction fluid No.
2 (pH = 2.88 ± 0.05). CN(T) in two TCLP leachates was found to be below detection
529
-------
limits. Under proper conditions, the tow pH of the extraction fluid could have generat-
ed hydrogen cyanide. However, there was no qualitative evidence of hydrogen cya-
nide release during this study. The stabilized sample CN(T) values were not adjusted
for dilution; however, there seems to be a "binder effect" with the lowest 28-day CN(T)
resuft from the Facility A kiln dust matrix.
TABLE 7. SUMMARY OF F006 CYANIDE DATA
Analysis
Facility A
Facility B
Untreated F0068
Total waste analyses
TCLPb
TCLP residue
722 mg/kg
<0.02 mg/L
396 mg/kg
2392 mg/kg
<0.04 mg/L
1212 mg/kg
28-Day Cure
Total Waste Analyses
Cement0
Lime/fly ashe
Kiln dust1
535 mg/kg
234 mg/kg
163 mg/kg
NMd
TCLP
Cementc
Lime/fly ashe
Kiln dustf
0.39 mg/L
<0.02 mg/L
<0.04 mg/L
NM
8 Moisture content, 72.9 percent, average
b Extraction fluid No. 2 (pH - 2.88 ± 0.05}
* 1.9 Binder/water (b:w)
NM - Not measured
' 0.4/0.4 b:w
1 0.7 b:w
CONCLUSIONS
Nickel and cadmium were highest in the Facility B kiln dust leachate and failed
the TCLP standard for these metals. The other two stabilized samples from Facility B
also failed the TCLP for cadmium. The lime/fly ash binder gave the lowest teachable
530
-------
cadmium and nickel values. All of the Facility A samples passed the TCLP for regulat-
ed metals. These results indicate a need for a cadmium recovery process at Facility
B and proper binder selection for S/S.
Differences in the TCLP metals values due to the binders used (copper, zinc,
chromium, nickel plating) for the 24-hour and 28-day S/S tests for Facility A were mini-
mal; however, the cement binder was the more effective stabilizer for copper and zinc.
The chromium and other metals detected in the binder materials had little effect on the
final TCLP results. The binder selection may have an impact on the CN(T) results
obtained after S/S treatment.
Acknowledgment: Phil Utrecht, PEI Associates, Inc., is acknowledged for his assis-
tance in the preparation of this paper, and the Waterways Experimental Station is ac-
knowledged for performance of the stabilization/solidification tests.
REFERENCES
1. U.S. EPA. Land Disposal Restrictions for First Third Scheduled Wastes, Rnal
Rule. 53 FR 31138, August 17, 1988.
2. U.S. EPA. Draft Report for Stabilization of F006 at Waterways Experimental
Station, Vicksburg, MS. August 25, 1989.
3. U.S. EPA. Draft Report on Binder Characterization for Evaluation of the S/S as
a BOAT. Waterways Experimental Station, Vicksburg, MS. December, 1988.
4. U.S. EPA. Test Methods for Evaluating Solid Waste, SW-846, Third Edition,
November, 1986.
531
-------
FIELD ASSESSMENT OF AIR EMISSIONS FROM
HAZARDOUS WASTE STABILIZATION OPERATIONS
by: Thomas C. Ponder, Jr., PE, CCE
Diane Schmitt
PEI Associates, Inc.
Arlington, Texas 76012
ABSTRACT
Millions of tons of hazardous waste are generated every year. One treat-
ment method for hazardous waste is stabilization which creates a cement/waste
mix that may be disposed of at a landfill. Since the waste is derived from a
number of processes, a variety of volatile and semi-volatile organics may be
present. Stabilization operations which mix and heat the waste increase the
potential for the release of particulates and organics to the air. This paper
presents the results of a comprehensive study that was performed for the EPA
to quantify the release of volatile and semi-volatile organics as well as
particulates from stabilization operations. The study included a field test
of one stabilization operation in which material balance was used to calculate
air releases.
This paper has been reviewed in accordance with the U.S. Environmental
Protection Agency's peer and administrative review policies and approved for
presentation and publication.
532
-------
INTRODUCTION
Hazardous waste stabilization is a process used by industry to stabilize
or fix hazardous waste before it is disposed of on land. The sludge stabi-
lization process involves fixing the waste with organic or inorganic agents.
This fixation process mixes and heats the waste, possibly resulting in the
vaporization of organics.
The U.S. Environmental Protection Agency (EPA) is currently collecting
information to develop standards necessary for the control of air emissions
from hazardous waste treatment, storage, and disposal facilities (TSDFs). The
Office of Air Quality Planning and Standards (OAQPS) has identified stabiliza-
tion facilities as a source of organic emissions at TSDFs. These field tests
were conducted to aid in the development of standards for hazardous waste
stabilization operations. The main objectives of this study were to:
0 Characterize the fate of volatile and semi-volatile organics in the
stabilization of hazardous waste;
° Quantify particulate emissions released during the stabilization pro-
cess; and
° Quantify organic emissions released during the stabilization process..
This paper provides general information on the stabilization process,
discusses the stabilization operations at one facility tested, and discusses
the procedures and methods used to determine air emissions from the stabiliza-
tion process.
STABILIZATION METHODS
As mentioned earlier, the objective of hazardous waste stabilization is
to stabilize hazardous wastes to allow for land disposal. Stabilization, also
referred to as waste fixation, involves the addition of an organic or inorgan-
ic agent.
The vast majority of stabilization facilities operate on a batch system.
After a certain amount of waste is received, it is stabilized and sent for
disposal. Typically, waste is received in drums, trucks or rolloff boxes. It
is emptied either into a mixing unit where stabilizing agents are added
through the use of an auger type mechanism or into a holding pit. At some
facilities, drum waste is removed from the drums and the drums are recycled.
At other facilities, the drums are included with the waste to be stabilized.
Both the mixing and holding areas may serve as premixing areas where mixing is
performed with a backhoe to obtain a uniform waste feed. If a holding pit is
used, once enough waste is received and mixed, it is sent to the mixing unit.
The mixing unit is usually a large concrete or steel bin, tank or mechan-
ical mixing unit. The amount of agent added is determined by tests conducted
on the waste feed. Inorganic (cement and pozzolans) or organic (asphalt and
organic polymerization) stabilizing agents are added. Inorganic agents in-
cluding portland cement, lime kiln dust and fly ash are the more common addi-
tives. Bench-scale tests have found the mixing process to be a major source
533
-------
of volatile and semi-volatile organics.
After mixing, the cement/waste mix is allowed to cure for a period of
approximately 24 hours either in the mixing unit or in a curing unit. Exam-
ples of curing areas are tanks and rolloff boxes. Since the waste may remain
at elevated temperatures during curing, this point in the process may be a
source of volatile emissions. The curing process is a minor source of vola-
tile organic compounds according to bench-scale tests. After curing is com-
plete, the cement/waste mix is sent for disposal to an on or off-site land-
fill. The waste is removed with a backhoe or through the use of a conveyor
system and emptied into trucks for transport.
Particulate air emission control devices are most common at stabilization
facilities. Some facilities may also utilize scrubbers or carbon adsorbers
for volatile control.
STABILIZATION PROCESSES FOR ONE SELECTED FACILITY
Eight facilities across the country were visited to observe and review
their sludge stabilization processes. The capacity and method of stabiliza-
tion varied from site to site. The facility tested was selected because of
the presence of a ventilation system, good testing conditions, and the broad
spectrum of hazardous waste received.
The test facility has one stabilization unit with a total waste through-
put of 20,000,000 gallons a year. After stabilization, the cement/waste mix
is disposed of at a landfill. Any cement/waste mix not meeting established
disposal criteria is recycled.
The process utilizes a series of enclosed mixers which blend the waste
and stabilizing agent. The interior of the process building is vented to
either activated carbon adsorbers for control of organic emissions or to both
a venturi type (Wet Air Tumbler) scrubber for particulate control followed by
the activated carbon adsorber.
The flow diagram for the test facility's hazardous waste stabilization
process is shown in Figure 1. The waste is received at the site as bulk liq-
uid in tankers; bulk 'sludge in vacuum trucks; bulk solids in rolloffs or dump
trucks; and liquid, sludge or solids in drums. All of the solids, liquids,
and sludges are put into a 10,000 gallon open tank, called the Non-Pumpables
tank.
The Non-Pumpables tank is a large 10,000 gallon open tank. The wastes
exit the Non-Pumpables tank through an auger/grinder pump which is attached to
the end of a. backhoe arm. As the backhoe arm moves around the tank, the screw
auger sends the waste to the grinder pumps which both grinds the wastes and
pumps it through a flexible hose which runs along the arm of the backhoe to a
small surge tank. The auger can also be used to circulate the waste in the
tank while the pump is not operating.
The waste is pumped from the surge tank up to the B-Hopper which acts as
a surge protector for the #2 Mixer. The stabilizing agents, cement kiln dust
and lime kiln dust, are added to the waste from the B-Hopper in the #2'Mixer
534
-------
Receiving
and Storage
VJasle Stabilization
and Disposal
Figure 1. Stabilization Process.
Mixer
Curing
Landfill
which is a pug mill.
The cement/waste mix leaves the #2 Mixer as a pumpable slurry and is fed
directly to the Rotary Conditioner, a rotating drum approximately 10 feet in
diameter and 60 feet long. The Rotary Conditioner completes the blending
operation. The waste leaves the Rotary Conditioner and drops into the C-Hop-
per. The cement/waste mix goes from the C-Hopper to cement mixer trucks via a
screw auger and pump. The screw auger feeds the cement/waste mix from the C-
Hopper to the pump and the pump moves the cement/waste mix to a chute which
hangs over the cement mixer trucks. The cement mixer trucks empty the cement-
/waste mix into large roll-off boxes located outside where it- is allowed to
cure. Once curing is complete, the stabilized waste is removed from the roll-
off boxes and trucked to a landfill for disposal.
All of the process equipment and the building in which processing takes
place is vented to one of two air handling systems which are identified by
their capacity, 20,000 cfm and 40,000 cfm. The 20,000 cfm system vents the
air from the drum handling area of the building. In this area, dust is not a
problem so this vents directly into a carbon adsorption system. The remainder
of the areas are vented through the 40,000 cfm system. These sources produce
dust as well as organic vapors, so this stream is first sent through a venturi
scrubber, followed by an 8 foot thick, 8 foot diameter mist eliminator, an air
preheater, a disposable prefilter, and finally two parallel carbon adsorption
systems.
TESTING METHODS AND RESULTS
As discussed previously, the objectives of the field tests were to deter-
mine the percent of volatile and semi-volatile organi.cs emitted from the waste
and to estimate the particulate air emissions during the stabilization pro-
cess. To meet these objectives tests were performed on the waste feed, ex-
haust air and cement/waste mix. Samples of the waste feed were taken from the
535
-------
following locations: the surge tank, the chute for the cement/waste mix to the
cement mixer truck, the exhaust air system from the B-hopper vent, and the
exhaust air system from the #2 mixer/rotary conditioner vent. These samples
were collected and analyzed for volatile and semi-volatile organic compound
concentrations. Figure 2 shows the sample locations at the test facility.
EXHAUST
AIR
NON-PUMPABIESTANK
SAMPUNG LOCATIONS
1. WASTE FEED TO B-HOPPER
2. ORGANIC AIR EMISSION FROM
6-HOPPPER
3. ORGANIC AND PAHTICULATE AIR
EMISSION FflOM #2 MIXEfV
CONDfTIONER
4, CEMENT/WASTE MIX
ROTARY
CONDtTIONER
0
CEMENT
MIXER
TRUCK
TO
CURING
AREA
Figure 2. Sampling Locations.
Waste and Cement/Waste Mix Samplingand Analysis
Three 3-hour sampling runs were conducted over a period of three days.
Samples were collected and flow rates were measured at half-hour intervals
during each run. Flow rates for the inlet and outlet streams were measured by
reading flow meters and/or collection a known quantity of material for a fixed
period of time. Table 1 shows the measured feed rates for the field test.
TABLE 1. MASS FLOWS DURING THE FIELD TEST, Ib/h
Run
1
2
3
Avg.
Std. Dev.
Waste feed
(Ib/h)
24,054
17,694
20,466
20,738
3,189
Cement/was te
mix (Ib/h)
39,798
32,214
39,348
37,120
4,255
536
-------
Samples were collected as follows:
0 waste feed - dipping a bucket into the surge tank located between the
Non-Pumpables tank and the B-Hopper.
0 cement/waste mix - lowering a bucket into the chute feeding the ce-
ment mixer trucks. The cement/waste mix samples were collected 20
minutes after the waste feed samples to allow for the estimated Rota-
ry Conditioner residence time.
For volatile and semi-volatile analysis of the waste feed, four 40-ml VOA
vials were collected. Two vials were reserved for volatile analysis, and two
were reserved for semi-volatile analysis. Approximately four drops of HC1
were added to each vial. The vials were filled completely to eliminate head-
space and stored in an ice-chilled cooler for transport. The vials for vola-
tile analyses were analyzed separately. The vials for semi-volatile analyses
were composited in the laboratory and aliquots taken for analysis.
For volatile and semi-volatile organic analyses of the cement/waste mix,
duplicate grab bag samples were collected at half hour intervals. Each 40-ml
VOA vial was weighed empty, then partially filled with methanol, and re-
weighed. Approximately 20 grams of cement/waste mix was added to the vial and
the vial was weighed to determine the amount of sample. The vial was then
filled completely with methanol, weighed and shaken to disperse the solids.
In the laboratory, each sample was analyzed separately according to Methods
5030 and 8240 (GC/MS).
Exhaust Air Sampling and Analysis
Sampling of the air exhaust system was performed by accessing a platform
located at the vent from the chute between the #2 Mixer and Rotary Conditioner
Volatile organic content of the exhaust air was measured following the
guidelines of EPA Method 18 (40 CFR 60, Appendix A, July 1989). Duplicate
samples were collected at sites 2 and 3. At both sites, the sampling trains
consisted of a Teflon line into the stack and two 1-gram sized coconut shell
charcoal tubes in series. At Site 3 (#2 Mixer/Conditioner vent), the sample
volume was measured with a dry gas meter measuring 1 liter/revolution with ±2
percent accuracy. At Site 2 (B-Hopper vent), calibrated personnel sampling
pumps were used. The pump flow rate was measured before and after each test
with a bubble buret.
Samples at sites 2 and 3 were collected over a three-hour period at a
rate of about 200 ml/minute. After the test, each charcoal tube was capped
and stored in a plastic bag on ice. The front and back charcoal tubes were
analyzed separately to determine the extent of sample breakthrough. Each tube
was desorbed with 2 ml of carbon disulfide. Because volatile components such
as acetone and methylene chloride are major constituents in this waste, analy-
ses were conducted by GC/FID instead of GC/MS.
Gaseous semi-volatile organics in the exhaust air were also measured by
EPA Method 18. The sampling trains were identical to the volatile organic
train except the charcoal tubes were replaced with 3.5-gram capacity XAD-2
tubes. Duplicate samples were collected over three-hour periods at a rate of
537
-------
1 liter/minute. EPA Method 8270 was used to analyze the XAD-2 tubes for HSL
semi-volatile organies. Method 8270 uses methylene chloride to soxhlet ex-
tract the XAD-2 tubes. This sampling method was selected because it is equiv-
alent to the procedures used at the laboratory bench study. However, the
method could be biased by semi-volatile organies adsorbed into the particulate
matter. For this reason, a second set of samples was collected at Site 3 by
the standard EPA Modified Method 5. The XAD-2 tubes were prepared by PEL
Flow rate composition was determined using EPA Methods 1 through 4 at
Sites 2 and 3. Method 1 is used to select the sample points. Method 2 is
used to determine the gas velocity and temperature. Method 3 is used to de-
termine the oxygen content of the stream. Method 4 is used to determine the
moisture content of the stream. This information is needed to calculate ex-
haust air flow rates.
Mass Flow Rate and Emission Results
Tables 2 and 3 show the mass flow rates of the volatile and semi-volatile
organies, respectively.
TABLE 2. SUMMARY OF VOLATILE ORGANIC COMPOUND FEED RATES
Compound
Methylene chloride
Acetone
2-Butanone
4-Hethyl 2-pentanone
Toluene
Ethyl benzene
Xylenes
Benzene
1,1,1-Trichloroethane
Trichloroethene
Waste Feed
Compound Feed Rale, Ib/h
Run 1
0.8765
0.4226
0.3099
0.1684
0.9866
0.4199
2.5483
0.0145
0.1391
0.2065
Run 2
0.2678
0.5981
0.3214
0.0544
2.5214
3.5696
17.782
0.2011
0.1094
0.3206
Run 3
1.342
0.1729
1.1432
0.2941
2.0411
1.706
8.1659
0.0882
2.7161
0.6417
Cement/Waste Mix
Compound Feed Rate, Ib/h
Run 1
5.4854
0.4039
0.3271
0.0955
1.9931
1.0634
5.3735
0.0135
0.2491
0.5281
Run 2
3.4301
0.2545
0.3695
0.0522
2.0714
2.8719
14.661
0.0680
0.1340
0.2297
Run 3
5.2026
0.1472
0.4120
0.1094
1.7443
1.7270
9.0390
0.0523
2.0815
0.5103
TABLE 3. SUMMARY OF SEMI-VOLATILE ORGANIC COMPOUND FEED RATES
Compound
Phenol
Naphthalene
2-Kethylnaphthalene
1,2,4-Trichlorobenzene
Waste Feed
Compound feed rate, tb/h
Run 1
3.127
--
0.1443
--
Run 2
17.694
0.6724
0.5308
--
Run 3
4.2979
3.2746
2.2513
0.8596
Cement/Waste Mix
Compound Feed Rate, Ib/h
Run 1
0.109
--
--
--
Run 2
0.3772
0.0812
0.0573
--
Run 3
0.6984
0.4128
0.3939
0.0952
538
-------
Volatile and semi-volatile organic emission factors are shown in Tables 4
and 5. The emission factor for each chemical is the ratio of the mass flow
rate out the stack to the mass flow rate in the feed. This represents the
fraction of each organic that would be released to the air. '
TABLE 4. EMISSION FACTORS FOR VOLATILE ORGANIC COMPOUNDS
Compound
Hethylene chloride
Acetone
2-Butanone
4-Hethyl 2-pentanone
Toluene
Ethylbenzene
Xylenes
Benzene
1,1,1-Trichloroethane
Trichloroethene
Compound feed rate, Ib/h
Run 1
0.8765
0.4226
0.3099
0.1684
0.9866
0.4199
2.5483
0.0145
0.1391
0.2065
Run 2
0.2678
0.5981
0.3214
0.0544
2.5214
3.5696
17.782
0.2011
0.1094
0.3206
Run 3
1.342
0.1729
1.1432
0.2941
2.0411
1.706
8.1659
0.0882
2.7161
0.6417
Air emission rate,
Ib/h
Run 1
2.544
0.087
0.087
0.204
0.591
0.212
0.944
0.099
0.240
0.286
Run 2
4.016
0.254
0.248
0.230
0.986
1.049
4.571
0.120
0.348
0.330
Run 3
34.082
0.406
2.948
2.827
1.212
0.444
1.640
1.731
2.965
3.789
Air emission factor, Ib/lb of feed
Run 1
2.902
0.206
0.281
1.211
0.599
0.505
0.384
6.828
1.725
1.385
Run 2
14.996
0.425
0.772
4.228
0.391
0.294
0.257
0.597
3.181
1.029
Run 3
25.463
2.348
2.579
9.612
0.594
0.260
0.201
19.626
1.092
5.905
Overall
average
14.453
2.979
1.211
5.017
0.528
0.353
0.281
9.017
1.999
2.773
TABLE 5. EMISSION FACTORS FOR SEMI-VOLATILE ORGANIC COMPOUNDS
Compound
Phenol
Naphthalene
2-Methylnaphthalene
1,2,4-Trichlorobenzcne
Compound feed rate, Ib/h
Run 1
3.127
0.1443
--
Run 2
17.694
0.6724
0.5308
Run 3
4.2979
3.2746
2.2513
0.8596
Air emission rate,
Ib/h
Run 1
3.9E-04
3.5E-03
1.0E-03
•-
Run 2
5.6E-04
4.2E-03
7.7E-04
•-
Run 3
3.3E-04
1.8E-02
2.8E-03
4.1E-03
Air emission factor,
Ib/lb of feed
Run 1
1.2E-04
--
6.9E-03
--
Run 2
3.2E-05
6.2E-03
1.5E-03
--
Run 3
7.7E-05
5.5E-03
1.2E-03
4.8E-03
Total volatile organic concentrations were measured using a continuous-
flame ionization analyzer (FIA) and the procedures of EPA Method 25A at Sites
2 and 3. The sampling system consisted of a heated Teflon sample line main-
tained at 280°F and a Beckman Model 402 FIA. The analyzer was calibrated with
gas standards of methane in hydrocarbon-free air. The total organic content
of the sample was compared with the methane response factor and was then re-
ported as ppm methane equivalent. This procedure does not yield an exact mea-
surement of the total organic carbon due to the fact that the carbon response
on an FIA will vary depending on the molecular structure. Table 6 summarizes
the results of the FIA and OVA analyses. There is little difference between
the total VOC measured by each method.
539
-------
TABLE 6. COMPARISON OF TOTAL VOC EMISSIONS MEASURED BY
FIA TO TOTAL SPECIFIC COMPOUND EMISSIONS BY METHOD 18
Site 2-B Hopper Vent
Site 3-Conditioner/Mixer Vent
Total VOC by Total VOC by
Run No.
1
2
3
FIA.
lb/h{1)
0.81
1.46
3.31
Method 18.
Ib/h
0.88
1.66
4.54
Total VOC by
FIA.
lb/h{1)
5.26
11.70
43.16
Total VOC by
Method 18,
Ib/h
4.41
10.49
47.60
(1)Pounds per hour as Methane.
Particulate emissions were measured at Site 3 (#2 Mixer/Conditioner) us-
ing EPA Modified Method 5 as described in Method 0100 of EPA SW846. The Modi-
fied Method 5 sample train was also used at Site 3 to sample for semi-volatile
organics. The filter and the probe rinse were analyzed for particulate. One
three-hour Modified Method 5 sample was collected during each run. Particu-
late emissions were sampled only from the #2 Mixer/Conditioner vent. The par-
ticulate emissions were sampled using an EPA Modified Method 5 as described in
Method 0010 of EPA SW8A6. The results of the particulate emission sampling
are shown in Table 7.
TABLE 7. UNCONTROLLED PARTICULATE EMISSIONS
FROM SITE 3 - CONDITIONER/MIXER VENT
Run No.
1
2
3
Average
Particulate
concentration.
gr/dscf{1)
0.064
0.210
0.065
0.113
Particulate
emission rate,
Ib/h
1.20
4.84
1.42
2.49
(1)Grains per dry standard cubic foot at 68 F,
29.92 in. Hg, and zero percent moisture.
Quality Assurance/Quality Control
Quality assurance and quality control procedures were employed during
the laboratory analysis of all samples obtained during this project. All sam-
ples of the exhaust air were collected in duplicate and desorption efficiency
for each compound was determined. The quality assurance objectives for preci-
sion (± 10 percent) and accuracy (± 30 percent) were met for all compounds.
Serai-volatile samples were spiked with the surrogates and duplicate samples
were collected by Method 18. The relative standard deviation between dupli-
cate semi-volatile samples found above detection limits met the QA objective
of ± 30 percent with few exceptions. The volatile compound QC samples con-
sisted of matrix spikes and method blanks. The QA objective for precision (±
30 percent) and accuracy as percent recovery of matrix spikes (40 to 160 per-
cent) were met for all compounds.
540
-------
DISCUSSION OF RESULTS
Air emission results obtained in this study indicate the extent of air
emissions from the mixing process. For both volatile and semi-volatile com-
pounds, the precision for the air sampling method as measured by the relative
percent difference between duplicate samples were quite good. Also, the accu-
racy of the air stream measurements as measured by recovery of target com-
pounds was good. In addition, a comparison of results obtained by Method 18
and Method 25A (Table 6) supports the accuracy and consistency of the air
emissions data. It should be noted that liquid wastes added to the sludge in
the Non-Pumpables tank were highly variable during this study. Thus, cross-
averaging the runs is not necessarily meaningful.
Material balance closure was not attainable for all compounds, especial-
ly semi-volatile compounds. The method for volatile and semi-volatile organic
analyses for the cement/waste mix was untried and the relative accuracy is
unknown. The major factor influencing the mass balance was the inability to
accurately test the cement/waste mix samples. Analytical methods for testing
cement are not established. In addition, the analytical methods used in sam-
pling the exhaust air streams were much more precise than those available for
analyzing the waste feed and cement/waste mix. As a result, a much higher
degree of accuracy and certainty was achieved from the air sampling, creating
higher effluent rates than feed rates.
The semi-volatile compounds listed in Tables 3 and 5 are those measured
above the detection limits in one or more exhaust air runs. Both tables 4 and
5 show the results of the tests in terms of pounds of volatile organics emit-
ted to the air per pound of feed sent to the stabilization process. The emis-
sion factors for some of the volatile organics were greater than one. This is
not a realistic occurrence since the emission factor represents the ratio of
the compound emission rate to the compound feed rate. A value less than or
equal to one is expected as the amount of a particular compound exiting the
system cannot be greater than the amount entering the system. In particular,
emissions factors for methylene chloride, 4-methyl 2-pentanone and benzene
were significantly greater than one. The inconsistent emissions factors are
due to the following: 1) variability in waste stream composition, 2) the meth-
ods used to analyze the waste feed and the cement/waste mix were not as accu-
rate as those used to analyzed the exhaust air stream, 3) fluctuations in the
flow rate of the waste feed and cement/waste mix.
During stabilization operations, organics may be released to the air.
Test results showed a large fraction of the volatile organic compounds in the
hazardous waste stabilized are emitted into the air during the mixing stages.
In addition, particulate emissions measured using EPA Modified Method 5 were
an average of 2.49 Ib/h over all three runs. Finally, the organic vapor ana-
lyzer showed a mass flow rate of total volatile organic hydrocarbons of 21.9
Ib/h averaged over all three runs. Based on an average of 20,738 pounds per
hour of waste was fed to the system, 0.11 percent over three runs, by weight,
of the feed is emitted to the air.
541
-------
REFERENCES
1, PEI Associates, Inc. Field Evaluation of a Hazardous Waste Stabiliza-
tion Operation at (test facility). Prepared for the U.S. Environmental
Protection Agency Office of Research and Development Risk Reduction En-
gineering Laboratory, Contract No. 68-02-4284, September, 1990.
2. Research Triangle Institute. Organic Emissions from Waste Fixation,
Characterization of Nationwide Waste Fixation Practices for Facilities
Subject to RCRA Subtitle C. Prepared for the U.S. Environmental Protec-
tion Agency Office of Air Quality Planning and Standards Emission Stan-
dards Division, August 17, 1989.
542
-------
ASSESSMENT OF THE PARAMETERS AFFECTING THE MEASUREMENT OF
HYDRAULIC CONDUCTIVITY FOR SOLIDIFIED/STABILIZED WASTES
by: DJ. Conrad, S.A. Shumborski, L.Z. Florence, AJ. Liem
Alberta Environmental Centre
Vegreville, Alberta, Canada TOB 4LO
C. Mashni, Project Officer
U.S. Environmental Protection Agency
Cincinnati, Ohio, USA 45268
ABSTRACT
A series of experiments conducted at the Alberta Environmental Centre examined the
variation in hydraulic conductivity (K) within and among three matrices formed by steel mill
baghouse dust treated with 8%, 9% and 10% Normal Portland Cement at a water/cement ratio
of 1:1. Within the 8% and 9% matrices, test gradient (i) and back pressure (P) were combined
into 3x3 factorial treatments.
A permeant-matrix interaction was indicated by K decreasing with time at a rate which
increased with higher cement contents. After hydraulic conductivity testing, the samples were
examined by scanning electron microscopy and energy dispersive x-ray analysis. A cement
hydration product, identified as ettringite had formed in the solidified/stabilized waste pores.
This product may reduce hydraulic conductivity by two orders of magnitude by restricting
conducting pores. Four to seven weeks of testing were required before an acceptable
equilibrium had been reached and statistical comparisons among the i x P treatments could
be made. Within each matrix, gradient was statistically the most significant parameter
accounting for 60% of the variation in results. The overall mean hydraulic conductivity was
significantly greater for the eight percent matrix (10 ± 5 x 10"6 cm.sec"1) compared to the nine
percent matrix (0.06 ± 0.03 x 10"6 cm.sec "') (p<0.01). Therefore, temporal effects, gradient
and cement content were identified as important factors affecting hydraulic conductivity
measurements and have implications for regulatory tests. Bulk density was a useful quality
control criterion for minimizing sampling variance within each matrix.
This paper has been reviewed in accordance with the U.S. Environmental Protection
Agency's peer and administrative review policies and approved for presentation and
publication.
543
-------
ASSESSMENT OF THE PARAMETERS AFFECTING THE MEASUREMENT OF
HYDRAULIC CONDUCTIVITY FOR SOLIDIFIED/STABILIZED WASTES
INTRODUCTION
Solidification and stabilization (SS) technologies are often used to treat hazardous
wastes to reduce the environmental impact of their disposal. Solidification removes free
water, usually by the hydration reactions of lime or cementitious materials, producing a
monolithic solid with reduced surface area. Stabilization with cementitious materials reduces
the solubility of wastes by the alkaline precipitation of metal hydroxides or metal
incorporation into the hydration products of cement.
The long term behaviour of these treated wastes is the subject of much concern.
Depending on the disposal scenario, treated wastes are eventually subject to leaching by
ground water, precipitation, or leachate. If the treated waste is relatively permeable, leachant
flow will be through the bulk of the SS matrix rather than being confined to the external
surface area. Thus, a major benefit of SS treatment, the reduction of surface area available
for leaching, is compromised.
The flow of liquid through a porous medium is described by Darcy's Law. The liquid
superficial velocity (Flowrate/Area) per unit gradient is defined as hydraulic conductivity,
which is a function of the properties of the medium and the liquid. Gradient is defined as the
headloss which occurs over the sample (cm of H2O) divided by the sample length (cm).
Darcy's Law may be written as:
where K is hydraulic conductivity (cm.sec"1), Q is flow rate (cm3.sec4), i is gradient
(dimensionless) and A is cross-sectional area (cm2).
544
-------
Permeability, a property of the medium alone, is related to hydraulic conductivity by
the following relationship:
where K is permeability (cm2), H- is absolute viscosity of the liquidCg.cm^.sec"1), p is density
of the Hquid(g.cm"3) and g is acceleration of gravity (998 cm.sec"2).
When there is no medium-liquid interaction, permeability is an intrinsic and useful
property of a medium. The flow rates of different liquids through a medium can be readily
predicted from its permeability and the properties of the liquids. However, when there are
changes in liquid properties, due to dissolution and in the internal structure of the medium,
as shown in this paper, the meaning of permeability becomes obscure. Since it is the flow
rate of aqueous permeant through SS waste which is of environmental interest, hydraulic
conductivity is the proper terminology and is used herein.
The literature available on hydraulic conductivity measurement with environmental
implications deals predominantly with clay and soil liners. Researchers are interested in the
effects of permeants, specifically inorganic salt solutions (1), organic fluids (2) and landfill
leachates (3). Test parameters such as saturation (4), temporal effects (5, 6) and gradient (6,
7) have been studied. Parker et al. (5) found that the hydraulic conductivity of fly ash-
stabilized soils decreased over time possibly due to fly ash-soil interaction. Carpenter and
Stephenson (6) and Edil and Erikson (7) both noted that hydraulic conductivity declined for
clays as the gradient was increased when flexible wall permeameter cells were used.
Bryant and Bodocsi (8) collected historical data on hydraulic conductivity
measurements for clay liners and analyzed them for the effects of sample variation,
preparation, equilibration and gradient, among other effects. They noted many confounding
effects^ and suggested that suitable experimental designs should be chosen to properly estimate
parametric effects. Longer test periods and statistical approaches to determine equilibrium,
as indicated by stable hydraulic conductivity, were suggested. Soil hydraulic conductivity was
found to be very sensitive to preparation technique. Some results showing decreasing
hydraulic conductivity were explained by sample consolidation resulting from increased
gradients.
Pierce et al. (9), who conducted ruggedness tests using both rigid wall and triaxial cell
permeameters, found that water content, lift thickness and back pressure had the greatest effect
on the measured hydraulic conductivity of a clay liner material. The first two factors pertain
to sample preparation while the third is an instrument measurement parameter. Gradient was
not found to be significant at relatively high levels (i = 100, 200), typical of laboratory tests.
The hydraulic conductivity results exhibited large variability. Thus, inter-laboratory results
could exhibit large variation due to individual laboratories performing hydraulic conductivity
tests at different levels of these sensitive parameters.
545
-------
Stegemann and Cote (10) provided comprehensive information on the results of an
inter-laboratory study evaluating test methods for SS treated wastes. Commercially available
SS treatment processes provided hydraulic conductivities between 3 x 10~6 and 7 x 10"10
cm.sec"1. The inter-laboratory variance was too large to determine differences between
individual products.
The work done on clay liners and soils suggests that if hydraulic conductivity
measurement is to be used as a regulatory test for SS wastes, several factors should be
studied: sample preparation, temporal effects, or time to equilibration, and the instrument
measurement parameters of gradient and back pressure. These factors may affect
reproducibility in terms of intra and inter-laboratory variance. Experiments should also be
conducted across the range of hydraulic conductivities typical of SS wastes. The effects of
gradient and back pressure on measured hydraulic conductivity may also provide the bases for
evaluating the potential of laboratory tests as predictors of field results. Thus, experiments
should include trials at low levels of gradient and back pressure approximating field
conditions.
This study is necessary as SS wastes behave differently than clay liners and soils
during hydraulic conductivity measurement due to their higher compressive strength and
water-reactive cementitious matrix. This paper reports the results of experiments carried out
at the Alberta Environmental Centre which were designed to address the possible effects of:
gradient, back pressure (the outlet pressure used to ensure liquid saturation at a given
gradient), time and matrix.
MATERIALS AND METHODS
SAMPLE PREPARATION
Samples for hydraulic conductivity measurement were prepared from steelmill
baghouse dust, 16-30 mesh silica sand (Badger), ASTM Type 1 Portland Cement (Canada
Cement LaFarge) and tap water. The sample formulations, shown in Table 1, were chosen
to give a range of hydraulic conductivities typical of SS wastes (10'6 to 10"9 cm-sec"1).
546
-------
TABLE 1. HYDRAULIC CONDUCTIVITY SAMPLE FORMULATIONS*.
Matrix
8%
9%
10%
Steelmill
Foundry Dust
42.0
41.0
40.0
Silica
Sand
42.0
41.0
40.0
Typel
Portland
Cement
8.0
9.0
10.0
Water
8.0
9.0
10.0
* Sample formulations are given in weight percent.
Samples were compacted and bulk densities measured according to ASTM D558-82
(11), modified for sample size to ensure similar compaction energies. Samples were prepared
in plastic molds 7.62 cm (diameter) x 15.2 cm (length) and cured for a minimum of 28 days
at ambient temperature and a minimum relative humidity of 95%.
Samples were saturated in the triaxial cell prior to hydraulic conductivity testing by
evacuating with a vacuum pump, isolating the sample, and allowing permeant flow from the
inlet to restore ambient pressure.
HYDRAULIC CONDUCTIVITY
Cured samples were tested for
hydraulic conductivity by a constant
head method based on USEPA SW 846
Method 9100 (12), Samples were
tested concurrently on three Geotest
Model No. S5425 permeameters
(Figure 1). Each permeameter
consisted of a triaxial sample cell with
a flexible membrane confined at a
pressure 21 kPa higher than the inlet
pressure. The permeant inlet and outlet
interfaces utilized a piston and linear
transducer calibrated for volume and
accurate to ±0.01 ml across the
interface volume of 50 ml. Data
collected concurrently by the data
loggers were downloaded to a
computer for calculating hydraulic
conductivities.
DATA LINE
PERMEANT LINE
Figure 1. Hydraulic Conductivity Test Apparatus
547
-------
ELECTRON MICROSCOPY
Electron microscopical analyses were performed on a Hitachi S510 scanning electron
microscope (SEM), a Hitachi X-650 (SEM) with energy dispersive x-ray spectrometer and a
Hitachi H-600 scanning transmission electron microscope (STEM) equipped with a Kevex Be
window x-ray detector.
DETERMINATION OF HYDRAULIC CONDUCTIVITY EQUILIBRIUM
Simple, linear regression was used to identify the time (x = day = independent
variable) when changes in hydraulic conductivity (y = K • 106 = dependent variable) had
attained sufficient stable equilibrium that the regression coefficient, or slope, was not different
than zero, as suggested by Pierce and Witter (13).
Equilibria were reached by 59, 34 and 27 days for 8, 9 and 10 percent matrices
respectively; the null hypothesis, that the estimated slope equalled zero, could not be rejected
for any of the three matrices with a probability of Type I error less than 0.13.
EXPERIMENTAL DESIGN
After equilibrium had been reached as indicated above, gradient and back pressure
were varied and treated as independent continuous variables at three levels in a completely
randomized 3x3 factorial design (see Table 2). Gradient levels corresponded to pressures
of 10 kPa to 340 kPa. Two independent experiments were performed at 8%, and 9% matrix
to provide a range of hydraulic conductivities. Second-degree polynomial coefficients were
calculated using response surface regression (SAS 1988 Release 6.03 (14)) to model the
response of hydraulic conductivity to varying levels of gradient and back pressure.
TABLE 2. TEST PARAMETERS AND LEVELS*
Parameter Level
-1 0 +1
Gradient 8 116 227
Back Pressure
(kPa) 14 69 124
"Performed at 8%, and 9% matrix
548
-------
RESULTS AND DISCUSSION
SAMPLE PREPARATION
The samples had the consistency of soil-cement and thus were compacted with standard
soil-cement procedures (11). The sample bulk densities are shown in Table 3, The percent
relative standard deviation ranges from 0.431% to 0.924%.
TABLES. SAMPLE BULK DENSITIES
Matrix
n+
8%
6
9%
10
10%
6
Bulk Density
(g.cnr3)
Mean
S.D.
RSD (%)
2.270
0.0137
0.604
2.327
0.0215
0.924
2.390
0.0103
0.431
Tn is the number ot samples
Bulk density is a function of gross sample porosity and has been implicated as a source
of variance in hydraulic conductivity measurements. The degree of compaction affects the
measured permeability of clayey silt (15). Stegemann and Cote (10) felt that sample
differences were a major source of variation in their hydraulic conductivity study.
Error introduced by weighing out reagents and sample mold variation was estimated
to be 0.42%. Thus, a quality assurance criterion was instituted which rejected all samples
with bulk densities not within ±0.5% of the sample mean. By instituting strict quality control
for sample preparation, a potential source of within matrix sample variation is reduced.
TEMPORAL EFFECTS DURING EQUILIBRATION
Hydraulic conductivity measurements were recorded when inflow and outflow rates
were determined to be equal (within 5%). The results summarized in Figure 2 show the
temporal effects over 80 days of testing at constant gradient and back pressure. Also shown
are the predicted hydraulic conductivities derived from a nonlinear regression. Hydraulic
conductivity decreased by nearly two orders of magnitude for the highest cement content
sample.
549
-------
100
0,0
10%
20 50 40 50
Elapsed Time (days)
60
70
Figure 2. Variation of Hydraulic Conductivity with Time at 8%, 9% and 10% Matrix
550
-------
The decrease in hydraulic conductivity was modelled according to the equation:
Kx 106 = A(T+ 1)B
where T is elapsed time measured from the first day of testing (T=0) and A and B are the
intercept (initial condition) and slope of the power function, respectively. The regression
results are summarized in Table 4, showing the statistical significance of the models.
TABLE 4. MODEL COEFFICIENTS* FOR THE CHANGE IN
HYDRAULIC CONDUCTIVITY WITH TIME
Matrix
8%
9%
10%
Initial
Value (A)
82.400
1.571
1.356
Standard
Error
±4.361
±0.044
±0.148
Slope
(B)
-0.413
-0.743
-1.476
. Standard
Error
±0.025
±0.032
±0.070
* regressions are significant at p < 0.001; the models explain 98% (minimum) of the variation
in hydraulic conductivity.
The models predict that with time hydraulic conductivity approaches "zero", but in fact
this will never be achieved However, if no future dissolution occurred, hydraulic conductivity
for the system studied would not be the most important waste parameter affecting
environmental loading. Leaching from sample surfaces and gross defects (cracks) would have
the largest impact on waste transport, rather than transport from within the interior by
permeant flow.
The phenomenon of reduced hydraulic conductivity with time has been observed with
hardened cement pastes. Powers et al, (16) noted that the hydraulic conductivity of cement
pastes cured underwater was reduced by six orders of magnitude due to increased hydration.
Powers et al. (17) described the mechanism of this phenomenon whereby volume of hydrated
paste is 2.1 times greater than unhydrated paste and hydration products fill pores and cavities,
causing discontinuities, effectively reducing the number of flow channels. The number and
radii of conducting capillary pores appear to define hydraulic conductivity in cement pastes.
Nyame and Illston (18) compared their hydraulic conductivity data to that predicted by
hydraulic radius theory and suggested a correlation. Hughes (19) considered the effects of
pore characteristics (isotropy and tortuosity) in developing a model for cement-paste hydraulic
conductivity which considered conducting channels as Poiseuille tubes.
The change in hydraulic conductivity with respect to time (slope), the "B" coefficient
(Table 4), shows that hydraulic conductivity decreases more rapidly as the cement content
increases. This indicates aqueous permeant-matrix interactions, specifically cement hydration
during hydraulic conductivity testing. Cement hydration may be promoted by passing aqueous
551
-------
permeant through the sample during hydraulic conductivity measurement. Samples of tested
material were examined by scanning electron microscopy. Typical micrographs are shown in
Figures 3a and 3b. Inlet portions show profuse "fibrous" growth in sample pores (Figure 3a).
Sample portions covered by the latex membrane sidewall where permeant flow is restricted
do not show this profuse fibrous growth. Patel et al. (20) noted a similar phenomenon that
large pore fractions decreased while gel (paste) porosity increased during cement paste
hydration.
X-ray analyses by STEM of individual fibres, dispersed on a carbon coated copper
grid, yielded atomic ratios of Ca/S ranging from 1.75:1 to 2.67:1, as well as significant Al and
traces of Fe. This suggests that the fibres are a calcium sulphoaluminate (AFt) phase of
hydrated cement related to ettringite, 3 CaO.Al2O3.3 CaSO4.31 H2O. The morphology of the
observed fibres is similar to that of the AFt fibres described by Dalgliesh and Pratt (21).
The results of hydraulic conductivity and electron microscopy analyses show that
cementitious wastes will react intimately with aqueous permeants and that cement hydration
reactions continue for some time. This suggests that hydraulic conductivity regulations should
consider temporal effects, by specifying sufficient replication over time, to measure permeant-
matrix interactions. These interactions wiU likely vary between SS treatment techniques and
should be studied on a case by case basis.
Caution is in order when predicting the long term behaviour of wastes based on short
term tests. The hydration reactions discussed above are no doubt promoted by passing water
through the samples during testing in an accelerated test. Not all hydration products may
prove ultimately beneficial. Ettringite for example is formed during the sulphate attack of
cements destroying its monolithic properties.
RESPONSE SURFACE REGRESSION ANALYSES
Hydraulic conductivity data were collected for 8% and 9% matrix and showed that the
overall mean hydraulic conductivity was significantly greater (p^O.Ol) for the 8% matrix (10
± 5 x 10"6 cm.sec"1) compared to the 9% matrix (0.06 ± 0.03 x 10"6 cm.sec"1). For the waste
system and measurement method, chosen hydraulic conductivity can differentiate between
matrix treatments even when the confounding factors of time and instrument measurement
parameters are retained.
The variation in hydraulic conductivity with gradient and back pressure was modelled
by a second order polynomial:
K x 106 = b0 + bjXi + b^ + buXi2 + b^x/ + b12XiX2
where xt and % are the normalized (-1 to 4-1) gradient and back pressure, respectively.
552
-------
a. Permeant Inlet Portion
b. Sidewall Portion
Figure 3. Scanning Electron Microscopy Photographs of SS Waste After Hydraulic
Conductivity Testing
553
-------
The results of the regression analysis are summarized in Table 5 using coded levels
of the test parameters. The predicted hydraulic conductivity contours are shown in Figure 4
using decoded levels of the test parameters.
The linear effect of gradient was significant in both matrices but the effect of back
pressure was significant at only 9% matrix, A negative quadratic component for gradient was
noted for the 8% matrix.
The largest variation in hydraulic conductivity was less than five-fold as a result of
increasing gradient and back pressure from low levels close to field conditions to the high
levels used in accelerated testing. (This insensitivity to pressure factors suggests that the
samples were highly liquid saturated.) Therefore laboratory measurements conducted to
accelerate testing are a reasonable approximation of field conditions.
The hydraulic conductivity contour intervals suggest that hydraulic conductivity is less
sensitive to changes in gradient and back pressure at medium and high levels. Regulatory
tests could minimize variance by specifying relatively high levels of gradient (i = 227, 343
kPa across a 15 cm sample) and back pressure (124 kPa) for testing hydraulic conductivity.
CONCLUSIONS AND RECOMMENDATIONS
For the SS waste analyzed in this study:
1) Hydraulic conductivity was very sensitive to matrix, a difference of only one percent
in cement and water content yielded statistically different results.
2) Sample bulk density was a practical and useful quality assurance parameter to
minimize variability in hydraulic conductivity measurements. By applying an
acceptance criterion of ± 0.5% from the mean, a difference of three orders of
magnitude in hydraulic conductivity was observed between samples differing only by
1% in nominal cement content.
3) A permeant-matrix interaction occurred, which reduced hydraulic conductivity as the
test progressed. A plausible explanation, based on electron microscopy and dispersive
x-ray analysis, is the formation of hydration products which 'plug' the conducting
pores in the matrix.
4) The temporal effects resulting from the above interaction can be described by
mathematical models, which suggest that in the absence of other effects, such as
matrix dissolution, the impact of hydraulic conductivity on contaminant leaching will
be reduced with time.
554
-------
TABLES. RESPONSE SURFACE REGRESSION ANALYSES
Matrix
(wt%)
8
(s.e.)
9
(s.e.)
b0
13.01"
(0.826)
0.06"
(0.005)
b,
2.33"
(0.691)
0.02"
(0.005)
b2
0.18
(0.728)
0.01"
(0.005)
b«
-4.30"
(1.106)
-0.01
(0.008)
b*
-1.21
(1.101)
0.003
(0.0077)
bia
-0.32
(0.846)
-0.001
(0.0063)
n
22
26
i»
0.68
0.58
LOFS
n.s.*
n.s.
*
P
0.001
0.002
en
tn
01
V = coefficient of determination (variance described by the model)
*LOF = (Lack of Fit Test)
*p = probability of a Type I error
= Significant Effect (P<0.01)
*n.s. = not significant at P<.05
-------
.1
•o
2
O
8%
.
•a
8
227.0
205.1
183.2
161 .3
139.4
117.5
95.6
73.7
51 .8
29.9
8.0
I I
9%
I I I I I I I 1 I
14 25 36 47 58 69 80 91 102 113 124
Back Pressure (kPa)
Figure 4. Contour Plots from Response Surface Analysis for 8% and 9% Matrix
556
-------
5) The effects of gradient and, for the 9% matrix, back pressure were statistically
significant, and the sensitivity of hydraulic conductivity was less at higher levels of
these parameters.
General recommendations for regulatory test development:
1) Bulk density should be used as a quality assurance parameter to minimize sample
preparation variability.
2) To estimate maximum field hydraulic conductivity, measurement should be made as
soon as the sample is cured.
3) To minimize variability, temporal effects should be taken into account and
measurements carried out at high levels of gradient and back pressure when
'equilibrium' is reached.
ACKNOWLEDGEMENTS
We gratefully acknowledge the assistance of Dr. Doug Ivey, Dr. Randy Mikula and
Dr. Maria Neuwirth in providing SEM, STEM and x-ray analyses, Mr. Philip Henry in
providing graphics, and Ms. Connie Jackson and Ms. Patricia Soldan in preparing the
manuscript
557
-------
REFERENCES
1. Ather, G., Evans, J.C., Fang, H.-Y., and Witmer, K. Influence of Inorganic Permeants
upon the Permeability of Bentonite. In: Hydraulic Barriers in Soil and Rock, ASTM
STP 874. American Society for Testing and Materials, Philadelphia, 1985. pp 64-73.
2. Acar, Y.B., Olivieri, L, and Field, S.D. The Effects of Organic Fluids on Compacted
Kaolinite. In: Hydraulic Barriers in Soil and Rock. ASTM STP 874. American
Society for Testing and Materials, Philadelphia, 1985. pp 203-212.
3. Fang, H.-Y. and Evans, J.C. Long-Term Permeability Tests Using Leachate on a
Compacted Clayey Liner Material. In: Ground-Water Contamination: Field Methods,
ASTM STP963. American Society for Testing and Materials, Philadelphia, 1988. pp
397-404.
4. Elzeftawy, A. and Cartwright, K. Evaluating the Saturated and Unsaturated Hydraulic
Conductivity of Soils. In: Permeability and Groundwater Contaminant Transport.
ASTM STP 746. American Society for Testing and Materials, 1981. pp 168-181.
5. Parker, D.G., Thornton, S.I., and Cheng, C.W. Permeability of Fly-Ash Stabilized
Soils. Paper presented at 1977 Specialty Conference of the Geotechnical Engineering
Division. American Society of Civil Engineers. Ann Arbour, Michigan. June 13-15,
1977.
6. Carpenter, G.W. and Stephenson, R.W. Permeability testing in the Triaxial Cell.
Geotechnical Testing Journal. 9:1, March 1986. pp 3-9.
7. Edil, T.B. and Erickson, A.E. Procedure and Equipment Factors Affecting
Permeability Testing of a Bentonite-Sand Liner Material. |n: Hydraulic Barriers in
Soil and Rock. ASTM STP 874. American Society for Testing and Materials,
Philadelphia, 1985. pp 155-170.
8. Bryant, J. and Bodocsi, A. Precision and Reliability of Laboratory Permeability
Measurements. EPA Contract No. 68-03-3210-03 U.S. Environmental Protection
Agency, Cincinnati Ohio, 1985. 177 pp.
9. Pierce, J.J., Salfors, G. and Peterson, E. Parameter Sensitivity of Hydraulic
Conductivity Testing Procedure. Geotechnical Testing Journal. 10:4, "December 1987.
pp 223-228.
10. Stegemann, J.A. and Cote, P.L. Summary of an Investigation of Test Methods for
Solidified Waste Evaluation. Waste Management, 10:1990. pp 41-52.
558
-------
11, ASTM D-558-82. Standard Test Methods for Moisture-Density Relations of Soil
Cement Mixtures. American Society for Testing and Materials. Philadelphia, 1988.
V 04.08, pp 108-111.
12, US EPA Method 9100. Saturated Hydraulic Conductivity, Saturated Leachate
Conductivity and Intrinsic Permeability. In; Test Methods for Evaluating Solid Waste.
SW 846, Vol 1., Sect. C. United States Environmental Protection Agency.
Washington D.C., Nov. 1986. pp 9100 1-57.
13. Pierce, JJ. and Witter, K.A. Termination Criteria for Clay Permeability Testing.
Journal of Geotechnical Engineering. 112:1,1986. pp 841-854.
14. SAS Institute Inc. SAS/STAT User's Guide. Release 6.03 Edition. Gary, NC:SAS
Institute Inc., 1988. 1028 pp.
15. Garcia-Bengochea, I. and Lovel, C.W. Correlative Measurements of Pore Site
Distribution and Permeability in Soils. In: Permeability and Ground Water Transport.
ASTM STP 746. American Society for Testing and Materials. Philadelphia, 1981.
pp 137-150.
16. Powers, T.C., Copeland, L.E., Hayes, J.C. and Mann, H.M. Permeability of Portland
Cement Paste. Journal of the American Concrete Institute. 26:3, November 1954. pp
285-298.
17. Powers, T.C., Copeland, L.E. and Mann, H.M. Journal of the Portland Cement
Association Research and Development Laboratories. 1:2, May, 1959. pp 38-48.
18. Nyame, B.K. and Illston, J.M. Relationships Between Permeability and Pore Structure
of Hardened Cement Paste. Magazine of Concrete Research. 33:116,1981. pp 139-
146.
19. Hughes, D,C. Pore Structure and Permeability of Hardened Cement Paste. Magazine
of Concrete Research. 37:133, 1985. pp 227-233.
20. Patel, R.G., Killoh, D.C., Parrott, LJ. and Gutteridge, W.A. Influence of Curing and
Different Relative Humidities upon Compound Reactions and Porosity in Portland
Cement Paste. Materials and Structures. 21:123, 1988. pp 192-197.
21. Dalgleish, BJ. and Pratt, P.C. Fractographic Studies of Microstructural Development
in Hydrated Portland Cement. Journal of Materials Science. 17:8, 1982. pp 2199-
2207.
559
-------
OZONE-ULTRAVIOLET LIGHT TREATMENT OF IRON CYANIDE COMPLEXES
Sardar Q. Hassan and Mark J. Briggs
University of Cincinnati
Cincinnati, Ohio 45221-0171
Mary Beth Foerst
XT Environmental Programs, inc.
Cincinnati, Ohio 45246
Dennis I*. Timber lake
US Environmental Protection Agency
Cincinnati, Ohio 45268
ABSTRACT
Recent concern over the quantity of total cyanide being
disposed of in landfills and discharged to surface water has
prompted the US Environmental Protection Agency Office of Solid
Waste to begin investigating alternative treatment technologies
for the destruction of total cyanide present in electroplating
wastewaters. Combined ozonation and ultraviolet light
irradiation was evaluated at the US EPA Test & Evaluation
Facility for the destruction of ferricyanide complex, a stable
metal cyanide complex that is not destroyed by conventional
treatment technologies. Effects of temperature, uv light
intensity, and reactor configuration on the destruction rate
were studied in this work.
Results from this study confirm that an initial
ferricyanide concentration of ISO mg/L can be reduced to less
than 1 mg/L in 4 hours using an UV intensity of 3 W/L,
temperature of 63.2°c and an ozone dose of 56.5 mg/min/L in a
reactor where the ozone bubbles are dispersed right below the
UV lamps submerged in the solution. Ozone saturation
concentrations and mass transfer coefficients for ozone were
found to decrease with temperature and UV light intensity. At
22.2*C the ozone saturation concentration was 10.7 mg/L and the
mass transfer coefficient was 0.3 min . In a simultaneous
study it was found that irradiating ozone in an external
chamber (gas phase) with subsequent introduction of the
irradiated ozone into the liquid phase produces very little
destruction of the ferricyanide complex.
560
-------
BACKGROUND
Ferricyanide, [Fe(CN)6]3~, is the most abundant metal
cyanide complex remaining in the electroplating and metal
finishing waste streams following alkaline chlorination or
ozonation (1). Other iron cyanide complexes such as
ferrocyanide are quickly oxidized to the stable ferricyanide
during such treatment (2). The resistance of ferricyanide to
conventional treatment is not completely understood, but
previous work has shown it can be destabilized and oxidized in
the presence of ozone and ultraviolet light (1,3,4,5,6).
A review of the literature on the decomposition of
ferricyanide by simultaneous UV irradiation and ozonation
(O3/DV) uncovered some inconsistencies between the
conclusions of previous researchers. There is also a shortage
of data required for further process design. Prober and Melnyk
(6), have shown that reaction rates for the oxidation of
ferricyanide by O3/DV are dependent on UV intensity and not
temperature, whereas Prengle and Mauk (5) have shown that
temperature has a significant effect on the rate of
ferricyanide decomposition by O3/UV (three-fold increase in
the reaction rate when the temperature is raised from 25°C to
66.5°C). Prober and Melnyk, however, show an insignificant
change in the reaction rate when the temperature is raised from
25°C to 50"C.
Another discrepancy found in the literature was the actual
function of UV light. Garrison and Mauk (1) believed that
ferricyanide was destabilized by UV light, making it more
susceptible to ozone oxidation. On the other hand, research by
Prober and Melnyk, using various reactor configurations have
shown significant decreases in reaction rates when ozone is
shielded from UV light within the same reactor. These results
suggest that the primary function of UV light is to activate
the ozone (forming hydroxyl radicals) which then oxidizes the
ferricyanide complex. The following mechanism has been
suggested by Ashmore (7), Barker (8), Peyton (9), and Wallace
(10), for the activation of ozone by UV light:
O3 + hv ——> O3*
°3* * H2° > H2°2* * °2
H202* > 2OH*
The relative reactivity of hydroxyl radicals is illustrated in
Table 1 where reaction rates with classes of organic compounds
are compared to those of ozone (8). As shown, reaction rates
increase by orders of magnitude for hydroxyl radicals as
561
-------
compared to ozone.
It has been postulated that activation of ozone by UV light
(excitation of electrons from the 3-P orbital to the 1-D
orbital of the oxygen atoms) takes place in the gaseous phase,
with hydroxyl radical formation occurring at the bubble
interface. If this postulation is correct/ then ozone can be
activated by UV light outside the reacting solution/ with
hydroxyl radical formation occurring once the activated ozone
contacts water. Also/ if activation of ozone can be achieved
in the gaseous phase then typical problems such as hindrance by
wastewater turbidity can be eliminated. In addition/ since air
has less resistance to penetration by UV light than water/ less
electrical power would be required to achieve the same hydroxyl
radical formation. It is/ therefore/ necessary to study the
effects of ozone activation/ in the gaseous phase/ on the
destruction of ferricyanide.
In addition to the above process parameters/ the mass
transfer rate of ozone into the aqueous solution is also a
critical parameter/ since often the low mass transfer rate is
the primary obstacle to the propagation of the reaction.
Therefore/ it is necessary to develop a reactor which will
provide a high mass transfer rate of ozone.
OBJECTIVES
The objectives of the study were as follows:
A. Develop a reactor which will provide a high mass transfer
rate of ozone and intimate contact of ozone bubbles and UV
light. Determine the saturation concentrations of ozone in
water and the mass transfer rates at two different
temperatures and two different UV light intensities.
B. Determine the rate at which O3/UV can oxidize an iron
cyanide solution containing 150 mg/L of cyanide.
1. Determine the effects of UV light intensity on the
reaction rate for the oxidation of ferricyanide.
2. Determine the effects of temperature on the reaction
rate for the oxidation of ferricyanide.
C. Determine the effects of ozone activation by UV light in
the gas phase prior to contacting the liquid phase.
EXPERIMENTAL SETUP
The complete study was performed in a semi-batch mode in
the laboratory using a 2-liter stainless steel reaction
562
-------
vessel. Figure 1 presents a diagram of the reaction system
including the ozone generator, the stainless steel reactor and
the gas washing bottles. Figure 2 is a cross sectional diagram
of the reactor showing the UV lamps/ thermometer/ and ozone
sparser. The mixer was a centrifugal pump from which the
plastic stationary housing containing the suction and discharge
ends had been removed in order to expose the impeller. The
pump was mounted vertically through an orifice in the reactor
base plate such that the impeller was located on the floor of
the reactor. Various degrees of mixing were achieved by
changing the power input to the mixer motor using a voltage
controller.
Samples from the reactor were obtained through a rubber
septum fitted into a Swagelock fitting in the reactor lid.
Syringes/ varying in sizes from 1 ml to 10 mis, and fitted with
6-inch piercing needles were used to draw samples from the
mid-section of the reactor. Both filling and emptying of the
reactor was accomplished by removing the sample port Swagelock
fitting and inserting a funnel or 1/2 inch plastic siphon hose,
depending upon the operation.
Heating tape was wrapped around the entire length of the
reactor to supply heat to the reactor for the high temperature
operations. The tape was insulated to reduce heat loss. The
power to the heating tape was controlled through a voltage
controller. When necessary ice packs were used to maintain the
reactor at 22.2°C.
The ozone generation system was a PCI Model 6-2 ozone
generator rated at a maximum output of 2.0 Ibs/day at 2%
concentration in combination with a PCI Model ADP-1 air
preparation unit. The ozone generator was fed prepurified lab
air at ambient temperature. Two liters of the ozonated air was
delivered into the reactor and the excess was vented into the
exhaust system. Using absorption in potassium iodide solution
and subsequent titration, the ozone delivery rate to the
reactor was determined to be 113+11.3 mg/min. Ozone which
entered the reactor traveled down a glass tube to a fritted
bubble stone positioned 3/4 inches above the mixing impeller.
Exhaust gases from the reactor could be passed through
potassium iodide traps to determine the quantity of the unused
ozone.
The UV lamps, selected for this study, were submersible,
short wave (254 nm), low pressure mercury ultraviolet lamps
with a maximum power rating of approximately 3 watts per lamp.
Each UV lamp was powered by a separate power supply. UV light
intensity in the reactor could be varied by changing the number
of UV lamps used during the reaction. The lamps were
positioned in such a way that ozone bubbles, coming out of the
563
-------
sparger at the reactor bottom, traveled along the length of the
UV lamps.
All of the experiments were performed at the U.S. EPA Test
and Evaluation Facility located in Cincinnati/ Ohio.
PROCEDURES, RESULTS AND CONCLUSIONS
A. OZONE SATURATION CONCENTRATIONS AND MASS TRANSFER
COEFFICIENTS
Procedure
To determine the transfer of ozone to the stainless steel
reactor, two liters of distilled water were added and the
reactor purged for 15 minutes with prepurified air to remove
any residual volatile organics. The temperature of the water
and reactor were adjusted to the predetermined set point
(22.2*C or 63.2°C) using heating tapes or ice packs. The
mixing rate was adjusted to 1600 rpm using the variable voltage
controller. The ozone charged air leaving the ozone generator
was fed to the bottom of the reactor, where it was dispersed
into the liquid phase through a fritted bubble stone. One set
of test runs was performed without any UV lamps and the other
set was performed with an UV intensity of approximately 3 W/L.
For each set of conditions the absorption runs were made in
triplicate.
Samples of the ozonated solution were collected at 1 minute
intervals for the first 5 minutes of the ozone absorption study
and at 5 minute intervals for the remaining 60 minutes of the
run. The samples were immediately spiked into 20 mis of 2% w/w
potassium iodide solution to trap the dissolved ozone. This
solution was then titrated with sodium thiosulfate and the
ozone concentration in the distilled water was determined.
Results
Figure 3 presents a plot for the typical average ozone
concentration vs. time data for one of the four above mentioned
absorption runs. Ozone saturation concentrations in the
distilled water were determined from the cocentration vs. time
plots. The ozone saturation concentration for three of the
above mentioned four sets of conditions were 10.7 mg/L (at
22.2*C), 7.9 mg/L (at 63.2°C) and 5.4 mg/L (at 22.2°C with 3
W/L UV intensity). At 63.2°C with 3 W/L UV intensity the ozone
concentration in the water was below the detection limit (1.5
mg/L) for the entire period of the absorption run.
564
-------
From tbe ozone saturation concentrations and the absorption
data, the mass transfer coefficients (KLa) for ozone were
determined. By relating the first few minutes of absorption
data for the various conditions (UV intensity and temperature)
and the maximum solubility of ozone for that condition, the
data required to determine the coefficient were calculated.
The KLa values for the above mentioned conditions were 0.30
min~j- (at 22.2'C), 0.22 min'^at 63.2'C), and 0.18
min-1{at 22.2"C with 3 w/L UV intensity).
Conclusions
The absorption data obtained for ozone in the bench-scale
reactor correlated well with the values obtained by previous
researchers. Ouederni (11), using a semi-batch reactor,
obtained an ozone saturation concentration of approximately
12.5 mg/L at 20°C. Also, from the data it was concluded that
saturation concentration of ozone decreases with increase in
temperature or UV light intensity. This is most likely due to
the faster rate of ozone decomposition at higher temperature
and higher UV light intensity.
The mass transfer coefficient of 0.3 min"1 at 22.2'C is
significantly higher than the values obtained by other
researchers (12) which are presented in Table 2. Assuming that
there were no errors in the measurements and calculations it
was concluded that the fabricated reactor provided a high mass
transfer rate of ozone. The results also show that the mass
transfer coefficient decreases with increase in temperature and
UV light intensity.
B. EFFICTS OF TEMPERATURE AND UV LIGHT INTENSITY ON THE
DESTRUCTION OF FERRICYANIDE
Procedure
In order to determine the effects of increased temperature
and UV light intensity a systematic testing program involving
both parameters was devised. Table 3 presents the various UV
light intensities and temperatures tested.
The initial total cyanide concentration for this portion of
the study was 150 mg/L. Examination of the various reaction
rates published in the literature for the oxidation of
ferricyanide showed the probable rate constant (K) to be 0.017
min"*1. Assuming the reaction to be first order and the
reaction system to act as a completely stirred tank reactor
(CSTR) it was calculated that a reaction time of 230 minutes is
reguired for a 98 percent removal of total cyanide. The actual
run time was extended to 240 minutes to ensure a complete 96
percent removal.
565
-------
Treatment of ferricyanide in the bench-scale reactor began
by starting the reactor mixer and adjusting the temperature to
either 22.2'C or 63.2'C, depending upon the specific run
conditions* Once the proper temperature was attained, the
required number of 0V lamps were turned on and the ozone flow
to the reactor was initiated. Ozone off-gas was passed through
the potassium iodide solutions to determine the quantity of
ozone exiting with the off gas. Samples were collected from
the reactor at various times throughout the treatment run as
shown in Table 4. Samples were immediately diluted to 1 liter
in a volumetric flask with 0.01 N sodium hydroxide solution
(pH>ll) to halt any further reaction (13) and to provide the
required volume for analysis.
Results
Table 5 summarizes the total cyanide concentrations in the
reactor following the 4-hour test period for each set of
conditions. These concentrations are averages of three runs
conducted for each set of conditions. Cyanide concentration
histories for all of the different sets of conditions are
plotted in Figure 4.
Conclusions
From the final cyanide concentration data, for runs under
different sets of conditions, it was concluded that changing
the temperature from 22.2"C to 63.2"C resulted in a significant
increase in the reaction rate while changing the UV light
intensity from 1.5 W/L to 3 W/L produced only a moderate
increase in the reaction rate. Neither UV light nor ozone,
used by itself, produced any significant destruction of the
ferricyanide complexes.
The results of this study confirmed that ozone/UV oxidation
is a potentially effective treatment technology for iron
cyanide complexes. Ferricyanide concentrations of 150 mg/L (as
cyanide) can be reduced to less than 1 mg/L in 4 hours of
treatment at a temperature of 63.2+0.5°C, an ozone dose of 56.5
mg/min/L, and UV light intensity of 3 W/L.
From the results of this study and some other previous
investigations it is concluded that destruction of ferricyanide
will occur only if ozone bubbles are exposed to 17V irradiation
in the reactor. Irradiating the ferricyanide solution only
prior to ozonation or ozonated ferricyanide solution does not
achieve any significant destruction of the iron cyanide
complexes. •
566
-------
C. EFF1CTS OF OZONE ACTIVATION BY UV LIGHT III THE GAS PHASE
Procedure
For this study, tbe bench-scale reactor was modified such
that a single DV lamp was suspended above the reactor in an air
tight glass chamber. Ozone was passed from the top of the
chamber, along the light and out the bottom, where it was
routed into the reactor. All gas flow rates, ozone delivery
rates and reactor mixing rates were identical to those
performed earlier, since only one light was used in this
study, uv intensity was 1.5 w/L. A temperature of 63.2°c was
used for this run.
Results
The results of this run showed that little or no
destruction of cyanide was achieved.
Conclusions
The test failed to confirm the hypothesis that ozone
activation by W irradiation in the gaseous phase can provide
similar results as ozone activation in the solution. One
possible reason is the delay between exposure of ozone to DV
light in the external chamber and its entrance into the reactor
where it is supposed to form the hydroxyl radicals. According
to Ashmore (7) the time of excitation and subsequent loss of
radical activity is in the order of microseconds whereas the
lapsed time in this system was approximately 0.5 seconds.
RECOMMENDATIONS
Since it is important to irradiate the ozone bubbles by the
UV lamps, the reactor design should ensure such irradiation.
In a large reactor UV lamps with ozone spargers right below the
lamps should be distributed in a grid system along the cross
section of the reactor. Distances between the UV lamps should
be inversely proportional to the amount of solids formed during
the reaction. Another way of reducing the number of UV lamps
with ozone spargers is to incorporate solids removal devices in
between stages of a multistage reactor.
Since ozone transfer to the liquid is never complete it is
desirable to utilize the ozone in the exhaust gas rather than
to destroy it or release it into the atmosphere. One possible
way of utilizing the ozone in the exhaust gas would be to use a
dual-stage reactor with fresh ozone entering the second stage,
where the highly stable metal-cyanide complexes will be
destroyed, and exhaust gas from the second stage going into the
567
-------
first stage where free cyanides and some less stable
metal-cyanide complexes will be destroyed. This setup would be
highly applicable to the electroplating industry wastes which
contain both free and completed cyanides.
REFERENCES
1. Garrison, R.L., Mauk, C.E./ "Advanced Ozone-Oxidation for
Complexed Cyanides11. Proceedings of the First
International Symposium on Ozone for Water and Wastewater
Treatment, pg 551, 1973.
2. Sober, T.W./ Dagon, T.J., "Ozonation of Photographic
Processing Waste". J. Water Pollution Control Fed., Vol
47, Mo. 8, pg 2114, 1975.
3. Garrison, R.L., Prengle, E.W., "Ozone Based System Treats
Plating Effluent". Metal Progress, 108(6), pg 61, 1975.
4. Hauk, C.E., "Chemical Oxidation of cyanide Species by Ozone
with Irradiation from Ultraviolet Light". Trans. Society
of Hining Engineers, Vol. 260, pg 279, 1976.
5. Prengle, H.W., Mauk, C.E., Ozone/UV Process Effective
Wastewater Treatment. Hydrocarbon Processing, October, pg
82, 1975.
6. Prober, R., Melnyk, P.B., Ozone-Ultraviolet Treatment of
Coke Oven and Blast Furnace Effluents for Destruction of
Ferricyanides. Dept. of Chemical Engineering, Case Western
University, Cleveland, OH, 1977.
7, Ashmore, P.G., Photochemistry and Reaction Kinetics.
Cambridge University Press, pg 36, (1967).
8. Barker, R., Jones, A.R., "Treatment of Malodorants in Air
by the UV/Ozone Technique". Ozone Science and
Engineering, Vol. 10, pp. 405 - 418, 1988.
9. Peyton, G.R., Glaze, W.H., "Destruction of Pollutants in
Water with Ozone in Combination with Ultraviolet Radiation.
3. Photolysis of Aqueous Ozone". Environmental Science and
Technology, Volume 22, Mo. 7, pg. 761, 1988.
10. Wallace, J.L., Vahadi, B., "The Combination of
Ozone/Hydrogen Peroxide and Ozone/UV Radiation for
Reduction of Trihalomethane Formation Potential in Surface
Water". Ozone Science and Engineering, Vol. 10, pp.
103-112, 1988.
568
-------
11. Ouederni, A., Mora, J.C., "Ozone Absorption in water: Mass
Transfer and Solubility". Ozone Science and Engineering,
Volume 9, pp. 1-12, 1987.
12. Roustan, M., Duguet, J.P./ "Mass Balance Analysis of Ozone
in Conventional Bubble Contactors", ozone Science and
Engineering, Vol. 9, pp 289-297, 1987.
13. staehelin, J., Decomposition of ozone in Water: Rate of
Initiation by Hydroxide Ions and Hydrogen Peroxide.
Environ. Sci. and Technol., Vol 16, Ho. 10, pg 676, 1982.
569
-------
FIGURE 1 OzoNE/UV TREATMENT SYSTEM
®
Gas Flow Meter
Spent O3
UV Lamp
Kl Traps
Ozone Generator
Stainless Steel Reactor
FIGURE 2 Cross Section of Reactor
na
Sample Port with
Syringe Septum r±
Thermometer
Gas Delivery Tube
Ozone Bubble Stone —
UV Power Supply
_3 Watt UV Lamps
Mixer
570
-------
FIGURE 3
Average Ozone ADsorpiion in Water
Ozone (mg/U
o
20
30 40
Time (min)
Temperature: 22.2 + 0,9 oC
so
60
70
FIGURE 4
Effect of Temperature and UV Intensity
on the Removal of Ferricyanide
Total Cyanide (mg/U
Ozone Only (22 C)
Ozone * UV (63 C)
o so
All UV Intensities at 3 W/L
1OO ISO
Time (min)
571
ZOO
250
-------
TABLE 1
COMPARISON OF REACTION RATES OF OZONE AND HYDROXXL RADICALS
WITH CLASSES OF ORGANIC COMPOUNDS (8)
compound k (L/nole/s)
°3 OH
Olefins 1 to 450 x IO3 109 to 1C11
S-containing organics 10 to 1.6 x io3 IO9 to IO11
Phenols 1C3 IO9
N-containing organios 10 to IO2 IO8 to 1010
Aromatics 1 to IO2 IO8 to I0io
Acetylenes 50 IO8 to IO9
Aldehydes 10 109
Ketones l 109 to 1010
Alcohols 10~2 to l io8 to IO9
Alkanes 10~2 106 to 109
Carboxylic acids 10~3 to 10~2 IO7 to IO9
TABLE 2
MASS TRANSFER COEFFICIENTS FOR OZONE INTO WATER DETERMINED BY
VARIOUS RESEARCHERS {12)
..-_
Researcher K^a (min~A)
Praserthdam 0.067 - 0.165
Tsuno 0.028 - 0.039
Jackson 0.044 - 0.048
Deckwer 0.028 - 0.144
572
-------
TABLE 3
RUN CONDITIONS FOR TEMPERATURE AND ULTRAVIOLET LIGHT
INTENSITY OPTIMIZATION
Run
A
B
C
D
E
F
Temperature
<°C)
21.1
21.1
21.1
21.1
65.5
65.5
UV Intensity
(W/L)
0
3.0
3.0
1.5
1.5
3.0
Ozone
Dose
(mg/min/L)
56.5 + 6
0
56.5 + 6
56.5 + 6
56.5 + 6
56.5 + 6
Number of
Samples
54
54
54
54
54
54
TABLE 4
SAMPLING FREQUENCY AND SAMPLE VOLUMES FOR TEMPERATURE AND
ULTRAVIOLET LIGHT OPTIMIZATION RUNS
Portion of Run
(min)
Sampling Interval
(min)
Sample Volume
(ml)
0-60
60-120
120-180
180-240
10
10
20
20
1
2
4
6
TABLE 5
TOTAL CYANIDE CONCENTRATIONS AFTER 4 HOURS OF TREATMENT
Ozone
(mg/min)
UV Light
(W/L)
Temperature
-------
DEVELOPMENT OF LDR STANDARDS
FOR CONTAMINATED SOIL AND DEBRIS
by: Carolyn K. Offutt
U.S. Environmental Protection Agency
Office of Solid Waste and Emergency Response
Washington, DC 20460
Joan O*Neill Knapp
Teresa A. Pagano
CDM Federal Programs Corporation
Fairfax, Virginia 22033
ABSTRACT
The RCRA Land Disposal Restrictions (LDRs) require that soil
and debris that are contaminated with hazardous waste be treated
prior to land disposal. Soil and debris from Superfund remedial
actions, RCRA corrective actions, and possibly underground
storage tank (UST) sites that are destined for land disposal will
be affected by the LDR regulation. Recognizing that Superfund
soil and debris would be much more difficult to treat than many
of the industrial process wastes that were used to develop the
existing LDR best demonstrated available technology (BOAT)
standards, the EPA Office of Emergency and Remedial Response
(OERR) launched an effort to investigate the treatability of
soil. Existing soil treatment data were collected in 1987 and
1988. The data collected were used in the development of
alternate treatability variance levels for soil and debris and
were summarized in the "Summary of Treatment Technology-
Effectiveness for Contaminated Soil" (1). EPA is actively
involved in the collection of additional data on the treatment of
soil and debris which will be used in the development of the LDR
regulations for soil and debris. The purpose of this paper is to
summarize what OERR has learned about soil treatment
effectiveness and to present the technical issues that EPA is
facing in the development of the final treatment standards.
INTRODUCTION
Section 3004(m) of the Resource Conservation and Recovery
Act (RCRA) mandates that the EPA require treatment of hazardous
wastes prior to land disposal. Known as the "Land Disposal
Restrictions" (LDRs), these regulations may apply to hazardous
industrial process wastes as well as contaminated soil, sludge
574
-------
and debris from Superfund and RCRA facilities that are destined
for land disposal.
The 1989 Superfund Management Review (also known as the
90-Day Study) by the Office of Solid Waste and Emergency Response
(OSWER) acknowledged that Superfund response actions may not be
able to meet existing RCRA treatment standards based on "best
demonstrated available technology" (BDAT) under the LDRs. The
existing LDR regulations may limit the potential treatment
technologies available for Superfund clean-ups, with technologies
such as soil washing, stabilization, and biological treatment,
being precluded because they may not meet the highest level of
performance required by LDRs. In contrast, the 90-Day Study
encouraged the greater use of innovative technologies and urged
the reduction of non-technical barriers, such as regulatory and
policy constraints, that inhibit the use of treatment
technologies, while preserving the intent and spirit of
applicable RCRA regulations.
OSWER recognized the potential limitation on treatment
technologies for Superfund actions and developed a process to use
LDR treatability variances for soil and debris. Guidance was
issued to the Regions through the Superfund LDR Guide 6A,
"Obtaining a Soil and Debris Treatability Variance for Remedial
Actions," (OSWER Directive 9347.3-06FS) in July 1989 and revised
in September 1990 (2). Superfund LDR Guide 6B, "Obtaining a Soil
and Debris Treatability Variance for Removal Actions," (OSWER
Directive 9347.3-07FS) was issued in December 1989 and revised in
September 1990 (3). These guides describe the treatability
variance process, include alternate treatment levels to be
obtained under treatability variances, and identify treatment
technologies which have achieved the recommended levels.
A memorandum issued on November 30, 1989 by OSWER entitled
the "Analysis of Treatability Data for Soil and Debris:
Evaluation of Land Ban Impact on Use of Superfund Treatment
Technologies," (OSWER Directive 9380.3-04) provides support for
decisions by the Regions to use treatability variances, when
appropriate (4). The analysis identifies some of the key
technical considerations to be evaluated in obtaining a
treatability variance.
OSWER recognizes that the use of treatability variances
represents an interim approach and is actively in the process of
acquiring additional data for developing separate treatment
standards for contaminated soil and debris.
The collection of data which supports the development of
regulations for contaminated soil and debris is a joint effort by
the OSWER's Office of Emergency and Remedial Response (OERR),
Office of Solid Waste (OSW), and Technology Innovation Office
575
-------
(TIO), and the Office of Research and Development (ORD) Risk
Reduction Engineering Laboratory (KREL). The initial data
collection effort by the OERR that produced the data for the
development of the treatability variance levels also identified
the types of data needed to develop treatment standards for soil.
This paper describes both the conclusions drawn by OERR to date
as well as the unique considerations of soil treatment which
require further investigation. Ongoing data collection and
evaluation activities are also described.
ANALYSIS OF TREATMENT EFFECTIVENESS
OERR launched an extensive effort in 1987 and 1988 to
collect existing data on the treatment of soil, sludge, debris,
and related environmental media. The results from several
hundred studies were collected and reviewed.
All applicable treatment information from the best
documented studies was extracted, loaded into a data base, and
analyzed to determine the effectiveness of technologies to treat
different chemical groups (1).
Based on this analysis, a number of technologies commonly
used in the Superfund program provide substantial reduction in
mobility and toxicity of wastes as required in Section 121 of the
Superfund Amendments and Reauthorization Act (SARA) of 1986. For
example:
o Thermal destruction has been effective on all organic
compounds, usually accomplishing well over 99%
reduction of organics.
o Although the data indicate that PCBs, dioxins, furans,
and other aromatic compounds have been dechlorinated to
approximately 80%, more recent data indicate that
removal efficiencies may approach 99.9%.
o Bioremediation successfully treats many halogenated
aliphatic compounds, non-halogenated aromatics,
heterocyclics, and other polar compounds with removal
efficiencies in excess of 99%.
o Removal efficiencies for low temperature thermal
desorption have been demonstrated with averages up to
99% for non-polar halogenated aromatics and with
treatment often exceeding 90% for other polar organics.
o Soil washing and chemical extraction data on organic
compounds indicate average removal efficiencies of
approximately 90% for polar non-halogenated organics
and 99% for halogenated aromatics, with treatment often
576
-------
exceeding 90% for polynuclear aromatics. The soil
washing process, with optimized solvent selection, has
demonstrated removal efficiencies often exceeding 90%
for volatile and non-volatile metals.
o Immobilization can achieve average reductions in
mobility of 93% for volatile metals, with reductions in
mobility often exceeding 90% for non-volatile metals.
Immobilization processes, while not actually destroying
the organic compounds, reduce the mobility of
contaminants an average of 99% for polynuclear aromatic
compounds. Immobilization may not effectively
stabilize some organic compounds, such as volatile
organics, and the long-term effectiveness of
immobilization of organics is under evaluation.
CONCLUSIONS REGARDING SOIL TREATMENT TECHNOLOGY EFFECTIVENESS
Contaminated soils can be treated via three basic
mechanisms: (1) destruction of the contaminants through
alteration to a less toxic compound; e.g., thermal destruction,
dechlorination, bioremediation; (2) physical transfer and
concentration of the contaminants to another waste stream for
subsequent treatment or recovery; e.g., low temperature thermal
desorption and chemical extraction, soil washing; and (3)
permanent bonding of the contaminants within a stabilized matrix
to prevent future leaching; e.g., immobilization and
vitrification. In general, the destruction technologies are
effective in reducing the toxicity of many organic contaminants.
The physical transfer technologies reduce the toxicity and often
the volume of selected organic and inorganic contaminants. While
the bonding technologies are most effective at reducing the
mobility and, therefore, the toxicity of inorganic contaminants,
some increasing effectiveness is being demonstrated on selected
organic contaminants as well. Figure 1 presents a summary of
these basic conceptual conclusions. A more detailed discussion
follows.
The technologies that have been widely demonstrated on soils
are thermal destruction for organic contaminants and
immobilization for inorganic contaminants. While these two
technologies may be highly effective in treating particular
classes of compounds, neither provides an ideal solution to
complex mixtures of organic and inorganic contaminants, which are
common at Superfund sites. The inherent difficulty in treating
contaminants in a soil matrix, where waste conveyance and mixing
are in themselves complicated unit operations, contributes to the
need to find special solutions. Other issues, such as landfill
capacity and cost, cross-media impacts, and natural resource
conservation, also support the need to develop and use
577
-------
en
»^J
CO
Contaminant
Volatile
Organics
Semi-Volatile
Organics
Metals
Technology
Destruction
•
•
X
Physical Transfer
•
«
Q
Stabilization
X
Q
•
Demonstrated Effectiveness X Not Effective, Not Advised
Potential Effectiveness
(More Data Required)
Figure 1. Soil Treatment Effectiveness - Conceptual Approach
-------
alternative and innovative treatment technologies for
contaminated soil.
Because of EPA's ultimate goal of developing LDRs for
contaminated soil and debris, this study evaluates a number of
treatment options that are applicable to excavated soils.
In-situ soil techniques, such as some types of bioremediation,
soil vapor extraction, in-situ immobilization, and combined
ground water and vadose zone soil treatment were not included in
the scope of this evaluation. In-situ techniques should also be
considered when researching remediation measures for a
contaminated soil problem. When in-situ technologies are used at
Superfund sites, the LDRs may not be applicable because the waste
has not been excavated and subsequently "placed" in a landfill or
other RCRA unit.
Based upon the data collected and evaluated by OERR from
more than 200 soil treatment tests, conclusions were developed
regarding the effectiveness of six soil treatment technology
groups for each of eleven contaminant treatability groups. For
destruction and physical transfer technologies applied to organic
contaminants, the removal efficiency was analyzed. This
evaluation factor was replaced by the reduction in mobility for
the following technologies: immobilization, chemical extraction,
and soil washing. The principles of operation and the
effectiveness of treatment on organic and inorganic contaminants
are presented below.
THERMAL DESTRUCTION
Principle of Operation
o Thermal destruction uses high temperatures to
incinerate and destroy hazardous wastes, usually by
converting the contaminants to carbon dioxide, water,
and other combustion products in the presence of
oxygen.
Effectiveness onOrganics
o This technology has been proven effective on all
organic compounds, usually accomplishing well over 99%
removal.
o Thermal destruction technologies are equally effective
on halogenated, non-halogenated, nitrated, aliphatic,
aromatic, and polynuclear compounds.
o Incineration of nitrated compounds such as
trinitrotoluene (TNT) may generate large quantities of
nitrous oxides.
579
-------
Effectiveness on Inorganics
Thermal destruction is not an effective technology for
treating soils contaminated with high concentrations of
some metals.
High concentrations of volatile metal compounds (lead)
present a significant emissions problem, which cannot
be effectively contained by conventional scrubbers or
electrostatic precipitators due to the small particle
size of metal-containing particulates.
Non-volatile metals (copper) tend to remain in the soil
when exposed to thermal destruction? however, they may
slag and foul the equipment.
DECHLORINATION
Principle of Operation
o Dechlorination is a destruction process that uses a
chemical reaction to replace chlorine atoms in the
chlorinated aromatic molecules with an ether or
hydroxyl group. This reaction converts the more toxic
compounds into less toxic, more water-soluble products.
The transformation of contaminants within the soil
produces compounds that are more readily removed from
the soil. An evaluation of the end products is
necessary to determine whether further treatment is
required.
gffectiveness on Qrganics
o PCBs, dioxins, furans, and other aromatic compounds
(such as pentachlorophenol) have been dechlorinated to
approximately 80% removal, with more recent data indicating that
removal efficiencies may approach 99.9%.
o Other limited laboratory data suggest potential
applicability to other halogenated compounds including
straight-chain aliphatics (such as 1,2-dichloroethane).
The removal indicated by the data may be due in part to
volatilization.
o Although no data were available for halogenated cyclic
aliphatics (such as dieldrin), it is expected that
dechlorination will be effective on these compounds as
well.
o When non-halogenated compounds are subjected to this
process, volatilization may occur.
580
-------
Effectiveness on Inorganics
Dechlorination is not effective on metals, and high
concentrations of reactive -metals (such as aluminum),
under very alkaline conditions, hinder the
dechlorination process.
BIOREMEDIATION
Principle of Operation
o Bioremediation is a destruction process that uses soil
microorganisms including bacteria, fungi, and yeasts to
chemically degrade organic contaminants.
Effectiveness on Organics
o Bioremediation appears to successfully treat many
halogenated aliphatic compounds (1,1-dichloroethane),
non-halogenated aromatics (benzene), heterocyclics
(pyridine), and other polar compounds (phenol) with
removal efficiencies in excess of 99%; however, the
high removal implied by the available data may be a
result of volatilization in addition to bioremediation.
o More complex halogenated (4-4'DDT), nitrated
(triazine), and polynuclear aromatic (phenanthrene)
compounds exhibited lower removal efficiencies, ranging
from approximately 50% to 87%.
o Poly-halogenated compounds may be toxic to many
microorganisms.
Effectiveness on Inorganics
o Bioremediation is not effective on metals.
o Metal salts may be inhibitory or toxic to many
microorganisms.
LOW TEMPERTURE THERMAL DBSPORTION
Principle of Operation
o Low temperature thermal desorption is a physical
transfer process that uses air, heat, and/or mechanical
agitation to volatilize contaminants into a gas stream,
where the contaminants are then subjected to further
treatment. The degree of volatility of the compound
rather than the type of substituted group is the
limiting factor in this process.
581
-------
Effectiveness on Organics
o Removal efficiencies have been demonstrated by these
units at bench, pilot, and full scales, ranging from
approximately 65% for polynuclear aromatics
(naphthalene), to 82% for other polar organics
(acetone) and 99% for non-polar halogenated aromatics
(chlorobenzene).
Effectiveness on Inorganics
o Low temperature thermal desorption is not generally
effective on metals.
o Only mercury has the potential to be volatilized at the
operating temperatures of this technology.
CHEMICAL EXTRACTION AND SOIL WASHING
Principleof Operation
o Chemical extraction and soil washing are physical
transfer processes in which contaminants are
disassociated from the soil, becoming dissolved or
suspended in a liquid solvent. This liquid waste
stream then undergoes subsequent treatment to remove
the contaminants and the solvent is recycled, if
possible.
o Soil washing uses water as the solvent to separate the
clay particles, which contain the majority of the
contaminants, from the sand fraction.
o Chemical extraction processes use a solvent which
separates the contaminants from the soil particles and
dissolves the contaminant in the solvent.
Effectiveness onOrganics
o The majority of the available soil washing data on
organic compounds indicates removal efficiencies of
approximately 90% for polar non-halogenated organics
(phenol) to 99% for halogenated aromatics
(chlorobenzene), with lower values of approximately 71%
for PCBs to 82% for polynuclear aromatics (anthracene).
o The reported effectiveness for these compounds could be
due in part to volatilization for compounds with higher
vapor pressures (such as acetone).
582
-------
This process is least effective for some of the less
volatile and less water soluble aromatic compounds.
Effectiveness on Inorganics
The chemical extraction process, with optimized solvent
selection, has demonstrated removal efficiencies of 85%
to 89% for volatile metals (lead) and non-volatile
metals (copper), respectively.
IMMOBILIZATION
Principle of Operation
Immobilization processes reduce the mobility of
contaminants by stabilizing them within the soil
matrix, without causing significant contaminant
destruction or transfer to another medium.
Volatile organics will often volatilize during
treatment, therefore an effort should be made to drive
off these compounds in conjunction with an emission
control system.
Effectiveness on Oraanics
Reductions in mobility for organics range from 61% for
halogenated phenols (pentachlorophenol) to 99% for
polynuclear aromatic compounds (anthracene).
Immobilization is also effective (84% reduction) on
halogenated aliphatics (1,2-dichloroethane).
Some organic mobility reductions of the more volatile
compounds may actually be removals as a direct result
of volatilization during the exothermic mixing process
and throughout the curing period.
The immobilization of organics is currently under
investigation, including an evaluation of the
applicability of analytical protocols (EP, TCLP, total
analysis) for predicting long-term effectiveness of
immobilization of organics. The preliminary available
data indicate that significant bonding takes place
between some organic contaminants and certain
organophilic species in the binding matrix? however,
immobilization may not effectively stabilize some
organic compounds, such as volatile organics.
583
-------
E f fectivenes s on Inorganics
o Immobilization can accomplish reductions in mobility of
81% for non-volatile metals (nickel) to 93% for
volatile metals (lead).
The effectiveness of the six technologies to treat soil was
classified as having demonstrated effectiveness, potential
effectiveness, or no expected effectiveness for the eleven
contaminant groups (Figure 2) . The ratings were based on removal
efficiency, scale of operation, and potential for adverse effects
as follows:
o Demonstrated Effectiveness: A significant percentage
of the data, at least 20%, is from pilot or full scale
operations, the average removal efficiency for all of
the data exceeds 90%, and there are at least ten data
pairs.
o Potential Effectiveness: The average removal
efficiency for all of the data exceeds 70%.
o No Expected Effectiveness: The average removal
efficiency for all of the data is less than 70% and no
interference from the contaminants in the soil is
expected.
o No Expected Effectiveness: Potential adverse effects
to the environment or the treatment process may occur.
For example, high concentrations of metals may
interfere with biological treatment.
In some cases, a different rating was selected when
additional qualitative information and engineering judgment
warranted. Two ratings were selected if the compounds within a
treatability group were so variable that a range of conclusions
could be drawn for a particular technology.
Although some of the data upon which the analysis is based
have limited quality assurance information, the data,
nevertheless, do indicate potential effectiveness (at least 90%
to 99% reduction of concentration or mobility of hazardous
constituents) of treatment technologies to treat Superfund
wastes. Some reductions in organic concentrations or organic
mobility of more volatile compounds may actually represent the
removal of those compounds as a direct result of volatilization.
Technologies where this is most likely to occur include
dechlorination, bioremediation, soil washing, or immobilization,
and consideration of appropriate emission controls is required.
Percentage removal reductions (removal efficiencies) are not
always a good measure of effectiveness, especially when high
584
-------
in
CO
en
^^* ^TECHNOLOat
TfteATABiLiTY OROU?""*-^
NON-POLAR HALOGENATED
AflOMATlCS
9*01)
PCB«, HALOGENATED
OIOXINS,FUHANS,ANO
THEIB PRECURSORS
(WfM)
HALOGENATED PHENOLS,
CRESOLS, AMINES, THIOLS.
AND OTHER POUR
AROMATICS (W03)
HALOGENATED
ALIPHATIC COMPOUNDS
(VW4)
HALOGENATED CYCLIC
ALIPHAT1C8, ETHERS,
ESTERS, AND KETONES
(VMS)
NtnWTED COMPOUNDS
(WOS)
HETEROCYCUCSAN0
SiMPLE NON-HALOGENATED
AROMAT1CS
own
POLYNUCLEAH
AROMATICS
C*W!
OTHER POUR
NON4WLOOENATEO
ORGANIC COMPOUNDS
(W08)
NON-VOLATILE
METALS
(W10)
VOLATILE
METALS
0*fi)
THERMAL
DESTRUCTION
•
•
•
•
•
•
•
•
•
o1
x1
DECNLORINATION
O
e
Q
e2
o1
o1
o2
o2
o2
V
o1
4
BIOREMEDIATION
O3
o3
o
o2
e1
~ e
o2
o
o2
ox1
ox1
LOW TEMPERA TUBE
THERMAL DESOflPTION
• O
1
o
o
•
o1
o1
•
o
o
o1
o1
CHEMICAL EXTRACTION
AND SOIL WASHING
Q
O
e
o
e1
o
o
o
o
o
o
4
HI MOBILIZATION
Q
Q1
e3
Q2
O1
o1
e2
o
o2
•
•
• Demoosinued iffectwmw
O Mo Expected Effecttvenees (no expected Interference to process)
X No Expected Effecttvensss (potential adverse effects to environment or
process)
1 Delaware not available tor ttitelreatabillty group. Conclusions are drswn from data for
compounds with simitar physical and chemical characteristics.
2 High removal efficiencies Implied by the data may be dua to volatilization or soil washing.
3 The predicted effectiveness may be different than the data Imply, due n limitations in the
test conditions.
4 These technologies may have limited applicability to high levels of organka.
Figure 2. Predicted Treatment Effectiveness for Contaminated Soil
-------
concentrations remain in the residuals. Some of the performance
observations are based upon a relatively small number of data
points and may not extrapolate well to the broad array of soils
requiring treatment.
QUANTIFYING TECHNOLOGY EFFECTIVENESS AND LIMITATIONS
TECHNOLOGY LIMITATIONS
A variety of potential limitations to the effective
treatment of Superfund wastes were identified in the analyses of
data from OERR's original survey. The EPA offices of OERR, OSW,
TIO, and ORD are now working together to identify technology
limitations and their impact on technology effectiveness.
The data suggest that the treatment of soil and debris with
organic contamination, by technologies other than thermal
destruction, will not be able to consistently achieve BDAT
standards previously developed for industrial process wastes.
The difficulty in treating soil and debris is a direct result of
the levels of contaminants, the types/combinations of
contaminants, the type of matrix, particle size, and other
physical and chemical characteristics of the soil and debris.
The residual concentrations in contaminated soil treated by
technologies other than thermal destruction is highly dependent
upon the concentrations in the untreated soil. Therefore, when
evaluating technologies other than thermal destruction, the
ability of those technologies to treat high concentrations of
organics should be considered. The number and types of
contaminants must also be carefully screened. Organic and
inorganic contaminants may require different treatment
technologies, thus requiring a treatment train. In some cases,
different technologies may be necessary for soils and sludges.
In addition, the distribution of contaminants often is also very
non-homogeneous and is dependent on patterns of contaminant
deposition and transport.
The complex nature of solid waste matrices, such as
contaminated soil from a Superfund site, severely complicates the
treatment process. Soil is a non-homogeneous living medium, and
the proportion of clay, organic matter, silt, sand, debris, and
other constituents can affect the treatability of a contaminated
soil. For example, the complex bonding forces that are exhibited
by various soil fractions, particularly clays and organic matter,
can be difficult to counteract and can affect the treatability of
contaminated soil. To further complicate these circumstances,
the age of many of these sites has allowed significant
opportunity for environmental weathering of the contaminants and
the medium.
586
-------
Collectively, these conditions make the treatment of
contaminated soil, "old" sludge, and debris a formidable
technical challenge. EPA intends to quantify the effects of
these factors, and the approach is to analyze the existing
treatment data for the effects of these factors. Specific
parameters affecting performance will be identified from existing
data; parameters include: soil morphology (particle size
distribution), clay content, permeability, total organic carbon,
cation exchange capacity and as many as twenty other parameters.
Differences in treatment performance among different
technologies, contaminants and soil and debris types will be
investigated.
DATA COLLECTION
EPA is in the process of developing the final regulations
for contaminated soil and debris. The initiatives EPA has taken
involve collecting all existing information on the treatment of
soil and debris to supplement the first data collection effort
and conduct experimental tests, when necessary, to better
understand the process (Figure 3). The EPA offices of OERR, OSW,
TIO, and ORD are working together in these efforts due to the
complexity of developing standards for soil and debris.
Discussion of the initiatives follows.
Existing Data Collection
The targets for existing soil and debris treatment data
include recent remedial/removal actions, Department of Defense
(DOD) and Department of Energy (DOE) actions, Superfund
Innovative Technology Evaluation (SITE) program demonstrations,
and activities conducted by private research organizations and
vendors. The information that is being requested includes data
on performance as well as other information important for
technology transfer. Parameters of interest include:
contaminants treated, scale of the test, measured contaminant
concentrations before and after treatment, QC protocols, design
and operating parameters of the treatment system, methods to
improve performance and problems encountered in treatment. The
information that is collected will be entered in the soil and
debris data base, designed specifically for storing and managing
this information.
Soil Treatment Tests
The treatment tests that are being performed are tests on
contaminants and technologies that lacked adequate treatment
performance data, but would be available technologies for
treating contaminated soil and debris (CS&D). Ten treatment
tests are planned; the technologies that will be tested include
bioremediation, low temperature thermal desorption, chemical
587
-------
Remedial &
Removal
Action Data
SITE
Program Data
Treatability Variance
Levels for CS&D
Original OERR Soil Treatment
Data Report 500 Documents
Treatment
Train Data
CS&D
Data Base
Illliiiiiiiiiiiimimimmimiimiiimmiiiiilllll
Data From
10 Soil Treatment
Tests
Variability
Data
BDAT Standards
for CS&D
"Old"
Sludge Data
KEY:
DataBase
Planned
Data Source ^ ^ Planned Output
— — — - Data Generated by CS&D program
Figure 3. Development Of LDRs For CS&D Data Collection Approach
588
-------
extraction, soil washing, and stabilization. The technologies
will be applied to different types of soils and contaminants.
For example, the biotreatment tests will be conducted on three
soil types with soil classifications ranging from sandy to
clayey. In addition, different types of contaminants, including
soils high in PNAs, PCBs and metals, will be tested. The
stabilization technology will be tested as both a primary
technology and as a residual treatment.
The treatability tests will be conducted according to the
OSW "Quality Assurance Program Plan for Characterization Sampling
and Treatment Tests Conducted for the Contaminated Soil and
Debris Program" (5) and site specific Sampling and Analysis
Plans. The individual sampling plans specify holding times,
analytical methods, chain of custody, and quality control
measures, such as blanks and spikes. The tests will include
measurements of contaminant concentrations before and after
treatment, and measurements of the waste characteristics that
affect the performance of soil treatment technologies. Examples
of waste characteristics that affect treatment performance
include but are not limited to moisture content,
oxidation/reduction potential, and particle size distribution;
the parameters that affect performance are listed in the QA
Program Plan,
Treatment Trains
OERR recognizes that much of the soil and debris from
Superfund sites are mixtures of contaminants and that individual
contaminants may need to be treated differently. Treatment
trains may be utilized in these cases. EPA wants to know the
types of technologies applied to mixtures of contaminants and the
effectiveness of the system. The major source of this type of
data will be from existing data, however, several of the
treatment tests will involve treatment trains. The treatment
trains used in the tests will be a technology for treating the
organic contaminants and stabilization/solidification to treat
the inorganics (metals) remaining in the soil residues.
Debris Treatment
Parallel with the effort to collect data on soil is an
effort to collect existing information on the treatment of
debris. The first data collection effort obtained very limited
data on debris treatment. The studies indicated that debris
could constitute as much as fifty percent of the contaminated
media, such as might be found at a wood preserving site. OERR
also recognized that the sampling procedures used to provide
representative samples of debris contamination were not well
documented. Recognizing the importance of debris, the CS&D
Program has implemented a comprehensive review of debris
589
-------
sampling, analysis and treatment. The treatment technologies
that require treatment performance data will be tested by the
CSSD Program. The characteristics of debris that have been
determined to affect treatment include permeability and
destructibility. The potential treatment technologies that have
been identified for debris to date are destruction, chemical
extraction, physical removal, and sealing/solidification.
"Old" Sludge
The OERR data survey identified the existence of large
quantities of "old" sludges on Superfund sites. These sludges
have aged or weathered, and are different than typical RCRA
sludges. The data on "old" sludge indicated that sludges are not
consistently defined in the literature. Furthermore, these
sludges, when identified, had higher concentrations of
contaminants than soils, and as a result, did not meet
treatability variance levels as frequently as soil. Of the OERR
survey data, 55% of the sludge treatment tests met variance
levels, while 78% of the soil treatment tests met variance
levels. These results indicate that sludge may require separate
treatment standards. In order to quantify the treatability of
sludges for regulatory development purposes, more data will be
collected on the characteristics and treatability of sludges.
Existing data will be collected as part of the data collection
effort, and characterization tests will be conducted on sludges
from Superfund sites to obtain the physical and chemical
characteristics of "old" sludge.
Variability Factors
Soil Morphology
Because the variability of the soil matrix may have
significant effects on the ability of a technology to perform,
EPA is conducting a special project to test the effects of soil
morphology or composition on treatment technology performance.
Preliminary data indicate that clayey soils are treated less
effectively than silty or sandy soils by some technologies. To
evaluate this finding, experimental treatment tests will be
conducted on three different soil types - sandy, silty, and
clayey. Each soil type will be subjected to low temperature
thermal desorption, solvent extraction, and
solidification/stabilization. Data generated in this study and
the available treatment data will be used to develop a
correlation between soil type and treatment effectiveness.
Materials Handling, Preprocessing and Treatment
Additional factors influencing treatment performance involve
materials handling, pretreatment processing, and design and
590
-------
operation of the treatment system. The previous OERR data survey
indicated that all three of these factors can have important
effects on treatment performance and therefore they are being
evaluated in the current study.
A critical element in soil treatment is materials handling.
Special approaches to waste transfer throughout the treatment
system are particularly important for solids and viscous sludges,
where traditional conveyance methods are frequently ineffective.
Slugs of material or debris tend to jam treatment equipment,
resulting in breakage, downtime, and the potential for
uncontrolled releases to the environment.
The preprocessing of waste to maximize homogeneity and
modify the waste characteristics is also important to successful
treatment technology operation. Any treatment technology will
operate most efficiently and cost effectively when it is designed
and utilized to treat a homogeneous waste with a narrow range of
physical/chemical characteristics. If contaminant types and
concentrations, waste viscosity, BTU content, moisture content,
acidity, alkalinity, etc. vary widely, control of the system can
be difficult and costly to maintain. Many of these waste
characteristics can be modified and improved with appropriate
preprocessing.
In addition, the most effective technology performance is
achieved when the soil particle size is small and the maximum
amount of surface area is exposed. This condition facilitates
adequate contact between the contaminant sorption sites and the
driving force of the technology (i.e., microorganism, solvent,
warm air, etc.). The key to achieving this contact, and
subsequent contaminant destruction, transfer to another medium,
or bonding, is often achieved only through significant
preprocessing.
Materials handling and preprocessing technologies with
potential application to contaminate soil are currently in use in
industries such as construction, agriculture, and mining. All of
these industries routinely handle large quantities of soil or
rock. The use of technologies from these industries should be
considered during all soil remediation activities. Materials
handling and preprocessing techniques should also be incorporated
in treatability testing programs. The results of such tests will
better define the range of waste characteristics which the full-
scale technology will have to treat.
EPA is obtaining results on preprocessing effects from
mixing studies performed on various uncontaminated soils. The
tests are designed to quantify the mixing of soil and test the
effects of soil homogeneity on treatment performance. A
selection of soil types and physical conditions will be combined
591
-------
to provide a matrix of samples commonly encountered during
treatment. Mixing experiments will be conducted on
uncontaminated soil at three scales (bench, pilot and full) to
establish trends in the degree of mixing as a function of soil
type, physical condition, and scale. Similarly, treatment tests
will be performed on contaminated soil at the bench and pilot
scale on a select set of samples from this matrix. Data
generated from the treatment tests could be used to establish a
correlation between treatment effectiveness,and degree of mixing.
CONCLUSIONS
EPA has launched a comprehensive and aggressive effort to
develop LDRs based upon best demonstrated available technologies
for treating soil and debris. The technical issues that need to
be considered in the development of LDRs for soil and debris have
been identified and are being investigated in testing programs
and by analyses of existing data.
Timely and complete technology transfer is an important key
in collecting data and developing land disposal restrictions for
contaminated soil and debris. Therefore, EPA will continue to
seek and evaluate all treatment results, and use the results for
evaluation for regulatory development and technology transfer.
In this vein, the data and conclusions presented in this paper
represent the most current information available in the Superfund
program. EPA recognizes that with each additional treatment test
performed more valuable information will be generated regardless
of whether the test was successful or not.
It is important that the research, remediation, and vendor
experts have an opportunity to participate in the development of
the Land Disposal Restrictions for contaminated soil and debris.
Two options exist for this participation. First, EPA requests
that all available information on the treatment of contaminated
soil, sludges, and debris be forwarded to EPA OERR or to COM
Federal Programs Corporation. Second, EPA plans to publish a
Notice of Data Availability in the Federal Register in the spring
of 1991. This notice will formally notify the public of EPA's
regulatory development approach and request the submission of
comments and additional data.
The data, experience, and opinions of members of the
hazardous waste treatment community, will be valuable additions
to this crucial regulatory development effort. Participation in
this process is strongly encouraged and will be greatly
appreciated. Please send all available information and any
592
-------
comments or suggestions to EPA OERR or to COM Federal Programs
Corporation at the following addresses:
Carolyn K. Offutt/Richard Troast
Hazardous Site Control Division (OS-22Q)
U.S. Environmental Protection Agency
401 M. Street, S.W.
Washington, D.C, 20460
(703)308-8330/308-8323
Joan O'Neill Knapp
COM Federal Programs Corporation
13135 Lee Jackson Memorial Highway
Suite 200
Fairfax, VA 22033
(703)968-0900
REFERENCES
1. U.S. Environmental Protection Agency. Summary
of Treatment Technology Effectiveness for
Contaminated Soil. EPA/540/2-89/053,
Washington, D.C., 1990.
2. U.S. Environmental Protection Agency. Superfund
LDR Guide #6A, "Obtaining a Soil and Debris
Treatability Variance for Remedial Actions."
OSWER Directive 9347.3-06FS, Washington, D.C.,
1989, Revised 1990.
3. U.S. Environmental Protection Agency. Superfund
LDR Guide #6B, "Obtaining a Soil and Debris
Treatability Variance for Removal Actions,"
OSWER Directive 9347.3-07FS, Washington, D.C.,
1989, Revised 1990.
4. U.S. Environmental Protection Agency. November
30, 1989, Memorandum on "Analysis of
Treatability Data for Soil and Debris:
Evaluation of Land Ban Impact on Use of
Superfund Treatment Technologies." OSWER
Directive 9380.3-04, in response to Superfund
Management Review: Recommendation 34A,
Washington, D.C., 1989.
5. U.S. Environmental Protection Agency, Office of
Solid Waste. Quality Assurance Project Plan for
Characterization Sampling and Treatment Tests
for the Contaminated Soil and Debris (CSD)
Program. Washington, D.C., 1990.
593
-------
FATE AND TREATABILITY OF DYES IN HASTEWATER
Thomas J. Holdsworth, Glenn M, Shaul,
Clyde R. Dempsey, and Kenneth A. Dostal
Risk Reduction Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
L. Don Betowski
Environmental Monitoring and Support Laboratory
Las Vegas, Nevada 89114
David A. Gardner
Radian Corporation
Milwaukee, Wisconsin 53E14
ABSTRACT
The first objective of this study was to determine the partitioning of
water soluble azo dyes in the activated sludge process (ASP). Specific azo
dyes were spiked at 1 and 5 mg/L to pilot-scale treatment systems with both
liquid and sludge samples collected. Samples were analyzed by high
performance liquid chromatography (HPLC) with an ultraviolet-visible
detector. Of the 18 dyes studied, 11 compounds were found to pass through
the ASP substantially untreated, 4 were significantly adsorbed onto the
mixed liquor solids (ML), and 3 were apparently biodegraded.
Upon completion of the above study an additional study was begun to
determine the fate of C.I. Disperse Blue 79, one of the largest production-
volume dyes, and select biodegradation products in a conventionally operated
activated sludge process and in an anaerobic sludge digestion system. To
achieve this objective, a pilot study was conducted with two continuous-feed
pilot-scale wastewater treatment systems, one control and one experimental.
The experimental treatment system was fed screened, raw municipal wastewater
594
-------
dosed with a target concentration of 5 mg/L of active ingredient in the
commercial formulation of C.I. Disperse Blue 79 and analyzed for the dye and
related compounds. The control system was fed only the screened, raw
municipal wastewater. A bench-scale activated sludge system was also
operated to assess the fate of dye degradation products arising from the
anaerobic digestion of sludges produced in the experimental aerobic
treatment system. This system was operated to simulate the recycle of
digester supernatant to the head-end of a typical wastewater treatment
system.
This paper has been reviewed in accordance with the U.S. Environmental
Protection Agency's peer and administrative review policies and approved for
presentation and publication.
INTRODUCTION
The U.S. Environmental Protection Agency's (EPA) Office of Toxic
Substances (OTS) evaluates Premanufacture Notification (PMN) submissions
under Section 5 of the Toxic Substances Control Act (TSCA). Azo dyes
constitute a significant portion of these submissions. Generally, azo dyes
contain between one and three azo linkages (-N=N-), linking phenyl and
naphthyl radicals that are usually substituted with some combination of
functional groups including: ami no (-NH2); chloro (-C1); hydroxyl (-OH);
methyl (-CH3); nitro (-N02); and sulfonic acid, sodium salt (-S03Na). OTS
is concerned because some of the dyes, dye precursors and/or their
degradation products such as aromatic amines, which are also dye precursors,
have been shown to be or are suspected to be, carcinogenic (1).
One aspect of the PMN review process is to estimate the release of a
new chemical. The industrial manufacturing and processing of azo dyes
generates a wastewater contaminated with azo dyes, which is typically
treated in a conventional wastewater treatment system. The effectiveness of
this treatment must be known in order to estimate the release from this
source. Therefore, EPA's Office of Research and Development undertook a
study to determine the fate of specific water soluble azo dye compounds in
the activated sludge process (ASP).
The study was approached by dosing the feed to the pilot ASP systems
with various water soluble azo dyes and by monitoring each dye compound
through the system, analyzing both liquid and sludge samples. The fate of
the parent dye compound was assessed via mass balance calculations. These
data could determine if the compound was removed by adsorption, apparent
biodegradation, or not removed at all. The results for 18 dye' compounds
tested are presented.
Upon completion of this research a follow-up study was implemented at
the request of the TSCA Interagency Testing Committee (ITC). The ITC is
comprised eight member agencies which determine areas or chemicals which
require investigation under the TSCA. The member agencies are:
595
-------
Council on Environmental Quality;
Department of Commerce;
Environmental Protection Agency;
National Cancer Institute;
National Institute of Environmental Health Sciences;
National Institute for Occupational Safety and Health;
National Sciences Foundation; and
Occupational Safety and Health Administration.
The ITC requested that C.I. Disperse Blue 79 be investigated because of "The
lack of measured values on physical and chemical properties for C.I.
Disperse Blue 79 increases the uncertainty with respect to chemical fate
predictions . . . Disperse Blue 79 released to the environment is likely to
partition to both water and sediments. In sediments, it may degrade
anaerobically and release 2-bromo-4,6-dinitroanaline. No data have been
found to substantiate or refute these predictions. Since the dye has
widespread large use in the United States and is likely to be released to
the environment during both manufacture and use, it is recommended that
biodegradation studies be conducted to determine (1) the potential for
aerobic and anaerobic biodegradation, and (2) the identity of relatively
persistent intermediates, if any, resulting from biodegradation" (2).
The purpose of this study was to determine the fate of C.I. Disperse
Blue 79 and select biodegradation products in a conventionally operated ASP
and in an anaerobic sludge digestion system. Two continuous-feed, pilot-
scale wastewater treatment systems (one control and one experimental) were
operated at the Milwaukee Metropolitan Sewerage District South Shore
Wastewater Treatment Plant. In addition to these pilot-scale systems, a
bench scale activated sludge system was operated to assess the fate of dye
degradation products from a digester in an anaerobic treatment system. This
system was operated to simulate the recycle of digester supernatant to the
head-end of a typical wastewater treatment system.
FATE STUDY FOR VARIOUS AZO DYES
EXPERIMENTAL PROGRAM
Screened raw wastewater from the Greater Cincinnati Mill Creek Sewage
Treatment Plant was used as the influent (INF) to three pilot-scale
activated sludge biological treatment systems (two experimental and one
control) operated in parallel. A diagram of an aerobic system is presented
in Figure 1. The system used for this study did not include the thickening
or digester stages that are pictured as part of the anaerobic treatment
system used to study Disperse Blue 79. Each system consisted of a primary
clarifier (33 L), complete-mix aeration basin (200 L), and a secondary
clarifier (32 L). Each water soluble dye was dosed as commercial product to
the screened raw wastewater for the two experimental systems operated in
parallel at targeted active ingredient doses of 1 and 5 mg/L of influent
flow (low and high spike systems, respectively). The principal focus of
this work was on the ASP, and, as such, the primary sludge was not sampled.
596
-------
FIGURE 1. AEROBIC AND ANAEROBIC TREATMENT SYSTEM
DYE
AERATION BASIN
LIQUIDS
TO DRAIN
DRAIN
TO DRAIN
597
-------
Table 1 presents a summary of the average operating conditions of the pilot
plant systems,
TABLE 1. SUMMARY OF OPERATING CONDITIONS
Parameter Value
Influent flow rate, L/d 720
Primary sludge flow rate, L/d 6
Primary effluent flow rate, L/d 714
Mixed liquor wastage flow rate, L/d 67
Secondary effluent flow rate, L/d 647
Solids retention time, days 2.7
Hydraulic retention time, days 0.28
Dissolved oxygen, mg/L 2.0-4.0
Target influent spike dosages, mg/L
Low 1
High 5
Influent pH, pH units 7.0-8.0
Aeration basin temperature, °C 21-25
Before each data collection phase, dye analytical recovery studies were
conducted by dosing the purified dye compound into organic-free water,
influent wastewater, and mixed liquor. These studies were run in duplicate
and each recovery study was repeated at least once to ensure that the dye
compound could be extracted. Purified dye standards were analytically
prepared from the commercial dye product by repeated recrystallization.
The analytical technique used to recover the dye from each source is
presented in reference 3.
All systems were operated for at least three times the solids retention
time to ensure acclimation prior to initiation of data collection. All
samples were 24 hr composites made up of 6 grab samples collected every 4 hr
and stored at 4"C. The 18 water soluble, acid and direct azo dyes studied
are listed below in Table 2 by Colour Index (4) name and number.
RESULTS AND DISCUSSIONS
Before a compound was judged acceptable for testing, spike recovery
studies were performed for each dye. The recoveries for all 18 dyes were
generally very good and with low standard deviations. Recovery for most
dyes was within 80% and 120%; thus, it appeared little or no chemical
transformation was occurring. The spike recovery method and results are
available in reference 3.
598
-------
TABLE 2. DYE COMPOUNDS SPIKED TO THE ACTIVATED SLUDGE PROCESS
Colour Index Name
Colour Index Number
C.I.
C.I.
C.I.
C.I.
C.I.
C.I.
C.I.
C.I.
C.I.
C.I.
r T
t . i .
C.I.
C.I.
C.I.
C.I.
C.I.
C.I.
C.I.
Acid Black 1
Acid Blue 113
Acid Orange 7
Acid Orange 8
Acid Orange 10
Acid Red 1
Acid Red 14
Acid Red 18
Acid Red 88
Acid Red 151
flriH RoH 717
HLIU r\cU Jo/
Acid Yellow 17
Acid Yellow 23
Acid Yellow 49
Acid Yellow 151
Direct Violet 9
Direct Yellow 4
Direct Yellow 28
20470
26360
15510
15575
16230
18050
14720
16255
15620
26900
*
18965
19140
18640
13906
27885
24890
19555
*Not assigned as of 01/91. Chemical Abstracts Number = 67786-14-5.
In addition to the spike recovery study, no photodegradation of the
dyes was found in laboratory studies. Moreover, the estimated .Henry's law
constant for each dye tested was less than 10"15 atm-m3/mol, and, as such,
air stripping was very unlikely (5). Therefore, adsorption and/or
biodegradation appeared to be the only removal mechanisms.
Eleven of the 18 azo dyes studied passed through the ASP substantially
untreated with the data from the low and high spike systems in excellent
agreement for these dyes. These were:
C.I. Acid Black 1
C.I. Acid Orange 10
C.I. Acid Red 1
C.I. Acid Red 14
C.I. Acid Red 18
C.I. Acid Red 337
C.I. Acid Yellow 17
C.I. Acid Yellow 23
C.I. Acid Yellow 49
C.I. Acid Yellow 151
C.I. Direct Yellow 4
The relatively high sulfonic acid substitution of these dyes may
explain why. they were not removed. If the azo dye has high sulfonic acid
substitution, then little or no adsorption of the dye by the microbial cell
or cell byproducts would occur, thus limiting the chance of aerobic
biodegradation (6). Ten of the 11 above dyes have at least two sulfonic
acid functional groups, C.I. Acid Red 337 has one.
599
-------
The positioning of the sulfonic acid functional group(s) and the
molecular weight of the compound also appeared to have an affect on how the
compound partitions. Four compounds were adsorbed onto the WAS and
apparently not biodegraded. These were:
C.I. Acid Blue 113
C.I. Acid Red 151
C.I. Direct Violet 9
C.I. Direct Yellow 28
C.I. Acid Blue 113, C.I. Acid Red 151, and C.I. Direct Violet 9
represent three of the four diazo (two azo bonds) structures. Although
these dyes are sulfonated compounds with two of the three having two
sulfonic acid functional groups, they also have a greater molecular weight
than the other compounds. Further investigations into the effect of
sulfonation (both in number of groups and position) versus molecular weight
are necessary before a relationship, if any exists, could be developed.
Three compounds appeared to be biodegraded. These were:
C.I. Acid Orange 7
C.I. Acid Orange 8
C.I. Acid Red 88
The conclusion that these compounds were apparently biodegraded comes
from an inspection of the mass balance data (3); for each compound, very
little of the dye was recovered during sampling. However, the preliminary
recovery studies showed that the compound could be recovered without
difficulty from wastewater and sludge matrices. Since the compounds were
not found in the activated sludge effluent (ASE) or mixed liquor solids
(ML), samples and chemical transformation appeared not to be occurring, then
biodegradation would account for the apparent loss of the parent compound.
The partitioning of the dye between the influent (INF), primary effluent
(PE), ASE, and ML along with the mass balance data summary for each dye are
available in reference 3.
In addition to the 18 dyes thus far discussed, 11 other azo dyes were
Investigated during this study but the analytical recovery methodology did
not produce satisfactory recoveries from the various matrices for these
dyes. Table 3 identifies these dyes.
600
-------
TABLE 3. DYE COMPOUNDS NOT TESTED DUE TO POOR RECOVERY
Colour Index Name Colour Index Number
C.I.
C.I.
C.I.
C.I.
C.I.
C.I.
C.I.
C.I.
C.I.
C.I.
C.I.
Acid Blue 92
Acid Blue 158
Acid Brown 14
Acid Red 114
Direct Black 80
Direct Blue 15
Direct Blue 78
Direct Blue 80
Direct Red 24
Direct Red 80
Direct Red 81
13390
14880
14720
23635
31600
24400
34200
24315
29185
35780
28160
FATE OF DISPERSE BLUE 79 IN AEROBIC AND ANAEROBIC TREATMENT
EXPERIMENTAL PROGRAM
Two pilot-scale treatment systems were operated for the entire study,
while only during the later phase was the bench-scale system operated.
Because a reliable method for dye analysis was needed to determine the
fate of C.I. Disperse Blue 79 in the treatment system, an analytical
procedure was developed. Various extraction methods and solvents were
investigated to develop a suitable extraction procedure to prepare samples
for high-performance liquid chromatography (HPLC) analysis.
A diagram of the aerobic and anaerobic treatment system used for this
study is presented in Figure 1. Both pilot-scale activated sludge systems
included a contact tank, a conical-shaped primary clarifier, an aeration
basin, and a conical-shaped secondary clarifier. The contact tanks were
installed to ensure the dye was mixed with the feed and to obtain ai 30-min
contact time between the raw wastewater and the dye. The primary and
secondary clarifiers were approximately 49 L, and the aeration tanks were
approximately 185 L.
The activated sludge basins were separated into three cells to operate
as a plug-flow system. Peristaltic pumps supplied the screened, raw
wastewater to the contact tanks. Gravity moved the wastewater from the
contact tanks to the primary clarifiers, then to the aeration basins, and on
to the secondary clarifiers. Activated sludge was wasted from the aeration
basins via peristaltic pumps. Primary sludge was wasted manually once each
day. The target hydraulic retention time (HRT) in both activated sludge
units was 5.5 hr and the solids retention time (SRT), 7 days.
601
-------
The anaerobic digesters were cylindrical-shaped vessels constructed of
clear PVC. Each digester had a total volume of 70 L with an operating
volume of 39 L. The digesters were completely mixed and heated to maintain
an operating temperature of 35*C. Gas production from the digesters was
monitored with gas meters.
Waste activated sludge and primary sludge from each activated sludge
unit were mixed, thickened, and fed to the respective anaerobic digesters.
The target SRT of the anaerobic digesters was 15 days and the target loading
was 1.2 kg total volatile solids (TVS)/nf»day.
The experimental treatment systems received screened, raw wastewater
dosed with a target concentration of 5 mg/L of the active ingredient in C.I.
Disperse Blue 79. The control system received only the screened, raw
wastewater. After acclimation and steady state conditions were reached, the
following samples from each system were analyzed for the dye and related
compounds: INF, PE, ASE, primary sludge, ML, digester feed, digester
supernatant, and digester effluent.
The bench-scale system was an activated sludge unit that was fed a
mixture of primary effluent from the experimental system, supernatant from
digester feed preparation (primary and waste activated sludge thickening),
and centrate from centrifuging digested sludge from the anaerobic digester.
The mixture was prepared to simulate the recycle of digester supernatant and
primary and thickened waste activated sludge supernatant to the head-end of
a treatment plant. The activated sludge unit consisted of a 6-L conical
reactor, which served as the aeration basin; a 2-L inner cone for solids
recycle; and a 125-mL clarifier tube for effluent clarification.
Peristaltic pumps were used to deliver the feed and remove waste activated
sludge from the unit.
RESULTS AND DISCUSSION
The recovery of C.I. Disperse Blue 79 was completed by an extraction
technique followed by HPLC and spectrophotometric analysis. The recovery of
the dye was acceptable and the method and results are presented in reference
7.
Operating and analytical data for the pilot-scale activated sludge
units are summarized in Table 4. The data for both Units 1 and 2 were
similar. The average final effluent TCOD value for Unit 2 was 73.5 mg/L and
that for Unit 1 was 59.2 mg/L. Although the slightly higher effluent TCOD
value for Unit 2 may have been caused by adding dye to the unit, the data
indicate that the overall performance of the experimental activated sludge
system was not affected by this addition.
The anaerobic digester's operating and analytical data are summarized
1n Table 5. The feed, effluent, and operational data indicate no
significant difference between the two units. No adverse affect was
detected on the operation of the experimental digester as a result of adding
dye.
602
-------
TABLE 4. SUMMARY OF THE PILOT-SCALE ACTIVATED SLUDGE SYSTEM'S OPERATIONAL
AND ANALYTICAL DATA FOR THE DYE TESTING PERIOD
Parameter
Feed data
TSS (mg/L)
TCOD (mg/L)
TBOD (mg/L)
NH3-N (mg/L)
Operational data
HRT (hr)
SRT (days)
Unit 1*
238
364
182
22,5
5.28
5.94
Unit 21"
211
375
177
20.7
5.28
5.87
Mixed liquor data
Temperature (°
pH (range)
DO, Cell 1 mg
DO, Cell 2 mg
DO, Cell 3 mg
TSS (mg/L)
05 Uptake rate
SSVI (ml/g)
Primary effluent
TSS img/Lj
NH3-N (mg/L)
C) 20.0
6.8-8.0
/L) 2.4
/L 3.6
A) 3.6
3,030
(mg/L hr) 6.8-8.0
/ «5 * L
data
134
22.7
20.0
6.8-7.6
2.8
3.5
3.9
3,060
6.8-7.6
58.0
139
21.7
Final effluent data
TCOD (mg/L)
TBOD mg/L)
TSS (mg/L)
NH3-M (mg/L)
*Control
fSpiked
59.2
16
27
0.26
73.5
21
31
0.18
603
-------
TABLE 5. SUMMARY OF THE ANAEROBIC DIGESTER'S OPERATIONAL AND ANALYTICAL
DATA FOR THE DYE TESTING PERIOD
Parameter Unit 1* Unit 2f
Feed data
TSS (mg/L) 21,300 21,400
TS (%) 2.42 2.42
TVS (%) 1.81 1.82
Effluent data
pH (range) 6.6-7.0 6.6-7.0
Temperature (eC) 35.0 35.0
TSS (mg/L) 12,700 12,200
TS (%) 1.46 1.48
TVS (%) 0.94 0.97
Operational data
Alkalinity (mg/L) 2,930 2,820
Volatile acids (mg/L) < 51 < 50
Loading (kg TVS/m* day) 1.21 1.22
TVS reduction (%) 47.8 46.4
Gas production 0.76 0.87
(mykg TVS destroyed)
Percent CH4 in gas 58.9 57.9
*Control
fSpiked
The influent and effluent streams from the experimental activated
sludge systems were sampled and analyzed for C.I. Disperse Blue 79 and any
related compounds (see reference 7) to determine the fate of the dye in the
treatment system. The average dye and TSS concentrations from the Unit 2
samples are summarized in Table 6. Influent and waste mixed liquor samples
were analyzed from Unit 1.
The average dye concentration in the Unit 2 feed to the primary
clarifier was 4.40 mg/L and the average final effluent concentration was
< 0.93 mg/L, so that the average dye removal was greater than 79%. Although
5 of 19 analyzed effluent samples were below the 0.25 mg/L detection limit,
the effluent dye concentration varied from < 0.25 mg/L to 3.70 mg/L. The
variation in effluent dye concentration may have been caused by the
variation in effluent TSS (9 - 72 mg/L) concentration.
604
-------
TABLE 6. AVERAGE C.I. DISPERSE BLUE 79 AND TSS CONCENTRATION IN THE
UNIT 2 EXPERIMENTAL ACTIVATED SLUDGE UNIT SAMPLES
Sample C.I. Disperse Blue 79 TSS
Location (mg/L)
Feed 4.40 212
Primary effluent 4.71 133
Primary sludge 31.8 14,500
Waste activated sludge 93.5 3,060
Final effluent < 0.93 28
In developing an analytical procedure (7) for C.I. Disperse Blue 79, it
was determined that the dye has a high affinity for the activated sludge
solids present in the experimental unit. In fact almost all the dye present
in a mixed liquor sample was extracted off the solids (7). The correlation
coefficient between TSS and dye concentrations in the Unit 2 effluent was
determined to be 0.78. Calculations performed on Table 6 data show that
each gram of suspended solids in the waste activated sludge (WAS) contained
30 rug of dye. Since the average final effluent TSS concentration was 28
mg/L, the final effluent dye concentration can be calculated to be 0.84 mg
C.I. Disperse Blue 79 per liter (assuming TSS is all WAS). Comparing 0.84
mg dye/L to the measured value of < 0.93 mg dye/L presented in Table 6
indicates that the calculated value is very close to the measured value and
supports that the dye has a high affinity for the activated sludge.
Lowering the final effluent TSS concentration by improving solids removal in
the final clarifier will result in a lower dye concentration in the final
effluent.
Mass balance calculations were performed with the use of the measured
dye concentrations and measured flow rates for each process stream. Mass
balance calculations across the entire activated sludge system showed that
an average of 86.5% of the dye contained in the feed stream was accounted
for in the effluent streams. The primary sludge contained an average of
3.6% of the dye fed to the system; waste activated sludge, 62.3%; and final
effluents, 20.4% (the percentages of the three streams do not equal 86.5%
because of rounding off the individual values). Since most of the dye fed
to the system was recovered and no other related compounds were detected, it
can be concluded that no significant biodegradation of C.I. Disperse Blue 79
occurred in the activated sludge system.
Feed sludge and effluent (digested sludge) samples from both the
control and experimental digesters were analyzed for dye content.
Detectable concentrations of dye were identified by HPLC-UV in 5 of 10
control-unit feed samples and in 4 of 10 effluent samples. The average
concentrations were low, however, at < 1.45 mg/L for the feed samples and
605
-------
< 1.22 mg/L for the effluent sample. The low level of dye in these control
unit samples is negligible when compared with the much higher concentrations
of dye in the experimental unit samples.
The average experimental unit feed dye concentration was 443 mg/L, and
the average effluent concentration was 7.86 mg/L. On average, 98.2% of the
dye contained in the feed sludge was degraded in the anaerobic digester.
Thermospray ionization mass spectrometry was used to identify
degradation products of C.I. Disperse Blue 79 in the anaerobic digester
effluent. With this ionization technique, the parent dye was observable,
but because of the electronegativity of many of the functional groups on the
molecule (e.g., NO^, Br), the sensitivity of the technique for this compound
was poor. Four major degradation compounds were tentatively identified and
found in significant amounts in the digester effluent. Their exact identity
and amounts have not been verified because appropriate analytical standards
were not available. Information about the degradation products is contained
in reference 7.
During normal operation of a wastewater treatment system, the
supernatant from sludge lagoons or other digester sludge thickening
operations is returned to the head-end of the plant for treatment. The
bench-scale activated sludge system (Unit 3) was operated to study the fate
of dye degradation products from the anaerobic digester in an activated
sludge system. The feed for Unit 3 was the supernatant from the sludge
thickening operation mixed with centrate from centrifuging digester effluent
and primary effluent. The supernatant was added to simulate the effluents
produced from thickening waste activated sludge in a typical plant.
The operating and analytical data from Unit 3 are presented in Table 7.
The average HRT was 6.04 days, which was slightly higher than the Unit 2
value of 5.28 days; the average SRT for Unit 3 was 4.83 days, which was
lower than the Unit 2 average of 5.87 days. The Unit 3 average SRT was
lower than the target value of 7 days because of a relatively high average
effluent TSS value of 40 mg/L. The bench-scale unit settling performance
was more variable than that in the pilot units.
The average effluent TCOD and TBOD values were also higher than the
pilot unit values. The higher effluent values probably resulted from the
higher TSS levels in the final effluent. The performance of Unit 3 with
respect to TSS, TBOD, and TCOD removal was not as good as that of the pilot
units but was typical of a bench-scale unit.
Table 8 summarizes the dye data from the bench-scale unit feed, waste
activated sludge, and final effluent sample analyses. The average feed dye
concentration was 3.43 mg/L and the average effluent concentration was 1.32
mg/L, for a removal efficiency of 62%. The effluent concentration was
probably high because of the relatively high TSS concentration in the final
effluent.
606
-------
TABLE 7. SUMMARY OF ACTIVATED SLUDGE UNIT 3'S OPERATIONAL AND
ANALYTICAL DATA
Parameter Average Value
Feed
TSS (mg/L) 130
TCOD (mg/L) 336
TBOD (mg/L) . 158
Operation data
HRT (hr) 6.04
SRT (days) 4.83
Mixed liquor data
Temperature (eC) 21.5
pH (range) 6.8-8.1
DO (mg/L) 5.6
TSS (mg/L) 1,650
Finaleffluentdata
TSS (mg/L) 40
TCOD (mg/L) 116
TBOD (mg/L) 31
TABLE 8. BENCH-SCALE ACTIVATED SLUDGE SYSTEM'S C.I. DISPERSE BLUE 79
ANALYTICAL RESULTS
Sample
Location
Feed
Waste activated sludge
Final effluent
C.I. Disperse Blue 79
(mg/L)
3.43
37.6
1.32
TSS
(mg/L)
14.5
2,150
53
Mass balance calculations of the dye across Unit 3 showed that an
average of 75.3% of the dye fed to the unit was accounted for in the
effluents from the unit. The mass balance for Unit 2 showed 86.5% of the
dye was recovered. Although the recovery from Unit 3 was slightly lower, it
does not appear that significant degradation of the dye occurred in the
bench-scale activated sludge system.
607
-------
Degradation products of C.I. Disperse Blue 79 were also monitored in
the influent, effluent, and waste sludge from Unit 3. Because no positive
identification was made of the by-products, quantification was not possible.
Some general observations can, however, be made concerning the degradation
products based on relative amounts. The observed trend indicated that the
concentration of these compounds decreased across Unit 3. The final
effluent samples always contained the lowest concentrations of the
degradation products, but because of limited data, further conclusions
cannot be drawn. Further evaluation of the degradation products and their
fate in biological treatment systems may be subject for further project
work.
CONCLUSIONS
1. A total of 18 water soluble azo dyes were successfully monitored in
wastewater and sludge samples collected from pilot-scale ASP systems.
The study of 11 additional dyes was attempted but could not be
accomplished due to poor analytical recovery.
2. Increased sulfonation seemed to be a major factor in preventing an azo
dye compound from being either apparently adsorbed or biodegraded by the
ASP.
3. Of the 18 dyes studied, 11 compounds were found to pass through the ASP
substantially untreated, 4 were adsorbed onto the WAS and 3 were
apparently biodegraded.
4. The addition of C.I. Disperse Blue 79 did not adversely affect the
operation of the pilot activated sludge unit or that of the anaerobic
digester. Both the control and experimental activated sludge units
produced effluents typical of municipal wastewater treatment systems.
The anaerobic digesters achieved volatile solids reductions within the
normal operating range for municipal digesters.
5. Little evidence of C.I. Disperse Blue 79 degradation in the activated
sludge systems was found. Mass balance calculations showed that, on
average, 86.5% of the dye contained in the feed to the system was
present in the effluent streams.
6, The majority of the C.I. Disperse Blue 79 fed to the activated sludge
system was removed in the waste activated sludge. The average dye mass
balance obtained around the system was 86.5%; the dye was partitioned in
the effluent streams as follows: 3.6% in the primary sludge, 62.3% in
the waste activated sludge, and 20.4% in the final effluent.
7. The C.I. Disperse Blue 79 was degraded in the anaerobic digester. The
dye concentration was reduced from an average feed value of 566 mg/L to
an average effluent value of 15.0 mg/L, or a 97.4% reduction.
608
-------
8. Possible degradation products of the dye were detected in the digester
effluent. Although some preliminary measurements were made to identify
the structure of these compounds, no positive identification or
quantification of the compounds was made.
9. Based on limited semi-quantitative results, some of the dye degradation
products from the anaerobic digester were destroyed when treated in a
bench-scale activated sludge system.
REFERENCES
1. Helmes, C.T., et al. A study of azo and nitro dyes for the selection of
candidates for carcinogen bioassay. J. Environ. Sci. Health A19(2):
97-231, 1984.
2, Nineteenth Report of the TSCA Interagency Testing Committee to the
Administrator. U.S. Environmental Protection Agency, Interagency
Testing CoTnmittee, Washington DC, 1986.
3. Shaul, G.M., Dempsey, C.R., and Dostal, K.A. Fate of Water Soluble Azo
Dyes in the Activated Sludge Process. EPA-6QO/88/030, U.S.
Environmental Protection Agency, Cincinnati, Ohio, 1988.
4. Colour Index. Revised 3rd Edition, Seven Volumes. The Society of Dyers
and Colorists.
5. Lyman, W., Reehl, W., and Rosenblatt, D. Handbook of Chemical Property
Estimation Methods. McGraw Hill Book Company, New York, 1982.
6. Wuhrmann, K., Mechsner, K., and Kappeler, T. Investigation on
Rate-Determining Factors in the Microbial Reduction of Azo Dyes.
European J. Appl. Microbiol. Biotechnol. 9: 325, 1980.
7. Gardner, D.A., Holdsworth, T.J., Shaul, G.M., Dostal, K.A., Betowski,
L.D. Aerobic and Anaerobic Treatment of C.I. Disperse Blue 79.
EPA-600/89/051, Cincinnati, Ohio, 1990.
609
-------
CARBON-ASSISTED ANAEROBIC TREATMENT OF HAZARDOUS LEACHATES
by: A, T. Schroeder, M. T. Suidan, and R. Nath
Department of Civil and Environmental Engineering
University of Cincinnati
Cincinnati, OH 45221-0071
E. R. Krishnan
PEI Associates, Inc.
Cincinnati, OH 45204
R. C. Brenner
U.S. Environmental Protection Agency
Cincinnati, OH 45268
ABSTRACT
Two anaerobic granular activated carbon (GAG) expanded-bed bioreactors were
tested as pretreatment units for the decontamination of hazardous leachates
containing volatile and semivolatile synthetic organic chemicals (SOCs). The
different characteristics of the two leachate feed streams resulted in one
reactor operating in a sulfate—reducing mode and the 'second in a strictly
methanogenic environment. Both reactors were operated with a 6—hr unexpanded
empty—bed contact time and achieved SOC removal acceptable for pretreatment
units. In both reactors, the majority of the SOCs were removed by biological
activity, with GAG adsorption providing stability to each system by buffering
against load fluctuations.
KEYWORDS: leachate, synthetic organic chemicals, anaerobic, activated carbon,
sulfate reduction, methanogenic, expanded—bed bioreactor.
This paper has been reviewed in accordance with the
U.S. Environmental Protection Agency's peer and
administrative review policies and approved for
presentation and publication.
610
-------
INTRODUCTION
Rainfall and surface runoff infiltrating sanitary and industrial landfills
and hazardous waste dump sites are exposed to a complex variety of pollutants.
Water percolating through a landfill is contaminated with a number of organic
and inorganic compounds, and direct discharge of exiting leachate to municipal
wastewater treatment plants can result in inadequate removal of many hazardous
substances. Mandated by the Comprehensive Environmental Response, Compensation,
and Liability Act (CERCLA), physical, biological, or chemical pretreatment
options may be necessary for removing hazardous substances from leachates,
depending on site-specific contamination.
Many volatile and semivolatile synthetic organic chemicals (SOCs) used
as solvents, degreasers, or components in industrial products are present in
leachates originating from hazardous waste sites. These SOCs are often
inadequately treated in aerobic wastewater treatment processes as volatiles are
subject to air stripping, many semivolatiles simply pass through untreated, and
highly chlorinated compounds are difficult to degrade aerobically. Leachate
pretreatment alternatives are not well defined because leachate chemical
profiles vary significantly from site to site, Leachate quality is affected by
a variety of factors, including the material in the fill, the site age, and
precipitation and climate patterns (1). These factors not only influence the
types and concentrations of hazardous compounds in water percolating through a
site, but also determine the amounts of background biodegradable substances in
a leachate. This will in turn influence the environment within any biological
pretreatment process.
With recent advances in anaerobic biological treatment, many leachates
contaminated with SOCs can be successfully pretreated. Although leachates are
highly complex and variable, many SOCs commonly found in leachates are partially
or completely biodegradable in anaerobic systems (2—11), Also, anaerobic
processes are generally effective in treating high strength wastes. The
anaerobic granular activated carbon (GAC) expanded-bed bioreactor is an
anaerobic pretreatment option that appears to be particularly well suited for
treating SOC-contaminated leachates. The combined processes of anaerobic
degradation and carbon adsorption provide a means for removing a variety of SOCs
during pretreatment. Compounds that resist degradation can be controlled with
carbon adsorption and replacement.
In this study, leachates containing a mixture of 14 SOCs were treated with
anaerobic GAC expanded-bed bioreactors. SOC influent concentrations were
maintained at levels typical of leachates from CERCLA waste sites.
MATERIALS AND METHODS
Two anaerobic GAC expanded-bed bioreactors were used to treat municipal
leachates rendered hazardous by the addition of 14 SOCs. The reactors were
operated in parallel using identical mechanical feed systems and piping
611
-------
networks. The treatment systems were operated at the U.S. EPA Test and
Evaluation Facility in Cincinnati, Ohio. A schematic representative of the two
treatment systems is shown in Figure 1.
EXPANDED-BED REACTORS
Each expanded-bed reactor consisted of a 10,2-cm ID, 96.5-cm long,
Jacketed Plexiglas* tube fitted at the top and bottom with headers to convey
influent and effluent streams. Warm water was circulated through the jacket
surrounding each reactor to maintain each system at 35°C. Each reactor was
charged with 1.0 kg of 16X20 U.S. Mesh F400 GAG (Calgon Carbon Corp.), underlain
by a bed of graded gravel that served to distribute flow evenly along the
reactor cross-section.
Effluent recycle was maintained at a rate sufficient to achieve completely
mixed conditions within each reactor and to sustain a bed expansion of
approximately 30%.
INFLUENT AND EFFLUENT SYSTEMS
Raw municipal leachate was fed to each reactor from a sealed, chilled,
mixed, stainless steel reservoir. A positive pressure head (4-8 in. H20) was
maintained in each reservoir using a nitrogen gas blanket. The blanket aided
leachate feeding by preventing the development of a vacuum in each sealed
reservoir and more importantly prevented oxidation of the leachate. Leachate
was fed using stainless steel lines, except around feed pumps, and was
introduced into each recycle loop on the suction side of each recycle pump.
A stock solution of the SOCs was fed into the suction side of each recycle
loop along with the leachate. By pumping the leachate and the SOC solution
independently, better spiking control was possible than if the SOC stock was
mixed into the leachate feed reservoirs. Also, volatilization and adsorption
losses were prevented. The SOC solution was pumped using a syringe pump and
gas—tight, volatile organic chemical (VOC) approved, glass syringes with teflon-
tipped barrels. Stainless steel tubing was used for all SOC spiking lines.
Effluent from each reactor was collected in a 50-L graduated vessel so
that throughhput flow (10 L/day at design flow) could be measured accurately
without disturbing the system. Off gases were volumetrically measured using
wet—tip gas meters connected to the top of each reactor.
MUNICIPAL LEACHATES
Raw municipal leachates were obtained periodically from two sanitary
landfills operated by the Delaware Solid Waste Authority (DSWA) . Both leachates
contained background contamination (e.g., COD, metals, sulfate and sulfide
sulfur, ammonia nitrogen, and suspended solids) similar to that expected for
many hazardous leachates.
612
-------
Gas Mew
Effluent
Rcdrcularmg
WanrSaifa
Organic Solnrion
and Syringe Pnmp
Reservoir
Recycle Pump
RuiffizcdGAC
I ConpSng
H Valve
Figure 1. Schematic of Anaerobic GAC Expanded-Bed Treatment: System
613
-------
Reactor A was fed a relatively weak, biologically stabilized leachate
(Leachate A) pumped from a collection well at the Cherry Island municipal
landfill near Wilmington, DE. This leachate was characterized by low COD levels
(approximately 1,100 mg/L) and a low volatile fatty acids (VFAs) content (15%
of total COD). Sulfate concentrations averaged 89 mg SQ4/L. Reactor B received
a moderate strength, low sulfate leachate (Leachate B) obtained from the DSWA
Southern Solid Waste Facility near Seaford, DE. Leachate B averaged 3,800 mg
COD/L, with the majority (60-90%) of the COD comprised of VFAs.
Because of the longer age and accompanying stabilized nature of the
landfill that produces Leachate A, COD levels in this leachate were relatively
stable over the course of the project (except for one early shipment). Most of
the easily biodegradable material in the fill has been consumed, leaving behind
a greater fraction of complex refractory compounds such as humic substances,
Leachate B COD levels (produced by a younger, less stable landfill) varied
significantly, with the highest COD levels occurring in the colder, wetter
months. During these months, biodegradation of solubilized organic material is
slowed within the landfill and increased flushing rates remove material before
it can be consumed internally.
SOC SUPPLEMENT
The SOC solution was fed simultaneously with each leachate to obtain
leachate contamination levels represented in Table 1. Selection of the nine
volatile and five semivolatile SOCs and their corresponding target
concentrations was based on a U.S. EPA review of CERCLA leachates (12).
Chloroform, although identified as a one of the compounds often found in
hazardous leachates, was not included in the SOC solution because of its
reported toxic/inhibitory effects on anaerobic processes (13).
The SOC solution was prepared in 70-mL batches, and each batch was
analyzed several times to ensure that the spike matrix did not change
significantly with time. With the particular recipe of compounds used in this
project, batches were prepared without the use of a solvent as the compounds
were mutually miscible.
PROJECT OPERATION
After startup and acclimation (accomplished over 160 days at incremental
feed rates of 2 to 10 L/day) to the raw municipal leachates, the leachate flow
rate to each reactor was maintained at 10 L/day for the duration of the project.
This resulted in an unexpanded empty-bed contact time of 6 hours in each
reactor. SOC spiking was initiated at 30% of target levels on Day 67 and was
increased to approximately 60% of target levels on Day 105. On Day 133, feed
concentrations reached 100% of desired levels. Although the spiking rates were
nearly constant during each period, the presence of some SOCs in the raw
municipal leachates caused fluctuations in actual compound concentrations, as
would be seen in any real hazardous leachate. Throughout the study, the pH in
each reactor remained near neutral ranging between 6.8 and 7.2.
614
-------
fable 1: Typical CERCLA Leachate Profile (1)
Compound . Concentration (ag/L)
VOLATILE ORGANIC COMPOUNDS
Acetone 10,000
Methyl Ethyl Ketone , 5,000
Methyl Isobutyl Ketone 1,000
Trichloroethylene 400
1,1-Diehloroethane 100
Methylene Chloride 1,200
Chlorobenzene 1,100
Ethylbenzene 600
Toluene 8,000
SEMIVOLATILE ORGANIC COMPOUNDS
Phenol 2,600
Nitrobenzene 500
1,2,4-Trichlorobenzene 200
Dibutyl Phthalate 200
Bis(2-Ethylhexyl)Phthalate 100
Beginning on Day 65, GAG in Reactor A was replaced with prewashed virgin
GAG at a rate of 10 g/day (1%/day). This precautionary measure was later deemed
unnecessary, and carbon replacement was halted on Day 214, . Carbon replacement
was not used in Reactor B.
Liquid flow and volumetric off—gas production rates were measured daily.
Samples were analyzed weekly for SOCs, total and soluble COD,.sulfate, off-gas
composition (CH4, C02), metals, ammonia nitrogen, and nitrate nitrogen, VFAs,
total and volatile suspended solids (TSS and VSS), and .SOCs in the off gas were
measured weekly during selected sampling periods.
ANALYTICAL METHODS
Isotope dilution analyses of VOCs were carried out according to EPA Method
1624B (14). A Model 5890 Hewlett Packard (HP) gas chromatograph (GC) and a
Model 5970 HP mass selective detector were used in conjunction with a Tekmar
LSC-2 purge and trap. 5-mL samples containing isotopically-labelled analogs of
each of the VOCs were purged (helium, 40 mL/min) for 12 minutes to a Supelco 2—
0293 trap. Compounds were desorbed (4.0 rain, 210°C) .to.a packed column (2 m
long, 2 mm ID; 1% SP-100Q, Carbopak B). The GC column temperature ramped from
45°C to 210°C at 8°C/min. The detector scan range was 35-250 amu with injection
and interface temperatures of 210 and 275°C-, respectively.
Isotope dilution analyses of the semivolatile compounds were performed
according to EPA Method 1625B (14) after continuous liquid-liquid.extraction
(Method 3520). Isotopically-labelled internal standards were injected into each
sample prior to extraction with methylene chloride. The GC (HP Model 5890) was
615
-------
fitted with a SPB—5 capillary column (30 m long, 0.25 mm ID), The temperature
was ramped from 50°C to 280°C at a rate of 8°C/min. The detector (HP Model 5970)
scan range was 35-500 amu, with an injection temperature of 270°C and an
interface temperature of 280°C,
COD samples were analyzed using COD reagent vials, a block digester, and
a spectrophotometer. Prior to analysis, COD samples were acidified and purged
with nitrogen to minimize sulfide interference. Sulfate was analyzed by ion
chromatography (15), and gas composition was measured using a Fisher 1200 gas
partitioner. All other analyses were conducted according to U.S. EPA methods.
All liquid SOC analyses were performed at the U.S. EPA Andrew W.
Breidenbach Environmental Research Center (AWBERC). All samples had to be
collected, transported to AWBERC, and stored for a short time until analysis,
Acidified semivolatile samples were extracted upon arrival at AWBERC. However,
volatile samples could not always be analyzed immediately and required a
preservation procedure that would allow for some lag time between sample
collection and analysis. A VOC sample preservation study, carried out with the
analytical support staff of the Risk Reduction Engineering Laboratory,
demonstrated that acidified, chilled, VOC samples could be stored for at least
2 weeks without significant losses. In this study, effluent samples (raw
leachate feed only) from each reactor were spiked with acetone, methylene
chloride, 1,1-dichloroethane (DCA) , and toluene at concentrations of 400, 120,
120, and 120 MS/L, respectively. They were acidified and chilled, then analyzed
over a period of 2 weeks and compared to a SOC solution made in methanol on the
first day. Table 2 shows that the amount of each compound recovered did not
vary significantly with time.
Table 2: VOC Preservation Study Results
% Recovery of Analyte Compared to Time Zero
VOC-Methanol Solution
Dav Methvlene Chloride Acetone DCA Toluene
3
6
8
10
14
94
109
117
102
101
115
138
116
124
128
99
104
102
95
93
101
104
102
110
116
RESULTS AND DISCUSSION
The primary objective in pretreating CERCLA leachates contaminated with
SOCs is to remove the hazardous organic compounds before leachates are blended
with municipal or industrial wastewater targeted for treatment in aerobic
wastewater treatment facilities. However, other contaminants in the leachate
need to be at sufficiently low levels so that blended waste streams do not
616
-------
significantly inhibit the efficiency of aerobic systems. Also, these secondary
contaminants, varying with the site-specific nature of each fill, need to be
examined as they will strongly influence the pretreatment environment.
REACTOR A'ENVIRONMENT
The biological activity in Reactor A, treating the weaker strength
leachate, was substantially limited by the amount of available substrate
entering the reactor. The degradable fraction of the influent COD was
consistently low with a VFA content averaging less than .15% of the total COD.
With the significant presence of sulfate in this leachate, a competition between
methanogens and sulfate-redueing organisms developed. Although biological
activity in this reactor was limited, a stable environment persisted throughout
the course of the project.
During the entire period of operation at 100% SOC spike levels, Reactor
A averaged only 26% COD removal. Until Day 214, carbon replacement accounted
for approximately 45% of the COD removal as highly refractory compounds were
removed by adsorption. However, after GAC replacement was halted, methane
production and sulfate reduction continued to remove the same amount of COD, by
mass, as during the period when GAC replacement was practiced. Overall COD
removal dropped to an average of 14% with no GAC replacement, but only because
the physical removal mechanism of GAC replacement was no longer being used.
Although several metals were present in significant amounts, metal toxicity was
not observed. Likewise, the SOCs did not appear to inhibit biological activity.
Influent and effluent COD concentrations are plotted in Figure 2, and influent
and effluent background contaminant concentrations are given in Table 3.
REACTOR B ENVIRONMENT
Reactor B, treating the moderate strength leachate, operated principally
under methanogenic conditions. During the 100% SOC spiking phase, Reactor B
averaged 82% COD removal, with 99% of this converted to methane. COD removal
was closely related to the amount of VFAs in the influent leachate. Influent
COD levels fluctuated with each leachate shipment, but accompanying responses
in methane production indicated a healthy system. As in Reactor A, metal
toxicity and SOC inhibition were not observed. Figure 3 shows COD levels
entering and exiting Reactor B, and Table 4 lists important influent and
effluent background parameter values.
SOC REMOVAL
With two uniquely different leachates, SOC removal data can be evaluated
for two different anaerobic environments. In both reactors partial or complete
mineralization of the hazardous organic compounds appeared to be the predominant
removal mechanism, while GAC adsorption provided stability to each system by
buffering against loading perturbations until organisms could adapt and respond.
Because of the number of compounds being spiked, the length of the project, and
the complex nature of each leachate, carbon adsorption could not be the
617
-------
o
O
2800
2400
2000
1600
120O
800
400
0
o Influent Total COD
* Influent Soluble COD
v Effluent Total COD
Effluent Soluble .COD
0
100
200
Day
300
Figure 2" Influent and Effluent GOD Concentrations for Reactor A,
7000
6000
5000
tU3 4000
Ji
3000
Q
O
O 2000
1000
o Influent Total COD
* Influent Soluble COD
v Effluent Total COD
Effluent Soluble COD
100
200
Day
300
400
400
Figure 3: Influent and Effluent COD Concentrations for Reactor B.
618
-------
Table 3: Reactor A Influent and Effluent Parameters
During 100% SOC Spiking Phase
Parameter
Influent A
Effluent A
Total COD
Soluble COD
Acetate
Propionate
Butyrate
Sulfate
Ammonia Nitrogen
Nitrate Nitrogen
TSS
vss
Copper
Iron
Magnesium
Manganese
Nickel
Lead
All concentrations in mg/L.
Standard deviation.
1,131 (130)*
1,016 (92)
23
14
66
89 (41)
306 (99)
0.6 (0.2)
159 (35)
80 (28)
0.06 (0.08)
13.4 (7.1)
187.4 (40.8)
2.0 (0.5)
0.2 (0.1)
0.1 (0.3)
836 (164)
775 (154)
14
6
51
21 (17)
304 (97)
0.5 (0.2)
195 (60)
84 (24)
0.04 (0.02)
9.1 (4.9)
189.9 (39.2)
(0.6)
(0.1)
1.7
0.2
0.1
(0.3)
predominant removal mechanism.
Reactors A and B, respectively.
Tables 5 and 6 summarize SOC removals in
High removal efficiencies were observed for the volatile compounds. The
three ketones, acetone, methyl ethyl ketone, and methyl isobutyl ketone (MIK),
were efficiently removed in both reactors with removal rates in excess of 95%.
In most cases, this would be ecceptable for a pretreatment process. In Figure
4, influent and effluent MIK concentrations are plotted as an example from this
group.
Of the chlorinated aliphatic compounds, trichloroethylene (TCE) and
methylene chloride were consistently reduced to low levels. TCE concentrations
decreased 98% in Reactor A and 99% in Reactor B, while methylene chloride
concentrations decreased 95% and 96% for Reactors A and B, respectively. One
chlorinated aliphatic compound, DCA, required a longer period of microbial
acclimation before high removal rates were obtained. In Figure 5, effluent
concentrations of DCA are shown to have increased as breakthrough occurred
between Days 125 and 250. With a higher concentration of DCA in the bulk
liquid, organisms utilizing DCA were able to acclimate to the compound. After
a number of weeks at the ultimate (100%) spike feed rate, DCA concentrations
began to drop and beyond Day 268 DCA removal averaged 91% in both reactors.
619
-------
Table 4: Reactor B Influent and Effluent Parameters
During 100% SOC Spiking Phase
Parameter* Influent B Effluent B
Total COD 3,846 (1,093)* 704 (165)
Soluble COD 3,759 (1,059) 622 (117)
Acetate 1,037 46
Propionate 467 22
Butyrate 284 56
Sulfate 32 (34) 3 (7)
Ammonia Nitrogen 311 (47) 296 (49)
Nitrate Nitrogen 0.9 (0.4) 0.6 (0.2)
TSS 370 (169) 335 (211)
VSS 188 (81) 140 (65)
Copper 0.06 (0.10) 0.02 (0.02)
Iron 55.9 (40.3) 19.7 (11.7)
Magnesium 145.7 (31.7) 139.8 (23.8)
Manganese 3.0 (1.2) 0.6 (0.2)
Nickel 0.1 (0.1) 0.1 (0,1)
Lead 0.1 (0.3) 0.2 (0.5)
All concentrations in mg/L.
* Standard deviation.
Table 5: Reactor A Influent and Effluent SOC Concentrations
During 100% SOC Spiking Phase
Influent Effluent Percent
Compound Average Average Removal
Acetone 10,169 (758)# 189 (216) 98
Methyl Ethyl Ketone 5,027 (282) 70 (68) 99
Methyl Isobutyl Ketone 1,006 (53) 35 (16) 97
Trichloroethylene 397 (20) 8 (10) 98
Methylene Chloride ,1,239 (84) 65 (50) 95
1,1-Dichloroethane 101 (5) 20 (17) 80
Chlorobenzene 1,094 (58) 67 (42) 94
Ethylbenzene 607 (34) 34 (19) 94
Toluene 7,960 (410) 436 (303) 95
Phenol 2,628 (171) 22 (23) 99
Nitrobenzene 514 (61) 6 (16) 99
1,2,4-Trichlorobenzene 203 (17) 10 (15) 95
Dibutyl Phthalate 212 (26) 26 (29) 88
All concentrations in
Standard deviation.
620
-------
1400
1200
1000
tlfl 800
$4 600
HH
400
200
0
v Influent A
* Influent B
o Effluent A
• Effluent B
0
50 100 150 200 250 300 350 400 450
Day
Figure 4: Methyl Isobutyl Ketone Concentrations,
175
-—•>
on 150
0) 125
ri
cd
^ 100
v Influent A
* Influent D
o Effluent A
• Effluent B
100 150 200 250 300 350 400 450
Day
Figure 5: 1.1-Dlchloroethane Concentrations.
621
-------
Table 6: Reactor B Influent and Effluent SOC Concentrations
During 100% SOC Spiking Phase
Compound
Acetone
Methyl Ethyl Ketone
Methyl Isobutyl Ketone
Trichloroethylene
Methylene Chloride
1 , 1— Dichloroethane
Chlo robenzene
Ethylbenzene
Toluene
Phenol
Nitrobenzene
1,2, 4-Trichlorobenzene
Dibutyl Phthalate
Influent
Average
12
9
1
. 1
1
8
2
,077
,570
,065
408
,284
110
,120
619
,243
,929
525
203
215
(2,900)*
(5,016)
(104)
(29)
(120)
(13)
(80)
(45)
(604)
(504)
(100)
(15)
(21)
Effluent Percent
Averase Removal
410
220
58
5
46
14
165
85
1,102
93
8
14
36
(577)
(150)
(25)
(3)
(44)
(8)
(86)
(41)
(744)
(63)
(17)
(16)
(30)
97
98
95
99
96
87
85
86
87
97
99
93
84
All concentrations in
Standard deviation.
Removal of aromatic volatile compounds was noticeably greater in the
sulfate—reducing conditions of Reactor A than in the methanogen—dominating
environment of Reactor B, although removal rates in both were encouraging.
Reactor A averaged 94% removal of chlorobenzene and ethylbenzene and 95% removal
of toluene. For the same compounds, Reactor B removals averaged 85, 86, and
87%, respectively. For all three compounds, effluent concentrations for Reactor
B slowly increased with time initially, but eventually decreased or stabilized,
This pattern is similar to that of DCA, where acclimation occurred only when
bulk liquid concentrations were sufficient to force acclimation. Influent
concentrations of these three compounds were slightly greater for Reactor B
because of differences in the two raw leachates. However, the modest
differences in influent concentrations do not appear large enough to contribute
significantly to the differences in removal efficiencies. Rather, the
contrasting reactor environments may be responsible.
For all of the VOCs except chlorobenzene and ethylbenzene, effluent
concentrations from Reactor A did not change significantly after GAG replacement
was halted. This indicates that compounds were degraded in the reactor and that
GAG replacement did not control the high removal efficiencies. Although
chlorobenzene and ethylbenzene effluent concentrations were slightly higher with
no GAG replacement, they were still adequate for most situations.
Off gases from Reactors A and B were analyzed for VOCs to investigate the.
importance of stripping losses. In Reactor A, none of the VOCs were seen in
significant amounts in the off gas, further indicating that biological activity
was responsible for SOC removal. With the substantial amount of gas produced
in Reactor B (14 L/day during gas sampling periods), a higher potential for VOC
622
-------
stripping was present. Ethylbenzene and toluene were lost in Reactor B off
gases at average rates equal to 18% and 11% of the influent loading rates,
respectively. These were two of the compounds with the lowest overall removal
rates in Reactor B.
Most of the semivolatile compounds were effectively removed from both
leachates. Because of analytical problems, results for bis (2-ethylhexyl)
phthalate are inconclusive regarding the fate of the compound and are not
included in Tables 5 and 6.. Phenol removal rates remained nearly constant at
99% for Reactor A and 97% for Reactor B (Figure 6). Likewise, nitrobenzene was
essentially nonexistent in the effluents, with both reactors achieving 99%
removal of the compound. Reactor A achieved 95% removal of trichlorobenzene,
while 93% was removed in Reactor B. Dibutyl phthalate removal rates were
adequate but not outstanding, averaging 88% in Reactor A and 84% in Reactor B.
CONCLUSIONS
Because of variations in raw leachate characteristics, two different
anaerobic environments were encountered in the expanded-bed reactors. One
reactor receiving a feed stream with a low biodegradable COD level exhibited a
symbiotic competition between sulfate-reducing bacteria and methanogens, while
the second reactor receiving a high biodegradable COD feed operated strictly as
a methanogenic unit. Both reactors achieved acceptable SOC removal rates for
a pretreatment process.
A comparison of the two systems indicates that a sulfate—reducing
environment may produce equal or better performance than a methanogenic
environment in removing a consortium of hazardous chemicals from waste streams.
All three volatile aromatic compounds in the SOC consortium were removed at
higher rates in the sulfate-reducing environment.
Effluent COD concentrations from both reactors were sufficiently low that
discharges of the pretreated leachates would not overload aerobic wastewater
treatment systems. Both systems responded well to seasonal variations in
leachate strength, indicating the adaptability of the anaerobic GAC expanded-bed
reactor.
ACKNOWLEDGMENTS
Funding for this work was provided by U.S. EPA under Contract Nos. 68-03-
4038 and 68-C9-0036. The authors wish to thank Dolloff F. Bishop, the technical
support staff at the U.S. EPA Risk Reduction Engineering Laboratory, personnel
at the U.S. EPA Test and Evaluation Facility, and personnel at the Delaware
Solid Waste Authority.
623
-------
00
O
rt
(D
6000
5000 -
4000 -
3000 -
2000 -
1000 -
Influent A
Influent B
o Effluent A
• Effluent D
50 100 150 200 250 300 350 400 450
Day
Figure 6: Phenol Concentrations.
624
-------
REFERENCES
1. Boyle, W.C. and Ham, R.K.; Treatability of Leachate from Sanitary
Landfills. Proceedings of the 27th Industrial Waste Conference, Purdue
University, May 2-4, 1972, pp. 687-704.
2. Platen, H. and Schink, B.; Methanogenic Degradation of Acetone by an
Enrichment Culture. Archives of Microbiology, v. 149, 1987.
3. Vogel, T.M. and McCarty, P.L.; Biotransformation of Tetrachloroethylene
to Trichloroethylene, Dichloroethylene, Vinyl Chloride, and Carbon Dioxide
Under Methanogenic Conditions. Applied Environmental Microbiology, v. 49,
1985.
4. Baek, N.H. andJaffe, P.R. ; Anaerobic Mineralization of Trichloroethylene .
Proceedings of the International Conference on Physiochemical and
Biological Detoxification of Hazardous Wastes, U.S. EPA. May 3-5, 1988.
5. Vargas, C. and Ahlert, R.C.; Anaerobic Degradation of Chlorinated
Solvents. Journal WPCF, v. 59, n. 11, 1987.
6. Bouwer, E.J. and McCarty, P.L.; Transformations of 1- and 2-Carbon
.Halogenated Aliphatic Organic Compounds Under Methanogenic Conditions.
Applied Environmental Microbiology, v. 45, 1983.
7. Grbic-Galic, D. and Vogel, T.M.; Pathways of Transformation of Toluene,
Benzene, and o-Xylene by Mixed Methanogenic Cultures. Applied
Environmental Microbiology, v. 53, 1987.
8. Wang, Y.T., Suidan, M.T., and Rittmann, B.E.; Anaerobic Treatment of
Phenol by an Expanded-Bed Reactor. Journal WPCF, v. 58, n. 3, 1986.
9. Suidan, M.T. ; Treatment of Coal Gasification Wastewater with Anaerobic
Filter Technology. Journal WPCF, v. 55, n. 10, 1983.
10. Shelton, D.R., Boyd, S.A., and Tiedje, J.M.; Anaerobic Biodegradation of
Phthalic Acid Esters in Sludge. Environmental Science and Technology, v.
18, n. 2, 1984.
11. Ng, A.S., Torpy, M.F., and Rose, C. ; Control of Anaerobic Digestion
Toxicity with Powdered Activated Carbon. Journal of Environmental
Engineering, v. 114, pp 593-605, 1988.
12. Personal communication with Richard C. Brenner and Richard A. Dobbs, Risk
Reduction Engineering Laboratory, U.S. EPA, Cincinnati, OH.
13. Narayanan, B. , Suidan, M.T., Gelderloos, A.B., and Brenner, R.C.;
Anaerobic Treatment of VOC's in High Strength Wastes. Submitted to
Journal WPCF, December 1990, Publication pending.
14. Federal Register Part VIII, 40 CFR Part 136; Guidelines Establishing Test
Procedures for the Analysis of Pollutants Under the Clean Water Act: Final
Rule and Interim Final Rule and Proposed Rule. October 26, 1984.
15. Standard Methods for the Examination of Water and Wastewater. 15th
Edition. American Public Health Association, Washington, DC, 1980.
625
-------
ANAEROBIC PRETREATMENT OF AN INDUSTRIAL WASTE CONTAINING SEVERAL VOC'S
by: B. Narayanan
John Corollo Engineers
Walnut Creek, CA 94598
M. T. Suidan
Department of Civil and Environmental Engineering
University of Cincinnati
Cincinnati, OH 45221
A. B. Gelderloos
Malcolm Pirnie, Inc.
Newport News, VA 23606
R. C. Brenner
U.S. Environmental Protection Agency
Cincinnati, OH 45268
ABSTRACT
The effectiveness of an anaerobic granular activated carbon (GAG) expanded^bed
bioreactor was evaluated relative to the pretreatment of high strength
industrial wastes containing RCRA volatile organic compounds (VOC's). A total
of six VOC's, methylene chloride, chlorobenzene, carbon tetrachloride,
chloroform, toluene, and tetrachloroethylene, were fed to the reactor in a high
strength matrix of background organic compounds. Operation of the reactor
resulted in excellent removals of all VOC's (> 97%). Chloroform, while itself
removed at levels in excess of 97%, was found to inhibit the degradation of
acetate and acetone, two of the base organic compounds. Without any source of
chloroform in the feed (either chloroform or its precursor carbon
tetrachloride), excellent COD removals were obtained in addition to near-
complete removal of all the other VOC's.
KEY WORDS; Granular activated carbon (GAG); anaerobic processes; expanded-bed;
volatile organic compound (VOC); chloroform; carbon tetrachloride;
chlorobenzene; methylene chloride; tetrachloroethylene (PCE); trichlorethylene
(TCE); vinyl chloride; toluene.
This paper has been reviewed in accordance with the
U.S. Environmental Protection Agency's peer and
administrative review policies and approved for
presentation and publication.
626
-------
INTRODUCTION
Many volatile organic compounds (VOC's) are known or suspected
carcinogens. They are extensively used in industry as degreasing and cleaning
fluids, refrigerants, and solvents. As such, they are often present in
appreciable concentrations in industrial wastewaters and their fate in treatment
systems is an issue of significant concern. Conventional aerobic processes are
often unable to satisfactorily detoxify VOC's due to the extreme volatility of
these compounds and because the high aeration rates commonly used in aerobic
biological processes result in excessive stripping into the gas phase1.
Furthermore, while non—chlorinated VOC's seem to be readily biodegraded
aerobically2, chlorinated VOC's for the most part resist aerobic breakdown and
stripping tends to be the dominant mechanism for their removal1-2.
Anaerobic treatment offers two distinct advantages for the treatment of
VOC's: first, the effect of stripping is substantially diminished compared to
that in aerobic processes. Stripping in an anaerobic process could occur only
due to the production of methane gas, and, typically, the amount of gas produced
is significantly smaller than the normal aeration rates employed in aerobic
processes. For example, 1 kg of COD fed to an anaerobic reactor would produce
about 395 L of methane under mesophilic conditions, which, assuming a gas
composition of 60% methane and an appropriate moisture content, would translate
to a gas production of 697 L. To treat the same 1 kg of COD aerobically,
typically, 0.7 kg of 02 would be required3, which, assuming an air composition
of 23.2% by weight of Q2, an oxygen transfer efficiency of 10%, and an operating
temperature of 20°C, yields an aeration flow rate of about 25,000 L, This is
approximately 38 times the gas production rate in the anaerobic treatment
process treating the same strength COD. Thus, stripping of VOC's will occur to
a much greater extent when wastewater is treated aerobically than when it is
treated anaerobically.
The second distinct advantage of anaerobic treatment of VOC's over aerobic
treatment is that biodegradation of chlorinated compounds under anaerobic
conditions occurs by reductive dehalogenation, and, as such, the greater the
number of chlorine atoms"on a compound the more easily it will be anaerobically
degraded. Several recent studies have shown that many of the VOC's appearing
on the Resource Conservation and Recovery Act (RCRA) list of compounds are
amenable to biodegradation under anaerobic conditions. Tetrachloroethylene
(PCE), for example, has been shown to be biotransformed by reductive
dehalogenation to trichloroethylene (TCE), dichloroethylene (DCE), and vinyl
chloride in a continuous—flow fixed film column*. Similar tests under aerobic
conditions exhibited no biotransformation of PCE5. Carbon tetrachloride has
been shown to be biologically transformed to chloroform and methylene chloride
in the presence of Clostridium sp. , a strictly anaerobic bacterium6. Chloroform
biodegrades under anaerobic conditions; however, higher concentrations have been
shown to inhibit the cultures7'8. Removal of chloroform by aerobic methods has
proven unsuccessful9. Methylene chloride, on the other hand, has been
demonstrated to readily degrade both anaerobically10 and aerobically11'12'13.
Chlorobenzene and other chlorinated benzenes have been shown to degrade
aerobically but not anaerobically5. Toluene has been shown to degrade under
both aerobic1'1'15'16 and anaerobic conditions17. Thus, anaerobic treatment appears
627
-------
to be a promising technology for the detoxification of many chlorinated VOC's.
When treating high strength industrial wastewaters, the toxicity of the
compounds involved is also a major concern as inhibitory concentrations or
accidental discharges of toxic compounds in a waste stream can completely
destroy a microbial community. There is, therefore, considerable merit in
combining carbon adsorption with anaerobic biological treatment while treating
high strength industrial wastes. Such a combination should effect efficient
removal of toxic substances via adsorption as well as biological degradation of
VOC's, many of which are biodegradable only under anaerobic conditions. Also
associated with such a system would be the cost—saving advantages of anaerobic
treatment resulting mainly from the conversion of organic chemical oxygen demand
(COD) to valuable methane gas with minimal sludge production.
Investigations of the anaerobic granular activated carbon (GAG) expanded-
bed bioreactor have demonstrated the effectiveness of the technology in treating
high strength industrial wastewaters. Suidan et al.18'19 and Khan et al.20'21
showed the effectiveness and resiliency of the process in treating synthetically
prepared phenol and catechol solutions. In subsequent studies, this process was
used to treat a mixture of biodegradable and biologically resistant and toxic
polycyclic nitrogen compounds22. The anaerobic GAG reactor has also been used
for the treatment of coal gasification wastewaters23'27 and for the treatment of
high strength industrial wastes containing chlorinated organic solvents28, COD
removal efficiencies of over 90% were consistently obtained in these studies for
COD volumetric loading rates that sometimes exceeded 28 kg/m3*d 27.
This study was designed to assess the potential of the anaerobic GAG
expanded-bed bioreactor in treating VOC's present in high strength industrial
wastewaters. A simulated wastewater consisting of several RCRA VOC's in a base
flow of non—RCRA organic compounds served as the system feed. The unit was
designed as a pretreatment system with emphasis on treatment of the VOC's and
not on the reduction of COD, which was contributed almost entirely by the
non-RCRA organic compounds.
MATERIALS AND METHODS
EXPERIMENTAL APPARATUS
The reactor system consisted of a jacketed column, recycle system, feed
system, and gas collection system. The anaerobic chamber (10.2—cm ID)(Figure
1) had a volume of 11 L and was maintained at 35°C. An internal recycle stream
at a recycle rate of 5 L/min provided an initial bed, expansion of 25% and helped
maintain a completely-mixed regime in the reactor. The reactor was charged with
1.5 kg of 16x20 U.S. mesh F400 GAG (Calgon Carbon Corp.) resulting in an
unexpended bed height of 43.2 cm and an apparent density of 0.43 g/cm3.
A gas collection system was devised to collect the gas, measure its
production rate, and then vent it to a hood. A wet tip gas meter was used to
measure the gas production rate. The system feed was prepared as three separate
solutions consisting of an organic feed, a buffer solution, and a nutrient
628
-------
Gas Meter
Effluent
Effluent
Header
Flow Meter-7
nr*l*
Influent
Header
=*=
=§=*
Organic Solution
and Syringe Pump
Vitamin and
Salt Solution
and Pump
Buffer Solution
and Pump
Recycle Pump
Fluidized GAC
Gravel Pack
| Coupling
H Valve
Figure 1. Schematic of Anaerobic Expanded-Bed Treatment System
629
-------
solution. The organics, buffer solution, and nutrient solution were all fed
directly into the suction side of the column recycle loop to thoroughly mix the
contents prior to entry to the reactor. The VOC's were fed to the reactor from
a 25-raL syringe connected to a variable-speed syringe pump. The speed of tfrje
pump was adjusted to run continuously over a 24—hour period. The pump syringe
was refilled on a daily basis by connecting another syringe with new organic
feed to a three-way valve near the tip of the pump syringe. The pump syringe
was connected to the column using stainless steel tubing.
The buffer and nutrient solutions were fed to the reactor from separate
55—L plastic reservoirs, calibrated at 1—L divisions. The solutions were pumped
with separate fixed—speed positive displacement pumps connected to separate
timers. Tygon tubing was used to feed the solutions to the reactor.
SYNTHETIC INDUSTRIAL WASTEWATER AND NUTRIENTS
The target influent concentrations of the organic feed constituents are
given in Table 1. Acetone, methanol, and acetic acid represent the background
organic compounds typically found in industrial wastes. Due to the volatility
of the solvents and organics, the organic feed to the reactor was prepared every
2 to 4 weeks and stored in 25-mL amber septum vials with a PTFE-faced septum at
4°C. The total daily volume of the three background organic compounds with the
RCRA VOC's was greater than the volume of the syringe (25 mL). Attempts to use
a larger syringe (50 mL), however, resulted in pump failure. To reduce the
volume to below 25 mL, acetic acid was taken out of the organic feed and added
to the buffer solution.
Table 1. Target Composition of Synthetic Industrial Wastewater
(All concentrations in mg/L)
Parameter Concentration
Volatile RGRA Compounds
Carbon Tetrachloride 20
Chlorobenzene 20
Chloroform 20
Methylene Chloride 20
Tetrachloroethylene 20
Toluene 20
BackgroundOrganic Compounds
Acetic acid • 1565
Methanol 1110
Acetone 755
The buffer solution containing sodium carbonate (6000 mg/L) was added to
maintain a near neutral pH. Ammonium carbonate (775 mg/L) and sodium sulfide
(300 mg/L) were added to the buffer solution as nutrients. Acetic acid at a
630
-------
concentration of 3130 mg/L was also added to the buffer solution. The buffer
solution was fed into the system at a flow rate of 4 L/day, which represented
49.9% of the system influent flow.
To supply the necessary micronutrients required for the growth of
microorganisms, a solution containing inorganic nutrients and vitamins was fed
to the reactor. The flow rate was identical to that of the buffer solution (4
L/day). Stock salt and vitamin solutions were made periodically and stored at
4°C until needed. The composition of the nutrient solution is given elsewhere29.
ANALYTICAL METHODS
Daily measurements were taken of the volume of gas produced and the flow
rates of the buffer, nutrient, and organic stock solutions. Room temperature,
pH, and water bath temperature were also monitored daily. Effluent samples were
analyzed for VOC's on a weekly basis. Analyses for acetate, COD, and gas
composition were also performed weekly.
Four of the six RCRA VOC's, methylene chloride, chloroform, carbon
tetrachloride, and tetrachloroethylene (PCE), were analyzed using a purge and
trap (Model 14-000-40, Tekmar Co.) followed by electron capture detection on a
gas chromatograph (GC) (Model 5980, Hewlett Packard). A megabore capillary
column (J&W Scientific, DB—624) was used for separation. The purge and trap was
run with a 15-minute purge at room temperature, which included a 4-minute dry
purge, followed by a 4-minute desorb at 180°C, and finally a 7—minute bake at
225°C. The GC was programmed from 35°C to 90°C at 5°C/minute with a 5-minute
hold at 35°C, then from 90°C to 150°C at 30°C/minute with a 4-minute hold at
150°C. The carrier gas for the purge and trap and the GC was ultra-high purity
helium. The purge flow rate was 40 mL/minute, and the carrier flow rate was 8
mL/minute. P—5, used as a make—up gas, was applied at a flow rate of 50
mL/minute. Injector and detector temperatures were 120°C and 260°C,
respectively.
Sample injection volume for aqueous samples of these four compounds was
5 mL using a glass hypodermic syringe with a Luer—Lok tip. Water used for
blanks and dilutions was prepared by purging nitrogen gas through distilled
water for at least 10 minutes. Standards were prepared following EPA Method
502.2, Section 6.3.1. At least four different standard concentrations were
analyzed, with one standard containing concentrations slightly higher than its
detection limits and the rest in the linear output range.
Effluent samples were filtered using a 10—mL syringe with an attached
0.45-^m membrane filter. This filtration step was found to cause no detectable
losses of the VOC's. Samples were analyzed within 2 hours of. being taken. If
samples could not be analyzed immediately, the samples were acidified (pH< 2.0)
and stored without headspace in 40—mL amber septum vials with PTFE—faced
silicone septum at 4°C. The release of carbon dioxide as a result of
acidification was found to cause, no VOC losses. Stored samples were analyzed
within 1 week.
To determine what fraction, if any, of the different VOC's was being
631
-------
stripped into the gas phase, the product gas was sampled and analyzed by direct
injection into the purge and trap. The gas sample was obtained directly from
a septum placed in the product gas line. The volume of the gas sample to be
used was adjusted to maintain a response within the linear output range of the
compounds on the chromatogram. For the analysis of the gas from this reactor,
a sample size of 1 mL ensured that all the compound peaks were in the linear
range. The gas phase concentrations thus determined were used in conjunction
with the daily gas flow rate to determine the total mass of each of the VOC's
being stripped daily. These were then compared with the daily influent and
effluent mass flow rates of the VOC's to provide a better assessment of the
fate of the various compounds.
Because a PID/Hall detector was not available at the time of the study,
analysis for two of the VOC's, toluene and chlorobenzene, was performed using
a Hewlett Packard 5890A GC and a flame ionization detector (FID). A DB-5
megabore column (J&W Scientific) was used for separation of the compounds.
Ultra high purity helium, at a flow rate of 15 mL/min, was used as the carrier
gas. Hydrogen and air at flow rates of 30 mL/min and 400 mL/min, respectively,
were used to fuel the flame.
Preparation of this sample for injection involved two steps. First, a
250—mL sample was acidified (to a pH<2) and passed through a packed bed of XAD—2
resin at a flow rate not exceeding 10-15 mL/min. The compounds were
subsequently extracted off the XAD—2 resin using methylene chloride, A known
mass (0.1 mg) of the internal standard, 2,4,6-trichlorophenol (TCP), was added
at this stage (100 fiL of 1 g/L stock TCP) prior to injection into the gas
chromatograph.
Standard curves were prepared for three different concentrations of each
compound. The standards used in preparing these curves were processed in
exactly the same manner as were the reactor effluent samples. The mass ratios
of the compounds to the internal standard were plotted versus the ratio of the
areas of the compound peak and the internal standard. Using these standard
curves, the area ratio obtained from the injection of the sample was used to
determine the mass ratio of the compound to the internal standard. Since the
mass of the internal standard in the injection sample was known, the mass of the
compound could be calculated. From the initial volume of sample passed through
the resin, the concentration of the compound was obtained. The efficiencies of
extraction of the VOC's were determined by comparing the chromatograms from
standards that were prepared directly in methylene chloride to aqueous samples
that had been extracted on the resin. The extraction efficiency for toluene was
50%, while that for chlorobenzene was 49%.
Anberlite XAD-2 resin was used for the extraction process. The resin was
cleaned thoroughly prior to use by the following procedure. Eighty grams of
resin were washed with a 2% ammonium carbonate solution for 20 minutes. The
resin was then rinsed with 500 mL of distilled water. Sequential extraction in
a soxhlet apparatus using water, methanol, and diethyl ether for -24 hours in
each solvent followed. The volume of each solvent was 300 mL. At the end of
the ether reflux, the resin was rinsed successively with 200 mL of methanol and
1000 mL of distilled water. The aqueous slurry of resin was stored in methanol
in glass stoppered bottles until use.
632
-------
Acetic acid -was measured by gas chromatograpby using a. Hewlett Packard
5710A GC equipped with an FID detector. Samples were diluted, if necessary, in
1% formic acid dilution water. The analysis was done at a constant oven
temperature of 100°C using a 60/80 Carbopack, C/0.3% Carbowax 20M/0.1% H3P04
glass column. The carrier gas was nitrogen applied at a flow rate of 30 mL/min.
The injection temperature was 180°C, and the detector temperature was,250°C..
Product gas was sampled weekly through a septum in the gas line. The gas
samples were analyzed for methane, nitrogen, carbon dioxide, and oxygen as
percentages of the total gas volume, A gas partitioner and certified gas
standards were used for the gas analysis.
COD was measured using a Hach COD Reactor (Model 16500-10) with prepared
digestion vials.
•EXPERIMENTAL RESULTS AND DISCUSSION
REACTOR OPERATION
Prior to the addition of the VOC's, the reactor was gradually acclimated
to the background organic compounds fed at the desired COD loadings over a 4-
month period. Day 1 corresponds to the time when steady—state conditions were
achieved in the reactor. Steady-state operating conditions were maintained for
approximately 2 weeks, Day 1 to Day 15, prior to the addition of the VOC's. The
total flow to the system during this period was 8.02 L/day, consisting of 4
L/day vitamin and nutrient solution, 4 L/day buffer solution containing acetic
acid, and about 20 mL/day of the remaining two background organic compounds.
The three background organic compounds were added in equal COD proportions
totalling 5000 mg/L. Gas production during the steady—state background- period
averaged 18 L/day, and the filtered effluent COD was approximately 100 mg/L.
Both these numbers correspond to COD removal efficiencies of greater than 90%.
No partial replacement of the carbon medium with virgin GAG was practiced during
this period.
The RCRA VOC's were added to the background feed on Day 16 at 25% of the
target strength listed in Table 1 (5 mg/L). Carbon replacement at 15 g/day,
or 1% of the total carbon bed, was initiated on Day 25. Table 2 lists the
subsequent changes in both VOC influent concentrations and carbon replacement
rates. As is the case with any methanogenic reactor, gas .production and
effluent COD were taken to be good indicators of reactor performance. These
parameters were monitored closely for indications of inhibitory effects of the
VOC's. No such effects were observed, and the daily gas production rate and
weekly filtered effluent COD remained stable at 18 L/day and 180 mg/L,
respectively. Effluent VOC concentrations were not analyzed until Day 123 due
to delays associated with analytical method development.
633
-------
Table 2. Date of VOC Concentration or GAC Replacement Rate Variations
Date
Day
CT>
w
Carbon Chloroform Remaining VOC GAG
Tetrachloride cone. (mg/L) Cone. (mg/L) Replacement
cone. (mg/L) (9/day)
1/1/89
1/16/89
1/25/89
3/18/89
4/30/89
5/30/89
9/15/89
6/30/89
7/26/89
8/17/89
2/15/90
0
16
25
77
120
150
164
181
207
229
411
0
5*
5
10
20
20
20
20
20
20
o.
0
5
5
10
20
10
10
o
0
0
0
0
1
5
It
20
20
20
20
20
20
20
0
0
15
15
15
15
30
30
15
0.
0
Bold type with underline indicates the change made to the operating conditions,
Plain type indicates a continuation of the current operating conditions.
-------
Consequently, on Day 77, the influent concentration of each of the VOC's
was doubled to 10 mg/L while maintaining a constant 15—g/day carbon replacement
rate. Stable operating conditions were observed; however, the daily gas
production rates were higher than those of the previous phase at around 20 L/day
and filtered effluent COD values were slightly lower at about 70 mg/L. This gas
production rate represents a conversion of 102% of the theoretically attainable
gas production if all the influent COD were converted to methane. A possible
explanation for the higher gas production rate is an increase in biomass within
the reactor that lead to the utilization of previously adsorbed compounds.
Since no inhibitory effects were observed by Day 120, the influent
concentration of all the VOC's was again doubled to 20 mg/L. Weekly analysis
of effluent VOC concentrations was begun on Day 123 and continued throughout the
rest of the study.
Figure 2 presents the influent and effluent COD and the COD equivalent
of the methane gas produced versus time. On approximately Day 140, the daily
gas production exhibited a steady decrease from approximately 19 L/day to a new
plateau of 13 L/day over a time period of 1 week. Accompanying the decrease in
gas production was a corresponding increase in effluent COD values. The
filtered effluent COD concentration increased from 145 mg/L on Day 129 to 1370
mg/L by Day 147. The effluent concentrations of all VOC's remained low and
stable, except for chloroform, which showed a sharp increase from 65 /zg/L on Day
141 to 102 fJ-g/L on Day 150. Chloroform was, therefore, believed to be the
primary cause of inhibition.
Figure 3 shows the concentration of influent and effluent chloroform and
effluent COD versus time from Day 123 onward. As can be seen, there is a
definite relationship between the effluent COD and effluent chloroform
concentrations, further confirming the hypothesis that chloroform was indeed the
cause of inhibition. Two strategies were attempted to return the system to
previous operating conditions. First, the influent concentration of chloroform
was decreased to 10 mg/L on Day 150, leaving the other VOC's at 20 mg/L.
Lowering the influent chloroform concentration did not alleviate the inhibition
and, subsequent GC analysis showed no decrease in effluent chloroform
concentrations. Rather, the effluent chloroform concentration increased further
to 265 /ug/L by Day 163, while effluent COD remained above 1000 mg/L and daily
gas production remained at 13 L/day,
An attempt was also made to determine whether carbon adsorption could
suppress the inhibitory effects of chloroform. The carbon replacement rate was
doubled to 30 g/day on Day 164. Prior to initiating the new replacement rate,
175 grams of carbon was removed from the column and replaced with 100 grams of
virgin carbon and 75 grams of biological carbon from a similar column treating
acetate and phenol. The batch carbon replacement and the new replacement rate
did not suppress the effluent chloroform concentration; rather, it continued to
gradually increase to a value of 308 ^g/L on Day 180. Effluent COD also did not
decrease; rather, values increased to 1580 mg/L. It is believed that
replacement rates greater than 30 g/day (50-day solids retention time (SRT))
would not be able to control chloroform inhibition any better because chloroform
is such a weakly adsorbing compound. Also, higher replacement rates decrease
the SRT and may adversely affect the biological removal capabilities of the system.
635
-------
500
Influent COD
Effluent COD
Gaseous COD
100 200 300 400 500
Time, Day
Figure 2, Influent, LiguM Effluent, said Gaseous COD Variations
600
3000
B»
E
eT
8 2000
it«*
e
*»
a
•*m
£3
£ 1000
ti
2%
GAC Replacement % of Bed
No CarbonTetraehloride
- Effluent COD
\-rHUrOlUMii
Ong/L
Influent Chloroform Cone,, mgll
fUV
-600 -J
-500
-400
e»
a.
O
300 g
U
-200 •£
4>
a
-100 £
0
100 150 200 250 300 350 400 450 500 550
Time, Day
Figure 3. Filtered Effluent COD and Effluent Chtoiofonn Concentration vs. Tme
636
-------
As a result of these upward trends of effluent COD -and chloroform
concentrations, chloroform was completely removed from the influent on Day 181
while the remaining VOC's were kept at 20 mg/L. Though the effluent chloroform
concentration dropped, a significant amount still appeared in the effluent,
leading to the conclusion that there was another source of chloroform. Previous
studies (5) have shown that carbon tetrachloride is anaerobically degraded to
chloroform. Therefore, it was concluded that the chloroform seen in the reactor
effluent after removal of chloroform from the feed was a product of carbon
tetrachloride degradation.
Accompanying the decrease in chloroform concentration, gas production
gradually increased from 13 L/day to a new plateau of 16 L/day. Effluent COD
values corresponding to this new plateau were approximately 1000 mg/L. Carbon
replacement was maintained at 30 g/day during this recovery period. On Day 207,
the carbon replacement rate was returned to 15 g/day. Effluent chloroform
concentrations remained stable at approximately 180 //g/L.
On Day 229, it was decided to discontinue carbon replacement completely
to allow the system to reach an equilibrium chloroform concentration. Effluent
chloroform concentration increased from approximately 200 fig/L to a maximum
value of 441 ^g/L on Day 296. Subsequently, the concentration dropped,
suggesting a trend towards a steady-state operating condition. At this stage,
PCE was the only other volatile compound being detected in the reactor effluent.
An additional compound, identified to be trichloroethylene (TCE), started
appearing in the effluent from approximately Day 200 onward. PCE has been shown
in the literature to degrade to TCE and further to dichloroethylene and vinyl
chloride. On Day 306, chloride analyses were begun in an effort to quantify the
biodegradation of the chlorinated VOC's. A significant increase in chloride
production was noted, indicating transformation of the VOC's within the reactor.
Around Day 348, the concentration of chloroform in the effluent began rising
rapidly. TCE and PCE concentrations also increased, though at a much slower
rate. These changes were accompanied by the appearance of methylene chloride
and carbon tetrachloride in the effluent. A new compound, identified to be
vinyl chloride, also started appearing in the effluent. Quantification of vinyl
chloride concentrations began on Day 348.
On Day 411, due to a power failure lasting 12 hours, reactor operation was
seriously disturbed. Gas production dropped drastically from 11 L/day to 1
L/day. The feed to the reactor was stopped for 1 day, and 200 g of GAC were
added to the reactor. Carbon tetrachloride was completely removed from the
feed, and the feed was restarted at 25% of the original rate. Gas production
was monitored carefully, and the feed rate was slowly brought back to its
original value. The effluent chloroform concentration, whi-ch was at 562 ng/T-.
before the accident, decreased rapidly once carbon tetrachloride was removed
from the feed. The effluent TCE concentration also, increased, with a
corresponding decrease in the PCE concentration". The effluent chloroform
concentration eventually dropped below detection levels, and gas production
increased to a new level of 20 L/day. A rapid decrease in the effluent PCE
concentration to levels below the detection limit, with an accompanying increase
in the effluent TCE concentration, was also observed in this phase. It is
suggested that the sudden increase in gas production following the complete
disappearance of chloroform from the effluent had the effect of shearing off the
637
-------
species responsible for the biotransformation of TCE, This theory of loss of
biomass from the reactor was borne out by a visual examination of the effluent
and by a comparison of the total and filtered effluent COD values.
Subsequently, the TCE concentration in the effluent levelled off at 350 /*g/L.
DEGRADATION OF BACKGROUND ORGANIC COMPOUNDS
Each of the background organic compounds, acetate, acetone, and methanol,
contributes 32.57% of the influent COD with the remaining 2.29% attributable to
the RCRA VOC's. Therefore, COD removal across the reactor is a good indicator
of the extent of degradation of these compounds. The cumulative COD balance is
presented in Figure 4. The total COD out includes the COD of the liquid
effluent and the COD equivalent of the methane gas produced. At any time, the
difference between these two lines represents the accumulation of COD that was
either adsorbed onto the GAC surface (including the COD associated with any GAG
that was removed from the reactor) or present in the form of attached biomass.
The slope of the influent COD line at any point represents the rate of COD
addition to the reactor, and the slope of the effluent COD line at any point
represents the rate at which COD is leaving the reactor. The COD leaving the
reactor represents the sum of the effluent aqueous COD and the COD equivalent
of the methane gas produced. Only during periods of bioregeneration will the
rate of COD release exceed the rate of COD addition. In a non-adsorbing system,
the rate of COD release cannot exceed the rate of addition except in cases where
sloughing of retained solids occurs. During stable operating periods, the rate
of COD addition will only slightly exceed the rate of release.
The total mass of COD retained in the reactor is presented in Figure 5.
Since the mass of carbon was kept constant, this figure also represents the
loading on the activated carbon. The slope represents the rate of COD
adsorption on the carbon. A negative slope indicates that bioregeneration is
occurring. As is clearly seen, the reactor experienced frequent periods of
bioregeneration, indicating an extremely dynamic system with a transition of
quasi steady—state conditions.
The cumulative mass balance in Figure 4 demonstrates good accountability
of the COD, even after inhibition began. The average difference between the
total COD in and the total COD out was 183 g/1500-g GAC, or 122 mg COD/g GAC.
During the phase of zero carbon replacement, 74.5% of the COD removal occurred
biologically. Adsorbed compounds and biomass accounted for 0.8% removal for a
total removal efficiency of 75.3%, leaving 24.7 % of the COD in the effluent.
Biological degradation of the background organic compounds was measured
by the amount of methane and carbon dioxide gas produced. The total loading
rate of these compounds was high enough (5000 mg/L as COD) that fluctuations in
daily gas production would give an indication of the degree of biological
inhibition. Theoretical and actual methane production with time are shown in
Figure 6. Chloroform inhibited the degradation of acetate and acetone, two of
the background organic compounds, following the increase in chloroform feed rate
to 20 mg/L. Hence, during the inhibitory phase, methane production was well
below the theoretically predicted rate. After the complete removal of all
638
-------
30000
§20000
O
+ TOTAL INFLUENT
o LIQUID EFFLUENT + GAS
a LIQUID EFFLUENT
100 200 300 400 500 600
Time, Day
Figme A. Cumulative Total Influent, Total Effluent, and Liquid Effluent COD
ouu
e»
. *
te
e
1 200
0)
CC
.£
•9
£ too
e
'3
Ite*
u
CC
§ o'
u
-mn
• . COD Retained
• ~*
• • «^%
* •
- .% * • '.
• * * * • *
» * *
^ * » *
• . .• • • ^ ' v
• ^f 9 * **
- V*
•**
.1 . i . i . i . i <
100 200 300 400 500 600
Time, Day
Figure 5. COD Retained in the Reactor as a Function of Time
639
-------
0179
Methane Production, L/day
o 01 o en c
• • | • • | • • | • • | • • | • *
10 go 10 o Influent Chloroform Cone., mg/L
a
Theorelical production = 14.2 Uday
DD Q D Q
''* "° °° B " •
D _D D
. D Qa a n BQ D n B0
V B 0°° o X°°,>0
a
D
QEQ
a
D
DQ
Accid^n? nH
r
T CCI4 Removed
EEJEI s_
^
Q Actual Methane Product! on
i . . i . . i . • i . . i . .
) 100 200 300 400 500 60
Time, Day
Figure 6. Theoretical vs. Actual Methane Production
-------
sources of chloroform from the feed, gas production increased rapidly to reach
the theoretically predicted rate, indicating excellent blodegradation of the
background organic compounds, which constituted almost all of the feed COD.
During the acclimation phase and prior to inhibition, acetate was the only
detectable background organic compound in the effluent. Neither methanol nor
acetone, or its intermediate 2—propanol, was identified in the effluent. The
inhibitory phase was characterized by an increase in acetate concentrations and
the appearance of acetone and 2-propanol in the effluent. Acetate accounted for
54,5% of the total effluent COD, with acetone and 2—propanol accounting for
40.7% and 3.2%, respectively. Methanol was not detected in the effluent,
suggesting that the mechanism degrading methanol to methane was not affected by
the levels of chloroform present in the system.
FATE OF VOC's IN THE REACTOR
VOC's fed to a GAC anaerobic reactor can theoretically be removed by
either one or a combination of the following mechanisms: adsorption onto the GAC
surface, stripping into the gas phase and biological degradation. Hence, these
three factors have to be considered when studying the fate of VOC's fed to an
anaerobic GAC reactor.
Adsorption on the GAC Surface
Adsorption of a compound on GAC is best characterized by an isotherm
equation. The isotherm equation for a compound gives the constant temperature
equilibrium relationship between the quantity of adsorbate per unit of
adsorbent, q, and the equilibrium concentration of adsorbate in solution, C.
An isotherm is valid only for the concentration range tested. Adsorption
Isotherms on the same GAC are available in the literature for the VOC's of
interest30 and are summarized in Table 3. These isotherms were derived for
single-adsorbate conditions at a constant temperature of 20 °C. The competitive
adsorption actually occurring in the reactor due to the presence of other
compounds and the higher reactor temperature would result in a reduced capacity
of the carbon for the compound from that predicted using the isotherm. The
presence of a biofilm around the carbon particle in the reactor may also reduce
adsorptive capacity. Hence, the isotherm cannot be used directly to calculate
the amount of adsorption of a. compound occurring in the reactor. However, it
can be used indirectly as an estimate of the maximum adsorption that would
occur if adsorption were the sole mechanism of removal. Starting with a
hypothetical effluent concentration of 100 jig/L (this value was chosen because
it fell within the valid concentration ranges for all the compound isotherms),
the equation can be used to calculate the mass loading of that particular
compound per unit mass of GAC. This value, when multiplied by the total mass
of GAC in the reactor, gives the mass of the compound on the GAC that is in
equilibrium with the 100—jtg/L effluent concentration. Using actual reactor feed
flow rates and effluent concentrations, the time required to load the
aforecalculated mass onto the carbon can be determined. This represents the
time it would take for the effluent concentration in the reactor to reach 100
/ig/L if adsorption were the sole mechanism responsible for removal. As
mentioned earlier, the isotherms overestimate the actual capacity of the reactor
641
-------
Table 3. Adsorption Isotherms for VOC's on Filtrasorb-400 GAG
(After Speth and Miltner30}
Temperature = 20°c
Compound Adsorption Isotherm. Valid Concentration Range (um/L)
Chloroform 92.5*C0<669 13.2 - 226
Carbon Tetrachloride 387*C°-594 9.1 - 429
Chlorobenzene 9170*C°-348 15.4 - 732
Methylene Chloride 6.25*C°'801 18.1 - 715
CT>
ji,
1X3 Toluene 5010*C°-429 2.3 - 104
Tetrachloroethylene 1070*C°-604 1.2 - 738
-------
carbon because they do not account for the higher reactor temperature and the
effect of competitive adsorption. The time value calculated, therefore,
represents an upper limit figure. Thus, applying the above assumptions, all of
which tend to overestimate the adsorption capacity of GAC for the compound, it
would at most take this calculated time for the reactor concentration to reach
100 pg/L. Since, in the absence of biological activity, all the assumptions
made in calculating the adsorptive capacity of carbon tend to overestimate
capacity, the concentration under actual conditions can only be higher than 100
pg/L at the calculated time. However, as can be seen from Table 4, for all the
VOC's except chloroform, effluent concentrations of 100 jig/L were never seen in
the reactor. Even for chloroform, the actual reactor concentration first
reached 100 jag/L long after theoretically predicted by the pure adsorption
calculations. On the basis of these calculations alone, there appears to be
sufficient justification for the argument that adsorption was not the sole
removal mechanism and that biodegradation of the compounds was also occurring.
Table 4, Comparison of Actual Reactor Performance
vs. Pure Adsorption
Reactor Operating Conditions:
Day 1. - Day 76: Feed Concentration of Each VOC = 5 mg/L
Day 77 - Day 119: Feed Concentration of Each VOC = 10 mg/L
Day 120 Onwards: Feed Concentration of Each VOC - 20 mg/L***
Feed Flow Rate - 8.0 L/day
Total Period of Reactor Operation = 550 days
Time to Reach Effluent Time to Reach 100
Cone, of 100 pg/L P8/L *n Actual Reactor
Assuming Only
Compound Adsorption
(days) (days)
Chloroform 76 150
Carbon Tetrachloride 134 Never Reached
Chlorobenzene 505 Never Reached
Methylene Chloride 9 Never Reached
Toluene 417 Never Reached
PCE 240 Never Reached
Except in cases of chloroform and carbon tetrachloride
(See Table 3).
643
-------
Stripping of VOC's
To determine whether any of the VOC's were being stripped from the
effluent by the gas produced in the system, analysis of the gas for VOC's was
started on Day 450. All the compounds appearing in the effluent, with the
exception of methylene chloride, were detected in the gas phase. This analysis
was performed on a regular basis for an extended period of reactor operation,
and sufficient data were gathered to enable derivation of a relationship between
the mass of compound in the effluent and the mass of compound in the, gas, The
results of this study are summarized in Table 5.
Table 5. Summary of VOC Stripping Analysis
Compound Mass in Gas/Mass in Effluent (g/day)/(g/day)
Vinyl Chloride 3.846
Methylene Chloride No Stripping in Gas
Chloroform 0.246
TCE 1.137
PCE 0.430
t
Biological Degradation
The effect of biodegradation can be best studied during the phase of no
GAG replacement. If no carbon is being replaced, then at steady state, no
removal by adsorption will be occurring, and any removal of VOC's is solely due
to biodegradation or stripping. The extent of removal by stripping can be
calculated as described in the previous section, knowing the effluent
concentrations of the VOC's. Hence, a mass balance on each VOC will yield
directly the removal that can be attributed to biodegradation. Mass balances
were performed for each of the VOC's separately and the results are summarized
below.
Methylene Chloride — Steady-state data from the period of zero GAC replacement
were utilized to perform a mass balance. The steady-state influent and effluent
concentrations were 20 mg/L and 25 Mg/L. respectively, which at a flow rate of
8 L/day through the reactor yield influent and effluent mass flow rates of
644
-------
160,000 /ig/day and 200 /ig/day, respectively. As described in the previous
section, no stripping was observed for methylene chloride. Furthermore, as no
carbon replacement was being practiced during this period, the effect of
adsorptive removal can be neglected, leaving biodegradation as the sole
mechanism responsible for the 99.9% removal observed.
Toluene, Chlorobenzene — Effluent toluene and chlorobenzene concentrations were
below detection limits throughout the period of no GAG replacement. Using the
same argument applied for methylene chloride suggests complete biological
removal of these compounds, as no trace of either of these compounds was
observed in the gas phase.
Carbon tetrachloride, Chloroform — Chloroform is a biodegradation by—product
of carbon tetrachloride; consequently, the mass balances have to be done
simultaneously for these compounds. One mole of carbon tetrachloride will
biotransform to one mole of chloroform. Thus, in terms of moles of chloroform,
the influent molar potential for chloroform equals the sum of the moles of
chloroform and carbon tetrachloride in the influent. Stripping characteristics
of carbon tetrachloride are not available as the stripping study was conducted
after carbon tetrachloride had been removed from the feed. The concentration
of carbon tetrachloride in the effluent was almost always below detection
levels, and the mass balance for the period when carbon tetrachloride but not
chloroform was being fed to the reactor is based on the not unreasonable
hypothesis that these negligible effluent concentrations are the result of
nearly all the carbon tetrachloride being transformed via chloroform.
Conversely, the same negligible effluent concentrations would render
insignificant the effect of any stripping that was to occur from the effluent.
Steady—state data for this period included a carbon tetrachloride influent
concentration of 20 mg/L and no chloroform at a flow rate of 8 L/day. This
corresponds to an influent chloroform potential of 1039 ^Moles/day. The
effluent steady—state concentration of chloroform was 350 £tg/L with no carbon
tetrachloride being detected. At the 8-L/day flowrate, this translates to a
molar effluent flow of 19.3 /xMoles/day. Using the factor obtained in the
stripping study for chloroform (Table 5) , the molar flow rate in the gas amounts
to 4.7 /uMoles/day. Thus, a total of 24 /jMoles/day of chloroform can be
accounted for in the effluent and gas. As this was a period of no carbon
replacement, adsorptive removal can be ruled out. During this period,
therefore, it is concluded that 97.7% of the influent chloroform potential was
being removed biologically.
Tetrachloroethylene (PCE) — The pathway of biodegradation of PCE is known to
be through TCE, DCE, and vinyl chloride. A mass balance on PCE would by
necessity, therefore, involve these other compounds as well. The transformation
occurs on a one—to-one molar basis; hence, one mole of either TEC, DCE, or vinyl
chloride in the effluent corresponds to an equivalent mole of PCE. The influent
concentration of 20 mg/L of PCE fed at rate of 8 L/day corresponds to a molar
flow rate of PCE of 964 /iMoles/day, The effluent quality corresponding to this
steady-state period shows effluent concentrations of 25 Mg/L of PCE, 15 /ig/L of
TCE, and 10 /ig/L of vinyl chloride. At a flow rate of 8 L/day through the
reactor, this corresponds to molar flow rates of 1.2 /Moles/day of PCE, 0,9
^Moles/day of TCE, and 1.28 /^Moles/day of vinyl chloride. The corresponding
flow rates in the gaseous phase can be calculated using the factors obtained in
645
-------
the stripping study and summarized In Table 5. These are 0.52 /iMoies/day of
PCE, 1.02 pMoles/day of TCE, and 4.92 /^Moles/day of vinyl chloride. The above
values add up to a combined equivalent PCE molar flow rate of 9.84 #Moles/day
leaving the reactor In the gas and liquid phases. Again neglecting adsorptive
removal based on the zero GAG replacement schedule practiced during this period,
the mass balance indicates that 98.9% of the PCE was being removed through
complete biological degradation.
SUMMARY AND CONCLUSIONS
The objective of this study was to examine the effectiveness of the
anaerobic GAG expanded—bed bioreactor as a pretreatment unit for the
detoxification of a simulated high strength industrial wastewater containing
several volatile RCRA compounds present In backgrounds consisting of non-RCRA
organic compounds. As a pretreatment unit, the goal was not to maximize COD
destruction but to reduce the VOC concentrations to acceptable levels. This
goal was achieved very satisfactorily. The reactor demonstrated excellent
treatment; removals of greater than 97% were achieved for all the VOC's.
Chloroform was found to be Inhibitory to the system at effluent concentrations
of about 100 /ig/L. It was found to inhibit the degradation of acetate and
acetone, two of the three base flow organic compounds. Chloroform itself,
however, was removed to greater than 97%. The only limiting factor in this
treatment study was the high effluent COD experienced during the inhibitory
phase, which was composed almost entirely of acetate and acetone and, as such,
could easily be removed by any of several treatment options. The amount of
stripping occurring was negligible compared to the amount of stripping
anticipated to occur in an aerobic biological process. The anaerobic GAG
expanded-bed bioreactor represents an excellent pretreatment unit for the
treatment of wastes containing VOC's.
REFERENCES
1. Dobbs, R,A. , 1990. "Factors Affecting Emissions of Volatiles From Wastewater
Treatment Plants." Workshop Report and Proceedings Air Toxic Emissions and
POTWs, Appendix J.
2. Melcer, H., D. Thompson, J. Bell, and H. Monteith, 1989. "Stripping of
"Volatile Organic Compounds at Municipal Wastewater Treatment Plants."
AWM/EPA International Symposium on Hazardous Waste Treatment: Biosystems for
Pollution Control.
3. Metcalf and Eddy, Inc., 1979. Wastewater Engineering-Treatment, Disposal,
Reuse.
4. Vogel, T.M., and P.L. McCarty, 1985. "Biotransformation of Tetrachloro-
ethylene to Trichloroethylene, Dlchloroethylene, Vinyl Chloride, and Carbon
Dioxide Under Methanogenic Conditions." Appl. Environ. MIcroblol.. 49: 1080-
1083.
5. Bouwer, E.J., and P.L. McCarty, 1985. "Utilization Rates of Trace
Halogenated Organic Compounds in Acetate—Grown Biofilms," Biotech, and
BioenEineering. 27:1564-1571.
646
-------
6. Galli, R. , and P.L. McCarty, 1989. "Biotransformation of 1,1,1-
Trichloroethane, Trichloromethane, and Tetrachloromethane by a
Clostridium sp." Appl. Environ. Microbiol.. 55: 837-844.
1. Bouwer, E.J., and P.L. McCarty, 1983. "Transformations of 1- and 2-
Carbon Halogenated Aliphatic Organic Compounds Under Methanogenie
Conditions." Appl. Environ. Microbiol.. 45: 1286-1294.
8. Bouwer," E.J., B.E. Rittmann, and P.L. McCarty, 1981. "Anaerobic
Degradation of Halogenated 1— and 2-Carbon Organic Compounds."
Environ. Science and Tech.. 15: 596-599.
9. Bouwer, E.J., and P.L. McCarty, 1982. "Removal of Trace Organic
Compounds by Activated Carbon and Fixed-Film Bacteria." Environ.
Science and Tech.. 16: 836-843.
10. Vargas, C., and R.A. Ahlert, 1987. "Anaerobic Degradation of
Chlorinated Solvents." JWPCF. 59: 594-968.
11. Brunner, W., D. Staub, and T. Leisinger, 1980. Bacterial Degradation
of Dichloromethane.'' Appl. Environ. Microbiol.. 40: 950-958.
12. LaPat-Polasko, L.T., P.L. McCarty, andA.J. Zehnder, 1984. "Secondary
Substrate Utilization of Methylene Chloride by an Isolated Strain of
Pseudomonas sp." Appl. Environ. Microbiol.. 47: '825-830.
13. KohlerStaub, D. , et. al., 1986. "Evidence for Identical
Dichloromethane Dehalogenation in Different Methylotrophic Bacteria."
J. of General Microbiology. 132: 2837-2843.
14. Dagley, S., 1971. "Catabolism of Aromatic Compounds by
Microorganisms." Adv. Microb. Physiol.. 6: 1-46.
15. Gibson, D.T., J.R. Koch, and R.E. Kallio, 1968. "Oxidative Degradation
of Aromatic Hydrocarbons by Microorganisms. I. Enzymatic Formation of
Catechol from Benzene." Biochemistry. 7: 2653-2658.
16. Rochkind-Dubinsky, M.L., G.S. Sayler, and J.W. Blackburn, 1987.
"Microbial Decomposition of Chlorinated Aromatic Compounds." Marcel
Dekker. Inc.. New York.
17. Grbic-Galic, D., and T.M. Vogel, 1986. "Transformation of Toluene and
Benzene by Mixed Methanogenie Cultures." Appl.Environ. Microbiol..
53: 254-260.
18. Suidan, M.T., et al., 1981, "Continuous Bioregeneration of Granular
Activated Carbon During the Anaerobic Degradation of Catechol." Prog.
Water Technol.. 12: 203-214.
19. Suidan, M.T., et al., 1981. "Anaerobic Carbon Filters for Degradation
of Phenols." J. of EnvironmentalEngineering. 107: 563-579.
20. Khan, K.A., M.T. Suidan, and W.H. Cross, 1981. "Anaerobic Activated
Carbon Filter for the Treatment of Phenol—Bearing Wastewater," J.
Water Poll. Control Fed.. 53: 1519-1532.
21. Khan, K.A., et al., 1982. "Role of Surface Active Media in Anaerobic
Filters." J. of Environmental Engineering. 108: 269-285.
22. Wang, Y.T., M.T. Suidan, and J.T. Pfeffer, 1984. "Anaerobic Activated
Carbon Filter for the Degradation of Polycyclic N-Aromatic Compounds."
J. Water Poll. Control Fed.. 56: 1247-1253.
23. Suidan, M.T., et al., 1983. "Anaerobic Filter for the Treatment of
Coal Gasification Wastewater." Biotechnological Bioengineering. 25:
1581-1596.
24. Suidan, M.T., et al. , 1983. "Treatment of Coal Gasification Wastewater
with Anaerobic Filter Technology." J. WaterPoll.Control Fed.. 55:
1263-1270.
647
-------
25. Pfeffer, J.T., and M.T. Suidan, 1985. "Anaerobic-Aerobic Process for
Treating Coal Gasification Wastewater." Proceedings, Industrial Waste
Symposium, WPCF Annual Meeting, Kansas City, Mo,
26. Suidan, M.T., et al., 1987. "Anaerobic Treatment of Coal Gasification
Wastewater." WaterSci. Technol.. 19: 229-236.
27. Suidan, M.T., et al., 1987. "Anaerobic Uastewater Treatment." Final
Report to Department of Energy Project No. DOE DE AC21-84MC21281.
28. Suidan, M.T., et al., 1991. "Anaerobic Treatment of a High Strength
Industrial Wastes Bearing Inhibitory Concentrations of 1,1,1-
Trichloroethane." Water Sci. Technol.. 23: 1385-1393.
29. Narayanan, B., et al., "Anaerobic Treatment of VOC's in High Strength
Wastes." submitted for publication in J. Water Poll. ControlFed..
30. Speth, T.F., and R.J. Miltner, 1990. "Technical Note: Adsorption
Capacity of GAG for Synthetic Organics." J.AWWA. 82; 72-75.
648
-------
DEVELOPMENT OF NONLINEAR GROUP CONTRIBUTION METHOD FOR PREDICTION OF
BIODEGRADATION KINETICS FROM RESPIROMETRICALLY-DERIVED KINETIC DATA
by: Henry H. Tabak
Risk Reduction Engineering Laboratory
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
Rakesh Govind, Xuanqiang Yu, Le Lai, and Sanjay Desai
Dept. of Chemical Engineering
University of Cincinnati,
Cincinnati, OH 45221
ABSTRACT
The fate of organic chemicals in the environment depends on their
susceptibility to biodegradation. Hence, development of regulations
concerning their manufacture and use requires information on the extent and
rate of biodegradation. Recent studies have attempted to correlate the
kinetics of biodegradation with the compound's molecular structure. This has
led to the development of structure-biodegradation relationships (SBRs) using
the group contribution approach. Each defined group present in the compound's
chemical structure is assigned a unique numerical contribution towards the
calculation of the biodegradation kinetic constants. In this paper, a non-
linear group contribution method has been developed using neural networks,
which is trained using literature data on the first order biodegradation
kinetic rate constant for a number of priority pollutants. The trained neural
network is then used to predict the biodegradation kinetic constant for a new
list of compounds, and the results have been compared with the experimental
values and the predictions obtained from a linear group contribution method.
It has been shown that the non-linear group contribution method using neural
networks is able to provide a superior fit to the training set data and
produce a lower prediction error than the previous linear method.
649
-------
INTRODUCTION
The number and amount of synthetic organic chemicals produced
commercially is large and increasing every year. The presence of many of
these chemicals in the ecosystem is a serious public health problem.
Biodegradation is an important mechanism for removing these chemicals from
natural ecosystems. The high diversity of species and the metabolic
efficiency of microorganisms suggest that they play a major role in the
ultimate degradation of these chemicals (Alexander, 1980). Biodegradation can
eliminate hazardous chemicals by biotransforming them into innocuous forms, or
completely degrading them by mineralization to carbon dioxide and water.
Kinetic data for calculating biodegradation rates in natural ecosystems
are important for several reasons. The computer-based overall fate models,
used to estimate the distribution and the concentration of organic compounds
in the environment, require rate constants to determine the importance of
biodegradation against other competing removal processes, such as
volatilization and adsorption. Information regarding the extent and the rate
of biodegradation of organic chemicals is very important in evaluating
relative persistence of the chemical in the environment, and for regulating
their manufacture and use. Due to the large number of chemicals, obtaining
this information is labor intensive, time consuming and expensive. Thus,
there is a need to develop correlations and predictive techniques to assess
biodegradability (Strier, 1980).
Structure-activity relationships (SARs) are used to predict intrinsic
properties of many chemicals and to estimate the kinetic constants for
important transformation processes. SARs approach can be effectively used to
shorten the list of thousands of chemicals to a few hundred key chemicals, for
detailed laboratory and field testing. In a recent review of SARs,
Nirmalakhandan and Speece (1988) concluded that application of SARs has great
potential in predicting the fate of organic chemicals and these techniques are
being accepted to a greater extent by regulatory agencies in decision making
and policy implementation.
In this paper, a quantitative structure-biodegradation relationship
(SBR) has been developed using the group contribution approach. This method
is widely used in chemical engineering thermodynamics to estimate pure
compound properties such as liquid densities, heat capacities and critical
constants.
The group contribution method is similar to the Free-Wilson model widely
used in pharmacology and medicinal chemistry. Using this method, a very large
number of chemicals can be constituted from perhaps a few hundred functional
groups. Using this method, the compound's property is predicted from its
molecular structure, which is structurally decomposed into groups or
fragments, each group or fragment having a unique contribution towards the
specific value of the property.
650
-------
TECHNIQUES FOR MEASURING BIODE6RADATION KINETICS
Techniques for evaluating biodegradation kinetics have been reviewed in
detail by Howard et al. (1981) and subsequently discussed by Grady (1985). It
would be inappropriate to repeat the vast literature incorporated into these
two reviews. The essential techniques have been summarized in Table 1 to
present their salient features. In the following section, the electrolytic
respirometric method will be presented in some detail, since this method was
used in obtaining experimental data on biodegradation kinetics utilized in the
development of the non-linear model,
ELECTROLYTIC RESPIROMETRY STUDIES
This study was conducted using an automated continuous oxygen uptake and
BOD measuring Voith Sapromat B-12 (12 unit system). The instrument consists
of a temperature controlled waterbath, containing measuring units, an on-line
microcomputer for data sampling, and a cooling unit for continuous
recirculation of waterbath volume. Each measuring unit consists of a reaction
vessel, containing the microbial inoculum and test compound, an oxygen
generator, comprised of an electrolytic cell containing copper sulfate and
sulfuric acid solution, and a pressure indicator which triggers oxygen
generation. The carbon dioxide produced is absorbed by soda lime, contained
in the reaction flask stopper. Atmospheric pressure fluctuations do not
affect the results since the measuring unit forms an air sealed system. The
uptake of oxygen by the microorganisms in the sample during biodegradation is
compensated by the electrolytic generation of oxygen in the oxygen generator,
connected to the reaction vessel. The amount of oxygen supplied by the
electrolytic cell is proportional to its amperage requirements, which is
continuously monitored by the microcomputer and the digital recorder.
Measurement of oxygen consumption through electrolytic respirometry has
been shown to be very promising for automatic data collection associated with
biodegradation (Tabak et aJ, 1984, 1989).
MATERIALS AND METHODS
The nutrient solution used in our studies was an OECD synthetic medium
(1983) consisting of measured amounts per liter of deionized distilled water
of (1) mineral salts solution; (2) trace salts solution; and (3) a solution
(150 mg/1) of yeast extract as a substitute for vitamin solution.
The microbial inoculum was an activated sludge from The Little Miami
wastewater treatment plant in Cincinnati, Ohio, receiving municipal
wastewater. The activated sludge sample was aerated for 24 hours before use
to bring it to an endogenous phase. The sludge biomass was added to the
medium at a concentration of 30 mg/1 total solids. Total volume of the
synthetic medium was 250 ml in the 500 ml capacity reaction vessels.
The test and control compound concentration in the media were 100 mg/1.
Aniline was used as the biodegradable reference compound, at a concentration
of 100 mg/1.
651
-------
In a typical experimental run, duplicate flasks were used for the test
compound, and reference compound, aniline, a single flask for toxicity control
(test compound plus aniline at 100 mg/1 each) and an inoculum control.
The reaction vessels were incubated in the dark at 25° C in the
temperature controlled bath and stirred continuously throughout the run. The
microbiota of the activated sludge were not pre-acclimated to the substrate.
The incubation period of the experimental run was between 28-50 days. A more
comprehensive description of the procedural steps involved in the
respirometric tests has been presented elsewhere (Tabak et al. 1984, 1989).
EVALUATION OF BIODEGRADATION KINETICS
In this study, biodegradation was measured by measuring the ratio of the
measured biological oxygen demand (BOD) values in mg/1 (oxygen uptake values
of test compound minus endogenous oxygen uptake values [inoculum control]) to
the theoretical oxygen demand (ThOD) of substrate i.e., the ratio BOD/ThOD.
The values of theoretical oxygen demand (ThOD) were calculated by using the
stoichiometric balanced oxidation equation.
The BOD/ThOD curves calculated from electrolytic respirometric data were
characterized by four indices (Urano and Kato, 1986) shown in Figure 1: [1]
the lag time (t,) which gives the adaptation time; [2] the rate constant (k);
[3] the biodegradation time (t^) before the endogenous respiration period;
and [4] ratio of BOD/ThOD at time td. The values of k can be calculated from
the slope of the straight line obtained by plotting log(BOD) vs time (t) for
values of t such that t, < t < td. The appropriate equations for calculating
the value of k are given as follows:
d(BOD)/dt = k'(BOD)
log (BOD) = (k'/2.3)t = kt + constant; t,< t < td
It should be noted that the above kinetic model for obtaining
biodegradation kinetics differs from the traditional Monod equation which has
been used extensively in the literature to analyze oxygen uptake data.
However, the above model was selected for several reasons: [1] it has the
ability to represent oxygen uptake data between time t, and td by a single
parameter (k); [2] it follows the method of Urano and Kato model (1986), so
that some of their kinetic constant values could be used in our training set;
[3] the simple model provided an acceptable fit with the experimental data;
and [4] the model results allowed the development of the prediction approach
(linear and non-linear) for estimating the kinetic constant values for a
variety of test compounds. This allowed us to compare the linear and non-
linear approaches using the same data set.
The kinetic constant (k) values for the compounds were divided into two
sets: [1] the training set; and [2] a testing set. This division of the
compound list was based on the criterion that the chemical groups or fragments
comprising the testing set compounds were all present in the training set and
there were at least 5 compounds in the training set for each chemical group
selected.
652
-------
The main motivation for constructing the training and testing sets was
to develop structure-biodegradation relationships using the training set data
and then test the relationship using the testing set compounds. A brief
review of the literature on structure-biodegradation relationships has been
presented in the following section followed by the presentation of the group
contribution approach, which was used for analyzing the experimental kinetic
constant values.
STRUCTURE-ACTIVITY RELATIONSHIPS
Structure-activity relationships have been widely used in pharmacology
and medicinal chemistry. In the field of biodegradation, interest in
structure-activity relationships between the biodegradability of the chemical
and its structure started many years ago (Ludzack and Ettinger, 1960). There
are several studies which have attempted to correlate some physical, chemical
or structural property of a chemical with its biodegradation. Literature
reveals both qualitative and quantitative structure-biodegradability
correlations. Lyman et al. (1982) summarized rules of thumb which may be used
to make qualitative predictions of biodegradability. These rules are based on
degree of branching, chain length, oxidation and on number, type and position
of substituents on simple organic molecules. Geating (1981) developed a
predictive algorithm based on the literature published between 1974 and 1981.
Based on the type and the location of substituent groups, the model predicts
biodegradability in qualitative terms. The algorithm was applied to group of
compounds of known biodegradability and it predicted correctly for 93% of
compounds, incorrectly for 2% of compounds and 5% could not be predicted at
all. When applied to known nonbiodegradable compounds it was not as
successful, predicting 70% correctly and of the remainder roughly half were
predicted incorrectly while half could not be predicted at all. Others
(Rothkopf and Bartha, 1984, Yoshimura et al., 1980) have also investigated
qualitative relationships for certain class of chemicals; however while these
studies are useful, none provide the kind of prediction power needed for
regulatory decision making, for which quantification is necessary.
Quantitative correlations relating either biodegradation rate constant
or 5-day biological oxygen demand (BOD) with different physicochemical
properties or rate of other transformation processes have appeared in
literature. Most of these correlations are linear single parameter models.
Paris and co-workers (1980, 1982, 1983, 1984, 1987) were among the first to
investigate quantitative correlations using microbial transformation rate
coefficient. They used Monod equation with the assumption that the substrate
concentration is less than the half saturation constant so that the
transformation rate becomes first order with respect to both substrate
concentration as well as microbial concentration. They called this resulting
rate coefficient a second order rate coefficient. They investigated several
groups of chemicals : pesticides, substituted phthalates, mono-substituted
phenols, carboxylic acid esters of 2,4-D, ethyl esters of chlorine substituted
acetic acids and substituted anilines. These studies were conducted either
with pure culture or mixed populations of organisms from natural environment.
In almost all cases they obtained good correlations with a particular
property of the chemicals. In case of substituted anilines and mono-
653
-------
substituted phenols, transformation rate was related with the van der Waal's
radii of the substituent groups. The transformation rate in case of
phthalates and pesticides was correlated with their alkaline hydrolysis rate
constant. They attempted to correlate the rates of carboxylic acid esters of
2,4-D and ethyl esters of chlorine substituted acetic acids with the
lipophilicity, specifically octanol-water partition coefficients. They were
successful with the first group but not with the second.
Reinke and Knackmuss (1978) studied di-oxygenation of substituted
benzole acids by two species of Pseudomonas and were able to obtain a good
correlation of the rate coefficient with the Hammett constant for one species
but not for the other. Fitter (1984) obtained a linear relation between the
logarithm of the biological degradation rate of substituted phenols and
anilines and the Hammett constant of the substituent. Of all the substituents
(OH, CHS, Cl, N02 and NH2), only the amino group led to deviations from the
linear correlations for mono-substituted phenols. He also attempted
correlations using the steric and lipophilic constants but failed.
Vaishnav et al. (1987) correlated biodegradation of 17 alcohols and 11
ketones as well as a series of alicyclic chemicals with octanol-water
partition coefficient, log P. Alcohols revealed a biphasic relationship with
an apparent change in slope at a log P of about 3. The relationship for
ketones was parabolic or bilinear with a peak at a log P value of about 1.
Statistically the difference between the parabolic and bilinear relationships
was marginal, but the bilinear model gives a closer fit to experimental data,
Degradability of the hydrophilic members of alicyclics was apparently not
related to log P but degradability of the more hydrophobic members decreased
with increasing lipophilicity. Banerjee et al. (1984) studied biodegradation
of phenol, resorcinol, p-cresol, benzoic acid and various chloro derivatives
of phenol, resorcinol and anisole. The biodegradation rate was related to
lipophilicity, where the rate increased with decreasing lipophilicity and then
levelled off for chemicals with log P less than 2.
Deardan and Nicholson (1986) studied aromatic and aliphatic amines,
phenols, aromatic and aliphatic aldehydes, carboxylic acids, halogenated
hydrocarbons and amino acids. They calculated different parameters for each
compound ; molecular connectivities up to seventh order, log P values,
molecular volume, accessible molecular surface area, Sterimol steric
parameters and atomic charges. They correlated 5-day BOD of these compounds
with the atomic charge difference across the bond(s) common to all compounds
in the series. The regression coefficient and the constant term for each of
the series of compounds were close enough to combine all the data into a
single, all embracing equation covering amines, phenols, aldehydes, carboxylic
acids, halogenated hydrocarbons and amino acids.
Another approach 1s to seek direct correlation between biodegradation
and molecular structure of the chemical. The structural features of a
molecule such as shape, size, branching and nature of atom-atom connections
are expressed in terms of numerical descriptors called topological indexes.
Many such indexes have been proposed, but the most successful of them in SAR
are molecular connectivity indexes, which were introduced by Randic (1975) and
then developed extensively by Kier and Hall (1976). Govind (1987) has
654
-------
correlated first order biodegradation rate constants of priority pollutants
with the first order molecular connectivity index. Boethling (1986) has
correlated log rate constant for 2,4-D alkyl esters, log percent degraded for
carbamates, log percent theoretical oxygen demand for dialkyl ethers, rate
constant for dialkyl phthalates and percent theoretical oxygen demand for
aliphatic acids with molecular connectivity indexes. All these were single
variable models. Two variable models substantially improved results for
aliphatic alcohols and acids.
Most of the correlations that have appeared in literature are single
parameter relations applicable to a particular class of compounds. This
demonstrates that correlations are possible, but also that single parameter
correlations are limited in their applicability. Babeu and Vaishnav (1987)
calculated 5-day BOD for 45 organic chemicals including alcohols, acids,
esters, ketones and aromatics. The BOD data were correlated with water
solubilities, log P, molar refractivities and volumes, melting and boiling
points, number of carbon, hydrogen and oxygen atoms, molecular weight and
theoretical BOD of chemicals. The experimental BOD values for 43 additional
chemicals were compared with values predicted by the model and for 84%-88% of
the test chemicals prediction was within 80% of the experimental values.
Desai et al . (1990) have predicted first order biodegradation rate constants
within 20% of the experimental values using group contribution approach. The
model based on group contributions was a first order linear model which
neglected the interactions between groups. Table 2 summarizes the work done
in SARs for predicting the biodegradability of chemicals.
GROUP CONTRIBUTION APPROACH
Using a group contribution approach, a very large number of chemicals of
interest can be constituted from perhaps a few hundred functional groups. The
prediction of the property is based on the structure of the compound.
According to this method, the molecules of a compound are structurally
decomposed into functional groups or their fragments, each having a unique
contribution towards the compound property. The advantage of this approach is
that the molecules of the compounds may be structurally dissected in any
convenient manner and no independently measured group constants are required
in the analysis.
The biodegradability rate constant, k, can be expressed as a function of
contribution a, of each group or fragment of the compound
Ln(k) =
In general, the above functional relationship can be classified into two
types: [1] linear function; and [2] non-linear function.
LINEAR GROUP CONTRIBUTION METHOD
The above general function can be expanded in terms of Taylor series.
If the terms from second order onwards are neglected, a linear first order
655
-------
model for biodegradation rate constant k (1/hr) is obtained and this can be
expressed as:
L
Ln (k) - I N,a,
j-1
where Nj is the number of groups of type j in the compound, aij is the
contribution of group of type j and L is the total number of groups in the
compound. For each compound a linear equation in a's is constructed. This
generates a series of linear equations for a given data set which are solved
for a's, using the method of least squares.
The above model, being first order approximation, will break down if
interaction between groups become important. The interaction of different
groups can be treated by considering second and higher order terms of the
series.
Data generated by Urano and Kato (1986) using electrolytic respirometry
was used in applying the linear group contribution method. This ensured that
the test conditions for obtaining the data were the same for all the
compounds. The experimental conditions used by Urano and Kato (1986) were:
temperature 20° C, pH of solution 7, sludge concentration 30 mg/1 and compound
concentration 100 mg/1.
Urano and Kato (1986) had obtained data [kinetic constant (k) values]
for 74 compounds in their study. However, in our analysis using the linear
group contribution method, it was necessary to ensure that each group,
considered in the analysis, occurred in at least 5 compounds in the training
set. This requirement prevented us from using the entire set of compounds
studied by Urano and Kato (1986), and only 18 compounds were used for
calculating the group contribution values for 8 groups. The compounds used in
our analysis (training set) were: ethyl alcohol, butyl alcohol, ethylene
glycol, acetic acid, propionic acid, n-butyric acid, n-valeric acid, adipic
acid, methyl ethyl ketone, hexamethylenediamine, n-hexylamine, mono-ethanol-
amine, acetamide, benzene, benzyl alcohol, toluene, acetophenone, and
aminophenol. The experimental values of the kinetic constants for these
chemicals have been tabulated in Table 3.
The group contribution parameters for all the groups considered in the
analysis are given in Table 4, which are modified since previous work (Desai
et al. 1990). Note these contribution values are used for calculating the
value of ln(k) rather than the kinetic constant (k) Itself.
To validate the results, experiments were conducted by the authors using
an electrolytic respirometer (Voith-Morden, Milwaukee, WI) for cresols,
phenol, 2,4-dimethyl phenol and butyl benzene. These compounds, obtained from
Aldrich chemical company, were of 99+% purity. Except for the source and
nature of biomass, the experimental conditions were the same as used by Urano
and Kato (1986).
656
-------
The oxygen uptake data was analyzed using the kinetic model presented
earlier, and the best fit kinetic constant (k) value was obtained for each
compound in the testing set. The experimental k values obtained for the test
compounds are given in Table 3.
NON-LINEAR GROUP CONTRIBUTION OR NEURAL NETWORK METHOD
To incorporate the effects of interactions between the chemical groups
used in the group contribution approach, it was necessary to develop a non-
linear method. It was important that the non-linear method included not only
all the interactions between the groups but also the algebraic form (type of
non-linearity, such as square, cubic, etc.) of such interactions. Since this
information was not known a priori, a neural network model was used to include
all possible interactions between the groups and the algebraic form of these
interactions was implicitly determined from the training set data using a
large number of adaptable parameters or network weights.
NEURAL NETWORK MODELS
There is extensive literature on mathematical models of artificial
neural networks, beginning with the work of McCulloch and Pitt (1943), Hebb
(1949), Rosenblatt (1959), Widrow (1960) and Posch (1968). Recent work by
Hopfield (1982, 1984, 1986), Rumelhart, et al (1986), Sejnowski and Rosenberg
(1986), Feldman and Ballard (1982), and Grossberg (1986) has revived interest
in the field of artificial neural nets. Neural networks have found
applications in image and speech recognition, on-line diagnosis of process
faults, process control and in optimization of complex functions.
Artificial neural network models consist of many nonlinear computational
elements operating in parallel and arranged in patterns similar to biological
neural nets. The nodes or computational elements are connected via weights
that are typically adapted during use to improve performance (Lippmann, 1987).
Superior performance is achieved via dense interconnection of simple
computational elements.
Computational elements or nodes used in neural net models are nonlinear.
In our model, each node, shown schematically in Figure 2, has a large number
of inputs and a single output. Each input value has an associated activation
and weight. Each node or computational element applies an activation function
to the sum of the products of the input activations and weights, and
thereby generates the output value. The output of each computational element
or node can be expressed as follows:
Opj = 1/[1 + exp[-(2 WJfOp, +
where Opj = output value of node j
0 - = output value of node i
Wj,- = connection weight between the ith and jth nodes
0j = bias of the jth node
Hence, each node or computational element forms a weighted sum of N inputs and
passes the result through a nonlinearity, mathematically expressed by the
657
-------
above equation. Figure 3 is a plot of this sigmoidal nonlinearity, inherent
in each node. More complex neural network models may include temporal
integration or other types of time dependencies and more complex mathematical
operations.
The neural network consists of interconnected nodes or computational
elements. In our study, a three layer neural network was used, with eight
input nodes, eight hidden or intermediate layer nodes, and one output node, as
shown in Figure 4. It consists of a hidden or intermediate layer between the
input and output nodes. Multi -layer neural networks overcome some of the
limitations of the single layer models. The capabilities of multi-layer
neural networks stem from the nonlinearities used within nodes, which allow
arbitrarily complex decision regions.
The neural network was trained by using a back-propagation algorithm,
which used a gradient search technique to minimize a cost function equal to
the mean square difference between the desired and the actual net outputs.
The desired output of all nodes is typically "low" (0 or < 0.1) unless that
node corresponds to the class the current input is from, in which case its
output is "high" (1.0 or > 0.9). The net is trained by initially selecting
small random weights and internal thresholds and then presenting the training
data (number of each type of chemical groups present and the experimentally
measured biodegradation kinetic rate constant) for each chemical repeatedly.
Weights are adjusted after each trial, until the weights converge and the cost
function is reduced to an acceptable value.
The output of the network depends on the weights assigned to each
connection between the layers. Training of the network corresponds to the
assignment of weights, which are determined in the back-propagation algorithm
by minimizing the error function, Ep, written as follows
Ep - 1/2 2 (tpj - Opj)2
where tpj- and Opj- are the desired and actual activation values of the output
node j, due to an input pattern p. The generalized back propagation algorithm
was used by Rumelhart et al (1986) to obtain the minimum value of Ep. In the
first step an input pattern is presented and propagated forward through the
network to compute the output value Opj for each node. The output is compared
with the desired value, resulting in an error 5 j for each output. For nodes
in the output layer, the error signal is given by
«PJ - (^ - Opj^Pjd-Qpj)
The error signal for a node in the intermediate layer is given by
5PJ - Opjd-Opj) SVw
In the second step, a backward pass is made through the network and the
error signal is passed to each node in the network and the appropriate weight
changes are made using the following equation
Wfj(t+l) » Wy(t) + B (Wfj(t) - Hf
658
-------
where 8 is the learning rate. In our study, a relatively fast learning rate (
6 ) of 0.5 was chosen.
In our study, there were eight input nodes, each node corresponding to a
specific chemical group, previously used in the linear group contribution
analysis, and listed in Table 4. The inputs to the neural network corresponds
to the number of each type of chemical group present in the chemical
structure. For example, in the case of ethyl alcohol, there are one methyl
group, one methylene group, and one hydroxy group. The number of each group
is entered correponding to the node representing the group. The output is the
biodegradation kinetic constant value. The data used to obtain the group
contributions in the linear group contribution analysis were used for training
the neural network.
RESULTS AND DISCUSSION
For the training set compounds, listed in Table 3, Table 5 presents the
results of the linear group contribution analysis and the nonlinear group
contribution or neural network approach. It shows the goodness of fit between
the experimental values and the computed values. It can be seen that the mean
error in the linear group contribution method is generally larger than the
neural network method. Figure 5 provides a comparison between the
experimental and predicted (calculated) biodegradation kinetic data by neural
network and linear group contribution.
Table 6 lists the results for the testing set compounds. This shows the
ability of the method to predict the kinetic constant value for compounds that
were not in the training set.
Both the linear group contribution and the neural network methods are
unable to distinguish between the ortho-, meta- and para-cresols since the
position of the hydroxy group was not considered in the analysis. The error
between the predicted and experimental values are generally lower by the
neural network method when compared with the linear group contribution method.
However, further testing of these methods is needed before any concrete
conclusions regrading these methods can be drawn.
REFERENCES
1. Alexander, M. 1980. Biodegradation of chemicals of environmental
concern. Science. 211, 132-138.
2. Babeu, L., and Vaishnav, D. D. 1987. Prediction of biodegradability
for selected organic chemicals. J. Industrial Hicrobiol.. 2, 107-115.
3. Banerjee, S., Howard, P. H., Rosenburg, A. M,, Dombrowski, A. E., Sikka,
H. and Tullis, D. L. 1984. Development of a general kinetic model for
biodegradation and its application to chlorophenols and related
compounds. Environ. Sci... Tech..... 18, 416-422.
659
-------
4. Boethling, R. S. 1986. Application of molecular topology to
quantitative structure-biodegradability relationships. Environ.
Toxicol. Chem.. 5, 797-806.
5. Chltra, S. P. and Govind, R. 1986. Application of a group contribution
method for predicting adsorbability on activated carbon. AICh.E.J..« 32,
167-169.
6. Deardan, J. C. and Nicholson, R. M. 1986. The prediction of
biodegradability by the use of quantitative structure-activity
relationships : Correlation of biological oxygen demand with atomic
charge difference. Pestici. Sci.. 17, 305-310.
7. Desai, S., Govind, R. and Tabak, H. H. 1990. Development of
quantitative structure-activity relationships for predicting
biodegradation kinetics. Environ. Toxicol. Chem. (accepted)
8. Feldman, J.A. and Ballard, D.H. 1982. Connectionist Models and their
Properties, Coqnitive Science, 6, 205-254.
9. Geating, J. 1981. Literature study of the biodegradability of
chemicals in water. U.S. Environmental Protection Agency, Report No.
EPA-600/2-81-175, Cincinnati, OH.
10. Govind, R. 1987. Treatability of toxics in wastewater systems.
Hazardous Substances. 2, 16-24.
11. Grady, C. P. L., Jr. 1985. Biodegradation: Its measurement and
microbiological basis. Biotech. Bioenqq.. 27, 660-674.
12. Grossberg, S. 1986. The Adaptive Brain I: Cognition, Learning,
Reinforcement, and ;Rhythm, and The Adaptive Brain II: Vision, Speech,
Language, and Motor Control, Elsevier/North Holland, Amsterdam.
13. Hebb, D.O. 1949. The Organization of Behavior, John Wiley & Sons, New
York.
14. Hopfield, J.J. 1982. Neural Networks and Physical Systems with
Emergent Collective Computational Abilities, Proc. Natl. Acad. Sci. USA,
79, 2554-2558.
15. Hopfield, J.J. 1984. Neurons with Graded Response Have Collective
Computational Properties Like Those of Two-State Neurons, Proc. Nat!.
Acad. Sci. USA, 81, 3088-3092.
16. Hopfield, J.J. and Tank, D.W. 1986. Computing with Neural Circuits: A
Model, Science. 233, 625-633.
17. Howard, P. H., Banerjee, S. and Rosenburg, A. 1981. Review and
evaluation of available techniques for determining persistence and rates
of degradation of chemical substances in environment. U.S.
Environmental Protection Agency, Report No. 560/5-81-011.
660
-------
18. Kier, L. B. and Hall, L. H. 1976. Molecular connectivity in chemistry
and drug research. Academic Press, New York, NY.
19. Lippmann, R.P. 1987. An Introduction to Computing with Neural Nets,
IEEE ASSP Magazine. 4, 4-22,
20. Ludzack, F. J. and Ettinger, M. B. 1960. Chemical structures resistant
to aerobic biochemical stabilization. J. Wat.Pollut. Control Fed.. 32,
1173-1200.
21. Lyman, W. J., Reehl, W. F. and Rosenblatt, D. H. 1982. Handbook of
chemical property estimation methods : Environmental behavior of organic
compounds, McGraw Hill Book Company, New York, NY.
22. McCulloch, W.S. and Pitts, W. 1943. A Logical Calculus of the Ideas
Imminent in Nervous Activity. Bulletin of Mathematical Biophysics. 5,
115-133.
23. Nirmalakhandan, N. and Speece, R. E. 1988. Structure-activity
relationships. Environ. Sci. Tech., 22, 606-615.
24. OECD. 1983. OECD guidelines for testing of chemicals. EEC Directive
79/831, Annex V, Part C : Methods for Determination of Ecotoxicity, 5.2
Degradation, Biotic Degradation, Manometric Respirometry, Method DGX1,
Revision 5. pp. 1-22.
25. Paris, D. F., Wolfe, N. L. and Steen, W. C. 1982. Structure-activity
relationships in microbial transformation of phenols. Appl. Environ.
Microbiol.. 44, 153-158.
26. Paris, D. F., Wolfe, N. L., Steen, W. C. and Baughman, G. L. 1983.
Effect of phenol molecular structure on bacterial transformation rate
constants in pond and river samples. Appl. Environ. Hicrobiol.. 45,
1153-1155.
27. Paris, D. F., Wolfe, N. L. and Steen, W. C. 1984. Microbial
transformation of esters of chlorinated carboxylic acids. Appl.
Environ. Microbiol., 47, 7-11.
28. Paris, D. F. and Wolfe, N. L. 1987. Relationship between properties of
a series of anilines and their transformation by bacteria. Appl.
Environ. Microbiol., 53, 911-916.
29. Pitter, P. 1984. Correlation between the structure of aromatic
compounds and the rate of their biological degradation. Collection
Czechoslovak Chem. Commum., 49, 2891-2896.
30. Posch, T.E. 1968. Models of the Generation and Processing of Signals
by Nerve Cells: A Categorically Indexed Abridged Bibliography, USCEE
Report 290.
661
-------
31. Randic, M. 1975. On characterization of molecular branching. J.Am.
Chem. Soc.. 97, 6609-6615.
32. Reinke, W. and Knackmuss, H. J. 1978. Chemical structures and
biodegradability of halogenated aromatic compounds: Substituent effects
on 1,2-dioxygenation of benzoic acid. Biochimica et Biophvsica Acta.
542, 412-423.
33. Rosenblatt, R. 1959. Principles of Neurodynamics, New York, Spartan
Books.
34. Rothkopf, 6. S, and Bartha, R. 1984. Structure-biodegradability
correlation among xenobiotic industrial amines. Am. Oil Chemists Soc.
jk, 61, 977-980.
35. Rumelhart, D.E., Hinton, 6.E. and Williams, R.J. 1986. Learning
Internal Representations by Error Propagation, in D.E. Rumelhart and
J.J. McClelland (Eds.) Parallel Distributed Processing: Explorations in
the Microstructure of Cognition. Vol. 1: Foundations, MIT Press.
36. Sejnowski, T. and Rosenberg, C.R. 1986. NETtalk: A Parallel Network
That Learns to Read Aloud, John Hopkins Univ. Technical Report JHU/EECS-
86/01.
37. Strier, M. P. (1980) Pollutant treatability : A molecular engineering
approach. Environ. Sci. Tech., 14, 28-31.
38. Tabak, H. H., Lewis, R. F. and Oshima, A. 1984. Electrolytic
respirometry biodegradation studies, OECD ring test of respiration
method for determination of biodegradability. Draft manuscript, MERL,
U.S. Environmental Protection Agency, Cincinnati, OH.
39. Tabak, H. H., Desai, S. and Govind, R. 1989. The determination of
biodegradability and biodegradation kinetics of organic pollutant
compounds with the use of respirometry. In Proceedings of EPA 15th
Annual Research Symposium : Remedial Action, Treatment and Disposal of
Hazardous Waste.
40. Urano, K. and Kato, Z. 1986. Evaluation of biodegradation ranks of
priority organic compounds. J. Hazardous Mtl., 13, 147-159.
41. Vaishnav, D. D., Boethling, R. S. and Babeu, L. 1987. Quantitative
structure-biodegradability relationships for alcohols, ketones and
all cyclic compounds. Chemosphere. 16, 695-703.
42. Widrow, B. and Hoff, M.E. 1960. Adaptive Switching Circuits, 1960 IRE
WESCON Conv. Record. Part 4, 96-104.
43. Wolfe, N. L., Paris, D. F., Steen, W. C, and Baughman, 6. L. 1980.
Correlation of microbial degradation rates with chemical structure.
Environ. Sci. Tech.. 14, 1143-1144.
44. Yoshimura, K., Machida, S. and Masuda, F. 1980. Biodegradation of long
chain alky! amines. Am. Oil Chem. Soc. J.. 57, 238-241.
662
-------
Table 1: Summary of Techniques for Evaluating Biodegradation Kinetics
Type of Reactor
Comments
Continuous
Batch
Fed-Batch
Requires an acclimated biomass
Time consuming to operate
Several steady-state conditions have to be
investigated
Requires several samples
Provides data only in the non-inhibitory
range for inhibitory substrates
Can be operated in several configurations
to simulate real plants
Can be used with either acclimated or
unacclimated biomass
Can be used for determining kinetic
parameters
Requires acclimated biomass
Requires specific assay for test compound
Provides only relative value for kinetic
parameter
663
-------
Table 2. Background Summary for SAR
Author(s)
QUALITATIVE
Lyman, et al .
Geating
QUANTITATIVE
Reineke and
Knackmuss
Wolfe, et al.
Paris, et al
Banerjee et al
Paris, et al
Pitter
Boethling
Deardan and
Nicholson
Babeu and
Vaishnav
Year
1982
1981
1978
1980
1982,
1983
1984
1984
1984
1986
1986
1987
Parameters
-
—
Hammett constant
of substituent
Alkaline hydrolysis
rate constant
van der Waal 's
radii of substituent
Hammett constant of
. substituent
octanol -water parti-
tion coefficient
Hammett constant of
substituent
molecular
connectivity
atomic charge
difference
theoretical BOD of
compounds, melting
Comments/Compounds Studied
summarized rules of thumb
developed qualitative
predictive algorithm
substituted benzoic acids
phthalate esters and
pesticides
substituted phenols
phenols and their chloro
derivatives
esters of chlorinated
carboxylic acids
substituted phenols and
anilines
2,4-D alky! esters,
carbamates, alky! ethers,
dialkyl phthalates and
aliphatic acids
amines, phenols, aide-,
hydes, carboxylic acids,
halogenated hydrocarbons
and amino acids
alcohols, acids, esters,
ketones and aromatics point
Govind
1987 molecular
connectivity
and number of Multipara-
meter model carbon atoms
priority pollutants
664
-------
Table 2 continued..
Author(s)
Year
Parameters
Comments/Compounds studied
Paris and
Wolfe
Vaishnav, et al
Desai, et al
1987
1987
1990
van der Waal's radii
of substituent
octanol -water parti-
tion coefficient
group contribution
substituted anilies
alcohols, ketones
alicyclics
various priority
and
parameters
pollutants
TABLE 3. Training and Testing Dataset
Compound
Ln(k)
Average Ln(k)
Training Set
Ethyl alcohol
Butyl alcohol
Ethyl ene glycol
Acetic acid
Propionic acid
n-Butyric acid
n-Valeric acid
Adipic acid
Methyl ethyl ketone
Hexamethylenediamine
n-Hexylamine
Mono-ethanolamine
Acetamide
Benzene
Benzyl alcohol
Toluene
Acetophenone
Aminophenol
Test Set
o-Cresol
m-Cresol
p-Cresol
Phenol
2,4-Dimethyl phenol
Butyl benzene
-2.90 -
-3.08 t-
-3.35 -
-2.60 -
-2.67 -
-2.70 -
-2.63 -
-2.81 -
-3.47 -
-4.34 -
-2.86 -
-3.32 -
-3.00 -
-2.86 ~
-2.78 ~
-2.60 -
-3.17 -
-3.24 ~
-2.57 -
-2.23 ~
-2.34 ~
-2.83 -
-2.63 -
-2.99 -
-3.15
-3.32
-3.65
-2.72
-2.97
-3.06
-2.66
-3.12
-3.69
-4.51
-3.06
-3.38
-3.06
-2.98
-3.17
-2.86
-3154
-3.30
-2.81
-2.51
-2.60
-3.17
-3.07
-3.27
-3.02
-3.19
-3.49
-2.66
-2.81
-2.87
-2.65
-2.96
-3.58
-4.43
-2.96
-3.35
-3.03
-2.92
-2.96
-2.73
-3.34
-3.27
-2.69
-2.37
-2.47
-3.00
-2.85
-3.13
665
-------
TABLE 4. Groups and Their Contribution Values
Group
Methyl
Methyl ene
Hydroxy
Acid
Ketone
Amine
Aromatic CH
Aromatic carbon
CH3
CH2
OH
COOH
CO
NH2
ACH
AC
ai
-1.28
-0.12
-1,54
-1.24
-0.59
-1.63
-0.48
0.93
666
-------
Table 5. Comparison of Experimental and Predicted 1n(k) for Training Set
Compound
Ethyl alcohol
Butyl alcohol
Ethyl ene glycol
Acetic acid
Propionic acid
n-Butyric acid
n-Valeric acid
Adipic acid
Methyl ethyl ketone
Hexamethylene diamine
n-Hexylamine
Mono-ethanolamine
Acetamide
Benzene
Benzyl alcohol
Toluene
Acetophenone
Aminophenol
-ln{k) -ln(k) %Error
Experimental Neural Network
3.02
3.19
3,49
2.66
2.81
2.87
2.65
2.96
3.58
4.43
2.96
3.35
3.03
2.92
2.96
2.73
3.34
3.27
3.01
3.16
3.45
2.68
2.81
2.83
2.70
2.93
3.63
4.22
2.97
3.38
3.01
2.94
2.94
2.70
3.31
3.29
0.33
0.94
1.15
0.75
0.06
1.39
1.89
1.01
1.40
4.74
0.33
0.90
0.66
0.68
0.68
1.10
0.90
0.61
-ln(k)' %Error
Linear Method
2.97
3.24
3.39
2.49
2.65
2.75
2.88
2.94
3.31
3.96
3.52
3.41
3.48
2.87
3.12
2.70
3.33
3.13
1.43
1.30
2.87
6.55
5.84
4.17
8.86
0.55
11.90
10.43
19.11
1.80
15.19
1.62
5.57
1.10
0.38
4.26
667
-------
Table 6. Comparison of Experimental and Predicted ln(k) for Testing Set
Compound
%Error
Experimental Neural Network
%Error
Linear Method
o-Cresol
m-Cresol
p-Cresol
Phenol
2,4-Dimethyl phenol
Butyl benzene
2,69
2.37
2.47
3.00
2.85
3.13
2.62
2.62
2.62
2.91
2.58
3.18
2.02
10.74
6.46
3.17
9.29
1.53
2.87
2.87
2.87
2.99
2.74
3.10
6.59
20.92
16.24
0.29
3.82
5.84
Key words - Structure-activity relationships, Biodegradation kinetics,
Respirometry, Group contribution method. Neural networks
668
-------
350
2 4 S • 10
Time, Days
1 -- Inoculum 2 - Phenol + Aniline
Figure 1. Oxygen Uptake of Phenol
12 14
669
-------
output j
input i
Figure 2. Processing element
1.0-
Ui
o.o
Figure 3. Sigmoid logistic function
670
-------
Ln(k)
t
o
Output layer
oooooooo
Hidden layer
OOOOOOOO Inputlayer
t t t t t t t t
CH3 CH2 CO NH2 COOH OH AC ACH
Figure 4. Neural network architecture used to evaluate -Ln(k)
4.5
4.0
asn
£
ra
3.0-
2.5
2.5
3.0
3.5
Experimental -Ln(k)
4.0
Neural network
Linear group
contribution
4.5
Figure 5. Comparison of experimental -Ln(k) by neural network and
linear group contribution method
671
-------
THE EFFECT OF CHLORINE ON NfX EMISSIONS FROM THE INCINERATION OF
NITROGEN-CONTAINING WASTES
by: William P. Linak
Air and Energy Engineering Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
Ravi K. Srivastava
Acurex Corporation
Research Triangle Park, NC 27709
Jost O.L. Wendt
Department of Chemical Engineering
University of Arizona
Tucson, AZ 85721
ABSTRACT
Combustion of organo-chlorides, chlorinated salts, and chlorinated acids
are unique to incineration systems. Many of these chlorinated species are
processed in hazardous waste incinerators as principal organic hazardous
constituents (POHCs), or in aqueous solutions containing other POHCs. In
municipal waste and hospital waste incinerators, chlorinated plastics
constitute sizeable fractions of the total waste streams. Numerous
examinations have been conducted in recent years that have sought to
characterize the incineration of chlorinated wastes as well as to determine
possible chlorinated products of incomplete combustion (PICs). However,
although the effect of halogens on flame chemistry has been studied
previously, little is known regarding chlorine's effects on other pollutants
such as NOX from fuel-bound nitrogen.
Experiments are being conducted on a 83 kW (280,000 Btu/hr) tunnel
combustor to quantify and understand any effect of fuel chlorine content on
NOX formation. The bench-scale combustor is a horizontal refractory-lined
cylinder with a quartz window for flame visualization, and multiple ports for
in-flame and post-flame sampling. A movable-block International Flame
Research Foundation (IFRF) type variable swirl burner with interchangeable
gaseous and liquid fuel nozzles is capable of providing near-burner zone
aerodynamic simulation of various flame types.
It is hypothesized that chlorine may act to inhibit NO formation through
its interaction with free radical species. Preliminary results co-firing
aqueous solutions of NH4OH (2.5% fuel nitrogen) and HC1 with natural gas
indicate a 30% reduction in NO emissions as fuel chlorine concentrations are
increased from 0 to 2.5%. Additional tests are currently being conducted to
examine pyridine (CsHsN) and chlorinated alkane (CCl,j and CaHivCl) combustion
in a No. 2 fuel oil flame. Other system parameters (furnace load, preheat air
temperature, stoichiometric ratio, and flame type) as well as air staging for
NOX control are also being examined. These results, and their interpretation
through kinetic modeling will be presented.
672
-------
PILOT-SCALE EVALUATION OF AN INCINERABILITY RANKING
SYSTEM FOR HAZARDOUS ORGANIC COMPOUNDS
by; Gregory J. Carroll
U.S. Environmental Protection Agency
Risk Reduction Engineering Laboratory
Cincinnati, Ohio 45268
W. Eddie Whitworth, Johannes W. Lee, and Larry R. Waterland
Acurex Corporation
Jefferson, Arkansas 72079
ABSTRACT
The U.S. EPA's hazardous waste incinerator performance standards specify
a minimum destruction and removal efficiency (ORE) for principal organic
hazardous constituents (POHCs) in the incinerator waste feed. In the past,
selection of appropriate POHCs for incinerator trial burns has been based
largely on their heats of combustion. Attempting to improve the system by
which trial burn POHCs are selected, the University of Dayton Research
Institute, under contract to EPA, has developed a thermal-stabi1ity-based
ranking of compound "incinerability". The subject study was conducted to
evaluate the laboratory-developed ranking system in a pilot-scale incinerator.
A mixture of 12 POHCs, spanning the ranking scale from most- to least-
difficult to destroy (Class 1 to Class 7, respectively), was combined with a
clay-based sorbent matrix, and fed into the rotary kiln incineration system at
the U.S. EPA Incineration Research Facility. In a series of 5 tests, the
following conditions were evaluated: nominal/typical operation; thermal
failure (quenching); mixing failure (overcharging); matrix failure (low feed
H/C1 ratio); and a worst-case combination of the 3 failure modes.
Under nominal conditions, mixing failure, and matrix failure, kiln exit
DREs for each compound were comparable from test to test. Operating
conditions in these 3 modes appeared to be sufficient to effect considerable
destruction (>99.99% ORE) of all 12 compounds. As a result, separation of the
highest ranked POHCs from the lowest ranked POHCs according to observed ORE
was not possible; a correlation between POHC ranking and DRE could not be
identified.
The correlation between predicted and observed incinerability was
stronger for the thermal failure and worst-case conditions. Kiln exit DREs
for the four most difficult-to-destroy POHCs (those in Classes 1 and 2) ranged
from 99% to 99.99% under these conditions, and were generally lower than DREs
for the higher class POHCs.
673
-------
U.S. EPA INCINERATION RESEARCH FACILITY UPDATE
by: J, W. Lee and D. J. Founder, Jr.
Acurex Corporation
Jefferson, Arkansas 72079
Robert C Thurnau
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
ABSTRACT
In FY'89, the physical plant of the EPA's Incineration Research Facility (IRF) in
Jefferson, Arkansas, underwent a major expansion. The building was enlarged and many
components of the rotary kiln incinerator system (RKS) were upgraded. In FY'90, the
IRF's RCRA permit was modified and reissued to include the expanded facility. The IRF's
capabilities were further upgraded to meet the increasing demand for incineration research
at the Facility. These improvements included installing an automatic process control
system; installing a new kiln ash removal system; installing a scrubber liquor heat exchanger;
and reconfiguring the venturi/packed-colurnn scrubber. Together with the single-stage
ionizing wet scrubber, the RKS now has the flexibility of two air pollution control systems
(APCSs) installed in parallel. Efforts are ongoing to upgrade the continuous emission
monitor sampling system, the kiln drive system, and the waste feed system. Onsite
laboratory capabilities were also expanded with the addition of an ion chromatograph.
With the completion of the facility upgrade, the pace of testing at the IRF has
accelerated. Completed tests include a program to evaluate the principal organic hazardous
constituent (POHC) incinerability ranking developed by researchers at the University of
Dayton Research Institute; low temperature desorption treatability testing with a soil from
a Region II Superfund site; an extensive incineration treatability test series with
contaminated soils from a Region III Superfund site; and a test program to evaluate the
applicability of incineration as best demonstrated available technology (BDAT) for treating
listed waste K088, spent potliner from the production of aluminum.
Planned programs include further testing to support EPA Regional Office Superfund
Site remediation efforts; a third series of trace metals tests using a Calvert Scrubber as the
APCS; further testing to evaluate the applicability of the POHC incinerability ranking; and
potential third party testing as provided for by the Federal Technology Transfer Act.
674
-------
INDUSTRY POLLUTION PREVENTION GUIDES
by: Teresa M. Marten
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
ABSTRACT
The Pollution Prevention Research Branch of the U.S. EPA's Risk
Reduction Engineering Laboratory is publishing a series of industry-
specific pollution prevention guides based on audit reports compiled for
the State of California, Department of Health Services. During 1990 EPA^
published seven guides covering the following industries and commercial
sectors: paint manufacturing, pesticide formulating, commercial printing,
fabricated metal, selected hospital waste streams, research and education
institutions, and printed circuit board manufacturing. Scheduled for
publication in early 1991 are guides for an additional six industrial
categories: photoprocessors, marine maintenance and repair,
pharmaceutical preparation, auto body refinishing, automotive repair, and
fiberglass reinforced and composite plastics manufacturing. The presenter
will discuss how the guides were developed, outline the waste minimization
assessment procedure for selected industry categories, and summarize key
findings from the 1991 guides.
675
-------
INORGANIC RECYCLING/DELCO REGION V HAZARDOUS WASTE RECYCLING DETERMINATION
by: Brian A. Westfall
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
ABSTRACT
Inorganic Recycling, Inc. (IR) proposed the use of a thermal process to
produce ceramic materials from industrial wastewater treatment sludges. In
conjunction with Delco Products in Kettering, Ohio, equipment was developed to
make abrasive material from a chromium-bearing electroplating sludge (RCRA
waste code F006). IR and Delco Products believed that the new process
represented recycling rather than treatment of hazardous waste, and thus was
exempt from RCRA permitting requirements.
USEPA's Region V and the Ohio EPA informed IR that the company needed to
substantiate the claim that their process constitutes recycling. Guidance was
provided from USEPA's Office of Solid Waste on criteria for legitimate
recycling of hazardous waste, and the Risk Reduction Engineering Laboratory
provided technical assistance in developing and monitoring a series of tests
for the process. Upon completion of the tests, Region V and Ohio EPA
determined that the system meets the criteria for recycling hazardous waste
under RCRA regulations.
676
-------
NEW JERSEY/EPA WASTE MINIMIZATION ASSESSMENTS
by: Mary Ann Curran
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
ABSTRACT
The use of pollution prevention concepts as an approach to
reducing the quantities of hazardous and non-hazardous wastes
which would otherwise require costly treatment or disposal is
receiving increased attention by industry, the public and
regulatory agencies. A cooperative effort between the EPA's Risk
Reduction Engineering Laboratory (RREL), the New Jersey
Department of Environmental Protection (NJDEP), and the New
Jersey Institute of Technology (NJIT) has been undertaken to
encourage the application of waste minimization approaches and
techniques at various facilities within the State of New Jersey.
The basis for the effort is the EPA's WasteMinimization
Opportunity Assessment Manual (625/7-88/003) which documents a
step-by-step approach to conducting assessments and leads to
identifying opportunities for improvement. NJIT is applying the
Manual at volunteer facilities and assisting them in conducting
the assessments in a self-audit approach. Assessments are being
conducted at thirty facilities covering ten industry categories.
This poster will present the findings of the first five
assessments which have been completed by NJIT. The five
assessments include 1) a printing operation, 2) a transportation
maintenance facility, 3) a finished leather manufacturer, 4) a
nuclear powered electrical generating station, and 5) a school
district (K-12).
677
-------
ASBESTOS CONTROL IN BUILDINGS
Thomas J. Powers
Risk Reduction Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
ABSTRACT
Asbestos Containing Material (ACM) has been used extensively in schools
and other buildings. The concern about exposure to asbestos in buildings is
based on evidence linking various respiratory diseases with occupational
exposure in the shipbuilding, mining, milling and fabricating industries.
The presence of asbestos in a building does not mean that the health of
building occupants is endangered. If ACM remains in good condition and is
unlikely to be disturbed, exposure will be negligible. However, when ACM is
damaged or disturbed, asbestos fibers may be released. These fibers can
create a potential hazard for building occupants. The methods for asbestos
abatement are removal, enclosure, and encapsulation. If ACM is found in a
building, a special Operations and Maintenance (O&M) Program should be
implemented as soon as possible. The Risk Reduction Engineering Laboratory
(RREL) Program for asbestos addresses the engineering evaluation of control
strategies and technologies used for asbestos removal and management
in-place. This program emphasizes (1) improvement of Operation and
Maintenance procedures and (2) evaluation of the long-term effectiveness of
removal on the decontamination of surfaces, equipment, HVAC systems and
building air. .
678
-------
SUPERCRITICAL WATER OXIDATION
DEEP-WELL REACTOR MODEL DEVELOPMENT
by: Dong-Soo Lee
The University of Texas at Austin
Center for Energy Studies
Austin, TX 78758
Earnest F. Gloyna
The University of Texas at Austin
Department of Civil Engineering
Austin, TX 78712
ABSTRACT
Supercritical water oxidation (SCWO) has the capability of destroying
toxic organic compounds. However, to critically evaluate the performance of
the SCWO process, extensive information is required on a wide variety of
compounds. In addition, engineering problems involving corrosion, heavy metal
speciation, ash Teachability, charring and encrustation need to be understood.
The main objective of this project is to expand the existing knowledge
for the development of the SCWO process. More specifically, the objectives
are to: (1) obtain destruction information on five organic compounds,
including acetic acid, pentachlorophenol, acenaphthene, 2,4-dinitrophenol
(2,4-D) and kepone (chlordecone); (2) conduct detailed kinetic studies for the
two most refractory compounds; (3) investigate the corrosion of various metal
alloys, and (4) establish the Teachability of chromium from effluent ash.
A batch reactor and a continuous-flow reactor system, respectively, are
being used to conduct the destruction tests and the detailed kinetic studies.
Both of these systems are being used to investigate corrosion and ash
Teachability. The effects of reaction temperature, waste concentration,
oxygen requirements, water density, and retention time are being evaluated.
Reaction temperatures range from 400 °C to 500 "C. Oxygen content and water
density, respectively, ranae from 150% to 300% (stoichiometric demand) and
from 0.3 g/cm3 to 0.4 g/cm . The waste and oxidant concentrations and these
relationships with retention time are being determined for each compound.
Typical batch test results for the destruction of the selected organic
compounds will be presented.
679
-------
EXPANSION OF RREL DATA BASE TO INCLUDE SOIL. DEBRIS AND SEDIMENT
by: Stephanie A. Hansen
Radian Corporation
Milwaukee, Wisconsin 53214
Kenneth A. Dostal
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
ABSTRACT
The U.S. Environmental Protection Agency Risk Reduction Engineering Laboratory
has developed a computerized database. The database contains.information on
the treatability of priority pollutants and other hazardous compounds regu-
lated under the Clean Water Act, Safe Drinking Water Act, Resource Conserva-
tion and Recovery Act, Toxic Substances Control Act, Superfund Amendments and
Reauthorization Act, etc.
The database was initially developed for aqueous wastes but is currently
being expanded to include treatability data on soils, sludge, sediment and
debris. Distribution of Version 3.0 was initiated in October 1990. This
Version contains over 1000 compounds with about 5800 sets of treatability
data. Very limited data on soil are contained in this Version. Several
sorting capabilities have been programmed into the database which markedly
increase its utility.
During the next six to nine months emphasis will be placed upon expanding
the data on soil, etc. The database can be obtained free of charge with a
written request to:
Mr. Kenneth A. Dostal
Risk Reduction Engineering Laboratory
26 W. Martin Luther King Drive
Cincinnati, OH 45268
680
-------
ABSTRACT
ORISMOLOGY:
THE NEXT STEP IN QUALITY CONTROL
by
Guy F. Simes
U.S. EPA
Cincinnati, OH 45268
Webster defines orismology as "the science of defining technical terms". The
area of quality control (QC) has become inundated with scores of new technical terms,
and the inconsistent use of these terms has become common. Given the potential for
confusion in the application and interpretation of environmental data, orismology has
become especially important.
Most QC terms associated with environmental measurements can be grouped
into five categories; (1) calibration/standardization, (2) quantitation limits, (3) quality
control samples and activities, (4) samples for monitoring measurement performance,
and (5) data assessment and reporting. In reviewing over 700 Quality Assurance (QA)
Project Plans, as well as experimental designs, sampling/analysis plans, and other
technical documents, the Risk Reduction Engineering Laboratory (RREL) has compiled
a list of QC-related terms and has attempted to define a uniform system of
terminology. Recommended usages of terms and their relationships to QA and QC
activities are presented. Also, a discussion of QC procedures and their uses and
limitations, is included.
As the next step in quality control, orismology has become important in
assuring the correct and comparable use of terminology among the many diverse
groups involved in the generation and use of environmental analytical data. Through
compilation, definition, and discussion of many of the important technical terms
associated with QA/QC of environmental measurements, RREL has attempted to
promote the idea of orismology in this area and to encourage the proper and standard
use of QC nomenclature.
681
-------
RESEARCH OPPORTUNITIES AT EPA's E-TEC FACILITY
by: Daniel Sullivan, P.E.
John Farlow
Frank Freestone
James Y,ezzi
U.S. Environmental Protection Agency
'• Edison, New Jersey 08837
ABSTRACT
The EPA Risk Reduction Engineering Laboratory {RREL} has proposed
construction of a research facility at Edison, New Jersey for developing,
evaluating, and improving technologies for cleaning up Superfund sites.
The Environmental Technology and Engineering (E-TEC} facility will be an
environmentally safe place for controlled, reproducible tests (up to full
scale) on hazardous and toxic materials. It will feature fully permitted,
state-of-the-art pollution control equipment to back up the control
technology of the hardware being evaluated. This backup capability is such
that even if the test equipment fails totally, then E-TEC will still meet
all permit conditions and fully protect all employees and citizens.
The facility will be operated by EPA in partnership with the Hazardous
Substance Management Research Center of the New Jersey Institute of
Technology. This consortium consists of the five major New Jersey
universities and over thirty industries involved in the cleanup of
hazardous waste sites.
Construction of E-TEC involves renovation of existing warehouse
buildings in Edison, New Jersey and the installation of state-of-the-art
pollution control equipment. Five separate wastewater treatment systems
and two air pollution control systems are proposed. All required federal,
state and county permits/approvals will be sought.
A full range of technologies will be able to be tested at
E-TEC. They include physical, chemical, thermal and biological testing.
No radiological or genetic biological work will be done. The E-TEC
testing/evaluation areas are very large, and can accommodate pilot-scale
and full scale demonstration projects.
Routine operations at E-TEC are expected to begin in 1994.
682
-------
ACCESSING LEAKING UNDERGROUND STORAGE TANK CASE STUDIES AND PUBLICATIONS
THROUGH THE EPA's COMPUTERIZED ON-LINE INFORMATION SYSTEM (COLIS)
by: R. Hillger
Risk Reduction Engineering Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Edison, New Jersey 08837
P. Tibay
T. Douglas
Foster Wheeler Enviresponse, Inc.
Edison, New Jersey 08837
ABSTRACT
The U.S. EPA's regulations for underground storage tanks (USTs) require
corrective action to be taken in response to leaking USTs. Recent developments
of UST programs nationwide as well as the introduction of new technologies to
clean up UST sites have increased the diversity of experience levels among
personnel involved with this type of work. The EPA's Computerized On-Line
Information System (COLIS) has been developed to facilitate technology transfer
among the personnel involved in UST cleanup. The system allows for the quick
and simple retrieval of data relating to UST incidents, as well as other
hazardous waste-related information. The system has been used by response
personnel at all levels of government, academia, and private industry. Although
it has been in existence for many years, users are just now realizing the
potential wealth of information stored in this system. COLIS access can be
accomplished via telephone lines utilizing a personal computer and a modem.
683
-------
METAL VALUE RECOVERY FROM ELECTROMACHINING SLUDGE WASTES
Larry G. Twidwell, D.Sc.
Advanced Minerals and Hazardous Waste Treatment Center of Excellence
Montana College of Mineral Science and Technology
Butte, Montana 59701
ABSTRACT
Electrochemical machining sludge wastes are high value materials containing
appreciable concentrations of chromium, nickel, cobalt and other elements,
Hydrometallurgical processing of these waste materials has been demonstrated to be
both technically feasible and economically favorable. Flowsheets, experimental data, and
economic evaluations will be presented. The process is presently being refined to
improve the selective separability of nickel and cobalt by cyanide complexation followed
by selective precipitation. Demonstration of this selective precipitation technique will
enhance the economic viability of the treatment process. The results of preliminary
experimental test work investigating nickel/cobalt separability by cyanide complexation
will be presented and discussed.
684
-------
BENCH-SCALE WET AIR OXIDATION OF DILUTE ORGANIC
WASTES AT THE ENVIRONMENTAL PROTECTION AGENCY'S
TEST AND EVALUATION FACILITY
by
Avi N. Patkar and Mary Beth Foerst
IT Environmental Programs, Inc., Cincinnati, Ohio 45246
and
Dennis L. Timberlake
U.S. EPA-RREL, Cincinnati, Ohio 45268
Wet air oxidation, in conjunction with activated carbon adsorption or biological
treatment, has been designated as the best demonstrated available technology (BOAT)
for many listed wastes banned from land disposal under the Resource Conservation and
Recovery Act (RCRA). A project has been initiated at the U.S. Environmental Protection
Agency's Test and Evaluation (T&E) Facility to develop information on the applicability,
effectiveness, and costs of wet air oxidation for treating RCRA category P and U wastes.
Wet air oxidation is the pxidative destruction of organics in wastewaters. This
technology operates on the principle that oxygen solubility in an aqueous waste is
greatly increased at higher pressures, and the oxidation rate of the waste is increased at
elevated temperatures. Wastewaters that have been shown to be amenable to such
treatment include pesticide production wastes, petrochemical wastes, spent caustic
wastewaters containing phenolic compounds, and wastewaters resulting from organic
chemical production. Wet air oxidation has typically been applied to the treatment of
waste streams containing dissolved or suspended organics with concentrations ranging
from 500 to 15,000 mg/L.
The purpose of this research project was to determine the ability to destroy listed
organic compounds in wastewaters using wet air oxidation. Bench-scale tests are being
conducted at the T&E Facility using a 1-liter autoclave to determine optimum conditions
(e.g., pressure, temperature) and obtain kinetic data for destruction of organic wastes by
wet air oxidation. The compounds that are being studied (toluene, cyclohexane, and n-
dipropylamine) are constituents of listed wastes for which wet air oxidation has been
proposed as BOAT by the U.S. Environmental Protection Agency. Cyclohexane and n-
dipropylamine have no published kinetic data for wet air oxidation. Studies on toluene
only report total batch time, reaction conditions and total destruction. The study will
provide the kinetic data required to model and design a wet air oxidation reactor.
Preliminary results from the bench-scale wet air oxidation tests are presented for toluene
and cyclohexane.
685
-------
CHARACTERISTICS OF BEVILL SMELTING WASTES
by: Henry Huppert
Science Applications International Corporation
McLean, Virginia 22102
Ronald J. Turner
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
ABSTRACT
In the Hazardous and Solid Waste Amendment of 1984, the Environmental
Protection Agency (EPA) was directed to develop regulations to restrict the
land disposal of untreated hazardous wastes. Under these requirements, EPA is
directed to establish treatment levels or methods based on the standards
achieved by treatment technologies for every hazardous waste listing.
In September 1988, the EPA amended its regulations under the Resource
Conservation and Recovery Act by reinstating the hazardous waste listings of
wastes generated from metal smelting and refining industries. Although these
wastes had been listed, the listings were suspended under Section 3001
(b)(3)(A)(ii) (the "Bevill Amendment"). Upon reinterpretation of the statute,
the suspensions were removed through court action. The reinterpretation of
the Bevill Amendment enacted the original listings and, thus, required the
development of treatment standards for the previously exempted listings.
Those reinstated wastes were the following:
K064; Acid plant sludges and slurries produced from the thickening of
acid plant blowdown in the manufacture of copper
K065: Surface impoundment solids contained in and dredged from surface
impoundments at primary lead production facilities.
K066: Sludge from treatment of process wastewater and/or acid plant
blowdown from primary zinc production
K088: Spent potliner from the primary reduction of aluminum.
K090: Emission control dust/sludge from ferrochromium-silicon production
in electric furnaces
K091: Emission control dust/sludge from ferrochromium production in
electric furnaces
The process of developing treatment standards for a hazardous waste
listing consists of characterization of the regulated industry and waste,
identification of candidate treatment technologies, performance analysis and
evaluation of plausible technologies, development of standards, and
promulgation of the rule. This poster presents the results (including data)
of the engineering site visits, sampling, and analytical information that will
be used to support the development of treatment standards for each waste code.
686
-------
TEST PROGRAM FOR EVALUATION OF FOAM SCRUBBING
FOR CONTROL OF SUPERFUND TOXIC GAS RELEASES
by: John E. Brugger
U.S. Environmental Protection Agency
Risk Reduction Engineering Laboratory
Edison, NJ 08837
Patricia M. Brown
Foster Wheeler Enviresponse
Livingston, NJ 07039
Ralph H. Hiltz
MSA Research Corp.
Pittsburgh, PA 15230
ABSTRACT
This poster describes a laboratory-scale test program to evaluate "foam scrubbing"
as a technique to mitigate emergency releases of airborne toxics.
In the "foam scrubbing" approach, foam is generated with conventional equipment,
actually using the contaminated air to form the foam. The foaming solution contains
neutralizing agents and may require a special surfactant system for compatibility. With the
airborne gases, vapors, and particulate materials encapsulated in the foam, a large
interior liquid surface area is available for their sorption. Neutralization agents present in
the bubble walls then react with the entrapped toxic gas or vapor to render it innocuous.
The self-collapsing foam yields a processable liquid that may be reusable.
Laboratory-scale tests were carried out using a 2-inch MSA hand-held generator,
and a 6-cubic-foot test box. To date, ammonia concentrations ranging from 1 to 20
volume percent have been treated with acid-modified foam, and chlorine concentrations
ranging from 1 to 10 volume percent have been treated with base-modified foam.
Ammonia removals of 90 - 100 % were obtained for all starting concentrations,
when using stoichiometric amounts of acid. Substantial chlorine removals were obtained
for the 1 % and 5 % concentrations, when using stoichiometric amounts of base. Future
plans include the addition of thiosulfate to further improve removal of chlorine.
687
-------
VAPOR-LIQUID EQUILIBRIUM STUDIES
AT THE U.S. ENVIRONMENTAL PROTECTION AGENCY'S
TEST AND EVALUATION (T&E) FACILITY
Avi N. Patkar, Sheryl A. Thurman, Gerry Henderson, and Mary Beth Foerst
IT Environmental Programs, Cincinnati, Ohio 45246
Franklin R. Alvarez
U.S. EPA-RREL, Cincinnati, Ohio 45268
The Resource Conservation and Recovery Act (RCRA), as amended by the
Hazardous and Solid Waste Amendments of 1984, prohibits the placement of RCRA-
regulated hazardous wastes in or on the land. Steam stripping has been designated
as the best demonstrated available technology (BOAT) for treatment and removal of
many of the RCRA-listed wastes that may have been land disposed in the past. The
purpose of this research is to provide data that can be used in the design and
operation of steam strippers for the treatment of specific RCRA wastes. The research
is being conducted at the U.S. Environmental Protection Agency's Test and Evaluation
(T&E) Facility in Cincinnati, Ohio.
The design of steam stripping columns requires a characterization of the vapor-
liquid equilibrium over the full range of liquid and vapor compositions for the
constituents of interest. Stripping columns can be designed by using Henry's Law
constants if the solution is sufficiently dilute. Henry's Law data are available for many
compounds at low temperatures (20 to 30°C) used in air stripping; however, vapor-
liquid equilibrium data are lacking for the high temperature range (90 to 100°C)
associated with steam stripping of aqueous wastes at atmospheric pressure. Several
chlorinated aromatic compounds and nitroparaffins are of special interest because
they are difficult to model accurately due to their limited miscibility with water.
A laboratory-scale Modified Othmer Still apparatus is being used in this
research to characterize equilibrium liquid and vapor compositions for organic
compounds in aqueous systems. The compounds being tested in the vapor-liquid
equilibrium studies are 2-nitropropand, 1,1 -dichloroethane, and 2,4-dichlorophenol.
Experimental vapor-liquid equilibrium data are unavailable for these industrially
significant compounds. The compounds 1,1-dichloroethane and 2,4-dichlorophenol
are constituents of EPA Hazardous Waste Codes U076 and U081, respectively. The
selected compounds can be analyzed by gas chromatography. Each compound is
being tested at atmospheric pressure at several concentrations below the compound's
solubility limit. Preliminary results from the vapor-liquid equilibrium studies are
presented.
688
-------
SORPTION ISOTHERMS FOR AZO DYES ONTO ACTIVATED SLUDGE BIOMASS
by: Richard J. Lieberman
Glenn M. Shaul
Clyde R. Dempsey
Kenneth A. Dostal
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
ABSTRACT
The Congress of the United States of America enacted the Toxic
Substances Control Act (TSCA) in response to the growing numbers of
chemicals in contact with humans and the environment, including some
substances suspected of presenting an unreasonable risk of injury to human
and/or environmental health. According to TSCA, adequate data should be
developed with respect to the effect of chemical substances and mixtures on
health and the environment. In addition, technological innovation is not to
be unduly hindered while ensuring that chemical substances and mixtures do
not present an unreasonable risk of injury to health or the environment.
The United States Environmental Protection Agency's Office of Toxic
Substances (OTS) is charged with the responsibility to carry out the terms
of the Toxic Substances Control Act. This research was conducted to provide
OTS with information about an important group of chemical substances known
as azo dyes. Sorption is generally considered as the most important process
in the removal of azo dyes from wastewater across an activated sludge
treatment system. Therefore, sorption isotherms were developed to be used
as indicators of the fate of azo dyes treated via the activated sludge
process. This poster presents the application of sorption isotherm
methodology as a potential tool for the evaluation of azo dyes under the
requirements of TSCA.
689
-------
EPA'S SYNTHETIC SOIL MATRIX (SSM) BLENDING FACILITY
by: Seymour Rosenthal
Foster Wheeler Enviresponse, Inc.
Edison, New Jersey 08837
Raymond M. Frederick
U.S. Environmental Protection Agency
Edison, New Jersey 08837
ABSTRACT
In 1987, a temporary blending facility was constructed at EPA's Center
Hill Research Facility in Cincinnati, Ohio for producing a synthetic soil
for use as a surrogate test material in waste treatability studies. Since
then, work has progressed with the establishment of a permanent blending
facility at the EPA's Edison, New Jersey laboratory for continuing the
production of this Synthetic Soil Matrix (SSM) for use as a standard test
medium in EPA's Superfund Innovative Technology Evaluation (SITE) Program.
The SSM is also available to the private sector for use in their development
of innovative treatment technologies. The SSM is formulated to. represent
typical soil types and contaminants found at Superfund sites. Clean soil
matrix is created by blending specified amounts of sand, gravel, silt, top
soil (for humic content), and a mixture of clays. Water and chemicals
(volatiles, semivolatiles, and metals) are added to the soil matrix to
simulate environmental contaminants. When performing treatability studies,
the synthetic soil matrix provides an effective means for comparing the
efficiencies of individual treatment technologies on a common and defined
soil matrix.
This presentation describes the Synthetic Soil Matrix Blending Facility
which includes a mixing room, personnel decontamination area, and a support
area. Capabilities of the facility to produce four standard blends of
soil with high and low concentrations of organics and metals, as well as
custom blends of other analytes are discussed. Future research efforts to
improve the synthetic soil matrix, and to develop a treatment efficiency
database from SSM user surveys are also discussed.
690
-------
RPM/OSC SUMMARY
REMEDIAL/REMOVAL INCINERATION PROJECTS
by: Laurel J. Staley
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
ABSTRACT
The Superfund On Scene Coordinator (OSC) and/or Remedial Project Manager
(RPM) must access a wide variety of information when supervising removal/remedial
activity at a Superfund site where incineration is to be used. In order to make
that information more readily available to the OSC/RPM, the QSC/RPM Summary
for Remedial/Removal Incineration Projects was prepared. The document was not
intended to be an all inclusive reference on incineration. The body of hazardous
waste incineration knowledge changes too rapidly for such a document to be useful
for very long. Rather, it was intended to document the OSC/RPM to experts within
and outside of the EPA who have the most current knowledge of incineration.
Background information on Incinerator Design and Operation, Incinerator
Manufacturers and Operators, Incineration Regulations, and Cost is presented
concisely and largely in tabular form. Extensive references are provided for
more detailed information on the topics discussed in the Document. Together with
the lists of incineration experts, the document should provide the OSC/RPM with
enough information to more effectively monitor and direct incineration related
activities at Superfund sites. This presentation discusses the contents of the
Summary document and some of the issues raised during its review.
691
-------
TREATABILITY STUDY RESULTS ON SOIL CONTAMINATED
WITH HEAVY METALS. THIOCYANATES. CARBON DISULFATE
OTHER VOLATILE AND SEMIVOLATILE ORGANIC COMPOUNDS
by: Sarah Hokanson
PEI Associates
Cincinnati, Ohio
ABSTRACT
Laboratory screening level treatability studies were performed to
support the ongoing RI/FS for the Halby Chemical Company site in Wilmington,
Delaware. The project team, in coordination with the START team project
leader, decided that several technologies were applicable, including
combinations of these technologies into treatment trains.
The results from these studies indicate that: (1) aerobic and anaerobic
carbon disulfide and aerobic thiocyanate degrading organisms are present in
site soils and biodegradation of carbon disulfide and thiocyanate compounds
(as indicated by microbial growth and oxygen consumption) can occur in the
laboratory with the indigenous microbial population under aerobic conditions
with sufficient amounts of nutrients; (2) while low temperature thermal
desorption may not be needed as a pretreatment step to solidification/
stabilization, it can successfully remove most volatile and semivolatile
organic compounds in the site soils at temperatures between 300°F and 500°F
and between 15 and 30 minutes residence time; and (3) the soils, themselves,
do not leach appreciable amounts of metals under TCLP test conditions and of
the binders studied (asphalt and cement), asphalt appears to be the better
binder for reducing leachate concentrations of arsenic and copper.
Significant flotation/separation of metals from soils using xanthate was not
achieved.
692
-------
EVALUATION OF TEMPORAL CHANGES IN SOIL BARRIER WATER CONTENT
by: R. J. Luxmoore
Environmental Sciences Division
Oak Ridge National Laboratory
Oak Ridge, Tennessee 37831
ABSTRACT
The integrity of soil caps and liners in shallow land burial facilities depends on the
longterm stability of the barrier water content. Periodic wetting and drying of soil
barriers will lead to crack formation and barrier failure. An evaluation of temporal
changes in compacted soil barrier water content in representative landfill operations in
humid and dry environments is being conducted by computer simulation modelling and
with an evaluation of longterm biological activity documented at abandoned landfill sites.
Six water-dynamics models were evaluated for simulation of temporal changes in
barrier water content and three were selected for application. The EPA-recommended
barrier designs are being evaluated with data from a humid-region site in Tennessee
and an arid site in New Mexico using HELP, UTM, and MIGRAT simulation programs.
These codes have differing modelling approaches that will be compared in the study.
Some initial simulations showed that the penetration of roots into the compacted soil
barrier resulted in significant drying, due to transpiration, below the original water
content at which the barrier was compacted. Simulation results illustrate the range of
variation in barrier water content associated with contrasting climatic regimes.
Ecological succession of local flora and fauna can be expected to result in
establishment of deep rooted vegetation and burrowing animals, eventually leading to
barrier failure. Different plant and animal species will be active in different climatic,
geologic, and ecologic regions of the United States resulting in differing rates of barrier
penetration. Measures that may prevent barrier failure by root penetration and animal
burrowing should be incorporated into the design and construction of soil-based barriers
in waste management structures.
693
-------
Abstract
Evaluation of Stress Cracking Resistance of Polyethylene Flexible Membrane Liners
Yick Halse-Hsuan, Robert M. Koerner, and Arthur E. Lord, Jr.,
GRI
Drexel University
Philadelphia, PA 19104
The stress cracking resistance of polyethylene flexible membrane liners (FMLs)
was evaluated using the notched constant load test (NCLT). The test materials
include both high density polyethylene (HOPE) and linear low density polyethylene
(LLDPE). For HOPE flexible membrane liners, the conventional test method ASTM
D 1693 "Bent Strip Test" was also performed.
The first part of the paper compares the sensitivity between NCLT and the Bent
Strip Test. The bent strip test is a qualitative technique. Also the stress relaxation of
the polymeric material can influence the outcome of the results, particularly for
long testing times. On the other hand, NCLT is a more quantitative test, from
which the ductile-to-brittle transition time is obtained. This test overcomes the
stress relaxation problem and subjects the test specimens to constant load during the
entire test period. The control of the test is more precise than the Bent Strip Test.
Also the NCLT is a quantitative test and Bent Strip Test is a qualitative one.
The second part of the paper describes two extrapolation methods used to
predict FMLs' behavior at site specific temperatures which are lower than practical
laboratory testing temperatures. The two methods are the Rate Process Method
(RPM) and Arrhenius modeling. NCLTs are performed at temperatures of 50°C and
75°C in a solution of 10% Igepal and 90% tap water. In addition, a 25°C test is
performed so that the predicted data can be compared with the actual experimental
results. The results of the second part of the paper further illustrated the power of
the NCLT, in that it can be used as a resin qualifying test for stress crack resistance
and as a predictive technique to assess polymer aging.
694
-------
CENTER HILL SOLID AND HAZARDOUS WASTE RESEARCH FACILITY
by: Gerard Roberto
University of Cincinnati
Department of Civil & Environmental Engineering
Cincinnati, Ohio 45221
Robert Landreth
Risk Reduction Engineering Laboratory
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
ABSTRACT
The Center Hill Research Facility provides technical support services in the geo-technical and geo-
scientific fields for the Agency's Superfund and Resource Conservation Recovery Act Program activities in
solid and hazardous waste testing and research. As one of several pilot-plants supporting the Risk
Reduction Engineering Laboratory in Cincinnati Ohio, the Facility houses specialized laboratories with a
high bay for testing services and the conduct of research projects at bench, pilot and field scale.
Technical support services by the on-site University of Cincinnati Contractor staff of engineers and
scientists focus on soil/chemical/microbiological interactions for pollution control, containment and
remediation of the geo-hydrological environment.
Superfund Program activities conducted at Center Hill include technical assistance to the USEPA
Regions for site characterization, assessment of proposed remediation technologies, and remedial action
program design and construction. In-house computer-aided-engineering services assist the Regions in
site-situation mapping and modeling for assessment and monitoring of remedial actions.
RCRA Program activities conducted at Center Hill include the study and evaluation of pollution
control technologies for the processing of municipal and industrial solid wastes to control the release of
pollutants to the land.
Current research and development activities at Center Hill address the performance of remediation
technologies in the Agency's Best Developed Available Technologies (BOAT), Superfund Innovative
Technology Evaluation (SITE) and Superfund Technical Assistance Response Team (START) Program
areas. Research and development activities are currently being conducted in the area of delivery and
recovery systems for in-situ and on-site remediation. A poster and slide presentation will provide a brief
description on current project activities.
695
U.S. GOVERNMENT PRIN I ING OFFICE: 19U1 — 548-187/20563
-------
-------
-------
United States
Environmental Protection
Agency
Center for tnvironmental Research
Information
Cincinnati OH 45268
Official Business
Penalty for Private Use, $300
Please make all necessary changes on the above label, detach
or copy and return 10 the address in the upper left-hand comer.
if you do not wish to receive these reports CHECK HERE il:
delsch, or copy this cover, and return to the address in the
upper left-hand corner.
EPA/600/9-91/002
------- |