TT TT ^ ^T A TT
Hazardous Wastes
and Hazardous
Material
Site Remediation • On-Site Treatment • Risk Assessment
Contaminated Groundwater Control • Permitting • Monitoring • Incineration
• Underground Leak Detection • Fixation • Cost/Economics
-------
PROCEEDINGS OF THE
NATIONAL CONFERENCE ON
HAZARDOUS
WASTES AND
HAZARDOUS
MATERIALS
Site Remediation • On-Site Treatment • Risk Assessment
• Contaminated Groundwater Control • Permitting • Monitoring •
Incineration • Underground Leak Detection • Fixation • Cost/Economics
March 4-6, 1986 • Atlanta, Georgia
AFFILIATES
U.S. Environmental Protection Agency
Hazardous Materials Control Research Institute
Department of Defense
Agency for Toxic Substances and Disease Registry
Portland Cement Association
National Environmental Health Association
National Lime Association
-------
Printed in the United States of America
Library of Congress Catalog No. 86-80269
Copyright © 1986
Hazardous Materials Control Research Institute
9300 Columbia Boulevard
Silver Spring, Maryland 20910
-------
Acknowledgement
The National Conference and Exhibition on Hazardous Wastes and Hazardous
Materials: Site Remediation • On-Site Treatment • Risk Assessment •
Contaminated Groundwater Control • Permitting • Monitoring • Incineration •
Underground Leak Detection • Fixation • Cost/Economics required dedication and
talent from many individuals and commitment from a number of organizations. We
gratefully express our thanks and appreciation to the following conference affiliates
for all their contributions to a successful conference.
U.S. Environmental Protection Agency
Hazardous Materials Control Research Institute
Department of Defense
Agency for Toxic Substances and Disease Registry
Portland Cement Association
National Environmental Health Association
National Lime Association
We also wish to express our gratitude to all of these knowledgeable individuals for
their advice and guidance in planning and producing a highly effective and infor-
mative program:
Gary F. Bennett, Ph.D., The University of Toledo
Hal Bernard, Hazardous Materials Control Research Institute
Ken Gutschick, National Lime Association
Robert Knox, U.S. Environmental Protection Agency
Charles Mashni, U.S. Environmental Protection Agency
Thomas Potter, National Lime Association
Jerry Steinberg, Ph.D., Water & Air Research, Inc.
Andres Talts, Department of Defense
Ralph Touch, Agency for Toxic Substances and Disease Registry
Producing a document of the magnitude of this proceedings requires a highly
skilled team, much cooperation and communication, and a tremendous amount of
effort by all involved. We are fortunate to have such a team and would like to
convey our special thanks to Dr. Gary Bennett, Professor of Biochemical Engineer-
ing, The University of Toledo, and Hal Bernard, Executive Director, HMCRI, for
the excellent editing; to the typesetters, proofreaders, and graphic artists who
completed a tremendous amount of work in an incredibly short period of time; and
to the staff of HMCRI for coordinating the myriad details and activities of this
conference.
-------
Preface
The RCRA Amendments of 1984 (the new RCRA) have
radically changed hazardous waste management through-
out the country and considerably improved the control of
hazardous wastes. Whereas it was estimated that the old RCRA
provisions would cost the regulated industry between $1 billion
and $3 billion per year, the new RCRA amendments, when com-
pletely implemented and fully in effect, may cost as high as $20
billion per year.
The new RCRA includes 58 congressionally mandated statu-
tory deadlines that go into effect by 1987 or early 1988. Congress
wrote, into the statute, provisions that are detailed and directive.
The bill does not wait for EPA to write regulations or issue
guidance; provisions go into effect automatically on prescribed
dates.
Land Application Management
The law establishes a hierarchy of management practices. There
is a very strong presumption against the disposal of hazardous
wastes on the land and a very strong preference for treatment and
destruction of hazardous wastes.
Site Remediation
The bill provides EPA with "corrective action authority" at
existing hazardous waste facilities which are very similar to the
enforcement authorities under Superfund. The bill also closes the
gap in the existing program by expanding the regulated com-
munity.
Leaking Underground Tanks
A major new subtitle of the legislation dealing with leaking
underground storage tanks went into effect on May 8, 1985. This
section probably affects two to five million underground tanks
across the U.S. including tanks storing petroleum products and
gasoline as well as very complex tank storage systems storing
hazardous chemicals.
As of May 8, 1985, there is a ban on new tanks that are not de-
signed to prevent releases due to corrosion or structural failure.
An estimated 100,000 new tanks must meet these requirements.
By May 1986, a nationwide registration program will require
state notification of the age, size, type and location of the tank, as
well as its uses.
Burning Hazardous Wastes
The provision concerning burning hazardous wastes and the
blending of hazardous wastes into fuel requires notification of
the fuel user that they are receiving hazardous waste. Some can-
not be burned in residential and commercial boilers.
RCRA Permit Program
The bill makes permitting of a hazardous waste facility far
more difficult than it has been. Permit applicants must now sub-
mit exposure information on the potential for public exposure to
hazardous substances from landfills and surface impoundments.
This exposure information will be used to write new permit con-
ditions. EPA has the authority to write any permit condition
necessary to protect human health in the environment, inde-
pendent of whether or not regulations are in place for that
purpose.
Congress has also set some very strict deadlines for the issuance
of permits. By November 1985, all applications for land disposal
facility permits must have been submitted to EPA or an author-
ized state. Final determination on these land disposal permit
applications must be made by November 1988.
By November 1986, all incinerator applications must be sub-
mitted and EPA must make final decision on these by November
1989. All remaining applications have to be in by November
1988, and EPA must make a decision on these by November
1992.
Small Quantity Generators
The first small quantity generator (SQG) requirement went into
effect in August 1985. Any SQG that ships wastes offsite must do
so by using the Uniform National Manifest Form. Effective in
April 1986, SQGs must send their wastes to fully regulated haz-
ardous waste facilities. These requirements will have a tremen-
dous impact on small generators; costs for their waste manage-
ment will be considerably greater.
Other RCRA Amendments
There are many other provisions of the new RCRA. There is a
set of provisions that deal with non-hazardous solid waste and
small quantity generator wastes disposed of in landfills and sur-
face impoundments. The bill also provides for federal enforce-
ment of the requirements for lagoons. There are a series of waste
minimization requirements for generators to certify that they are
doing everything economically feasible to reduce the amount of
wastes generated. There are requirements for listing additional
hazardous wastes. Delisting of wastes is much more difficult and
complex under the new bill. New tests for ascertaining toxicity are
included in the new RCRA.
-------
CONTENTS
MONITORING
Equipment for Data Collection at Hazardous Waste
Sites—An Overview for Environmental Management
Professionals 1
James P. Mack & Thomas J. Morahan
A Statistician's View of Groundwater Monitoring 8
Douglas E. Splitstone
Data Evaluation in a Groundwater Study of Waste
Management Practices in the Phosphate Processing
Industry 13
Edward W. Mullin, Jr., Jack S. Greber & William
E. Thompson
Plant Cuticular and Dendrochronological Features
as Indicators of Pollution 17
O.K. Sharma, Ph.D.
Detailed Stratigraphic and Structural Control: The Keys
to Complete and Successful Geophysical Surveys of
Hazardous Waste Sites 19
H. Dan Harmon, Jr., P.O.
Detection and Measurement of Groundwater
Contamination by Soil-Gas Analysis 22 /
H.B. Kerfoot, J.A. Kohout & E.N. Amick
Environmental Appraisals and Audits: A Case Study
of Their Application 27
Anthony R. Morrell
DETECTION OF RELEASES
Chlorinated Organics and Hydrochloride Emissions
Sampling from a Municipal Solid Waste Incinerator 31
Thomas A. Driscoll, James P. Barta, Henry J. Krauss,
David H. Carmichael & J. Maxine Jenks
CARE—Modeling Hazardous Airborne Releases 34
M. Gary Verholek
The Use of PC Spreadsheet-Based Graphics to Interpret
Contamination at CERCLA/RCRA Sites 39
George A. Furst, Ph.D.
CONTAMINATED GROUNDWATER CONTROL
Emergency Response to Toxic Fumes and Contaminated
Groundwater in Karst Topography: A Case Study 44
P. Clyde Johnston, Mark J. Rigatti & Fred B. Stroud
Use of Low Flow Interdiction Wells to Control
Hydrocarbon Plumes in Groundwater 49 >
John H. Sammons, Ph.D. & John M. Armstrong, Ph.D.
Computer Groundwater Restoration Simulation at a
Contaminated Well Field 58
Shih-Huang Chieh, Ph.D. & Jeffrey E. Brandow, P.E.
Enhancement of Site Assessments by Groundwater
Modelling 64
Joseph R. Kolmer, P.E. & John B. Robertson, P.G.
Alternative Treatment Techniques for Removal of Trace
Concentrations of Volatile Organics in Groundwater 69
Mark E. Wagner & Brian V. Moran
CONTAMINATED SOIL TREATMENT
Cost-Effective Soil Sampling Strategies to Determine
Amount of Soils Requiring Remediation 76
Gregory J. Gensheimer, Ph.D., William A. Tucker,
Ph.D. & Steven A. Denahan
Land Treatment of Wood Preserving Wastes 80
John R. Ryan & John Smith
Method for Determining Acceptable Levels of
Residual Soil Contamination 87
William A. Tucker & Carolyn Poppell
Objective Quantification of Sampling Adequacy and
Soil Contaminant Levels Around Point Sources
Using Geostatistics 92
Jeffrey C. Myers
ON-SITE TREATMENT
Innovative Application of Chemical Engineering
Technologies for Hazardous Waste Treatment 98
Robert D. Allan & Michael L. Foster
Field Studies of In Situ Extraction and Soil-Based
Microbial Treatment of an Industrial Sludge Lagoon 102
David S. Kosson, Erik A. Dienemann & Robert
C. Ahlert, Ph.D., P.E.
Cleanup of Contaminated Soils and Groundwater
Using Biological Techniques 110
Paul E. Flathman & Jason A. Caplan, Ph.D.
Physical/Chemical Removal of Organic Micropollutants
from RO Concentrated Contaminated Groundwater 120
L. Simovic, J.P. Jones & I.C. McClymont
State-of-the-Art Technologies of Removal, Isolation
and Alteration of Organic Contaminants Underground 124
Walter W. Loo & George N. Butter
-------
Assessment of Volatile Organic Emissions from a
Petroleum Refinery Land Treatment Site 127
Robert G. Wet herald, Ph.D., Bart M. Eklund,
Benjamin L. Blaney, Ph.D. & Susan A. Thorneloe
Technology for Remediation of Groundwater
Contamination 133
David V. Nakles, Ph.D. & James E. Bratina
TREATMENT OF HAZARDOUS WASTES
Anaerobic Biological Treatment of Sanitary
Landfill Leachate 136
A.K. Mureebe, P.E., D.A. Busch & P.T. Chen, Ph.D.
On-Site Versus Off-Site Treatment of Contaminated
Groundwater—An Evaluation of Technical Feasibility
and Costs 143
Kent L. Bainbridge & Daniel W. Rothman, P.E.
Optimization of Free Liquid Removal Alternatives in
the Closure of Hazardous Waste Surface Impoundments 148
David K. Stevens
Assessment of Chemical Treatment Technologies and
Their Restrictive Waste Characteristics 154
Hamid Rasiegar, Ph.D.. James Lit, Ph.D. &
Chris Conroy
Treatment Technologies for Hazardous Materials 157
Jeffrey M. Thomas & Phillip T. Jarboe
Update on the PACT Process 163
Harry W. Heath, Jr.
Treatment of PCB-Contaminated Soil in a
Circulating Bed Combustor 171
D.D. Jensen, Ph.D. & D.T. Young
Treatment of Hazardous Waste Leachate 175
Judy L. McArdle, Michael M. Arozarena, William
E. Gallagher, P.E. & Edward J. Opatken
BARRIERS & WASTE SOLIDIFICATION
Field Testing of a New Hazardous Waste
Stabilization Process 180
Richard H. Reifsnyder
Innovative Techniques for the Evaluation of
Solidified Hazardous Waste Systems 186
Harvill C. Eaton, Marty E. Tittlebaum & Frank K.
Cartledge
Site Characteristics and the Structural Integrity
of Dikes for Surface Impoundments
Jey K. Jeyapalan, Ph.D., P.E. & Ernest R. Hanna
Use of X-Ray Radiographic Methods in the
' Study of Clay Liners
Philip G. Malone, Ph.D., James H. May, Kirk W.
Brown, Ph.D. & James C. Thomas
Closure Design and Construction of Hazardous
Wastes Landfills Using Clay Sealants
John F. O'Brien, P.E., Lonnie E. Reese & Ian
Kinnear, P.E.
190
198
202
Soil Liners for Hazardous Waste Disposal Facilities 206
D.C. Anderson
Slurry Wall Economical in Dewatering of
Sydney Mine Disposal Site 210
Bruce J. Haas, Mark R. Nielsen & Norman N. Hatch
Utility of Soil Barrier Permeability Data 216
Walter E. Grube, Jr.
Quality Assurance and Quality Control Procedures for
Installation of Flexible Membrane Liners 221
James R. Woods, P.E. & Salvatore V. Arlotta, Jr., P.E.
Mechanisms for the Fixation of Heavy Metals in
Solidified Wastes Using Soluble Silicates 224
Ella L. Davis, James S. Falcone, Scott D. Boyce &
Paul H. Krumrine
.229
INCINERATION
Illinois Plan for On-Site Incineration of
Hazardous Waste
James F. Frank & Robert Kuykendall
An Introduction to EPA's New Trial Burn Data Book 233
M.P. Esposito & N.J. Kulujian
Low Temperature Thermal Stripping of Volatile
Compounds 2-**
John W. Noland, P.E.. Nancy P. McDevitt & Donna
L. Kolluniak
Comparisons Between Fluidized Bed and Rotary Kiln
Incinerators for Decontamination of PCB Soils/
Sediments at CERCLA Sites 242
Henry Munoz, Frank L. Cross, Jr., P.E. & Joseph
L. Tessitore, P.E.
UNDERGROUND LEAKING TANKS
New Requirements for Underground Storage Tanks 246
A nna O. Buonocore, P. E., Gerald F Kotas &
Kevin G. Garrahan, P.E.
Cost-Effectiveness Evaluation of Leak Detection and
Monitoring Technologies for Leaking Underground
Storage Tanks 251
James Lu, Ph.D.. P.E. <$ Wayne Barcikowski
Underground Storage System Assessment, Testing
and Remediation 269
Scott J. Adamowski, Angela J. Caracciolo, III
& G. David Knowles, P.E.
Case Study of Product Detection in Groundwater 273
Pratap N. Singh, Ph.D.. P.E.
Underground Storage Tanks—Leak Prevention,
Leak Detection, and Design 278
Jey A. Jeyapalan. Ph.D., P.E. & James B.
Hutchison. P.E.
SITE MANAGEMENT
Performance Evaluation of Commercial Hazardous
Waste Treatment Facility Operations 292
Ronald J. Turner & Joan V. Boegel
An Examination of Siting Problems for
Off-Site TSDFs 298
Douglas B. Taylor, P.E.
Spatial Data Research for Hazardous Waste Sites 301
Timothy W. Foresman & Lynn K. Fensiermaker
Sensitivity Analysis of Remedial Action Alternatives
for Hazardous Waste Sites 305
Elio F. Arniella, P.E. & E. Lawrence Adams, Jr., P.E.
RCRA SITE REMEDIATION & EXPANSION
Design of a Lateral and Vertical Expansion at an
Existing Interim Status Landfill 310
Rodney T. Bloese & Thomas G. Ryan, P.E.
-------
.314
.318
.327
.331
.336
.342
.346
Cost Estimating for RCRA/CERCLA Remedial
Actions
Michael R. Morrison & Gregory P. Peterson, P.E.
A Modular Computerized Cost Model for Remedial
Technologies at Superfund Sites
William Kemner, John Abraham, P.E., Jack
Creber & Jay Palmisano
RISK ASSESSMENT
Using Cost/Risk Analysis in Waste Planning:
The New England Pilot Project
Debora C. Martin
Assessment of Potential Public Health Impacts
Associated with Predicted Emissions of
Polychlorinated Dibenzo-Dioxins and
Polychlorinated Dibenzo-Furans from a
Resource Recovery Facility
David Lipsky, Ph.D.
Improving the Risk Relevance of Systems for
Assessing the Relative Hazard of Contamianted Sites
Ellen D. Smith, Lawrence W. Barnthouse, Ph.D.,
Glenn W. Suter II, Ph.D., James E. Breck, Ph.D.,
Troyce D. Jones & Dee Ann Sanders, Ph.D.
Soil Cleanup Criteria for Selected Petroleum Products
Sofia K. Stokman & Richard Dime, Ph.D.
Risk Management: Personnel, Equipment and Indemnity
Denny M. Dobbs & Arlene B. Selber
RCRA AMENDMENT EXPERIENCE
Innovative Approach to Site Remediation Involving
an RCRA Part B. Permit
Michael J. Conzett, P.E. & Michael E. Harris, P.E.
Overview of the Proposed Natural Resource
Damage Assessment Regulations
Richard J. Aiken, Willie R. Taylor, Ph.D.
Regulatory Impact Analysis for the Toxicity
Characteristic
John L. Warren
Preliminary Assessments and Site Investigations
Under the Corrective Action Authorities of RCRA:
Analysis of Early Experiences
John W. Butler & Robert D. Volkmar
Assessment of the Application of RCRA Part 264
Standards to CERCLA Site Remediation
Rebecca N. Fricke, P.E.
STATE, REGIONAL & LOCAL PROGRAMS
Small Quantity Generators: The Maryland
Approach to Regulation and Assistance 375
Alvin L. Bowles, P.E.
Siting Efforts in Southern California 378
Kieran D. Bergin
The California Site Mitigation Decision Tree 380
Paul W. Hadley & William Quan
.348
.353
.360
.364
.369
Hazardous Waste Management: The Role of the
Local Health Department 384
Michael J. Pompili & Philip G. Brown
PUBLIC COMMUNICATIONS
U.S. EPA's Initiatives for Expanded Public
Involvement in the RCRA Permitting Program 386
Vanessa Musgrave & Edwin Berk
Cleanup in the Sunshine: Florida DOT's Public
Information Program at the Fairbanks Site 390
Robert C. Classen & Charles C. Aller
Public Involvement in the RCRA Permitting Process—
A Facility Perspective 394
Gordon Kenna
Public Participation in Siting Hazardous Waste
Management Facilities in Alaska 397
Sharon O. Hillman
Developing a Comprehensive Public Affairs Policy
for New and Existing Industrial Operations 401
Ann E. Burke
Behind the "F" in Public Education for Hazardous
Waste Management: A Case for Special "Tutoring"
—Plus Some Tips for Better Grades 404
Howard A. Coffin, Patricia Hunt & L.T. Schaper, P.E.
REUSE & RECOVERY
Recycling of Dust from Electric Arc Furnaces—
An Experimental Evaluation 409
E. Radha Krishnan, P.E., William F. Kemner &
Copal Annamraju, P.E.
Trends in Used Oil Composition and Management 419
Jacob E. Beachey & William L. Bider
Used Solvent Elimination Program 424
Renato G. Decal, P.E.
WASTE MINIMIZATION PROGRAMS
Waste Reduction Audit Procedure—A Methodology
for Identification, Assessment and Screening of
Waste Minimization Options
Carl H. Fromm, P.E. & Michael S. Callahan, P.E.
Minnesota Technical Assistance Program: Waste
Reduction Assistance for Small Quantity Generators
Cindy A. McComas & Donna Peterson
Waste Minimization at Air Force GOCO Facilities
Douglas L. Hazelwood, Brian J. Burgher, P.E. &
Charles Alford
Defense Environmental Leadership Project Study
of Industrial Processes to Reduce Hazardous Waste
Thomas E. Higgins, Ph.D., P.E. & Drew P. Desher
Exhibitors' List 450
.427
.436
.440
.445
-------
Equipment for Data Collection at
Hazardous Waste Sites—An Overview for
Environmental Management Professionals
James P. Mack
Thomas J. Morahan
Fred C. Hart Associates, Inc.
New York, New York
INTRODUCTION
Industry and government are in the process of cleaning up
America. Industry is becoming increasingly concerned with ques-
tions of environmental liability. Regulatory agencies are actively
protecting public health and the environment. Future liability
can be minimized by the use of environmental audits. Present
liabilities, however, must be thoroughly understood in order to
develop cost-effective liability management plans that meet the
goals and objectives set for environmental cleanup.
This understanding generally is gained through environmental
investigations at individual sites or facilities. CERCLA, RCRA
and other statutes require regulatory compliance. In some cases,
the buyer or seller of a property may run into unforeseen en-
vironmental liabilities. Some states have recently passed legisla-
tion that requires environmental conditions be defined and
problems investigated prior to real estate transactions. In cases of
corporate mergers or other processes where ownership of a facil-
ity changes, new owners or sellers may be held accountable for
the environmental liability at a facility. While the government
agencies ensure that cleanups are effective, industries which fi-
nance the cleanups can encounter huge expenditures. Business-
men and legal professionals handling environmental affairs stay
close to technical events, carefully monitoring costs and cleanup
strategies. These environmental management professionals are
sometimes unfamiliar with techniques for data collection and the
creative use of equipment used for data collection activities. This
paper is directed toward business, legal and environmental man-
agement professionals seeking knowledge of various types of
equipment available for environmental data collection. An under-
standing of the use of this equipment will help in the develop-
ment of cleanup strategies for a proper, complete and cost-effec-
tive cleanup.
While each site investigation is unique and requires a certain
degree of site-specific modification, there is a logical sequence of
events that is followed in order to collect the desired informa-
tion. While the purpose of this paper is to discuss the field equip-
ment available to perform the investigation, it is important to
understand the sequence of events in order to define the situa-
tions in which the equipment should be used.
The first phase of the investigation should be devoted to
collecting enough background data on site conditions as possible.
While there may not be much data available on actual site con-
ditions, there is usually a certain amount of regional information
available. Data can be obtained from local health departments,
state and federal environmental and geological agencies and uni-
versities. This information is important in defining the general
environmental setting of the site.
The second phase of the investigation should be devoted to
understanding the source of the potential contamination. At this
point, certain indirect techniques such as surface geophysical
instruments or portable analytical instruments are very useful.
They can be used to define the possible horizontal extent of the
source, detect buried objects and initially determine the general
types of contaminants. Soil borings and soil sampling techniques
also are used to collect actual waste or contaminated soils for
chemical analysis to further define the characteristics of the con-
taminant.
Once it has been determined that there is a contaminant source,
the question arises as to whether the contaminants are migrat-
ing. This usually requires more extensive sampling of surface
water, sediment, soils and groundwater. Normally, groundwater
is an important issue and becomes the focus of further investiga-
tions. Wells usually are installed adjacent to the source to deter-
mine the subsurface conditions as well as to detect any immed-
iate migration. Proper soil sampling and well drilling techniques
are essential to the adequate completion of this task. Downhole
investigative equipment becomes useful in defining subsurface
geology. Water levels are measured in wells to define flow direc-
tions. As the scope of the potential problem develops, additional
wells may be required at further distances to define long-term mi-
gration pathways.
The final component is the data evaluation and risk analysis.
This phase is as important as the field effort. The data must be
properly organized and the actual risks correctly established in
order to arrive at a cost-effective remedial solution.
SURFACE GEOPHYSICAL TECHNIQUES
Surface Geophysical Techniques have been used for years to
give indications of subsurface rock and soil conditions. These
techniques vary greatly in theory and technology, resulting in very
specific conditions where their use is warranted. In recent years,
these techniques have been applied successfully to investigations
at hazardous waste sites.
Surface geophysical surveys are performed as indirect tech-
niques to provide information to investigators in advance of
direct sursurface investigatory techniques such as test pitting, test
borings or well drilling activities. At hazardous waste sites, these
techniques generally are used in three different ways. First, some
methods can be employed to provide information about subsur-
face conditions such as depth to bedrock, depth to the water
table or thickness of certain soil units. Next, some methods can
be employed to examine the migration of contamination in
groundwater. Finally, some techniques can be utilized to search
for buried metal, specifically containerized wastes such as drums
or tanks which could act as ongoing sources of contamination.
MONITORING 1
-------
Magnetometry
Magnetrometry has been used successfully in the past to iden-
tify trends in regional geology. More recently, it has been ap-
plied to search for buried metals at hazardous waste sites. The
proton precession magnetometer measures the total intensity of
the earth's magnetic field at one point. The earth's magnetic field,
which is fairly constant over a small site area, is affected by fer-
rous metals. Ferrous metal objects create their own magnetic
fields which change the intensity of the magnetic field over a
site. Using the magnetometer, these ferrous metal objects can be
located.
Several key steps in a magnetometry survey must be observed
to obtain quality data. First, all measurements must be referenced
to a grid system. Second, a point free of magnetic inferences
must be located, known as a reference point. The magnetic field
of the earth changes throughout the day. These changes arc
known as diurnal changes. A full record of diurnal changes must
be kept for corrections of the intensity measurements taken at
grid points during the day. Often it is necessary to have a base
station magnetometer dedicated to this task. These measurements
also can guard against the collection of poor quality data such as
that collected during magnetic storms.
Once correct data are plotted, it is sometimes possible to model
the depth, mass, shape and angle of orientation of an object.
Theoretically, that degree of data interpretation is possible only
under ideal conditions.
Variations on this technique are possible. Gradiometers can be
used to measure the point changes in a field, instruments can be
hand held, sensors can be placed on staffs and instruments may
be towed along traverse lines for continuous readings.
Limitations are critical to this type of survey. Magnetic induc-
tion caused by alternating current in power lines will affect read-
ings. In this case, or in the case of extensive surface metal inter-
ference, surveys should not be performed. Additionally, it is pref-
erable to have a map of shallow buried surface metal from a metal
detection survey to aid in data interpretation. With these precau-
tions in mind, magnetometry can be a useful tool to indicate the
presence of buried ferromagnetic objects.
Metal Detention
Metal detention has been used for years to locate pipes and
other buried metal objects. It relies on the effects of a conduc-
tive metal object or a radio signal which it broadcasts and re-
ceives. The effective depth of penetration is only a few feet, but
it can be used very accurately to locate shallow buried metal tar-
gets.
Electromagnetic Conductivity
Electromagnetic conductance, generally referred to as "EM,"
also responds to the conductive properties of materials. EM, also
called terrain conductivity, usually is used to indicate the presence
of a conductive contaminant plume.
EM does not require ground contact. The EM transmitter coil
radiates an electromagnetic field which induces eddy currents in
the earth. Each of these eddy current loops, in turn, generates a
secondary electromagnetic field which is intercepted by the re-
ceiver coil. An output voltage fis produced which is linearly re-
lated to subsurface conductivity.
Continuous survey instruments are available with transmitter
and receiver fixed in the same instrument, but their use is limited
to shallow depths. Instruments are available with separate trans-
mitter and receiver coils for deeper surveys. Initially, several EM
soundings are made by varying the intercoil spacing at each loca-
tion. After a determination of background conductivity and
appropriate intercoil spacing, EM profiles can be completed
along previously surveyed transverse lines.
Electrical Resistivity
Electrical Resistivity (ER) allows a measurement of subsur-
face resistivities in soil, rock and groundwater. Application of the
method requires that an electrical current be introduced into the
ground through a pair of surface electrodes. The resulting poten-
tial field is measured at the surface between a second pair of elec-
trodes. The electrodes can be arranged into several different con-
figurations.
Initially, several ER soundings can be made to establish sub-
surface background conditions and determine the appropriate
electrode ("A") spacing for the ER profile lines. The sounding
data will be collected by establishing a central reference point
and varying the spacing between the current and potential elec-
trodes.
ER profiles then can be assembled to indicate areas of low re-
sistance, indicating the presence of potential contaminant plumes.
Areas of low resistivity also could indicate perched water zones
or low lying topographic areas.
ER is best used, as with most geophysical techniques, in con-
junction with other indirect testing. Soundings always should be
done to determine the depth to the water table and depth of any
clay layers that might be present. ER is a quick and inexpensive
technique generally available for site investigations.
Ground Penetrating Radar
Ground penetrating radar (GPR) is a reflection technique in
which high frequency radio waves are radiated downward into
the subsurface and then reflected back to a receiving antenna.
Variations in the return signals occur when subsurface materials
have different electrical properties. An interface between two lay-
ers having sufficiently different electrical properties will show up
in the radar profile. The effectiveness of GPR can be limited by
the penetration depth of the radio waves. The effective pene-
tration is highly site specific and dependent upon subsurface
boundaries, water content and clay content.
The radar system electronics usually are mounted in a van, with
the antenna towed behind. An impulse radar transmits electro-
magnetic pulses of short duration into the ground from a broad-
band antenna. Pulses from the antenna are reflected from vari-
ous interfaces in the subsurface and then are picked up by the re-
ceiver section of the antenna.
Unfortunately, GPR systems are very expensive and usually
are used to locate contaminant plumes which are lighter than
water.
PORTABLE ANALYTICAL INSTRUMENTS
Hazardous materials emergency response has spawned the de-
sign of many field analytical instruments due to the added safety
requirements inherent in the performance of such dangerous
work. Many of these instruments have been readily adapted for
use in site investigations to collect data which can be used in the
assessment of a site. These instruments can be used either to
gather data directly or as a screening technique to indicate where
further investigation is needed. This section of the paper dis-
cusses the more common types of instruments which can be used
to gather field data and their uses and limitations.
Organic Vapor Analyzers
Two general types of instruments are currently used to analyze
for the presence of volatile organic compounds. Both detectors
rely on the ionization potential of the organic compounds. Flame
ionization detectors (FID) rely on a flame which burns com-
pounds to break chemical bonds to release energy. Photoioniza-
tion detectors (PIDs) rely on ultraviolet light which is capable of
breaking the chemical bonds to release energy. The major differ-
ence in their application is that methane and some very light or-
MONITORING
-------
ganic compounds cannot be ionized by ultraviolet light. As a re-
sult, PIDs can be used only if methane concentration is not a con-
cern. If qualification of methane is a concern, a flame ioniza-
tion detector should be used to measure its presence.
PIDs, such as the HNU portable photoionizer, and FIDs, such
as the Century Systems (Foxboro) Organic Vapor Analyzer
(OVA), have been used in hazardous waste site investigations to
survey for the presence of organic compounds. Since compound
separation and identification are not possible, total concentra-
tions are measured as the benzene response equivalent. When
used in the survey mode, the OVA is an efficient and accurate
indicator of total organic compound concentrations on a con-
tinuous sampling basis with a response time of one to two
seconds. If instantaneous information is required on the presence
of methane and non-methane hydrocarbons, both instruments
can be used in tandem. An example of this application is a meth-
ane-vinylchloride gas mixture which can be common at landfill
vents.
In cases where compounds identification or quantification is
required, portable gas chromatographs can be used. The Model
128 OVA is a portable FID gas chromatograph. The OVA has an
injection port which enables a sample to be injected from the air
intake line or by a syringe. As an initial approach to sample eval-
uation, this technique can quantify the amount of volative organ-
ics in a particular sample. This technique is useful when screen-
ing soil or water samples to select screen placement or selecting
worst case samples for laboratory analyses. This type of field
screening can lead to significant savings. If a gas chromatograph
(GC) column is present, separation of each of the individual com-
ponents in a mixture is possible for identification and quantif-
ication. The OVA provides data in the 1 ppm concentration
range.
Low ppb concentrations can be measured with a Photovac
portable PID gas chromatograph. Several Photovac models cur-
rently are available. The most advanced is the model 10S70 which
has two columns, is fully programmable, has a 100-compound
memory with computer peak search and integration and an inter-
face to allow communications with a home-base computer. This
type of instrument can separate non-methane organic com-
pounds, identify them and perform integration to provide accu-
rate concentrations. When used with prepared standards and lab-
oratory analytical checks, the Photovac can provide confident
analytical analysis on-site in a matter of minutes without the ex-
pense of an on-site laboratory.
Applications of Portable Analytical
Instruments
Portable analytical instruments typically are employed in sev-
eral different approaches during site investigations. One approach
is an investigation of the lateral of extent of shallow volatile
organic soil contamination. This is commonly done by opening a
hole in the ground with a slam-bar or other tool and inserting the
probe of a survey instrument to measure the concentration of
total organics. If measurements are referenced to a grid, they can
provide an indication of areas to be sampled for analytical analy-
ses or remediation.
Another common use is to screen samples taken during an in-
vestigation. To prevent costly analyses of every sample, only
certain samples are sent to the laboratory based on the sample
screening. A small amount of sample is placed in a VOA vial and
heated to volatilize the organics into the air, or "headspace," in
the vial. An aliquot of air is removed by a syringe and injected
into a GC column for analysis. This technique can also be used to
assist site geologists in deciding the depths at which to finish wells
designed to monitor for volatile organic contamination.
The photovac is gaining widespread use as a quantitative ana-
lytical tool to monitor drilling discharge water to prevent releases
of contaminants into the environment.
Portable units to scan for concentrations of trace metals are ex-
pensive and relatively insensitive. They can be useful for sites with
widespread metals contamination in the 100 ppm and up range.
The Colombia X-Met 700 portable x-ray fluorescence (XRF)
spectrophotometer recently has been used successfully for this
application. Although it is expensive, the portable XRF currently
is gaining acceptance as a field analytical tool.
Other direct-reading instruments which could be adapted from
safety-oriented tasks to portable analytical data gathering tools
include scintillometers, mercury vapor analyzers and Hydrogen
Sulfide indicators.
In addition to safety related instrumentation, some field ana-
lytical tools have been adapted from the laboratory and have been
around for some years. They include pH meters, conductivity
meters, specific ion probes for the measurement of dissolved
metals and colorimetric test kits for indicator parameters such
as chlorides and COD.
The advantages of using these techniques are time and cost
savings. Field analyses provide data rapidly and allow data to be
factored into an investigation while work is still taking place.
Plans can be developed concurrently for further investigation.
The number of field analyses need not be limited in these situa-
tions. A limited number of samples can be selected and analyzed
by the laboratory for confirmation of contamination, saving ex-
pensive laboratory time and reducing lab costs. Scheduling and
costs are crucial on any project but are paramount at sites re-
quiring hydrogeologic investigations or soil investigation and ex-
cavation projects.
SUBSURFACE SAMPLING EQUIPMENT
A large portion of the effort devoted to any site investigation
should be toward understanding the geologic and hydrologic
properties of the soils or rock, the extent of soil contamination
and the potential for contaminant migration through the unsat-
urated zones. Since most hazardous waste site investigations in-
volve surface or near surface disposal of liquid and solid ma-
terials, one of the primary questions that needs to be answered is
how much of the material remains in the soil and how much con-
tinues to leach. Source investigations involve the combined appli-
cation of surface geophysics and direct soil sampling.
Additionally, contaminant migration potentials within the un-
saturated zone and upper water table are important issues to be
addressed. This determination requires the collection and analysis
of subsurface soil and rock samples. Many methods are avail-
able for advancing boreholes to obtain samples or details of
strata; the selection of a method is dependent on the extent of
potential contamination, the complexity of the site geologic con-
ditions and the risk of contaminant migration. The principal
methods in use are: hand augers, power augers and rotary core
drilling methods.
Augers
The auger technique is useful as a tool to collect preliminary
samples, to use with an OVA, HNU or Photovac in a soil gas
survey or to collect near surface waste samples. At least six types
of light portable augers are available for sampling soft to stiff
rolls.
Hand augers may be used by one or two people. The hole is
advanced by pressing down on the cross bar as the tool is rotated.
Once the auger is full or has collected sufficient material, it is
brought back to the surface and the soil is removed. The most
commonly used augers are the post hole, the helical and the spiral
augers.
MONITORING
-------
Hand augering usage is commonly limited to 3 to 7 ft. In tills
or clays that contain gravels or cobbles, hand augering may be
impossible. In uncemented sands or gravels, it will not be pos-
sible to advance the hole below the water table since the hole will
continually collapse. Only samples of very limited size can be ob-
tained from the hole.
In addition to collecting shallow subsoil samples, hand augers
also are used to create small diameter holes for a soil gas survey.
The holes are installed based on a grid system and the probe of an
OVA, HNU or Photovac is inserted into the hole. A sample of
the gas is collected and analyzed. This technique is a useful initial
method to define the extent of soil contamination by volatile
organic compounds.
Because hand augers are limited by soil conditions, depth of
penetration, quantity of sample, size of borehole and thickness of
the unsaturated zone, they usually are confined to preliminary
surveys intended to generally define the limits of the contamina-
tion.
Power Augers
The most common power augering procedures used at haz-
ardous waste site investigations are the continuous flight solid
stem and hollow stem augers. Another type of augering tech-
nique is the bucket auger. This procedure uses open-topped
cylinders with base plates which have one or two slots reinforced
with cutting teeth which break up the soil and allow it to enter
the bucket as it is rotated. However, because of the limited depth
of penetration and the expense involved with operating a bucket
auger rig, it is rarely used.
Solid stem augers drill much deeper holes with fewer problems.
However, this technique presents a serious problem when sub-
surface soil samples are required. To obtain the sample, the auger
string must be pulled from the borehole. Unless the soils are
stiff clays or silts above the water table, the borehole is likely
to collapse and any sample obtained would not be representative
of the stratum.
Hollow stem augers consist of 3- to 6-in. diameter pipe with
continuous flights attached to the outside. Auger tailings are re-
moved from the hole by traveling up the flights and emerging at
the surface. When samples of the soil in advance of the augers
are required, the drilling is stopped and a sampling device is in-
serted into the pipe. The device is gently lowered to the bottom
of the hole and then either driven or pressed into the soil ahead of
the augers. Hollow stem augers allow drilling and sampling below
the water table, the acquisition of samples without pulling the
drill stem and the collection of "undisturbed" samples.
Auger rigs are mounted on four wheel drive trucks and all ter-
rain vehicles, allowing access to most locations. Auger rigs will
not drill through rock and certain cobbly or till soils. Augers
usually are restricted to 200 or 250 ft in depth and still require cas-
ing if the hole is to remain after auger removal.
Rotary Core Drilling
The most common use of rotary drilling in site investigations
is to obtain intact samples of rock. To do this, a "core-barrel"
fitted with a "core-bit" at its lower end is rotated and grinds
away an annulus of rock. The stick of rock in the center of the
annulus passes up into the core-barrel and is subsequently re-
moved from the borehole when the core-barrel is full. The length
of core drilled before it becomes necessary to remove and empty
the core-barrel is termed a "run."
In its simplest form, the core-barrel consists of a single tube
with an abrasive lower edge which is rotated against the rock
while fluid is passed under pressure. However, most rock drilling
in the United States utilized the double tube swivel type core-
barrel. The importance of this tool is that it contains an inner
barrel connected to the outer tube at the top via a swivel which
allows the inner barrel to remain stationary while the outer barrel
is rotated. The rock core "rides up" into the inner barrel.
Double tube rock core barrels normally obtain a core between
5 and 10 ft long. Rock cores should be at least 2 1/8 in. diameter
or larger. In soft to moderately hard rock, cores 12 in. in diameter
or larger can be taken. A core barrel that collects a 2 1/8-in.
diameter rock core will produce a 3-in. borehole.
Deep rock coring can be expedited with the use of the "wire-
line" technique. A wireline can produce rock cores without the
removal of drill rods on the outside core barrel. The inside core
barrel is retrievable as a separate unit with the use of a method
cable, or "wireline." An empty barrel can be sent down on a
messenger to continue coring. This type of technique saves time
and money for deep continuous coring operations.
Core barrels should be held horizontally while the cores are ex-
tracted onto a rigid surface. The core should be properly tagged
and then placed in a core-box. Wooden spacer blocks should in-
dicate the top and bottom of each run.
Soil Sampling Devices
The primary purpose of soil augering is to collect subsurface
samples for visual examination and chemical and physical tests.
In unconsolidated material capable of being drilled with an auger,
an important component of any investigation should be the
collection of soil samples ahead of the drill bit. Hollow stem
augers are the most common drilling technique, but soil sam-
ples also can be collected ahead of drilling while using hydraulic
rotary methods.
A large number of samplers are available, most adapted from
geotechnical or soil survey investigations. However, the most
common device used in conjunction with hollow stem augering is
the split spoon sampler. In this device, the sampler barrel is split
longitudinally into two halves. The device is lowered into the
inner space of the hollow stem auger and driven into the soil by
repeated blows of a 140 Ib hammer falling through 30 in. Dur-
ing driving, the longitudinal halves are held together by the shoe
and head which are screwed into each other. The split barrel
allows easy examination and extraction of soil samples.
Soil obtained with the split barrel sampler can be subject to a
wide variety of field and laboratory tests. Samples can be field
screened with an OVA to determine presence of volatile organic
compounds. Soil should be logged by describing its texture,
color, grain size distribution, moisture content and odor. If the
split spoon has been properly decontaminated before sampling,
soil can be obtained for chemical analysis.
Occasionally, large diameter "undisturbed" soil samples are re-
quired for complex physical tests such as hydraulic conductivity
determinations. Shelby or push tube samplers often are used to
obtain undisturbed samples of medium to stiff consistency co-
hesive soils. The sampler is pressed into the soil. Care should be
taken because the sampler can be damaged, either by buckling or
blunting or tearing the cutting edge, when driven into very stiff,
hard or stony soils. Undisturbed samples of soft or loose soils
must be taken with a piston sampler. Undisturbed samples con-
taining much gravel or soft rock must be obtained with a Deni-
son sampler.
SURFACE WATER/SEDIMENT SAMPLING
One of the primary waste migration pathways is via surface
runoff to rivers, streams and lakes. The contaminant may move
as a dissolved constituent in the runoff or attached to entrained
soil particles. One of the most cost-effective means of determin-
ing the extent of waste migration is to sample surface water and
sediment.
MONITORING
-------
Surface Water Sampling Techniques
Surface water sampling locations are selected on the basis of
their probability for showing contaminants migrating from the
site. Prior to sampling, the surface water drainage in and around
the site must be characterized using all available background
maps, topographic maps, serial photographs, river basin surveys
and other sources.
Either grab or composite samples may be collected. Grab sam-
ples are collected at one particular point and at one time. Flow
or time-weighted composited samples are composed of more
than one specific aliquot collected at various sampling sites and/
or at different points in time. Because of the unknown safety
risks, as well as the changes in chemical nature of the sample
that may occur through compositing, samples containing haz-
ardous materials at significant concentrations shall not be com-
posited.
If it is necessary to enter the water to obtain the sample, it
should be done carefully to leave the bottom sediments undis-
turbed. If the water is moving, a grab sample should be obtained
by pointing the open end of the container into the direction of the
flow. An attempt should be made to obtain the sample in the
middle of the stream and at mid-depth.
The choice of a particular sampling device is dependent upon
the size of the water body, the purpose of the sample and the
types and concentrations of wastes anticipated to be in the water
body. The types of samplers available include open tube, pond
sampler, manual hand pump, weighted bottle sampler, kemmerer
sampler and extended bottle sampler. Of these, the pond sampler
and weighted bottle are used most often.
The pond or dip sampler consists of a container attached to the
end of a long pole by an adjustable clamp. The pole can be of any
non-reactive materials such as wood, plastic or metal. The sam-
ple is collected in a jar or beaker made of stainless steel, glass
or non-reactive plastic. Preferably, a disposable beaker which can
be replaced at each station should be used.
The weighted bottle sampler consists of a glass bottle, a weight
sinker, a bottle stopper and a line that is used to open the bottle
and to lower and raise the sampler during sampling. The sam-
pler can be either fabricated or purchased. This sampler is used to
take discrete samples at predetermined intervals. The pond sam-
pler can be used to develop composite samples from a single loca-
tion by collecting individual samples at regular intervals.
Sediment Sampling Techniques
Sediment samples are valuable for locating pollutants of low
water solubility and high soil binding affinity. Heavy metals and
high molecular-weight halogenated hydrocarbons are examples
of contaminant groups which might be found in greater con-
centrations in sediment. A background sediment sample should
be obtained from sediments upstream from the suspected source
for comparison. This is especially important if contamination
with heavy metals is suspected, because they occur naturally.
Very simple techniques usually are employed for sediment
sampling. Most samples will be grab samples from one particu-
lar locations although, for preliminary studies, several locations
may be composited to reduce the analytical requirements. Sug-
gested techniques include:
• In small, low-flowing streams or near the shore of a pond or
lake, the sample container (typically an 8-ounce wide-mouth
glass jar) may be used to scoop up the sediments.
• TO obtain sediments from larger streams or farther from the
shore of a pond or lake, a Teflon beaker attached to a tele-
scoping aluminum pole by means of a clamp may be used to
dredge sediments.
• To obtain sediments from rivers or in deeper lakes and ponds,
a spring-loaded sediment dredge or benthic sampler may be
used by lowering the sampler to the appropriate depth with a
rope. The sediments thus obtained are then placed into the
sample container.
DOWNHOLE TECHNIQUES
Geophysical borehole logging has been used for oil, coal and
mineral exploration for years. Recently, these techniques have
gained widespread acceptance in the performance of hydrogeo-
logic investigations. There are many different techniques avail-
able. Lithologic logging can be used to measure properties of the
rock or soil in a borehole or well. Hydrologic logging can be used
to measure the properties of the fluids in a borehole. Downhole
television also can be used to examine downhole conditions. A
review of these techniques and their applications is presented
below.
Lithologic Logging
Lithologic logging is used to measure the properties of down-
hole soils and rock. As with all logging techniques, it is first ad-
visable to log at least one hole by physical examination of either
split spoon samples from soils or rock core samples for bedrock.
Rock core samples are more useful than chips because fracture
occurrence and widths (which are important conduits of ground-
water flow in bedrock) can be determined. Once lithology of at
least one hole is known, the information can be correlated to
other holes.
The simplest of the lithologic techniques is caliper logging.
The caliper tool measures the diameter of a rock borehole using
three spring loaded prongs. As the tool is pulled up the hole, the
prongs spread out indicating the width of the borehole. Results
ared recorded electrically at the surface. The caliper method in-
dicates the relative hardness of rock units. Caliper logging also
can indicate fracture location and occurrence. Caliper logging can
be the best tool to examine the presence of bedding plane frac-
tures.
Spontaneous Potential Logging
Spontaneous potential logging is one form of electric logging.
It serves to measure the natural electrical potential that develops
between the formation and the borehole fluids. This logging
technique must be performed in an open borehole filled with
fluid. The logging device consists of a surface electrode and a
borehole electrode with a voltmeter to measure potential.
Generally, SP logs are read in terms of positive and negative
deflections from an arbitrary base line which might correlate
either with permeable or impermeable zones. Information re-
garding zones of higher permeability will indicate likely contam-
inant pathways that may require further investigation. The rate
at which the particular logging tool is lowered into the borehole
is another important criterion in the data evaluation. The rate
should be sufficient to detail specifics of the geology and pos-
sibly detect isolated zones of contamination or likely contam-
inant pathways.
Resistivity Logging
Resistivity logging is another electric logging technique. Two
types of resistivity logging techniques are commonly employed
in downhole methods. The single point resistance log is the sim-
plest form of electric log. In this method, a single electrode is low-
ered into the hole and the return path for the current flow is furn-
ished by the ground electrode. The single point resistance log
measures the total resistance of earth materials.
MONITORING
-------
The second type of resistivity logging involves measurements
taken in the borehole in a similar manner as surface resistivities
investigations. In this system, four electrodes are commonly em-
ployed, two for emitting current (I) and two for potential meas-
urement (P). A number of different electrode configurations can
be used in resistivity logging to provide specific information. The
short normal spacing indicates the resistivity of the zone close to
the borehole where the drilling fluid might be an influence. The
long normal spacing has more distance between the electrodes
and thus measures the resistivity further away from the borehole,
presumably beyond the influence of the drilling fluid. Both con-
figurations, short and long-normal, measure a greater radius of
influence than the single point resistance.
A third configuration involving lateral devices utilizes widely
spaced electrodes for measuring zones that are far from the bore-
hole. Because of the wide spacings, lateral devices will not de-
tect thin beds of different resistivity. Boundaries of formations
having different resistivities are located most readily with short
electrode spacing, whereas information on fluids (e.g., contam-
inated groundwater) in permeable formations can be obtained
best with long spacings.
Density Logging Techniques
Many downhole techniques are available to measure the den-
sity of rock units. Natural gamma ray logging is perhaps the most
common. Natural gamma logging measures the quantity of
gamma rays naturally emitted by radioisotopes contained in clay
minerals. The higher the clay content of a rock, the more gamma
rays will be emitted. This information is useful in indicating the
permeability of rock units, since lower permeability units gen-
erally contain more clay size particles.
Other techniques which can be used to indicate density are high
resistivity density, gamma-gamma density and neutron logging.
As with natural gamma ray logging, these techniques can be run
in cased boreholes. In other words, if the geology at a well needs
to be investigated and no records exist, a cased hole can be logged
to gather qualitative information. Natural gamma ray logging
can be especially useful to locate the depth and thicknesses of
bentonite seals on grout packs in monitor wells.
Vertical Seismic Profiling
Recently, some success has been met utilizing vertical seismic
profiling, or "cross-hole seismic." This technique utilizes an
array of detectors set in a borehole to produce a three-dimen-
sional image of subsurface structures such as fractures. A seismic
source is used on the surface at some distance from the borehole
to create the seismic movement measured by the detectors. Al-
though this technique is still under investigation, it shows promise
for hydrogeologic investigation in the area of the development of
three-dimensional data.
Hydrologic Logging Techniques
The two most common hydrologic borehole logging techniques
are fluid conductivity and temperature logging. These open bore-
hole techniques examine the properties of fluids within a bore-
hole and can provide important information on the natural cir-
culation of fluids in a borehole or the presence of contamina-
tion.
These two probes are commonly available together and can be
run as a hydrologic suite. Fluid conductivity logs provide a con-
tinuous record of the conductivity or resistivity of fluid in the
borehole, which may be related to the conductivity of fluids in the
adjacent formation. A temperature log made simultaneously with
a fluid conductivity log allows the most accurate conversion to
specific conductance and also may identify contaminated or more
permeable zones in the formation.
Fluid conditions in the boreholes should stabilize prior to im-
plementing this technique. The longer the period of time between
the drilling of the borehole and the logging of the hole, the more
accurate the results will be. Hydrologic logging can detect geo-
thermal gradients in groundwater in holes on the order of 100 ft
thick. Fractures contributing to groundwater movement some-
times can be noted by changes in the geothermal gradient.
Changes in temperature or conductivity also may be an indication
of contamination.
Downhole Television
Downhole television has been utilized successfully to gather
in-situ information on boreholes and wells in several ground-
water monitoring programs. Borehole television surveys are a
viable alternative to other downhole instruments in the subsur-
face investigation stages of a groundwater monitoring program.
Miniature borehole television cameras, developed for use in the
examination of nuclear reactor cores, have been modified for
use in borehole investigations. The lens attachments are capable
of looking sideward or downward and include built-in lighting
assemblies.
The in situ characterization of fractures that can provide path-
ways for contaminant migration is critical in some investiga-
tions. Borehole television inspection can provide information on
the frequency, size and orientation of these fractures. Vertical
correlations of rock cores in areas where voids are present (i.e.,
deep mining or karst topography) also can be simplified by this
technique. Borehole television can be used to check monitoring
well integrity. Casing inspections are especially useful for con-
struction inspections when construction details are not known.
Well screens may be inspected in place to determine if rusting has
enlarged the screen openings or if screens have been damaged dur-
ing emplacement or well development operations. This informa-
tion may be invaluable in the decision to decommission a well.
This technique is quick, inexpensive and creates a permanent
record for potential court cases.
Packer Assemblies
Downhole packers were developed for in situ measurements of
permeability in geotechnical investigations such as dam building
where data are critical. The packer assembly consists of two sets
of packers, upper and lower, separated by a perforated pipe.
Each packer is surrounded by a rubber seal which is fastened
to the assembly. Packers are inflated with compressed air from
the surface to seal against the sides of the borehole. Particular
zones then can be isolated for testing.
For in situ permeability testing, water is pumped in at a low and
constant pressure, creating a constant artificial head. Operating
pressures are stabilized with a gate valve until they remain oon-
stant. Pressures then are checked with a pressure gauge and re-
corded for later use in permeability calculations. Each zone is
tested for a defined period of time, and the amount of water re-
ceived by the zone is recorded with a standard water meter.
Another recent development is the insertion of electric or blad-
der pumps inside the assembly to remove water from the bore-
hole. A particular zone can be isolated and tested with a portable
GC for identification of zones carrying volatile organic contam-
ination. Screens then can be set at optional levels for the study of
groundwater contaminant migration.
SPECIAL EQUIPMENT
Equipment which was not discussed in this manuscript includes
air sampling equipment. Generally, air sampling programs are
highly specific. Particulates such as asbestos can be collected on
filters by high volume air samplers, while organics can be col-
lected in packed tubes for thermal desorption by the use of low
MONITORING
-------
volume air samples. Air investigations generally include the use of
continuously reading portable meteorologic stations.
Air sampling technology recently has been extended to plume
tracking through the analyses of soil gas. Tubes are placed in the
ground, usually for several days. Organic compounds which vol-
atilize from groundwater or soil travel upward over time in the
unsaturated zone of the soil and collect a trace levels in the tubes.
The tubes then are removed, thermally desorbed and analyzed
for volatile organics. The technique is not yet accepted as a stan-
dard methodology, but has been useful in many cases and cur-
rently is gaining wide acceptance.
Groundwater pumps continue to be refined, especially dedi-
cated systems. These systems include small diameter electric sub-
mersible pumps and bladder pumps. Materials used in the pumps
are being researched. Many bladder pumps currently available
for purchase are built almost completely of Teflon.
Technological developments continue in the area of contin-
uous measurement devices for groundwater levels in monitor
wells. These devices range from single well units with solid-state
memories to full-scale 16-channel devices. Some multi-channel
devices can be operated up to one mile away from a well, making
them useful for aquifer tests. Some actually have built-in com-
puters, disk drives, monitors and printers and can analyze aquifer
test data, keep records on each well and even generate reports.
CONCLUSIONS
The general structure of an investigation requires an analysis of
data gaps, particularly with regard to source definition, pathway
examination and assessment of potential for contaminant migra-
tion. Cost and time constraints involved in collecting certain types
of data can be balanced against the usefulness of the results. For
instance, it sometimes may be less expensive to install additional
wells than to use a costly indirect geophysical technique.
Investigations generally rely on indirect investigatory tech-
niques to start, followed by necessary confirmation steps. Spe-
cifically, a technique such as magnetometry can be used to search
for buried containerized wastes, but this generally is confirmed
through test pitting and subsurface waste analyses. It is impor-
tant to note, however, that the limitations inherent in some tech-
niques may preclude their use in certain situations. A knowledge
of these limitations is critical to cost-effective investigations and
proper cleanup strategies.
Knowledge of indirect measurement techniques and field ana-
lytical equipment is critical to the collection of proper data and
the control of costs. Innovative technologies constantly are being
considered for their usefulness in these investigations. Use of this
type of equipment to paint a complete picture of site conditions
can help the environmental management professional reduce
long-term liability and provide useful data for risk assessments
and feasibility studies. We hope that we have provided a basic
overview of the equipment, its limitations and its usefulness to
environmental site investigations.
MONITORING
-------
A Statistician's View of Groundwater Monitoring
Douglas E. Splitstone
IT Corporation
Pittsburgh, Pennsylvania
ABSTRACT
Owners and operators of hazardous waste facilities are required
to monitor the groundwater around the hazardous waste facility
and perform prescribed statistical tests on the resulting data to
detect the existence, if any, of groundwater contamination. The
statistical test prescribed by the U.S. EPA, which is referred to as
Cochran's Approximation to the Behran's-Fisher Student's t-test,
has received increasing criticism as being inappropriate and
resulting in the allegation of groundwater contamination when, in
fact, contamination did not occur. In addition, its specified ap-
plication requires the existence of a well that clearly monitors
groundwater upgradient of the site. The U.S. EPA recognized
these problems. In publication of its regulations for owners and
operators of permitted waste facilities, the U.S. EPA indicates the
admissibility of statistical procedures alternate to the Student's
t-test.
The inadequacy of the Student's t-test is in part due to the lack
of recognition that background groundwater quality varies both
spatially and temporally. This paper describes a statistical model
appropriate for describing spatially and temporally varying back-
ground water quality. An alternate statistical test procedure for
assessing groundwater contamination as a deviation from this
background model is proposed. Use of this test procedure permits
the owner/operator to control the risk o' "false positive" test
results to a small specified probability. Thus, the risk of unfairly
being required to conduct an expensive groundwater quality
assessment program is controlled while assuring protection
against groundwater degradation.
CURRENT GROUNDWATER
MONITORING PROCEDURE
A sampling program to determine whether a particular hazar-
dous waste facility is contaminating groundwater should be
designed so that any contribution from the particular facility can
be distinguished from other sources. These other sources, which
may be both anthropogenic and natural, provide what is loosely
referred to as the background. The design of a sampling program
to distinguish the contribution of the site of interest from the
background is complicated by temporal and spatial variations in
the background.
Temporal variations in groundwater may, in part, be explained
by recharge rates which are weather related. Spatial variations
among wells may be due to a variety of reasons. Some of this
variability may be reduced by ensuring that the wells monitoring
the aquifer beneath a particular waste disposal site are located
relatively close together and drilled to the same depth. Neverthe-
less, spatial variations may occur even if there is no contribution
from the site being monitored. The variations might be expected
to be horizontal gradients across the site.
The Behrens-Fisher Student's t-test recommended by the U.S.
EPA for monitoring the groundwater near possible sources of
contamination1-2 has increasingly become the topic of critical
reviews.3-4-5 The U.S. EPA procedure is based on establishment of
a background level before the disposal facility is put into opera-
tion and is, therefore, insensitive to spatial and temporal varia-
tions in the background. It can be used, however, if variations in
the background levels are taken into account.
Considering the hypothetical situation in which measurements
of groundwater quality are obtained without sampling or
measurement error, the sample variance of upgradient well
measurements, SB2, reflects only real background variations.
Measurements on subsamples from the sample taken from a
downgradient well all have the same value Xm. The test statistic
given by U.S. EPA, which is referred to as Cochran's approxima-
tion to the Behrens-Fisher Student's t-test,' reduced to:
(xm ' XB)/(SB2/MB)1/2 (1)
where XB is the mean of the background values and NB is the
number of background values. This test is inappropriate because
in the absence of contamination, Xm can vary as widely as the
background values. The test is appropriate for the inappropn tte
null hypothesis that Xm equals the mean of the background
values. The appropriate null hypothesis for testing is that Xm is a
member of the same population as the background values. The
appropriate statistic for this test is:
Compared to the first test statistic, the estimate of the standard
deviation used in the second is considerably larger. Therefore, for
many sets of measurements, while the first test statistic may er-
roneously indicate that there is contamination, the second test
statistic may indicate the opposite.
The U.S. EPA has recognized this problem. In publication of
its regulations for owners and operators of permitted waste
facilities, the U.S. EPA acknowledges the admissibility of
statistical procedures alternate to the Students t-test.1 An alter-
nate statistical procedure is described below.
PROPOSED ALTERNATE MONITORING
DATA EVALUATION PROCEDURE
Although the background quality of groundwater varies both
spatially and temporally, temporal variations in background at
MONITORING
-------
nearby points are expected to be related. Extending this assump-
tion, temporal variations (apart from sampling and measurement
error) are the same at nearby points. This procedure leads to the
expression of a mathematical model for the background quality
of groundwater around a waste disposal site. In this model, the
background quality is given by the sum of a general mean (/*), a
spatial effect (Wi) and a temporal effect (Tj).
If contamination of the groundwater occurs from the waste
disposal facility, and if the monitoring wells surround the site so
that they are located both upgradient and downgradient from the
site, then contamination is unlikely to change the levels of a con-
stituent in all wells by the same amount. Thus, contamination
from a disposal site represents a deviation from the background
model, and such a deviation can be detected.
Monitoring groundwater quality around a disposal site is an ex-
periment with two factors: a spatial factor (or well factor) and a
temporal factor. In conducting such an experiment, samples are
collected and measurements made at each well at each sampling
time. If the background model holds, these measurements,
denoted by Yij, can be represented by:
M + W.
.
(3)
= 1
= 1
where N is the number of wells, M is the number of sampling
times and Ejj represents the random sampling and measurement
error. The measurements can be arranged in a two-way table and
can be analyzed by a two-way analysis of variance.6
Estimates of n, Wi and Tj may be obtained from the data.
Denoting these estimated by m, w; and tj, respectively, they are
given by:
(4)
(5)
(6)
The differences between the observed values, yij, and the values
predicted by the model are the residuals
rij = Xij - wi - tj - m (7)
As mentioned above, the contamination of groundwater by a
waste disposal site will result in a deviation from the background
model. Therefore, it is the residuals which contain the answer to
the question of whether the site is contaminating the ground-
water. However, the residuals also are affected by the sampling
and measurement error. Because the assumed background model
does not indicate whether the residuals occur as a result of sam-
pling and measurement error or are indicative of contamination,
the model is incomplete.
The analysis of groundwater data can proceed no further
because there is no way to decide whether the residuals from the
background model can be attributed solely to sampling and
measurement error. The solution to this problem is the collection
of replicate samples. Replicate samples differ from each other
only because they each contain an independent replicate of the
sampling and measurement error. For example, a sampling design
for groundwater monitoring might specify collection of two
replicate samples from each well at each sampling time. Replicates
are important in environmental sampling because the adequacy of
environmental models is always in question.7 Replicates allow the
variations unexplained by the model to be compared with the
variations caused by sampling and measurement so that model in-
adequacies can be detected.
The exact specification of the replicates that are needed is a
complex process since it depends on the environmental model.
For this reason, replicate laboratory measurements obtained by
splitting samples are rarely adequate because they do not reflect
sampling error. Specification of replicates often involves choosing
the time period between successive replicate samples. The model
should account for variations over long time spans with the term
Tj. Sampling error usually accounts for local and short-term
variations.
Assume that true replicate samples can be obtained by resam-
pling the wells after a specific short time period. Consider a sam-
pling program design which requires L replicate samples from
each well at each occasion that the wells are sampled. In most
cases, two (L = 2) replicates seem reasonable. The measurements
on the samples may then be described by the following model:
(8)
Tj
Eijk
where, as before,
i = 1 . . . N (number of wells)
j = 1 . . . M (number of sampling periods)
k = 1 . . . L (number of replicates)
This model is the same as Model (3) except that another
subscript has been added to index the replicates and a term (WT)jj
has been added to account for the various changes that occur
when the site is contaminating the groundwater. This term is
referred to as the well by period interaction and describes changes
in the relative relationship of the wells between time periods.
If the sampling and measurement errors Eijk are statistically in-
dependent and normally distributed with the same variance, then
measurements obtained under this model can be analyzed by the
two-way analysis of variance appropriate for models with
replicate sampling.6 Estimates of /t, Wi, Tj and (WT)jj, which are
denoted by m, wj, tj and ry, can be obtained as before with yij
replaced by:
The appropriate analysis of variance (ANOV) table is given in
Table 1.
The crucial comparison of the interaction with sampling error
which is indicative of groundwater contamination is made using
the F-statistic.6 The value of this statistic is given by:
NT(L-l) 45rij2 , (7)
F =
(N-D(T-l)
If the interaction is found to be significantly greater than the
sampling error, then one concludes that the wells do not vary in
the same way over time, and thus there is local influence on the
measured contamination.
True replicate samples can be obtained by sampling the same
well after a fixed period. The duration of this period may be as
short as hours or as long as days. The determi.iation of this period
can be established for a given well system with an initial sampling
experiment. Such an experiment has been conducted on the
monitoring wells of a major industrial facility.
DETERMINATION OF REPLICATE SAMPLES
The need for determining what constitutes a replicate sample
has been discussed above. It also was indicated that replicate
analyses of the same sample do not provide for the determination
of sampling variability. To determine true replicate sampling for
the monitoring wells, a sampling experiment was conducted over
a four-day period. This experiment required the sampling of each
well once a day on each of the four days. The samples were col-
MONITORING
-------
Table 1
Appropriate Analysis of Variance (ANOV)
for Groundwater Monitoring
Table 4
Groundwater Monitoring Study TOC (mg/l)
SOURCE OF
VARIANCE
Between Wells
Between Sampling
Interaction
Sampling Error
Total
DECREES OF
FREEDOM
N-l
Periods T-l
(N-lHT-1)
NT(L-l)
NTL-1
Table 2
Groundwater Monitoring Study pH
DAY
ANALYSES
B i
SUM OF
SQUARES
LT'w2
t 1
LN1 t2
1 1
, II 2
L. . r. .
'J ']
"J ("ijk" " ij.)
"i t"ijk m'2
(Standard Units)
WELL
B2 B3 B4
1 1
2
3
4
2 1
2
3
4
3 1
2
3
4
4 1
2
3
4
7.00
7.00
7.00
7.00
6.50
6.60
6.60
6.70
6.60
6.60
6.70
6.80
6.60
6.60
6.60
6.60
7.00
7.00
7.00
7.10
6.90
6.90
6.90
7.00
6.90
6.90
6.90
7.00
6.90
6.90
6.90
7.00
7.30
7.30
7.40
7.40
7.30
7.30
7.40
7.40
7.30
7.40
7.40
7.40
7.30
7.40
7.40
7.40
6.50
6.50
6.60
6.60
6.50
6.50
6.60
6.60
6.50
6.50
6.60
6.60
6.50
6.50
6.50
6.60
Table 3
Groundwater Monitoring Study Conductivity
OtMHOS @ 25 °C)
DAY ANALYSES
1 1
2
3
4
2 1
2
3
4
3 1
2
3
4
4 1
2
3
4
Bl
1039.00
1037.00
1012.00
1021.00
1062.00
1074.00
1068.00
1041.00
1080.00
1067.00
1071.00
1064.00
1045.00
1050.00
1033.00
1039.00
B2 B3
913.00
904.00
911.00
917.00
1097.00
1090.00
1087.00
1087.00
1062.00
1060.00
1050.00
1055.00
1011.00
1027.00
1005.00
966.00
599.00
587.00
592.00
599.00
688.00
683.00
686.00
672.00
698.00
696.00
695.00
696.00
712.00
712.00
712.00
717.00
B4
1355.00
1353.00
1265.00
1352.00
1339.00
1332.00
1337.00
1341.00
1339.00
1341.00
1339.00
1340.00
1327.00
1351.00
1345.00
1356.00
DAY
1
2
3
4
DAY
1
2
3
4
ANALYSES
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
Groundwater
ANALYSES
1
2
3
4
1
2
3
4
1
2
3
4
1
2
3
4
Bl
28.40
27.80
27.10
27.90
37.60
33.80
36.80
32.20
35.50
33.40
35.50
33.90
32.20
31.80
32.40
31.70
TableS
Monitoring
Bl
21.00
21.00
22.00
18.00
21.00
21.00
23.00
21.00
11.00
10.00
18.00
16.00
12.00
10.00
11.00
12.00
WELL
B2 B3
102.00
98.80
102.00
98.50
72.60
75.30
75.80
73.90
64.80
64.80
62.40
64.00
64.60
61.70
61.30
62.80
Study TOX
28.80
29.30
28.10
31.30
44.70
42.90
42.70
44.70
40.80
40.30
38.50
40.50
38.90
39.40
38.20
41.70
0*/1>
WELL
B2 B3
29.00
21.00
32.00
29.00
35.00
36.00
33.00
34.00
28.00
31.00
32.00
31.00
26.00
27.00
31.00
28.00
<10.00
<10.00
<10.00
<10.00
11.00
<10.00
<10.00
<10.00
16.00
16.00
19.00
17.00
<10.00
<10.00
<10.00
<10.00
B4
101.00
97.90
98.70
98.20
103.00
100.00
106.00
98.20
102.00
100.00
102.00
103.00
93.40
93.10
97.30
96.70
B4
19.00
16.00
17.00
18.00
19.00
22.00
21.00
25.00
25.00
22.00
25.00
27.00
17.00
21.00
19.00
23.00
lected after three well volumes of water were removed from the
well. Each sample was analyzed four times for pH, conductivity,
total organic carbon (TOC) and total organic halogens (TOX).
The data collected during this experiment are found in Tables 2
through 5.
The sampling plan was designed to compare the variability
among samples from a given well taken within a relatively short
time span (approximately 24 hr apart) to the variability among
repeated analyses on the same sample. The sampling plan is,
therefore, hierarchical, and the data were analyzed accordingly
using the appropriate ANOV technique.' The results of these
analyses are found in ANOV Tables 6 through 9.
The values of the F statistics shown in these tables are all
greater than the tabulated critical values of F for the appropriate
degrees of freedom.* They clearly indicate that highly significant
variability exists in all the measured constituents among the wells
and among samples from the same well during a comparatively
short period of time. The significance of variability along the
wells is as expected for reasons given above. The significance of
10
MONITORING
-------
the variation among analyses of a given sample clearly illustrates
that sampling variability is important.
To further illustrate this point, assume that the four samples
taken from Well Bl describe the background groundwater quality
for TOX. Thus, the background mean TOX is 16.19 /ig/1, and
standard deviation of the background TOX is 6.11 /ig/1. Now
assume that downgradient Well B4 was to be samples on one of
the four days of this experiment chosen at random. The mean and
standard deviation of the four repeated analyses for TOX of the
daily samples is given in Table 10. Also presented in this table are
the calculated and critical values of the Student's "t" statistic as
prescribed by Appendix IV of 40 CFR 264.
Comparing the calculated and critical values of the Student's
"t" statistic for these four consecutive days serves to illustrate the
importance of the sampling variability. On two of the four days
(Day 2 and Day 3), the calculated value of Student's "t" exceeds
the critical value. Thus, a conclusion of groundwater degradation
would be made. However, this is not true of the samples taken on
Day 1 and Day 4. If the quarterly sampling day were picked at
random from these four, there would be a 50% chance of con-
cluding that groundwater degradation occurred solely due to ran-
dom variation introduced by the act of sample collection.
VARIATION
SOURCE
Wells
Samples
Within
Well
Analyses
Within
Sample
TOTAL
VARIATION
SOURCE
Wells
Samples
Within
Well
Analyses
Within
Sample
TOTAL
VARIATION
SOURCE
Well
Samples
Within
Well
Analyses
Within
Sample
TOTAL
Table 6
pH ANOV
DEGREES SUM QF MEA[) p
FREEDOM SO-UARES SQUARES STATISTIC
3 6.0162 2.0054 51.03
12 0.4713 0.0393 12.61
48 0.1500 0.0031
63 6.6375
Table 7
Conductivity ANOV
DEGREES SUM QF MEAN F
™.?«L,., SQUARES SQUARES STATISTIC
FREEDOM
3 3,578,628 1,192,876 127.81
12 111,990 9,333 43.21
48 10,348 216
63 3,700,968
Table 8
TOC ANOV
DEGREES SUM op MEAN p
FREEDOM SQUARES SQUARES STATISTIC
3 48,326 16108.70 44.22
12 4,372 364.32 132.96
48 132 2.74
63 52,829
SIGNIFICANCE
LEVEL
.999
.999
SIGNIFICANCE
LEVEL
.999
.999
SIGNIFICANCE
LEVEL
.999
.999
Table 9
TOX ANOV
VARIATION DEGj>EES SUM OF MEAN F
SOURCE FREEDOM SQUARES SQUARES STATISTIC
Well 3 2,918 972.63 17.26
Sample 12 676 56.36 12.44
Within
Well
Analyses 48 218 4.53
Within
Sample
TOTAL 63 3,811
SIGNIFICANCE
LEVEL
.999
.999
Table 10
TOX Concentration for Downgradient Well B4
DAY
1
2
3
4
MEAN
Ug/l
17.50
21.75
24.75
20.00
STANDARD
DEVIATION
Pg/t
1.29
2.50
2.06
2.58
CALCULATED
STUDENT'S "t"
0.79
2.82
4.65
1.91
CRITICAL
STUDENT'S "t"
1.84
1.99
1.94
1.96
Table 11
Expected Mean Squares for Groundwater Monitoring
VARIATION
SOURCE
EXPECTED MEAN
SQUARES
Well
Samples Within
Well
Analyses Within
Sample
Table 12
Variance Components for Groundwater Monitoring Parameters
VARIATION
SOURCE
Analyses, o.
Sampling , o
PH
0.003
0.009
CONDUCTIVITY
215
2,279
TOC
2.74
90.39
TOX
4.53
12.95
The selection of a sampling plan for continued groundwater
monitoring can be made with the information generated by the
analysis of variance. By equating the mean squares estimated by
the analysis of variance to their theoretical expectations, one can
estimate the contribution of each component (e.g., sample collec-
tion or analysis) to the total variability of the measurement. If a%
and as are used to symbolize the variability due to analysis and
sample collection respectively, then the expected mean squares for
the analysis of variance are given in Table 11.
In Table 11, W2 symbolizes the natural background variation
among wells. The resulting estimates of a2, and a^are given in
Table 12 for the indicator parameters observed.
The characterization of the contaminant concentration for a
well undergoing quarterly monitoring is by the mean of the obser-
vations taken during the quarterly monitoring period. A sampling
plan should be designed to minimize the variation in this mean. If
S samples are taken during the quarterly monitoring period and A
analyses are run on each sample, the variance of the quarter mean
is given by: 2 2
SA
MONITORING 11
-------
The effect on this variance of varying the number of samples,
S, and analysis per sample, A, then can be investigated. If the
analytical cost is constrained by requiring the S + A £.4, then the
minimum variance, and hence maximum precision, of the
quarterly mean value for the indicator parameters for a given well
is obtained for the same analytical cost if the well is sampled four
times each quarter approximately 24 hr apart and one analysis is
performed on each sample. These results are given in Table 13.
The information on the last line of the table indicates that the
precision of the quarterly mean is approximately doubled (the
variance halved) from the U.S. EPA recommended sampling
scheme of one sample per quarter by sampling each well twice and
analyzing each sample once. This reduction in analytical cost may
be sufficient to compensate for increased sampling costs.
Table 13
Variance of Well Quarterly Mean
NUMBER OF
SAMPLES
1
2
4
2
NUMBER OF
ANALYSES
PER SAMPLE
4
2
1
1
0
0
0
0
pH
.00975
.00525
.00300
.00600
CONDUCTIVITY
2332
1193
.75
.25
623.50
1247
.00
TOC
91.
45.
23.
46.
08
88
28
56
TOX
14
7.
4.
8.
.08
61
37
74
CONCLUSIONS
The salient conclusions are summarized as follows:
• The portion of the total measurement variability due to the
act of sampling is much greater than that due to repeated chem-
ical analyses of the same sample. Thus, repeated chemical
analyses of the same sample do not necessarily improve quar-
terly characterization of groundwater.
• Adequate replicate samples may be obtained by repeated sam-
pling of each well approximately 24 hr apart.
• The precision of the quarterly measurement for groundwater
contamination will be greatly improved from that specified in
Federal Regulations if each well is sampled twice within a four-
day time span and only one analysis is performed on each
sample. This sampling scheme also will permit the use of the
analysis of the variance statistical procedure to adequately test
the hypothesis of groundwater contamination.
• Continued use of the statistical procedure specified in the Fed-
eral Regulations will, more than likely, yield results alleging
that groundwater contamination has occurred when the appli-
cation of adequate statistical procedures will conclude that
there is not contamination.
REFERENCES
1. U.S. EPA, "Interim Status Standards for Owners and Operators of
Hazardous Waste Facilities." 40 CFR 265.92.
2. U.S. EPA, "Regulations for Owners and Operators of Permitted
Hazardous Waste Facilities," 40 CFR 264.97.4.ii.
3. Liggett, W., "Statistical Aspects of Designs for Study Sources of
Contamination," Quality Assurance for Environmental Measure-
ments, ASTM STP867, June 1985.
4. Miller, M.D. and Kohoul. F.C., "RCRA Ground Water Monitoring
Statistical Comparisons: A Belter Version of Student's T-Test,"
Proc. of the NWWA/AP1 Conference on Petroleum Hydrocarbons
and Organic Chemicals in Ground Water Prevention, Detection, and
Restoration, National Water Well Association, 1984.
5. Ross, L. and Elton, R., "Maximizing the Statistical Performance of
Ground Water Monitoring Systems," Proc. of the NWWA/APl
Conference on Petroleum Hydrocarbons and Organic Chemicals in
Ground Water Prevention. Detection, and Restoration, National
Water Well Association, 1984.
6. Snedecor, G.W. and Cochran, W.G.. Statistical Methods, 5lh ed.,
The Iowa State University Press, Ames, I A, 1950.
7. Krumbein, W.C., "Experimental Design in the Earth Sciences,"
Transactions of the American Geophysical Union, 36, 1955, 1-11.
8. Kempthorne, O., The Design and Analysis of Experiments, John
Wiley and Sons, New York. NY. 1952. 104-109.
12 MONITORING
-------
Data Evaluation in a Groundwater Study
of Waste Management Practices
in the Phosphate Processing Industry
Edward W. Mullin, Jr.
Jack S. Greber
William E. Thompson
PEI Associates, Inc.
Cincinnati, Ohio
ABSTRACT
The phosphate processing industry has been the focus of
several groundwater monitoring programs within the past several
years. As a result of these studies, a relatively large data base has
been assembled which, in addition to evaluating the
characteristics of the various wastes generated by this industry,
also includes groundwater monitoring data which were used to
evaluate the impact of the industry on groundwater quality. The
interpretation of the monitoring data involved the use of several
different evaluation techniques which will be discussed.
Student's t-test is one traditional method of evaluation that was
used. This method compares the average and standard deviation
of a set of paired values. This method has been used extensively in
hazardous waste-related groundwater studies and has been
criticized for its lack of sensitivity.
Another traditional method of evaluation is the comparison of
Stiff diagrams. These diagrams are graphical representations of
the basic quality of a water sample. This method is very useful
when dealing with the wastes from the phosphate industry, but
the method is somewhat subjective.
Several newer methods also were used in the course of these
projects. These methods included the calculation of ion indices
and the use of cluster analysis.
The ion indices are calculated using basic water quality com-
ponents such as sodium and chloride. Through the comparison of
indices from impacted wells with indices from background wells,
an evaluation of contamination and attenuation can be made.
Cluster analysis was used to identify any trends or relationships
between monitoring wells. Cluster analysis is a collective term
covering a variety of statistical techniques for delineating natural
groups or clusters. Evaluations were made based upon the degree
and the order in which the individual wells were clustered. This
method proved to be quite useful in confirming trends which
other methods also identified.
INTRODUCTION
Over the past several years, PEI Associates has been involved
with evaluating the waste management practices in the phosphate
industry. These evaluations have involved the installation of
groundwater monitoring programs at phosphate processing
facilities. In the course of these projects, eight different facilities
with a total of 85 wells were monitored. Samples were collected
from these wells at varying times; each well was sampled at least
twice and analyzed for a wide variety of water quality parameters.
The interpretation of the monitoring data involved the use of
several evaluation techniques which will be discussed.
STUDENT'S T-TEST
Groundwater monitoring data collected for the purpose of ful-
filling RCRA groundwater monitoring permit requirements have
been evaluated using Student's t-test. The Student's t-test is the
principal criterion for testing hypotheses concerning two popula-
tion means. The t statistic is the deviation of a normal variable (X)
from its mean (/*) measured in standard error units (Sx).
t = (X - M)/SX (1)
This test compares the average and standard deviation of the
measurements for each parameter with the same parameter as
measured in another well, usually the background well. This test
was used on some of the data. However, it was not used on all the
data since there were not enough replicates to make the test
statistically valid. At least three or more replicates are required to
make the test valid. Due to various constraints on the projects,
only two sets of samples were collected at some of the sites. In ad-
dition to the lack of sufficient replicates, the Student's t-test has
been shown to generate false positive results as has been
documented in the Federal Register. Therefore, additional
methods for evaluating the data were used.
STIFF DIAGRAMS
Stiff diagrams are a traditional method of depicting the basic
water types. Only the major cations and anions typically are used
in the diagrams (Ca + 2, Mg+2, Na+, K + , C1-, alkalinity and
SO4-2). The concentrations of the ions are converted to milli-
equivalents/liter and then plotted on a Stiff diagram. The shape
of the resulting polygon is indicative of the type of water being
evaluated. For example, the Stiff diagram of a sodium carbonate
water is distinctly different from a calcium sulfate water.
Therefore the comparison of a Stiff diagram of a sample of liquid
waste from a processing facility to the diagrams from monitor
wells will enable one to graphically determine if there has been
any impact on the groundwater from the liquid waste.
Figure 1 shows such a comparison for one of the test sites. The
waste sample (GL-1) and the background well are compared to
two different wells from the site (PEI-2 and PEI-3). Neither of the
diagrams for PEI-2 or PEI-3 resemble the diagram for the back-
ground well, indicating that the water quality in these wells differs
from background conditions. The Stiff diagram for PEI-3 bears a
close resemblance to the diagram of the waste (GL-1), indicating
that some contamination has occurred in this well. The Stiff
diagram for PEI-2 does not resemble the diagram for GL-1 as
closely as PEI-3, but is more similar to GL-1 than to the
background well, and therefore, it can be concluded that PEI-2 is
contaminated but not as severely as PEI-3.
MONITORING 13
-------
MO+K
Mg
50 25 0 25 50
Cations (mg eq'l) Amons (mg eq/l)
BACKGROUND vs. PEI-2
Na+K
Mg
100 50 0 50 1UU
Cations (mg eq/l) Anions (mg eq/l)
GYP LIQUID vs. PEI-2
Na+K
Mg
S04
50 25 0 25 50
Cations (mg eq/l) Anions (mg eq/l)
BACKGROUND vs. PEI-3
Mg
100 50 0 50 100
Cations (mg eq/l) Anions (mg eq/l)
GYP LIQUID vs. PEI-3
Figure 1
Stiff Diagrams from the Test Site
ION INDICES
Another method which has been used with some degree of suc-
cess is the calculation of ion indices. These indices were developed
by the U.S. Geological Survey during an evaluation of phosphate
mining and processing in Florida.1 The indices are a numerical
evaluation of the concentration of the major anions and cations
within a sample and in some cases the sample concentrations were
compared to the concentrations in the background samples.
The two most useful indices are the index of Nonmineral Input
(INMI) and the Percentage of Ion Increase (PII). The INMI
measures the percentage increase in concentrations of sodium
above chloride and calcium plus magnesium above sulfate plus
alkalinity compared to the same parameters in the background
samples.
'-"«'MCl-CDl
C1 + A1
|(Ca+Ca'
Mg-Mg ') - (S04-S04 '+Alk-Alk ') I - 1C-A ] - | C' -A'
c, + Ai
(2)
x 100
where:
INMI
Alk
C
A
Index of Nonmineral Input, in percent above
background
Alkalinity, in milliequivalents/1
Cation sums, in milliequivalents/1
Anion sums, in milliequivalents/1
Background conditions
The PII is used to evaluate the increase in ions above back-
ground concentrations and is based solely on the sum of selected
cations and anions.
100
(3)
where:
PII = Percentage of Ion Increase, in percent above
background
C = Cation sums, in milliequivalents/1
A = Anion sums, in milliequivalents/1
' = Background conditions
The senior author of the U.S. Geological Survey report1 consid-
ered an INMI of approximately 5% to be above background
levels for the groundwater systems in Florida and to reflect an im-
pact from the waste management practices. The relationships be-
tween the INMI and PII also can be used to evaluate the extent of
seepage and whether or not the soil is attenuating the seepage
through ion exchange. For example, if the INMI and PII are both
significantly elevated above background conditions, this probably
indicates that the seepage has not been altered by ion exchange. If
the INMI is low and the PII is high, then this result may indicate
that the seepage is being altered through ion exchange with the
soil or that another source of contamination is influencing the
system.
14 MONITORING
-------
At the test site previously mentioned, ion indices were
calculated for each sampling period. Table 1 shows the indices
(calculated with PEI-1 as background) for the second sampling
round. As can be seen from the data, the conclusions concerning
PEI-2 and PEI-3 are supported by the indices. These two wells
have the highest indices for both INMI and PII. The lower INMI
for PEI-2 indicates that more attenuation has taken place at
PEI-2 than PEI-3, while the other wells show some degree of im-
pact. The high INMI for PEI-5, the second background well, is
most likely a result of it being completed in a different lithology
than PEI-1, rather than it showing the effects of seepage.
Table 1
Ion Indices at the Test Site
INMI
PII
32%
695%
411%
1332%
32%
123%
32%
5%
0%
100%
These methods have proven to be quite useful for evaluations
of the phosphate industry where the waste being studied consists
mainly of calcium and sulfate, both of which are included in the
Stiff diagrams and the indices. However, these methods would
not be very useful in evaluating seepage from wastes which did
not impact any of the major cations or anions, such as organic
contamination, and these evaluations often may be subjective
when there is not a great deal of dissimilarity in the diagrams or
the indices.
CLUSTER ANALYSIS
The ground water data collected during these projects also were
analyzed by cluster analysis to identify any water quality trends or
relationships between the monitoring wells and the waste samples.
Cluster analysis is a collective term covering a wide variety of
statistical techniques for delineating natural groups or clusters in
data sets.2 The objective of this type of analysis is progressive
grouping of the variables into clusters according to their degree of
statistical similarity.
The cluster procedure in the BMDP Statistical Software
package was used for this analysis.3 The exact procedure used was
Cluster Analysis of Cases (Method 2M). This procedure uses a
hierarchial approach to building the clusters. Initially, each case
(an individual sample from a monitoring well or waste sample) is
treated as a single member cluster; then, the two most closely
related cases are combined into a new cluster and the "distances"
between the new cluster and the remaining ones are computed.
The distance measure calculated is the Euclidean distance
radius between the clusters. The general formula used for this
calculation is as follows:
(4)
where:
djk = distance between two cases of clusters j and k
Xjj = value of the ith variable in the jth case
This process continues until all the cases are combined into one
large cluster. Figure 2 is a graphical representation of this process;
the initial cases are placed on a line in an order that is determined
by the program. Vertical lines then are drawn from each case until
it joins with another. By using the computed distance between
clusters as the vertical scale, the vertical location of each joint or
cluster shows the dissimilarity of the joined groups. Therefore,
clusters formed near the top represent homogeneous groups,
whereas clusters formed further down represent clusters formed
only because the process goes to its logical end of one large
cluster.
Cluster Diagram
Cases
E --
o
o>
e
O)
0>
Figure 2
Example of Hierarchial Clustering
For the purposes of these studies, each sample from a monitor-
ing well or waste sample was treated as one case, and all samples
from a given sampling period were analyzed together; i.e., no
clustering was conducted over the different sampling periods. The
variables used in the analyses included the basic water quality
parameters (Ca, Mg, Na, etc.) and several parameters chosen as a
result of their concentrations in the waste (sulfate, fluoride,
Ra-226, etc.).
The results of the cluster analysis were evaluated based on the
criterion that any increase in distance between clusters that was
twice the previous distance (a 100% increase) was considered to
be a significant difference between clusters.
Cases
PEI PEI PEI PEI PEI PEI 0L
1564231
~ 100-1-
39
= 150-f-
£ 200- -
a E 250- -
O (ft
S » 300- -
1000- -
2000- -
3000
a
10000
20000- -
30000- -
Figure 3
Test Site Cluster
MONITORING 15
-------
The cluster analysis of the data represented previously in the
Stiff diagrams supported the conclusions of the Stiff diagrams.
As shown in Figure 3, the six wells analyzed showed definite
clustering. The background wells (PEI-1 and PEI-5) were deter-
mined to be similar on the basis of the clusters, as were PEI-4 and
PEI-6 (these wells were furthest downgradient from the waste
source). Wells PEI-2 and PEI-3 also clustered together and were
determined to be the least similar to the other well clusters and the
most similar to the waste sample GL-1.
At this site, all the data evaluation methods provided the same
conclusions as to which wells were being affected by the waste
management practices. Cluster analysis is a worthwhile method to
assist in the evaluation of data when traditional methods may not
have the necessary statistical power to sort out all the variables.
The power of the cluster analysis may be increased if a factor
analysis is performed on the data set prior to the cluster analysis.
The factor analysis will identify which analytical parameters are
responsible for the majority of the experimental variance. Use of
factor analysis also will prevent the researcher from inadvertently
weighting the cluster analysis by the inclusion of two or more
parameters which represent the same effect. The primary variable
should be identified by the factor analysis, and only those
variables should be included in the cluster analysis.
CONCLUSIONS
As can be seen, there are a wide variety of methods available
for the evaluation of groundwater monitoring data. The use of
cluster analysis with or without factor analysis is a valuable tech-
nique which can help identify trends in the data that otherwise
may not be recognized.
ACKNOWLEDGEMENTS
The authors would like to thank the U.S. EPA and the Florida
Institute of Phosphate Research for their assistance in funding the
projects which were discussed in this paper.
REFERENCES
I. Miller, R.L. and Suicliffe, H.. Jr., "Effects of Three Phosphate In-
dustrial Sites on Ground Water Quality in Central Florida, 1979 to
1980," U.S. Geological Survey Water-Resources Investigations Re-
port 83-4256.
2. Andberg, M.R., Cluster Analysis for Applications, Academic Press,
New York. NY, 1973.
3. Dixon, W.J., Ed., BMDP Statistical Software, University of Cali-
fornia Press, Berkeley, CA, 1981.
16 MONITORING
-------
Plant Cuticular and Dendrochronological Features
as Indicators of Pollution
G.K. Sharma, Ph.D.
Biology Department
University of Tennessee
Martin, Tennessee
and
Harvard University
Cambridge, Massachusetts
ABSTRACT
In a highly industrialized world, it is imperative that scientists
explore different methods of monitoring environmental contam-
ination. One new field of environmental research, begun in re-
cent years, may do this. This field is the study of leaf cuticle and
annual growth rings in woody plants.
Several plant taxa growing in habitats characterized by varied
levels of environmental contamination were studied by the author
of this paper for dendrochronological and cuticular patterns.
Growth rings, ring and porous wood, and related features were
analyzed with a dendrochronograph. Cuticular patterns such as
stomatal frequency, stomatal size and trichome frequency and
type were studied in relation to environmental contamination in
various habitats. Most plant populations under investigation were
in the industrial areas of Nashville and Memphis, Tennessee—
known for their high levels of environmental pollution.
Statistical analysis of the data revealed the significance of plant
features as indicators of environmental pollution in the area.
Growth rings were, for example, extremely narrow in highly
polluted areas, while the plant populations exposed to relatively
low levels of pollution exhibited fully developed annual incre-
ments. Growth rings tended to be unusually irregular in plants
growing in highly polluted areas. Cuticular patterns such as tri-
chome frequency and type were quite diagnostic for determin-
ing pollution levels in the environment.
With an increase in pollution level, the stomatal frequency de-
creased and the trichome frequency increased—possibly a re-
sponse to offset the detrimental effects of pollutants on various
metabolic reactions and reproductive strategies in plants. In addi-
tion, floral productivity seemed to be adversely affected in
polluted areas. The results demonstrate that tree ring character-
istics and leaf cuticular patterns can be used as monitors for eval-
uating the growth responses of a plant species to varied levels
of environmental pollution.
Analysis of annual increment variation can be used to deter-
mine the degree of stress exerted by environmental pollution on
plant growth. Results of several cuticular and experimental stud-
ies conducted by the author substantiate the findings reported
herein. These results suggest that plants can be used as extremely
reliable indicators of environmental pollution, including hazard-
ous waste sites, provided long-term monitoring is done for a
variety of plant species.
INTRODUCTION
Studies3'5 have demonstrated the relationships between plants
and the environment. Damaging effects of fluorides and sulfur
dioxide on plants have been documented.'
In 1934, Chamberlin1 determined that industrial pollutants of a
large mid-western city in the United States were detrimental to
coniferous flora of the area. He especially referred to the fatal
effects of industrial pollution on Pinus banksiana. Hill and
Thomas3 observed that there was a decrease in alfalfa yield after
exposure to sulfur dioxide. Pyatf suggested lichens as possible
indicators of air pollution in a steel-producing town in Wales. His
studies revealed that thallus size decreased and the lichen flora
decreased in the number of species present with increasing prox-
imity of the pollution source.
While working in the forests of the northwestern United States,
Scheffer and Hedgcock5 found that sulfur dioxide from smelters
produced a characteristic mottling of leaves. Coniferous plants
were more affected than deciduous species.
Feder2 reported that geranium and carnation plants had re-
duced branching and retarded floral productivity after exposure
to low levels of oxidant-type pollutants. Although numerous
plant taxa have been studies to determine the significance of
various plant morphological features as indicators of environ-
mental pollution, relatively little work has been done to establish
relationships between cuticular dynamics and environmental
pollution, especially hazardous waste sites. Recent studies7'8 on
the subject suggest the potential of leaf cuticular dynamics in en-
vironmental pollution research.
The purpose of this study was to collect data for three plant
species in the relatively clean environment of Reelfoot Lake and
to be compared to samples collected in the industrialized areas
of Memphis and Nashville. The present investigation is, there-
fore, a continuation of a comprehensive study involving plant
cuticular features as indicators of environmental pollution with
special reference to hazardous waste sites.
MATERIALS AND METHODS
Plant populations of Polygonum pensylvanicum (smartweed),
Platanus occidentalis (sycamore) and Catalpa bignonoides (catal-
pa) were studied; comparisons of samples taken from the relative-
ly unpolluted sites of Reelfoot Lake and Martin in northwest
Tennessee were made with ones collected from the-contaminated
sites in middle and northwest Tennessee—the latter sites were
characterized t>y exposure to additional pollutants emitted by
vehicular traffic and the industrial complex of the surrounding
metropolitan areas.
Twenty-five randomly selected leaves were gathered from each
plant species. Their lengths, widths, internodal lengths and
length-width ratios were determined. Gross morphological meas-
MONITORING 17
-------
urements were supplemented with cuticular data.
For cuticular studies, representative leaves were washed with a
mild detergent and distilled water. After air-drying, the upper and
lower surfaces of the leaves were coated with Duco-cement™.10
Upon drying, fine layers of Duco-cement showing the cuticular or
epidermal complex of leaves were removed; slides were prepared
from the central portion of the layers for microscopic analysis.
Cuticular dynamics representing in stomatal frequency stomatal
size, epidermal wall undulation, trichome frequency, trichome
length and type and subsidiary cell complex were recorded
(N = 25) for the upper and lower surfaces representing upper and
lower surfaces of leaves using 43x objective and lOx oculars.
Dendrochronological studies were conducted on the woody
taxa by getting core samples in the two sets of habitats. Annual
increments were recorded from the core samples. The entire data
were analyzed by a computer.
RESULTS
Statistical analysis of the data revealed several significant dif-
ferences in the gross morphological, cuticular and dendrochron-
ological features of the two sets of samples (polluted versus rela-
tively unpolluted). Plants growing in the contaminated areas had
smaller leaves with both their length and width significantly re-
duced. In addition, internodal length also was reduced. Plants
growing in the rural, Reelfoot Lake area were healthier and had
larger leaves and longer internodes. These leaves were dark green,
while the leaves of plants from the contaminated sites had yellow-
ish, light green pigmentation. It is obvious that the photosyn-
thetic productivity of the plants at the contaminated sites was ad-
versely affected by the environmental pollution prevalent in the
area.
A reduction in the total biomoass of plant populations in the
contaminated areas was an additional evidence to suggest the
detrimental effects of environmental degradation on plant
growth. A reduction in metabolic activities, especially in photo-
synthesis, must mean reduced bimoass; this result was evident in
all the populations of contaminated sites. Cuticular measure-
ments revealed additional features affected by the environmental
contamination of the sites under investigation. Stomatal fre-
quency values of the lower leaf surfaces in all plants under study
were high in the rural, less polluted habitats, while the plants
from the contaminated sites exhibited extremely low stomatal
frequency values.
In addition, trichomes on the leaf surfaces of plants from the
latter sites had more numerous and much longer trichomes with
large bases. Similar results have been found in earlier studies6'7
in which affected plant populations were growing in sites char-
acterized by heavy vehicular traffic. Stomatal size differences in
the two sets of samples were not significant. Subsidiary cell com-
plex remained unaffected by environmental contamination in all
three plant taxa, and hence must be regarded as a reliable feature
for identification purposes of the species.
The fact that some cuticular and morphological features of
some plant species were different in the two habitats is indica-
tive of the evolutionary trend exhibited by plants growing in
areas exposed to various kinds of environmental contamination.
Monitoring of these plant features over a long period of time is
needed. This study is underway in order to determine if these
morphological and cuticular trends are permanent. If so, these
and similar plant species might be used as indicators of environ-
mental complex, especially at contaminated sites.
Preliminary investigations of the core samples of woody taxa
revealed a general decrease in the size of annual increment. Addi-
tional detailed studies are underway to determine the usefulness
of plant features as monitors of environmental pollution, espec-
ially al hazardous waste sites. Such investigations might reveal
plant species that are able to withstand specific contamination
and hence be of significance for environmental and commercial
considerations.
ACKNOWLEDGMENTS
Financial support for this work under the Research Grant Fund
of The University of Tennessee at Martin is gratefully acknowl-
edged.
REFERENCES
I. Chamberlain, C.J., Gymnosperrru: Structure and Evolution, Dover,
New York. NY. 1934.
2. Feder, W.A., "Plan! Response to Chronic Exposure to Low Levels
of Oxidant type Air Pollution," Environ. Pollut. I. 1970, 73-79.
3. Hill, G.R. and Thomas, M.D., "Influence of Leaf Destruction
by Sulfur dioxide and Clipping on Yield of Alfalfa," Plant Physiol.
8. 1933.334-345.
4. Pyatt, B.F., "Lichens as Indicators of Air Pollution in a Steel pro-
ducing Town in South Wales," Environ. Pollul. I, 1970, 45-55.
5. Scheffer, T.C. and Hedgcock, G.C., "Injury to Northwestern For-
est Trees by Sulfur dioxide from Smelters," U.S. Dept. Agr. Tech.
Bull. 1117. 1955.
6. Sharma, O.K. and Butler, J., "Leaf Cuticular Variation in Trifolium
repens L. as Indicators of Environmental Pollution," Environ.
Pollul. 5. 1972.287-293.
7. Sharma, O.K.. Chandler, C. and Salemi, L., "Environmental Pollu-
tion and Leaf Cuticular Variations in Kud/.u (Purrario lobalo
Willd.)," Ann. Botany. J5. 1980. 77-80.
8. Sharma, O.K. and Tyree, J., "Geographic Leaf Cuticular and Gross
Morphological Variations in Liquidambar slyraciflua L. and their
Possible Relationship to Environmental Pollution," The Botanical
Gazette, 1.1-1, 1973, 179-184.
9. Solbcrg. R.A. and Adams, D.F., "Histological Responses of some
Plant Leaves to Hydrogen fluoride and Sulfur dioxide," Am. J.
Botany. 43. 1956, 755-60.
10. Williams, J.A., "A Considerably Improved Method for Preparing
Plastic Epidermal Imprints," The Botanical Gazette. 134, 1973,
87-91.
18
MONITORING
-------
Detailed Stratigraphic and Structural Control:
The Keys to Complete and Successful Geophysical
Surveys of Hazardous Waste Sites
H. Dan Harman, Jr., P.G.
Engineering-Science, Inc.
Atlanta, Georgia
ABSTRACT
Since the passage of RCRA and CERCLA legislation, count-
less geophysical surveys have been conducted at hazardous waste
sites as part of site investigations in an attempt to delineate the
horizontal and vertical extent of groundwater contamination as
well as to delineate the limits of the sites. Of utmost importance
during a geophysical survey is the identification of Stratigraphic
and structural control features which influence the potential
migration of leachate away from a site.
Stratigraphic features such as sand zones, clay lenses, outwash
valleys and top of rock zones are critical in the overall hydro-
geological assessment of a site. Structural features such as faults
and fractures are equally critical in the assessment of a site.
Proper use and interpretations of geophysical data can result in a
better understanding of a site's subsurface characteristics even
before a drilling program begins.
The geophysical data interpretations explained in this paper,
using the "Modified Wenner" method, are empirical and have
resulted in the site-specific identification of subsurface details
which have been reasonably accurate as compared to subsequent
drilling results. The acquisition of detailed knowledge of a site's
subsurface is the key to a complete and successful geophysical
site survey.
INTRODUCTION
Geophysical surveys consist of surveying a site and its immed-
iate vicinity utilizing one or more remote sensing techniques. The
techniques may include earth resistivity, electromagnetics,
magnetometry, seismic, metal detection and ground penetrating
radar. Each technique has its own advantages and disadvan-
tages at a particular site. When groundwater contamination is
a suspected problem, either earth resistivity or electromagnetics
normally is chosen as the technique in a geophysical survey.
Among the many methods utilized within the earth resistivity
technique, the "Modified Wenner" method1 has been reasonably
accurate in terms of depth investigated as compared with sub-
sequent drilling results. By utilizing the "Modified Wenner"
method and empirical interpretations, site-specific identifica-
tions of Stratigraphic and structural control features are possible.
The identification of these features is the key to performing a
complete and successful geophysical survey of a hazardous waste
site.
STRATIGRAPHIC AND STRUCTURAL
CONTROL
Stratigraphic and structural control is the understanding of a
site's subsurface characteristics. When Stratigraphic and struc-
tural control is understood, a site's geological characteristic can
be understood. Furthermore, a site's hydrological characteristics
can be understood better in terms of aquifers, confining layers
and groundwater migration routes. Also, this control enables
the investigator to more effectively perform resistivity profiles
in an attempt to define the horizontal and vertical extent of iden-
tifiable contaminate plumes. By understanding the Stratigraphic
and structural control, the resistivity profile measurements can
be effectively placed in the appropriate subsurface zones in which
contaminant plumes are suspected.
Stratigraphic and structural control is normally established
only after an exploratory drilling program, but the control can
be established by the proper interpretation of resistivity sound-
ings. Soundings are measurements of the earth's resistivity at
various depths at a single land surface point.
RESISTIVITY SOUNDINGS
Resistivity soundings, utilizing the "Modified Wenner"
method and empirical interpretations, have resulted in site-spe-
cific subsurface identifications which have correlated well with
existing and/or subsequent drilling data. The key factor in the
interpretations of "Modified Wenner" method soundings is that
the potential or inner electrode spacing across the land surface
very closely approximates the depth below the land surface at
which a measurement is taken; or, in general terms, "electrode
spacing equals depth below ground."
Formula lor Apparent Resistivity
Current Meter Battery
(T) 1 ill
Volt Meter
(v)
P'
SOURCE' Carrtaglon A wmon. lost
Figure 1
"Modified Wenner" Array
Diagram of Electrode Spacing
MONITORING 19
-------
The electrode arrangement in the "Modified Wenner" method
is shown in Figure 1. In this arrangement, the current or outer
electrodes (C and C) are stationary while the potential or inner
electrodes (P and P') are moved at equal distances from the
center of the array. The arrangement of the electrode array can
be varied depending on the objective and depth to subsurface
targets.
Numerous investigators have discussed the advantages and dis-
advantages of empirical versus theoretical interpretations, and
a discussion here will not be attempted. A partial list of refer-
ences is included at the end of this paper.2"4 The author has ap-
plied the "Modified Wenner" method and empirical interpre-
tations in numerous subsurface investigations because of the
method's ease of operation and the relative simplicity of the inter-
pretations. The correlation between the sounding interpretations
and actual subsurface drilling data has been reasonably accurate.
EXAMPLE SOUNDINGS
Five example soundings are presented to illustrate the appli-
cation of the "Modified Wenner" method and empirical inter-
pretations. The method has proven to be an asset in solving strati-
graphic and structural control problems, the solutions of which
have improved the effectiveness of subsequent resistivity profil-
ing, monitoring well placements and groundwater contaminant
plume tracking.
SHALLOW AND DEEP SOUNDING
COMPARISONS
Resistivity soundings at a landfill in the mid-western states
were conducted to aid in the understanding of the stratigraphy
underlying the landfill as well as determining the thickness of the
landfill cap. Figure 2 illustrates a shallow sounding to 5 ft below
ground. An apparent resistivity change is evident at approxi-
mately a 2 ft depth which was interpreted as the thickness of the
landfill cap. Similar soundings over the landfill were conducted,
and the apparent resistivity responses varied only slightly (mainly
due to the variation of cap saturation and lithology). A limited
number of shallow borings on the landfill had confirmed the cap
thickness to be approximately 2 ft deep. Resistivity profiling iden-
tified variations across the landfill which aided in the determina-
tion of possible recharge occurring through the cap.
P-Pl SPACING (FEETI
A deep sounding to 100 ft over the landfill was conducted to
aid in understanding the deep stratigraphy underlying the site.
Figure 3 illustrates this sounding showing interpreted features
such as the water table, sand/clay lenses, limestone rock and pos-
sible fractures in the rock. The actual drilling logs from both land-
fill borings and a nearby water supply well are presented to show
the correlation with the interpretations. Note the depth correla-
tion of the apparent resistivity graph to the boring logs, especially
the deflection of the graph at 2 ft (cap-fill interface) and between
90 and 96 ft which correlates with the 90-ft water-bearing zone
identified by the water well log. The deep sounding at this land-
fill site aided greatly in the understanding of the site-specific
stratigraphic and structural control.
\
I 40
24
16
-—3 j-sf—-I
I
SI
« SO 60
P-Pl V1CI-* !<•££']
Figure 3
Deeper Sounding at Land Till
SOUNDINGS IN HIGHLY CONDUCTIVE
GROUNDWATER
Soundings were conducted downgradient of a leaking lagoon at
a site in the southeastern states. This site is underlain by crystal-
line rocks. The groundwater downgradient of the site contains
highly conductive salts which created special problems during the
soundings. Figure 4 illustrates -a 60-ft sounding conducted with
all four electrodes in the center of the highly conductive ground-
water plume. The graph is essentially flat, and interpretations
are impossible to make.
!4 JO 36
P-Pl SPACING tfEETI
Figure 2
Results of Resistivity Soundings at a Landfill
Figure 4
Results of Sounding Downgradient of a Leaking Lagoon
20
MONITORING
-------
Figure 5 illustrates a 200-ft sounding conducted with the cur-
rent electrodes placed outside the plume. This graph is more rep-
resentative of the subsurface than is the graph in Figure 4. Note
that the low apparent resistivity between 50 and 69 ft on Figure 5
is not seen on Figure 4. The low resistivity values at this zone
were confirmed to represent contaminated groundwater within
weathered and fractured crystalline rock. The sounding in Figure
5 aided in the understanding of the stratigraphy and structure
below a highly conductive groundwater plume.
200 ,
ISO
ISO
140
ui 120
I 100
a
i eo
&
60
ao
o
0 20 10 60 30 100 120 140 1EO 160 200
P-P1 SPACING (FEET)
Figure 5
Results of Sounding Near and Leaking L.agoon with Electrodes
Outside the Plume
CONCLUSIONS
Understanding the stratigraphic and structural control under-
lying and in the vicinity of a waste site is critical in the overall
hydrogeological assessment of a site. Surface resistivity sound-
ings aid this understanding by yielding data which can be corre-
lated with existing and/or subsequent drilling results. Subsequent
geophysical surveys, resistivity profiles for example, can be
planned by the proper interpretation of the soundings in terms of
site-specific stratigraphy and structural features.
1500
1350
1200
1050
5 900
u.
§ 750
in
£ 500
|
450
300
150
0
P-P1 SPACING (FEET!
Figure 6
Results of Soundings in Sedentary Rocks to Identify Solution Cavities
SOUNDINGS IN SEDIMENTARY ROCK
CONTAINING SOLUTION CAVITIES
Soundings were conducted to various depths at a waste site in
the southeastern states to aid in identifying solution cavities with-
in sedimentary rock. The rock at this site is limestone with solu-
tion cavities commonly occurring. The solution cavities are pos-
sible avenues of contaminant migration; therefore the under-
standing of the subsurface structure at the site is critical. Figure 6
illustrates one of the soundings conducted at the site. The sound-
ing graph correlates well with the actual drilling log. Note the
low resistivity between 11 and 29 ft and the zone of staining and
vugs present at 14.5,19.8 and 24.9 ft.
As the degree of weathering in the rock decreases with depth
to 30.3 ft, the resistivity graph shows an increase (more resistive
rock). Yet, between 40.1 and 48.0 ft where a solution pitted zone
and small vugs are present, the resistivity graph shows a corres-
ponding decrease (more solution and weathering). The monitor-
ing well subsequently installed 27 ft deep at the sounding sta-
tion yielded groundwater with higher conductivity levels than
background wells. The resistivity sounding in Figure 6 aided in
the understanding of the solution cavity distribution below the
site.
Application of the "Modified Wenner" method and empirical
interpretations has resulted in reasonably accurate correlations
between the interpretations and actual drilling data as well as
groundwater quality data. The ability to predict the presence of
subsurface features with reasonable accuracy is a real advantage
and time-saving element in the decision-making processes con-
cerning a hazardous waste site.
REFERENCES
1. Carrington, T.J. and Watson, D.A., "Preliminary Evaluation of an
Alternate Electrode Array for Use in Shallow-Subsurface Electrical
Resistivity Studies," Ground Water, Jan./Feb., 1, 1981.
2. Moore, R.W., "An Empirical Method of Interpretation of Earth-
Resistivity Measurements," American Institute of Mining and Metal-
lurgical Engineering, Technical Publication No. 1743, July, 1944.
3. Muskat, M., "The Interpretation of Earth-Resistivity Measure-
ments," American Institute of Mining and Metallurgical Engineering,
164,224-231.
4. Moore, R.W., "Geophysical Methods of Subsurface Exploration in
Highway Construction," Public Roads, 26, Aug., 1950, 49-64.
MONITORING 21
-------
Detection and Measurement of Groundwater
Contamination by Soil-Gas Analysis
H.B. Kerfoot
J.A. Kohout
E.N. Amick
Lockheed Engineering and Management Services Company, Inc.
Las Vegas, Nevada
ABSTRACT
A soil-gas sampling probe which can penetrate calcareously
cemented alluvium was developed and evaluated above a chloro-
form-contaminated groundwater plume. A linear correlation of
greater than 95% significance between groundwater and soil-gas
chloroform concentrations was observed. The precision of the
method is controlled by sampling, with relative standard devia-
tions between closely spaced samples ranging from 12 to 43%. A
linear chloroform depth profile was observed; the profile, as de-
lineated by the probe, agreed with calculations made assuming
diffusion-controlled vertical flex of soil gases through the vadose
zone.
INTRODUCTION
Contamination of groundwater is a problem of increasing con-
cern. In efforts to detect and measure groundwater contamina-
tion, groundwater sampling and analysis is the traditional method
of choice. Recently, preliminary surveys using remote-detection
methods have been used as a tool for planning more cost-effective
groundwater monitoring networks.' These surveys typically have
been made using geophysical instruments such as electromag-
netic techniques that can be effective for locating inorganic spe-
cies but are not useful in detecting organic compounds. For the
detection and measurement of subsurface contamination by vola-
tile organic compounds (VOCs), a new technology, soil-gas meas-
urement, has been developed.
Soil-gas surveying originally was developed for oil exploration.1
Applications of the technique to delineate subsurface contamina-
tion by VOCs have been developed by several workers.3'4
Measurement of soil gases for detection of groundwater con-
tamination takes advantage of Henry's Law, which states that
the concentration of a volatile compound in vapors that are at
equilibrium with a VOC solution is directly proportional to the
VOC concentration in solution. The relationship between these
two concentrations is quantitatively described by the Henry's Law
constant for that compound. The magnitude of this constant is
directly proportional to the vapor pressure of the compound and
inversely proportional to the compound's solubility in water.
Because of their relatively low water solubilities and high vapor
pressures, VOCs in contaminated groundwater tend to be present
in soil gases above the source. Table 1 lists the 17 substances most
frequently encountered at Superfund sites and their Henry's Law
constants; of these chemicals, 10 are amenable to soil-gas survey-
ing. In addition, major components of petroleum products which
can leak from underground storage tanks can be detected using
this technique.'
This paper describes a study undertaken to validate equipment
and procedures for soil-gas surveying. The validation of the sys-
tem involved assessing the bias of the technique for indicating
chloroform contamination of groundwater, as well as assessing
the precision and short-range variability of the method.
Table 1
Most Frequently Identified Substances at 546 Superfund Sites*
Sub«C4nc«
Henry*e L*" Constant Percent
(ppbv L/y9) of aitea
1 Trichloro«thyl«n*
2 Le«d and compound*
3 Tolu«n«
4 B«nz«ne
5 Polychlorinated blphcnyl* (PCB*)
6 Chloroform
7 Tctr*chloro«thylene
40
133
30
28b
26»>
22
20 b
8
9
10
11
12
1]
14
IS
16
17
Phenol
Araenlc and coapounde
Cadpdun and coe^xMinda
Chroadua and coapounda
1,1. 1-Trlchloro«chane
Zinc and coag>ounda
Bthy Ibeniene
«ylena
Hethylene chloride
trana- 1,2-dlchloroethylene
« 1
NA
HA
HA
30
NA
59
4]
21
580
15
IS
IS
IS
14b
14
Ub
Ub
12b
lib
"Source: Kcrfoot, H.B. and Barro*\, L.J.. "Soil Gas Measurement for ihc Detection of Sub-
surface Organic Contamination," U.S EPA, l.*u Vegas, NV, 1986,
Compound amenable 10 soil-gas surveying.
EXPERIMENTAL DESIGN
The objectives of the study were to assess the bias and the pre-
cision of this newly developed technique for detection and meas-
urement of groundwater contamination by VOCs and to study the
vertical distribution of VOCs in the vadose zone. The study loca-
tion has a known chloroform-contaminated groundwater plume
22 MONITORING
-------
Pittman
Figure 1
Survey Site Location, Henderson, Nevada
and an existing system of groundwater monitoring wells. These
monitoring wells are separated by 200 ft and lie along a line per-
pendicular to the direction of groundwater flow.
For evaluation of the method bias, soil-gas samples were taken
at a 4-ft depth at four locations 20 ft to the north, east, south and
west of four monitoring wells within the boundaries of the
groundwater plume. To delineate the edges of the plume, addi-
tional samples were taken at a 4-ft depth 20 ft to the east of a well
at the western edge of the groundwater plume and 20 ft to the
END CAP
1/8" TUBING
3/4" THREADS
TUBING NUT
1/2" THREADS
Figure 3
Sampling Probe Design
Benzene/Chlorobenzene
LEGEND
Well Location
Water Table
Sand and Gravel
Clay
Chloroform
Elevation (feet)
1660—,
1640 —
1620 —
1600
1580 —
o
o
~^^~ T— — • -r *
^ [ -] -j- •" 1 — r--T — '
14
655 650 645 640
West
•Sand 8. !
Gravel j
9' deep ! \
,,-S_ ~
^^^^ " .i
o
i ! I 1 ! i i i**^~
-_J H 1 IWate Tabe l__}-'l,^""
j(^^~TT^
—
635 630 625 620 615 610
Stations Teit Well Eagt
0 600
Scala in Feat
0 200
Scale in Meter*
Figure 2
Subsurface Hydrogeology at the Pittman Lateral
MONITORING 23
-------
west of a well on the eastern edge. Samples were taken in triplicate
and were analyzed by gas chromatography on site. The triplicate
analyses were used to assess the method precision.
At one sampling location, four points separated by 3 ft were
sampled as a check on the short-range variability of the method.
Halfway between two of the wells, three points separated by 3 ft
were sampled in duplicate at depths of 1, 2, 3,4 and 5 ft.
STUDY LOCATION
The geohydrology of the study area is relatively simple. Un-
confined groundwater occurs at a depth of 7 to 15 ft in calcified,
unconsolidated alluvium overlying a clay acquiclude. The ground
surface, water table and acquiclude all slope downward about 1
degree to the north, and the groundwater moves northward about
1.5 ft/day. Figure 1 shows the site location, and Figure 2 shows
the geohydrologic profile along the line of our survey.
EXPERIMENTAL
A shallow probe was used to obtain soil-gas samples from a
depth of 4 ft. The probe consists of an outer pipe constructed of
3/4-'m. o.d. high-strength steel that ends in a tapered head with six
horizontal sampling ports, each 1/8 in. in diameter. A 1/8-in.
o.d. stainless-steel tube is connected to the sampling ports and
runs through the pipe to the sampling manifold. The sampling
manifold is assembled from commercially available stainless-steel
fittings. Figures 3 and 4 show the probe design. Figure 5 shows
the sampling manifold.
After insertion into the soil, the sampling manifold was purged
with soil gas. Samples were then withdrawn from the manifold
using gaslight syringes and were transported to a mobile labora-
tory.
Soil-gas samples were analyzed on-site using an Analytical
Instruments Development Model 511 gas chromatograph with a
3H electron-capture detector. The gas chromatograph was oper-
ated at 43 °C, and the output was processed using a Shimadzu C-
R3A integrator. Calibration standards were prepared by serial
dilutions of chloroform headspace vapors.
Figure 6 shows the soil-gas sampling locations relative to the
monitoring wells. Groundwater samples were taken and analyzed
in April and August of 1985. The samples were analyzed by the
purge-and-trap gas chromatography/mass spectrometry (GC/
MS) method specified in the U.S. EPA Contractor Laboratory
Program.'
RESULTS
The bias of the method was evaluated by comparing the mean
soil-gas chloroform concentration measured to the results of the
groundwater analyses. The soil-gas chloroform concentrations
correlate strongly with the groundwater concentrations. Table 2
lists the soil-gas and groundwater chloroform concentrations at
each well. Figure 7 is a plot of these data. A linear regression of
the soil-gas concentrations on the groundwater concentrations in-
dicates a correlation of better than 95% significance (r = 0.85,
n = 6).' In addition, the spatial delineation of the chloroform
groundwater plume, as shown in Figure 8, is very good.
The mean chloroform concentration from the four locations
around each well was used in the above calculations; however, the
relative standard deviation (RSD) of these mean values was often
quite high (above 100%). We interpret this result to be an effect
of spatial variability in the soil-gas chloroform concentrations
over the 20- to 40-ft lateral distances separating these sampling
points.
'x/s */",?./"•:*! i.
D vl -...
SHAFT *4130 COLO DRAWN, CHROM-MOLY. CONDITION N STEEL
END CAP *400 CHROM-MOLY
"'i r
-I -«IO
INNER TUBING
Figure 4
Probe Construction Specifications
24 MONITORING
-------
Table 2
Groundwater and Soil-Gas Chloroform Concentrations
Figure 5
Sampling Manifold
Soil-gaa concentration (ppbv)a
Ground-water
concentration 20 ft. 20 ft. 20 ft. 20 ft.
Well (U9/L) West North East South
631
629
627
625
11
175
25(2)
5(0)
27(5)
5
23
28(5) 72.9(0.1) 124(53)c 45.6(0.2)d 67
(WSW) (WNW) (ENE) (25 ft. SSE)
266(6) 326(10)
376(6)
100 ft.
E of
625
623
621
150
555 115(6) 12(5) 6(3) 27(2) 40
28 10.5(0.3) 10.5
triplicate analyses; standard deviation in parentheses.
Not detected; 5 ng/1 used in regression.
^Mean of four closely spaced points (see text).
Duplicate determinations.
The effect of short-range geologic variability on analytical re-
sults was assessed in two locations. At one location, four sam-
ples were taken at a depth of 4 ft, at separations of between 3 and
7 ft. Figure 6 shows the pattern of the sampling points. The rsd
value of the chloroform concentrations among these points was
42%. At another location, three points 3 ft apart along a north-
south line were sampled at five depths in 1-ft increments between
1 ft and 5 ft. The mean rsd value of the mean chloroform con-
centration for all five depths at the three locations was 12%.
These results indicate the short-range geologic variability can be
a major factor in soil-gas surveys and that the magnitude of this
factor can vary.
The precision of the method was evaluated by analysis of mul-
tiple samples. The analytical precision, based on multiple daily
analyses of calibration standards, showed an rsd value below 4%.
The combined sampling/analysis precision, based on triplicate
samples from each location sampled at a 4-ft depth, was charac-
terized by a mean rsd of 12 %.
631
O
629
627
625
623
N
|—200ft.—|
O Well Location
• LGAS Probe Location
Figure 6
Soil-Gas Sampling Points (Sampling Depth = 4 feet)
S 100-
a 1
"o "H.
o -^
100
Log~Ground Water Chloroform Concentration
lug/L)
Figure 7
Plot of-Chloroform-Soil-Gas Concentration and
Groundwater Concentration
1000
MONITORING 25
-------
Ground-Watp r
Concent rat ion
631 629 627 625 623 621 t, I0
Table 3
Chloroform Concentration Vertical
629 627 62i 62] 621 619
Figure 8
Spatial Distribution of Chloroform in Soil-Gas and
Groundwater Samples
The results of an evaluation of the chloroform concentration
vertical profile (Table 3) showed a linear dependence of concen-
tration upon sampling depth. A correlation coefficient indicating
greater than 99% significance was obtained (r = 0.999, n = 5).
This result is in agreement with a model proposed by Swallow
and Gschwend' that attributes vertical transport of VOCs
through the vadose zone to gaseous diffusion. The water table
below the sampling locations was at a depth of approximately
13ft.
CONCLUSIONS
Soil-gas surveying accurately indicated groundwater chloro-
form contamination at a site in Pittman, Nevada. The precision
of the technique is controlled by sampling and short-range geo-
logic variabilities. At the site studies, results of a depth study
agreed along with a model for vertical transport of volatile organ-
ic compounds through the vadose zone by gaseous diffusion.
ACKNOWLEDGEMENTS
The authors would like to acknowledge contributions made by
J.W. Curtis, K.L. Ekstrom, and L.J. Barrows to this study.
Although this work was supported in part by the U.S. EPA, it
has not undergone review by that agency and does not reflect
agency policy.
REFERENCES
1. Walther, E.G., LaBrecque, D.J., Weber, D.D., Evans, R.B. and van
Ee, J.J., "Study of Subsurface Contamination with Geophysical
Monitoring Methods at Henderson, Nevada," Proc. Fourth Na-
tional Conference on Management of Uncontrolled Hazardous Waste
Sites, 1983, Washington, DC, 28-36.
Location
base
3 ft. N
6 ft. N
base
3 ft. N
r, ft. N
base
3 ft. N
6 ft. N
base
3 ft. N
6 ft. N
6 ft. N
base
3 ft. N
base
Depth
1
1
1
2
2
2
3
3
3
4
4
4
5
5
5
6
Mean chloroform
(ppbv) (SOa)
23.0 (0.1)
22.9 (0.8)
19 (1)b
76 (1)
70 (2)b
58 (3)
109 (1)
111 (1)
99 (14)*>
153 (9)
149 (1)
132 (14)
205 (2)b
183 (8)
167 (23)
236 (29)
Duplicate measurements unless noted
triplicate measurements
2. Horowitz, L., "Geochemical Exploration for Petroleum " Science
229,(1985), 821-827.
3. Marrin, D.L. and Thompson, G.M., "Investigation in the Unsatur-
ated Zone Above TCE Polluted Groundwater." U.S. EPA, Ada,
OK, U.S. EPA Project Number CR811018-01-0.
4. Voorhees, K.J., Hickey, J.C. and Klusman, R.W., "Analysis of
Groundwater Contamination by a New Surface Static Trapping/Mass
Spectrometry Technique," AnalyticalChem.. 56. 1984, 2604-2607.
5. LaBrecque, D.J., Pieretl, S.L., Baker, A.T., Scholl, J.F. and Hess,
J.W., "Hydrocarbon Plume Detection at Stovepipe Wells, Cali-
fornia," U.S. EPA, Las Vegas, NV, 1985.
6. McGhee. J.W., Introductory Statistics. West Publishing New York
NY, 1985.
7. U.S. EPA, "Chemical Analytical Services for Multi-Media Multi-
Concentration Organics GC/MS Techniques," WA-85J680 US
EPA, Washington, DC, 1985.
8. Swallow, J.A. and Gschwend, P.M., "Volatilization of Organic
Compounds from Unconfined Aquifers," Proc. National Symposium
on Aquifer Restoration and Ground water Monitoring^ National
Water Well Association, 1983, 327-333.
26 MONITORING
-------
Environmental Appraisals and Audits:
A Case Study of Their Application
Anthony R. Morrell
U.S. Department of Energy
Bonneville Power Administration
Portland, Oregon
ABSTRACT
The Bonneville Power Administration (BPA) has had an envir-
onmental audit and appraisal program for 2 years. BPA's unusual
approach in implementing its appraisal program resulted in the
program's quick acceptance and organizational effectiveness.
The first audits conducted under the appraisal program iden-
tified problems in the use of polychlorinated biphenyl (PCB)
equipment at BPA substations. Subsequent U.S. EPA inspections
confirmed these same findings. In response to this situation, the
U.S. EPA and BPA entered into a Memorandum of Agreement
(MOA) specifying corrective actions to be taken.
The U.S. EPA, recognizing the strength of BPA's appraisal
program, agreed to forego their normal ad hoc inspections of
BPA facilities. Instead, the U.S. EPA deferred to BPA's own
audits with U.S. EPA inspections limited to followup or verifica-
tion inspections.
In addition to ensuring compliance with environmental regula-
tions, the Agreement's reliance upon BPA's appraisal program
maximizes the effective use of limited staff resources in both
agencies. The Agreement interjects a large measure of predic-
tability and manageability into BPA efforts to achieve com-
pliance.
INTRODUCTION
As a power marketing agency within the U.S. Department of
Energy (DOE), the Bonneville Power Administration (BPA) sells
and transmits the electrical output from 30 federal hydroelectric
dams in the Columbia River Basin. The BPA service area includes
300,000 square miles primarily in the states of Oregon, Wash-
ington, Idaho and Montana, with small service areas in Califor-
nia, Nevada, Utah and Wyoming. The BPA transmission system
serves as the backbone for the interconnected utilities in the
Pacific Northwest and is connected with 18 other transmission
systems at over 100 locations. The BPA transmission system con-
sists of approximately 14,200 circuit miles of high voltage
transmission lines and about 400 substations. Approximately 100
of the BPA substations contain polychlorinated biphenyl (PCB)
equipment, primarily PCB capacitors. A total of about 140,000
PCB capacitors are in service at these substations.
In 1983 BPA instituted an environmental appraisal and audit
program. Upon adoption of the program, BPA decided as a mat-
ter of policy to use its appraisal and audit program as a
mechanism to ensure that BPA facilities comply with applicable
environmental standards.
BPA APPRAISAL AND AUDIT PROGRAM
Equal to the importance of its adoption is the special approach
BPA used in implementing its new program. BPA is convinced
that, given its own organizational culture, any other implementa-
tion approach would not have been as readily accepted and
therefore would have been much less successful in achieving the
objectives of the program.
BPA appraisal program objectives include the following:
• Assure that DOE environmental policy and requirements are
appropriately interpreted and implemented by BPA and BPA
contractors.
• Help line managers achieve BPA's compliance commitments.
• Increase employees' awareness of environmental regulations
and BPA's commitment to compliance.
• Provide management with objective, timely and reliable in-
formation on BPA and BPA-contractor environmental per-
formance, including significant achievements and efficiencies.
• Evaluate the effectiveness and efficiency of BPA's implemen-
tation of measures to avoid, minimize, rectify or otherwise re-
duce adverse impacts to the environment and measures to com-
pensate for impacts.
• Provide management with recommendations for improvement
of BPA's environmental program performance.
• Develop and recommend long-term solutions to current en-
vironmental problems in anticipation of future standards or
conditions.
• Evaluate the accuracy of environmental analysis or impact
predictions and identify methods for improvement.
These objectives are relatively standard. What distinguishes the
BPA program is the unusual approach taken in its implementa-
tion. Without compromising the objectives of the program, the
approach taken was a positive one designed to find the facilities
audited in compliance. As a practical matter, this simply meant
that every effort was made to communicate, beforehand, what
was required and what would be audited. This was accomplished
by providing advance copies of checklists to be used during the
audit and by meeting with the affected organizations prior to their
actual audits. This approach was very well received by the af-
fected field organizations. Given BPA's functional organization
and culture, any other approach would have met with resistance.
Other factors which ensured successful implementation were pro-
visions to: (1) resolve problems at the lowest organizational level;
(2) include representatives of affected offices on each appraisal
team; and (3) resolve smaller problems in the field (i.e., labeling
equipment), so that they need not be mentioned in the written
audit reports. Taken together, these provisions were viewed as a
sincere attempt by central headquarters staff to assist the field
facilities to come into compliance without unnecessarily embar-
rassing them organizationally.
MONITORING 27
-------
PHASE 1
PRE-PLANNING ACTIVITIES
PHASE 2
FIELD APPRAISAL
PHASE 3
POST-APPRAISAL ACTIVITIES
Figure 1
BPA Environmental Appraisal Process
DEVELOPMENT OF THE APPRAISAL PROGRAM
A useful strategy in the initial development of the Environ-
mental Appraisal Program was to seek management consensus
regarding both the scope of the program and the priorities of sites
to be appraised.
To begin the program, meetings were held with all affected
managers. The objectives and purposes of the appraisal program
were discussed. As a result of these discussions, it was agreed that
BPA's appraisals initially would focus on operations and
maintenance activities.
It was also agreed that the first year's program should examine
the requirements of those environmental laws and regulations
dealing with oil spill prevention, hazardous and toxic waste
management, noise pollution control, safe drinking water protec-
tion and commitments made to the public through documents
prepared to comply with the National Environmental Policy Act.
After a schedule was developed for the first year and appraisal
teams were identified, training was conducted for those who
would participate in the program. BPA sought contractor
assistance in conducting the initial training and obtained the ser-
vices of a nationally recognized firm that had published informa-
tion in the area of environmental auditing.
Through the training, the teams obtained skills regarding the
use of checklists and how to document audit findings in a report.
Also, environmental staff members contacted other electric
utilities who had completed environmental audits or appraisals in-
order to utilize their experience in performing environmental
audits on facilities associated with the electric utility industry.
In the program's first 2 years, 23 field appraisals were con-
ducted. Each of these appraisals required 1 -2 days of field work at
the site followed by the preparation of a written report; the
equivalent of approximately four full-time positions have been re-
quired to implement the appraisal program.
Figure 1 illustrates the basic steps used in conducting the ap-
praisal process.
STRUCTURE OF THE APPRAISAL PROGRAM
The BPA appraisal program contains three components. Each
component has a different focus (site versus program) or em-
Figure 2
BPA Audit/Appraisal Program TSCA Checklist
Polychlorlniled Blphenyli (PCBi) In Service
Definitions: PCBs in service al BPA facilities may fall into any of the following
categories: (I) non-PCB equipment, where the level of PCBs is below 50 ppm; (2)
PCB-contaminaled equipment, levels of JO to 499 ppm PCBs; and (3) PCB equip-
ment with levels above 500 ppm. This checklist addresses all PCB articles in service,
whatever the distinction in PCB levels. If appropriate, remarks regarding the levels
of PCBs should be made in the Remarks Column or on the back of the page.
Nole 1: Definition of Transformer: All transformers (e.g., power transformers,
potential transformers, current transformers) are defined as any piece of equipment
which contains the description "transformer" on the manufacturer's nameplate.
a) The transformers that have no PCB ppm tag, and which have the word "oil"
marked on the manufacturer's nameplate, arc assumed to be 50-499 ppm PCB
until these are lagged with the actual PCB ppm level.
b) The transformers that have no PCB ppm tag. which have markings such as
"liquid-filled" or "contains dielectric" (other than the word "oil") or do not
give any indication of the content of the transformer, are assumed to contain
more than 500 ppm PCB until these are tagged with the actual ppm level.
Yea No
Yes No
Yes No
Type
0. Does the facility have any equipment which contains any levels
of PCBs (cither in service or in storage)?
I PCB-Conlaining Equipment (Transformers, Circuit Breakers.
Reactors. Reclosers. etc.)
1.1 Has this facility ever conducted an inventory of its equip-
ment \thich contains PCBi (or an inventory of all equipment
including pieces containing PCBs)?
I.I.I If yes, when was the inventory conducted? (dale of
pnntoul)
1.1.2 If yes, is the inventory complete? (Check each piece of
equipment to see if it corresponds to the inventory.)
1.1.3 Describe each type of RGB-containing equipment,
quantity, status. PCB concentrations and volumes (or
include inventory list as working paper):
Quinllly
Stilus
PCB Cone. PCB Volume
Yes No
Yes No
1.2 Are all pieces of equipment marked with an appropriate
PCB ppm level tag and or the appropriate PCB label (ML)?
1.2.1 If no. list the ones that are not marked.
1.3 Are any pieces of equipment leaking? (Identify)
1.3 I What action is being taken to control these leaks?
Describe:
Note 1: Action to contain and cleanup leaks in 500+ ppm PCB Transformers must
be initiated within 48 hours of having knowledge of the leak. All leaks from any
(ransformer or other piece of equipment should be contained and cleaned as soon as
possible.
Yes No
Yes No
Yes No
Yrs No
1.4 Are all pieces of equipment inspected?
1.4.1 At » hat frequency?
1.5 Are inspection records kept?
Where?
Yes
Yes
No
No
Yes No
I.S.I Arc operators' inspection records maintained for three
(3) years after equipment disposal?
Where? . _
1.6 Arc pieces of equipment stored for use (spares) listed on the
inventory?
I .ft. 1 Arc these spares inspected regularly as described above?
1.7 Have any pieces of equipment been serviced or repaired
(rebuilt) at the facility?
1.7.1 Arc maintenance records for these pieces of equipment
kept? Where
28
MONITORING
-------
Figure 2 (continued)
Yes No 1.8 Does the facility ever top-off or refill any existing pieces of
equipment?
1.8.1 What materials are used? (include PCB ppm levels)
Note 3: No new oil will be procured with over 2 ppm PCB level.
Yes No
Yes No
Yes No
Yes No
Group
1.9 Have any pieces of equipment been removed from service
at this facility?
1.9.1 Are records kept on the disposition of these pieces of
equipment? Where?
2. Capacitors
2.1 Has this facility ever conducted an inventory of capacitors?
2.1.1 If yes, when was the inventory conducted? (date of
printout)
2.1.2 If yes, is the inventory complete?
2.1.3 Describe each type (or group) of capacitor, quantity,
status, PCB concentrations and volumes (or include
inventory list as working paper):
Quantity
Status
PCB Cone. PCB Volume
Yes No 2.2 Are all capacitors marked with an appropriate PCB label
(ML) or in an appropriately marked area?
2.2.1 If no, list the ones that are not marked.
Yes No
Yes No
Yes No
Yes No
Yes No
Yes No
Yes No
Yes No
Yes No
Yes No
Yes No
Yes No
2.4 Is there a designated area for capacitors stored for use?
2.4.1 Is this area appropriately marked?
2.5 Are all capacitors inspected?
2.5.1 At what frequency?
2.6 Where are inspection records kept?_
2.6.1 Are operators' inspection records maintained for three
(3) years after equipment disposal?
Where?
2.7 Are capacitors stored for use (spares) listed on the
inventory?
2.7.1 Are these spares inspected regularly as described above?
2.8 Have any capacitors been removed from service at this
facility?
2.8.1 Are records kept on the disposition of these capacitors?
Where?
3. General Procedures
3.1 Has or is a PCB analysis done before any oil handling?
3.2 How are solvents, filters or rags disposed of if used on
equipment of unknown PCB level?
3.3
Does the facility have the SPIFs on handling, inspecting,
storing PCBs and PCB equipment?
3.3.1 Where are they kept?
3.3.2 List the SPIFS found at the facility:
Yes No
4. Other Equipment Containing PCBs (Light Ballasts, Relays, etc.)
4.1 Does the facility have any other equipment containing
PCBs? List:
Yes No
2.3 Are any capacitors leaking?
2.3.1 How many?
2.3.2 What action is being taken to control these leaks?
Describe
Note 4: Action to contain and cleanup leaks in PCB capacitors must be initiated
within 48 hours of having knowledge of the leak.
Yes No 4.2 Has any of this other equipment leaked?
Yes No 4.3 Was cleanup initiated?
4.4 Describe how the material was contained, cleaned up and
disposed of:
Yes No 4.5 Has any of this other equipment been removed from service,
but retained for backup at this facility?
phasis and each is the responsibility of a different organization
(field personnel versus central staff). These three components are
listed below.
Functional Appraisal
The functional appraisal is an evaluation of the agency's overall
performance based upon a compilation of findings obtained from
the individual field appraisals. This appraisal is conducted an-
nually by BPA's Environmental Manager's office. The functional
appraisal focuses its attention on major or repetitive problems
identified in the field appraisals. In addition to apprising top
management of its findings, the functional appraisal is used as a
means for securing management support for implementing those
recommendations that require an organizational commitment of
resources.
Field Appraisal
The field appraisal is a documented, on-site appraisal of pro-
gram effectiveness for specific disciplines, namely hazardous and
toxic waste management, oil spill containment, noise and safe
drinking water regulations. Following the field appraisal, a writ-
ten action plan is prepared. The action plan outlines an agreed
upon course of action to implement recommendations designed to
correct problems identified in the field appraisal. The action plan
is prepared within 60 days after completion of the field appraisal.
Internal Audit
A self-audit is a less formal evaluation, conducted by the line
organization on a particular part of its environmental program.
These audits are scheduled on an as needed basis.
The three types of appraisals described above permit appraisals
to be specific to a particular audience or level within the organiza-
tion and to focus upon issues of greatest concern to that level.
MEMORANDUM OF AGREEMENT
WITH THE U.S. EPA
In February 1985, the BPA and the U.S. EPA entered into a
Memorandum of Agreement (MOA) to clarify each agency's
responsibilities and commitments for conducting actions required
and authorized by TSCA and CERCLA.
MONITORING
29
-------
Among other things, this Agreement specifically provides that
BPA will audit its field installations for compliance with TSCA
PCB regulations. The Agreement requires that these audits be
done on a scheduled basis and that they be administered by cen-
tral headquarters staff. Any deviations from the PCB regulations
that are found in the conduct of these audits are to be
documented, and a compliance plan is to be included in the audit
report. All reports are forwarded to the U.S. EPA. A copy of the
PCB audit checklist, which reflects U.S. EPA Region 10's inter-
pretation of the TSCA regulations, is included in Figure 2.
With the Agreement in place, the U.S. EPA agreed to forego
ad hoc inspections of BPA facilities. Instead, U.S. EPA inspec-
tions are limited to verification inspections to be conducted only
after the facilities have been audited by BPA.
CONCLUSIONS
The utility of BPA's environmental appraisal and audit pro-
gram was evident even before the Agreement with the U.S. EPA,
but its usefulness to the agency was certainly increased when the
U.S. EPA agreed to use the appraisal program as a substitute for
its ad hoc inspections of BPA facilities.
From the U.S. EPA standpoint, this ensured not only that
potential violations would be corrected, but that the underlying
causes would be corrected as well. From BPA's standpoint, rather
than reacting to the U.S. EPA's inspections, this arrangement
resulted in an opportunity to prioritize corrective actions at BPA
facilities based on environmental and operational considerations.
Consequently, a large measure of predictability and manageabil-
ity was introduced into what could have been a counterproductive
and possibly adversarial relationship with the U.S. EPA.
30
MONITORING
-------
Chlorinated Organics and Hydrochloride Emissions
Sampling from a Municipal Solid Waste Incinerator
Thomas A- Driscoll
James P. Barta
Henry J. Krauss
David H. Carmichael
J. Maxine Jenks
Texas Air Control Board
Waxahachie, Texas
INTRODUCTION
As the use of solid waste incinerators in Texas increases, so do
the questions regarding the environmental impact of their emis-
sions. Are the incinerators creating or releasing potentially toxic
materials from the waste?
Currently, there are eight municipal incinerators operating in
Texas, and permit applications are pending for several more. The
purpose of this paper is to report the levels of chlorinated organic,
chlorine and chloride measured during the stack sampling con-
ducted at the City of Waxahachie municipal waste incinerator.
The sampling was performed in August 1985, by staff from the
Texas Air Control Board (TACB) Sampling and Analysis Divi-
sion. Three sets of 3-hour samples were collected.
The results of this sampling will be reviewed to determine
whether or not additional preventive measures are needed to
reduce the emission of toxic chlorinated organic compounds or
hydrochloric acid. The hydrochloric acid emissions generally are
thought to result from the burning of plastics in the refuse.1 The
chlorinated organic compounds (i.e., polychlorinated biphenyls
(PCBs), polychlorinated dibenzodioxins (PCDDs) and polychlor-
inated dibenzofurans (PCDFs) may originate as contaminants in
the refuse or as byproducts of the combustion of precursors in the
incinerator.2 In addition, these data will be used by the Permits
Division of the TACB to develop emission factors to be used in
permitting new incinerators.
The stack sampling procedure was a modified version of U.S.
EPA Reference Method Five. The sample is collected by XAD-2,
florisil, alkaline arsenite solution and glass fiber filter media.
RESULTS
At this time, analysis of all the samples collected is not com-
plete. Results of the hydrochloric acid testing show the emissions
to be on the order of 2 Ib/hr. A preliminary list of compounds
detected (but not quantified) is presented in this paper. The com-
pounds that were detected are listed below with some information
about their use, toxicity and exposure limits in air:
• Dichlorobenzene—used as a process solvent and as an in-
termediate in the synthesis of dyestuffs and herbicides;3 causes
hemolytic anemia and liver necrosis in humans; can also cause
irritation of eyes and nose; the OSHA occupational exposure
limit (OEL) has been set at 50 ppm upper limit.4 The TACB
Health Effects staff uses 1 % of the OEL as a guideline for per-
mit and health effects review.
• Hepta and hexachlorodibenzofurans—similar properties and
health effects to dioxins of similar molecular weight. The On-
tario Ministry of Health has set 1 x 10~3 /*g/m3 as a standard.
• Hepta and hexachlorobiphenyls—used in electricity transmis-
sion transformers; can cause cancer and adverse skin, liver and
reproductive health effects. The NIOSH standard is 1 jtg/m3.
• Hexachlorobenzene—used as an herbicide, wood preservative
and also is a byproduct from chlorinated hydrocarbon produc-
tion;5 can cause death in breast-fed infants, skin sores, skin
discoloration and enzyme disruptions.' There are no standards
set for this pollutant, however, the TACB Health Effects staff
uses 1 % of the estimated OEL of < 1 ppm as a guideline for
permit and health effects review.
• Pentachlorobenzene—used as a precursor to fungicide produc-
tion and as a flame retardant; also occurs as a product of the
degradation of Lindane;7 may cause mutagenic and carcino-
genic effects in mice.' There are no air standards set for this
compound, but 1 % of the estimated OEL is used as a guideline
by the TACB.
• Tetrachlorodlbenzo-p-dioxin—occurs as a byproduct of herbi-
cide production;' causes liver cirrhosis, spontaneous abortions,
kidney disease and chloracne.10 The New York Air Pollution
Control has set a standard of 9.2 x 10-' jtg/m3.
• Trichloroethane—used as a degreaser, "cold" cleaner and dry-
cleaning agent;" acts as a narcotic to depress the central ner-
vous system; can cause dizziness, uncoordination, drowsiness
and death systematically. Locally, it is irritating to eyes and
causes dermatitis.12 Air limits include TLV of 10 ppm. The
short-term exposure limit (STEL) value is 20 ppm (tentative).13
The OSHA standard is 10 ppm.'4
The compounds detected in the analyses of these samples are
consistent with compounds measured in stack sampling con-
ducted in other areas.15'16 Stack sampling conducted by the
state of California and Scott Environmental (contracted by
EPA) indicates the small amounts of PCDDs and PCDFs were
detected in municipal solid waste incinerator emissions.
INCINERATOR DESCRIPTION
The municipal solid waste incinerator stack sampling was con-
ducted from Aug. 13 to 15, 1985. Three separate sets of samples
were collected over a 3-hr period each time. A representative por-
tion of the stack emissions was collected isokinetically using stan-
dard operating procedures described in the TACB Sampling Pro-
cedures Manual.
The incinerator sampled is located on Singleton Drive in North
Waxahachie, Texas. It has two dual-chambered incinerators that
operate at approximately 1600°F. The incinerators are permitted
to destroy 2088 Ib/hr of solid waste. The residence time is approx-
imately 45 min to 1 hr in the primary burn chamber and 1.22 sec
DETECTION OF RELEASES 31
-------
TC THERMOCOUPLE
Figure I
Chlorinated Organic*, Sampling Train
for the secondary burn chamber which is often used to burn efflu-
ent gases from the first chamber.
The California Air Resources Board recommends a flue gas
temperature of 1800°F + / - 190° with a residence time of 1 sec
under well-mixed conditions. These conditions should destroy
99.9% of the toxic organics compounds."
The facility is operated 24 hr/day, 6 days/week. There is down
time for maintenance and cleaning each day. The facility is per-
mitted to emit 12 Ib/hr or 53 tons/yr of particulate matter.
SAMPLING TRAIN
The sampling train used to collect chlorinated organic com-
pounds is shown in Figure 1. The sampling apparatus is made en-
tirely of glass except for the stainless steel nozzle and teflon tape
seals used between the ground glass joints. Following the stan-
dard front-half, the back-half is modified to consist of a water
jacketed condenser followed by two XAD-2 and one Florisil resin
traps in series. The remainder of the train consists of two im-
pingers. The first impinger contains 200 ml of sodium arsenite to
collect chlorides/chlorine and to protect equipment downstream.
The second impinger contains silica gel to insure the complete
dryness of the sampled gas. The Organic Laboratory Section per-
formed the analyses of the samples using the Finnigan triple
quadrupole mass spectrometer (TSQ).
The hydrochloric acid emissions were captured using the collec-
tion system as described in Chapter 5 of the TACB Sampling Pro-
cedures Manual. The stack exhaust gas entered the system
isokinetically through a stainless steel probe and traversed a non-
heated probe across a glass fiber filter which was maintained al
250 to 275 °F. The condenser train consisted of a series of im-
pingers containing deionized water, alkaline arsenite and silica
gel. The analyses followed the procedures described in the TACB
Laboratory Procedures Manual.
Leak checks were performed several times during the process to
insure sample integrity. If acceptable, the sampling process was
continued. The sample air was circulated through a condenser
filled with cold water. Due to the large volume (approximately
100 ft3) and the high moisture content of the air drawn from the
stack, a large volume of condensate was collected in the flask
following the condenser. As necessary, the condensate was
removed from the flask, the volume recorded, transferred to an
amber glass container and returned to the TACB laboratory for
analysis. A check for air leakage in the sampling system preceded
and followed the condensate transfer to insure sample integrity.
Following the completion of sampling, the three adsorption
tubes and impingers were weighed again and their final weights
were recorded to determine stack air moisture content. The probe
and glassware in front of the filter were rinsed with tetrahydro-
furan (THF). All rinsate was collected in amber bottles. The
amber bottles were used to prevent rinse solutions from reacting
with ultraviolet light. The glass fiber filter was returned to its pro-
tective alumi.ium foil and stored in plastic envelopes until
analysis. The connecting glassware and two impingers were rinsed
with deionized water and the rinsate was collected for analysis at
the TACB laboratory.
ANALYTICAL METHODS
The XAD-2 and Florisil adsorbers were extracted using hexane
in a Soxhlet extractor for 8 hr. The distillate then was evaporated
down to 2 or 3 ml using a rotary evaporator. The product was in-
jected into a gas chromatogram/mass spectrometer (GC/MS).
The pollutant results were interpreted by a TACB chemist who
used graphic responses compared to a chemical library to identify
the pollutants.
The front-half and back-half rinsates were concentrated by
evaporation. Separately, the products of the evaporation were in-
jected into the GC/MS. The results were interpreted as discussed
in the previous paragraph.
The impinger containing deionized water was used to collect
hydrochloric acid. The samples were analyzed turbidimetrically.
Chlorine was collected in the arsenite solution. The solution also
was analyzed turbidimetrically.
The glass fiber filter also was analyzed for chlorides and chlor-
inated organics. A 37 mm diameter circle was cut from each of the
three filters. The exposed filters were analyzed versus the field
blank for chlorine on an x-ray fluorescence spectrometer system.
Duplicates also were run to test the method's precision. In addi-
tion, the filters were analyzed for chlorinated organics. They were
extracted separately using hexane and a Soxhlet extractor. The
resulting solution then was reduced to 2 or 3 ml with nitrogen.
The product was injected into the TSQ and analyzed as described
previously.
32
DETECTION OF RELEASES
-------
CONCLUSIONS
More stack sampling of municipal solid waste incinerators in
Texas is needed. During this initial stack sampling exercise, there
were some problems with the plant operation and with the stack
monitoring process due to malfunctions in the incinerator.
Preliminary determinations of the hydrochloric acid test have
been made. The emissions of hydrochloric acid are approximately
1 Ib/hr. This value is lower than expected.
Particulate matter collected during the test was analyzed and
found to contain significant amounts of chloride salts. It is be-
lieved that the chlorides exiting this process were primarily
chloride salts with relatively low hydrochloric acid emissions. Fur-
ther testing is required to determine if this is the case.
Also, some of the pollutants detected in preliminary analysis
are very toxic and warrant additional attention. PCDDs and
PCDFs have been described by some sources as unsafe at any
level. This project has not produced enough information to
prescribe incinerating at higher temperatures, having longer
residence times in the burn chamber, requiring additional pollu-
tion control devices or sorting out plastics before incineration.
Whether or not the Waxahachie municipal solid waste in-
cinerator is representative of incinerators in Texas also is not
known. Stack sampling at other sites may be needed to adequately
characterize incinerator emissions. Other types of air quality
monitoring also may be warranted. Perhaps monitors placed
downwind may help correlate stack emissions levels to levels near-
by residents may eventually breathe. Therefore, developing con-
clusive emission factors is premature.
ACKNOWLEDGEMENT
The authors would like to acknowledge the Texas Air Control
Board Laboratory, Dallas Regional personnel, and S. Thomas
Dydek of The Health Effects Section for their assistance with this
project.
REFERENCES
1. Krakower, T., Resource Recovery Facilities Overview, Texas Air
Control Board (Health Effects and Research Division), Sept. 1984,
Section III.
2. Ibid., Summary.
3. Sittig, Handbook of Toxic and Hazardous Chemicals, Noyes Publi-
cations, Park Ridge, NJ, 1981, 227-228.
4. Ibid., 228.
5. Ibid., 358.
6. Ibid., 359.
7. Ibid., 522.
8. Ibid., 522.
9. Ibid., 632-633.
10. Ibid., 632.
11. Ibid., 669, 671.
12. Ibid., 670-671.
13. Ibid., 669, 671.
14. Ibid., 669.
15. Sheffield, A., "Sources and Releases of PCDDs and PCDFs to the
Canadian Environment," Chemosphere, 14, 1985, 811-812.
16. Nunn, A.B., III, "Gaseous HC1 and Chlorinated Organic Com-
pound Emissions from Refuse Fired Waste-to-Energy Systems,"
Scott Environmental Services (for U.S. EPA Environmental Sciences
Research Laboratory), Plumsteadville, PA, 1984, 30, 37.
DETECTION OF RELEASES 33
-------
CARE—Modeling Hazardous Airborne Releases
M. Gary Verholek
Environmental Systems Corporation
Knoxville, Tennessee
ABSTRACT
Emergency preparedness for airborne hazards means being able
to answer such questions as where the material will go, who it will
effect, what areas are safe as evacuation centers, how long the
chemical will take to disperse and which preplanned response op-
tion would be most effective in the event of an incident. To help
answer such questions, Environmental Systems Corporation
(ESC) has developed an assemblage of sophisticated
mathematical models into a user-friendly and truly capable com-
puter system called the CARE (Computerized Airborne Release
Evaluation) System.
CARE provides the capability to perform a variety of func-
tions, such as risk assessment, hazardous materials training,
monitoring and emergency response. In short, CARE provides
the tools needed to prevent an emergency from becoming a
disaster.
INTRODUCTION
Are you ready for a hazardous materials accident? What would
you do if a toxic cloud from a spill or a ruptured tank threatened
your facility? What if burning chemicals were threatening your
town today? Where would the cloud go? Who would be effected
and who would be out of danger? What areas would need to be
evacuated? Where could people go to be safe? What is the best
way to respond to the emergency? How do you get answers to
these questions in time to prevent the emergency from becoming a
disaster?
To answer such questions. Environmental Systems Corpora-
tion (ESC) has developed a system called CARE, which stands for
Computerized Airborne Release Evaluation System. CARE gives
you the capability to perform a variety of emergency preparedness
functions:
• Risk Assessment
• HazMat Training
• Real-Time Monitoring
• Emergency Response
CARE is a system that puts truly capable emergency
preparedness on your desk. In the event of an accident that
releases a hazardous material into the air, CARE will tell you
where it will go, who it will effect, which areas are safe as evacua-
tion centers, how long it will take to disperse and which pre-
planned response options would be most effective. CARE gives
you the ability to neutralize the danger, to assess alternative
defense strategies and to keep abreast of changing situations. In
short, CARE gives you the tools you need to prevent an emer-
gency from becoming a disaster.
CARE
Computerized Airborne Release Evaluation System
SOURCE Mode!
METEOROLOGY Model
Calculates source
cna'actensiics
r
Calculates winds
stability, lemp
over modeling region
DISPERSION Model
Calcinates pollulani
conce^ra'.o^ 'te'ds
EFFECTS Model
Describes lie Hazards
due to ambient
concentrations
RESPONSE Model
Provides protective
action guidance
Figure 1
CARE System Diagram
SYSTEM DESIGN
CARE uses sophisticated mathematical models, assembled by
ESC's meteorologists and engineers, to provide a realistic depic-
tion of the hazard due to a cloud of airborne material. These
models are grouped into five modules (Fig. 1):
• Source Module
• Meteorological Module
• Dispersion Module
• Effects Module
• Response Module
34 DETECTION OF RELEASE-IS
-------
The Source Module provides one of the two basic mathematical
models required to estimate atmospheric concentrations of air-
borne materials. For chemical spills, it calculates the evaporation
rate of the chemical into the atmosphere. This rate of evapora-
tion, determined through heat and/or mass transfer mechanisms,
it used as the emission rate in the second mathematical model, the
air dispersion model. The final product is then the calculation of
the geographical distribution of pollutant concentration at
selected times since the spill, via the atmospheric dispersion
model. Specific source modules are available for various applica-
tions such as chemical spills, fires and transportation accidents.
nuclear plant accidents, fossil plant emissions, airborne pesticide
releases, military smoke generators, etc.
The Meteorological Module is a group of several models that,
in combination, characterize the weather conditions that effect
the dispersion of hazardous materials. This module uses observed
meteorological data, such as wind speed, wind direction, tur-
bulence and temperature and provides the necessary inputs to the
dispersion module.
The Dispersion Module, using the source and weather inputs,
provides a realistic depiction of the movement of the cloud of
hazardous material and calculates its concentration with respect
to time and distance from the release. The model, which is the key
to systems realism, is a variable-trajectory, puff-advection model.
This kind of model treats the airborne release as a series of puffs
which are moved by the air flow that has been calculated by the
meteorological module. The puffs are diffused along the way ac-
cording to the character of the atmosphere. The resultant disper-
sion of the material is displayed on the computer as a meandering
plume which flows over and around obstacles, much as one would
expect a plume to move once released to the air.
The Effects Module assesses the critical concentration values
and provides information on the hazard to health and welfare
posed by the materials in the plume. This module examines the
concentration fields for critical levels based on the application.
For example, the effects module for chemical accidents deter-
mines exceedances of Threshold Limiting Values or Lower Ex-
plosive Levels; for nuclear releases, it calculates dose and dose
rate; for smoke obscurant releases, it calculates transmission or
visibility; for ambient air quality, it determines exceedances of the
National Ambient Air Quality Standards; and so forth.
The Response Module provides the user with recommended re-
sponse actions based on the effects assessment. The response
module is tailored by ESC to meet user requirements, special con-
ditions and capabilities.
The CARE system assembles the aforementioned capabilities
into a user-friendly system that is fully documented and tested.
The near real-time display of the location and concentration of
hazardous airborne materials provides the information needed to
make informed decisions that can save lives and property.
SOURCE MODULE
The Source Module for toxic releases is based on the SPILLS
model developed by Shell Development Company.' SPILLS is an
unsteady-state model which calculates the evaporation of a
chemical spill to determine the source strength of the vapor cloud.
Used in conjunction with the atmospheric dispersion model, it is
used to estimate concentrations of the vapors as a function of
time and distance downwind of the spill.
Three options depending on the nature of the spill have been in-
corporated in the model: (1) continuous spills, such as leaks from
tank cars, tanks or pipelines; (2) instantaneously-formed pools of
liquids or liquified gases; and (3) stacks, where the emission rate is
assumed to be known. For options 1 and 2, thermophysical pro-
perties (available as a subroutine in the computer program) of 36
potentially hazardous chemicals are used to calculate, through
heat and mass transfer mechanisms, the evaporation rate, which
becomes the emission rate for the atmospheric dispersion calcula-
tions.
The computer program was adapted for use with complex at-
mospheric dispersion models to provide a realistic prediction of
hazard caused by an airborne toxic cloud. The program now con-
tains the necessary properties of 36 potentially hazardous
materials commonly handled, used or transported by industry,
but the data easily can be expanded to include the thermophysical
properties of other substances. The air dispersion model takes in-
to account the atmospheric conditions, type of source and emis-
sion rate to calculate downwind concentrations. The source
strength can be specified by the user (stacks) or is determined by
the evaporation rate models if the chemical is known. At-
mospheric and soil conditions and chemical properties are utilized
to predict the effects of heat and mass transfer on the vapor cloud
formation.
Evaporation Rate Models
The objective of an evaporation rate model is to predict the
amount of material emitted in the case of a continuous or an in-
stantaneously formed spill from a transport line or a storage tank.
Although the basic theory is the same for both cases, the calcula-
tions are different, and thus they are presented separately.
Three evaporation rate models are discussed in this section.
First, the continuous leak case which corresponds to emissions
generated from small ruptures in transport equipment (e.g.,
pipelines, storage tanks or tank cars). In this case, the chemical is
assumed to flow at atmospheric conditions through the rupture to
form a quiescent pool on the ground. Since the chemical will be at
ambient temperature, no heat transfer will occur between the
material spilled and the surroundings, and the amount of mass
taken by the wind blowing over the pool will be limited by mass
transfer.
Two key variables involved in the mass transfer model are
simultaneously unknown: (1) the area of the spill and (2) the emis-
sion rate into the atmosphere. It was assumed that the area of the
pool would be the more difficult variable to predict during the ac-
cidental spill. Therefore, the computer program requests the
value of the emission rate through the rupture, which the model
takes as the emission rate into the atmosphere. The model then
determines the spill area by assuming that this emission occurs by
convective mass transfer. The evaporation then is assumed to occur
indefinitely at the same rate unless the user specified the elapsed
time from the beginning to the end of the flow through the rupture.
At this point, the program would set the emission rate to zero.
Evaporation from instantaneously formed pools, the second
option in SPILLS, corresponds to the spill of a chemical from a
storage tank or transport line. In this case, however, the spill and
the resultant pool at the ground are assumed to occur instan-
taneously. The area of the pool is assumed to remain constant
during the evaporation process so that its estimation at the scene
of the accident would be easier than the continuous spill case. The
computer program then requests the value of this spill area in ad-
dition to the total amount spilled.
Chemicals with a normal boiling point below ambient
temperature will first flash off due to the pressure drop between
the storage pressure (the storage temperature has to be specified
by the user) and atmospheric pressure. The evaporation rate due
to this adiabatic flash calculation is assumed to occur during the
first minute after the spill. The chemical will then form a pool of
liquified gas at its normal boiling point. The difference between
this temperature and the ambient temperature will cause heat to
be transferred from the ground (the soil is assumed to be at am-
DETECTION OF RELEASES 35
-------
bient temperature) to the pool by conduction and from the at-
mosphere to the pool by convection.
Mass transfer, due to the wind blowing over the pool, takes
over when the heat transfer evaporation rate becomes equal to the
mass transfer evaporation rate. Note that this occurs because the
heat transfer rate decreases as time increases, whereas the mass
transfer rate is independent of time.
Finally, the emission rate is equated to zero when all the
chemical is evaporated. Chemicals with a normal boiling point
higher than ambient temperature are assumed to form a liquid
pool at ambient conditions. Here, the chemical is assumed to
evaporate only by convective mass transfer.
Continuous Leaks
Continuous leaks are regarded in the present work as emissions
from a small rupture of a transport line (such as a pipeline) or a
storage tank (such as a tank car). It is assumed that the mass flow
rate through the orifice is known and that the chemical flows at
atmospheric conditions, i.e., as a gas or as a liquid depending on
the normal boiling point being below or above ambient
temperature, respectively.
The chemicals of interest in the present program are all more
dense than air so that a gaseous plume or a liquid pool will be
formed at ground level after the chemical exits the original car-
rier. Convective mass transfer is assumed to be the limiting pro-
cess for the chemical to be transported by the wind. The spill area
then is predicted by equating the mass transfer rate to the flow
rate through the rupture.
The pool size is determined for both laminar and turbulent flow
cases and is used to predict the corresponding Reynolds numbers.
The value that gives the consistent flow regime is chosen as the
final pool length.
The area of the spill is used as a ground level area source in the
air dispersion model. In addition, the flow rate through the rup-
ture is utilized as the emission rate in the air dispersion calcula-
tions. The user has the option of specifying the length of time
from the beginning to the end of the flow. This time can be
estimated by knowing the amount of material to be emitted. This
case is approached by the air dispersion model as an unsteady-
state case where the constant emission rate is dropped to zero at
the corresponding time. The program assumes a steady-state
operation if no time is specified by the user.
Instantaneously Formed Pools
The time-dependent evaporation rates of stationary, instantan-
eously formed pools are calculated by two different procedures
depending upon the chemical spilled: (1) liquified gases or (2) liq-
uids. However, both methods assume that the chemical pool does
not spread on the land as a function of time, such that the spill
area specified by the user remains constant.
Liquified Gases
An initial adiabatic flash calculation is performed based on the
pressure drop of the liquified gas (i.e., from the pressure of the
chemical while stored, to atmospheric pressure).
The original enthalpy of the chemical is evaluated as the en-
thalpy of the saturated liquid at the user's specified temperature.
For conservative purposes, the amount of vapor which is flashed
off initially is assumed to be emitted in a period of 1 min.
The rest of the liquified gas remains in the liquid pool specified
by the user at its normal boiling point. From this point on, the
chemical can undergo two different transport processes: heat
transfer or mass transfer.
Due to heat transfer, which occurs mainly by conduction from
the soil assumed at ambient temperature and by convection from
the atmosphere, the chemical evaporates off the boiling liquid
pool. It is assumed that solar radiation is negligible and that no
thermal resistance exists between the soil and the boiling liquid.
The mass range at which the chemical evaporates is then deter-
mined.
The air dispersion model uses the time-dependent mass rate
calculation to estimate the source strength during a heat transfer
evaporation mechanism based on the total amount of mass
evaporated between a time t(0) and any later time t.
Since all the chemicals treated in the present work are more
dense than air, the emission rate into the atmosphere is mass
transfer limited. This means that only a portion of the evaporated
mass will be transported by the wind. Mass transfer governs this
take-up flux into the atmosphere. The rest of the evaporated
chemical remains as a gas cloud above the liquid pool. For conser-
vative purposes, and due to the fact that the person at the scene of
the accident will report an estimate of the size of the pool, it is
assumed in the present work that all the mass evaporated by heat
transfer will be emitted into the atmosphere.
Mass transfer occurs predominantly by forced convection over
the liquid pool due to the wind. The amount of chemical emitted,
used in the air dispersion model, is obtained from integration with
respect to time of the amount of material emitted by mass transfer
during the specified interval of time.
Heat transfer dominates at the beginning of the process. As
time increases, the heat transfer emission rate decreases and
reaches a point in time where it becomes equal to the mass
transfer rate.
Liquids
Liquids are defined as those chemicals with a normal boiling
point, T(B), higher than the ambient temperature, T(0). These
chemicals are transported at conditions very close to ambient so
that emission into the atmosphere after a spill will occur only
through mass transfer by convection. The risk to health and life
from an atmospheric release of hazardous effluent is a function of
the ambient downwind concentrations and the length of the ex-
posure period.
THE DISPERSION MODEL
A vital element in such a consequence assessment is the ability
to estimate the location and concentration of an airborne effluent
plume. To provide this estimate, the dispersion models employed
must include both the transport and diffusion properties of the at-
mosphere. Consequently, the models must account for the
physical, temporal and spatial changes that the atmosphere
undergoes during the course of the release and during transport.
To estimate the downwind location and concentration of air-
borne effluents, CARE uses a highly sophisticated atmospheric
dispersion model. The model includes both transport and diffu-
sion components which account for the temporal and spatial
changes that the atmosphere undergoes during the course of the
release and during transport.
Atmospheric dispersion calculations are performed usin$ the1
ESC variable trajectory, puff-advection model—TRAJ.' The
model simulates a release as a series of superposed puffs which are
transported by the local atmosphere. Each specification of new
meteorology is considered a new transport interval. During such a
transport interval, the puffs are diffused according to standard
Gaussian theory. The dimensions of an individual puff are pro-
portional to its travel distance (or travel time). The model accom-
modates multiple point sources and includes algorithms for plume
rise and shoreline effects. The model provides a. rigorous treat-
ment of the spatial and temporal variations in the atmosphere and
accounts for influences due to local meteorological effects.
The advection of the puffs is accomplished through inputs of
local meteorological variables, wind speed and direction. The
36
DETECTION OF RELEASES
-------
model can accommodate several meteorological stations by apply-
ing a weighting to each station variable. For these applications in
areas of complex flow fields (e.g., mountains, valleys, shorelines,
etc.), a three-dimensional wind field model can be added to pro-
vide a rigorous, dynamic treatment of the spatial and temporal
variations in the atmosphere.
The dispersion model calculates the concentration over a grid
covering the modeling region and also at specified special receptor
locations within the grid. Special receptor locations are stored as
data to facilitate modification should additional stations be needed.
Site-specific run parameters, which are input as data, define the
various site-specific characteristics unique to this model applica-
tion. The calculation is performed in a time-wise, step-by-step
analysis of the plume path from the release point through the
limits of the calculation grid.
THE METEOROLOGICAL MODULE
Dispersion modeling requires the input of critical
meteorological data. Wind speed and direction are needed to
define the transport of airborne material; stability data are used
to determine the spread of the material as it travels downwind.
The depth of the layer of the atmosphere where mixing occurs is
called Mixing Depth; this parameter defines the maximum height
of the plume. Temperature data are used to determine the
bouyant rise of the plume and, where applicable, the existence of
a sea breeze.
At each modeling time increment (generally 15 min), input
meteorological data are used to determine:
• Transport wind (u,v,w) components
• Plume dispersion coefficients (sigma-y; sigma-z)
• Atmospheric temperature
The meteorological data can be acquired in real-time from an
on-site meteorological tower or, if data are acquired manually
from alternate sources such as National Weather Service stations
or United States Air Force bases, they can be input from the con-
sole via menu responses. Wind data, input as speed and direction,
are converted to transport wind (u, v and w) components at the
grid points. Horizontal and vertical plume dispersion coefficients
(sigma-y and sigma-z) can be derived from direct measurements
of turbulence (sigma-theta and sigma phi) or from routine me-
teorological observations using well-known stability classification
schemes, depending on the available input data.
Stability Classifications
The most commonly used method to determine dispersion coef-
ficients is the indirect method, requiring the determination of a
stability class which then determines the dispersion curve. The
Pasquill-Turner stability classification scheme or the NWS-
Turner method are typically used with the Pasquill-Gifford
dispersion curves. Stability categories cannot be measured directly
but must be determined from other meteorological parameters.
Any of the following approaches can be used:
• Sigma phi (standard deviation of the vertical wind direction)
• Sigma theta (standard deviation of wind direction)
• Delta temperature (the difference in temperature with height)
• Cloud conditions (i.e., clear, scattered, broken or overcast,
and ceiling height)
• Default conditions (e.g., F-nighttime; D-daytime)
• Manual override inputs (operator discretion)
When using sigma theta or sigma phi to determine stability clas-
sifications, and changes in dispersion are expected because of sur-
face roughness, a factor of (Zo/15 cm) A0.2, where Zo is the
average surface roughness in centimeters, may be applied to the
tabular values. Suggested Zo values that may be used as a guide to
estimating surface roughness are given in Hogstrom and
Hogstrom.2
The sigma stability classification schemes are adequate for
daytime use, but during nighttime (1 hour to sunset to 1 hour
after sunrise), adjustments adapted from Mitchell and Timbre3
are used.
If cloud cover inputs from NWS or USAF sources are used, the
values are compared to criteria published in the U.S. EPA
CRSTR model. Default dispersion classes (e.g., D-daytime;
F-nighttime) can be provided for the operator in the event no
other data are available.
Dispersion Coefficients
The purpose of establishing a stability class is to be able to
calculate the horizontal and vertical dispersion parameters,
sigma-y (Sy) and sigma-Z (Sz), which are used in the calculation
of downwind concentration. Sy and Sz are approximated by a
curve fitting process which describes the PGT curves with power
laws as a function of the total distance traveled, x:
Sy(x) = a*vAb
Sz(x) = c*vAd + e
(1)
(2)
where the coefficients (a,b,c,d,e) are a function of stability class.
The coefficients are taken from the POLYN routine found in
NUREG/CR-2919, the manual for the XOQDOQ model, written
by Sagendorf, et al.' but have been modified slightly to improve
the continuity of sigma-z from the region of x < 1000m to x < =
1000m.
Mixing Height
The boundary layer height (H) is used to limit the vertical
dispersion parameter (Sz) to values < 0.8*H. The mixing depth
will be determined by applying the climatological mixing heights
according to the method described in the U.S. EPA CRSTER
manual,' which is the standard Holzworth method.
Sea Breeze
In regions near large bodies of water where a sea breeze
phenomenon may be a significant meteorological feature, the
Thermal Internal Boundary Layer (TIBL) will be used as the max-
imum mixing height. The software will determine if a sea breeze
condition is likely to exist, using the following rules:
• The sun is above the horizon (based on the data and time)
• The lower-level air temperature is greater than the sea water
temperature, or monthly climatological temperature if tem-
perature data are not available
• The lower wind direction measurement indicates on-shore
winds within a 180 degree sector
If a sea breeze is determined to exist, the mixing height is set
equal to the height of the Thermal Internal Boundary Layer
(TIBL):
H = C * sqrt (d)
(3)
where H is the TIBL height (in meters) at distance d (in meters)
inland6'7 and C is a constant determined for the local area.
Wind Field Model
For applications in areas of complex flow (e.g., valleys, moun-
tains, seashores, etc.) a three-dimensional, mass-consistent flow
field model is used to provide a rigorous, dynamic treatment of
the spatial and temporal variations in the movement of the cloud.
The ESC windfield model (ESCWIND) was adapted from code
available in the public domain from several sources. It was
developed to provide the user with an economical windfield
DETECTION OF RELEASES 37
-------
predictor. The model produces three-dimensional, terrain-
dependent, divergence-free windfields given observed surface
and/or upper-air data as input. Given the wind components for
each meteorological station and at each level, a 1/r weighted in-
terpolation is used to "fill out" the rest of the windfield along
each horizontal plane.
The terrain surface boundary is approximated using obstacle
cells to represent the terrain in a stairstep fashion. For computa-
tional purposes, the irregular surface boundary is removed by a
coordinate transformation in which the terrain surface also
becomes a coordinate surface. This "sigma-space" computa-
tional domain has two main advantages. First, the bottom boun-
dary condition is defined more accurately; second, the option of
variable vertical zoning improves the model's accuracy and
economy.
ESCWIND follows a procedure of extrapolating and inter-
polating the input data to determine an estimated three-
dimensional windfield on a specified finite difference grid. This
prospective windfield then is adjusted to account for terrain
effects and atmospheric stability considerations constrained by
the condition that the resulting windfield be nondivergent.
The basic set of model equations was derived from the ap-
proach that minimizes the squared variation of the windfield sub-
ject to the constraint that the adjusted field be nondivergent.8'°
APPLICATIONS
The CARE Dispersion Modeling System has many applications
such as toxic spills, nuclear dose assessment, pesticide drift, am-
bient air quality assessments in complex terrain, chemical fires
and visibility obscurations. The dispersion model is designed to
provide physically realistic calculations of the position and con-
centration of downwind airborne effluents. These calculations
then are available for application in any scenario that relies upon
atmospheric dispersion calculations. The type of application for
which the system is to be used will dictate the type of source term
calculations and the effects and response modules that would be
useful in the CARE system.
Nuclear Accidents
The system is applicable for risk assessment and operational
dose assessments from radioactive releases. It has been im-
plemented for real-time, operational dose assessments using ac-
tual meteorological data signals, for manual and back-up dose
calculations and for training purposes. The model can be used for
dose projection estimates based on forecase or historical
meteorological data. For training exercises, the operator would
run the model using the interactive menus where input data and
run parameters have been selected to simulate the desired exercise
conditions.
Toxic Spills
For toxic spills emergency response, a chemical source term
model has been added which calculates the appropriate release
rate based on chemical properties, release mode and weather con-
ditions. The effects model provides recommended response ac-
tions based on the concentration Threshold Limiting Value or the
Lower Explosive Level.
Visibility Obscurations
By using a visibility model to calculate the effects of an aerosol
released to the air, the system has been used to calculate the
obscuration effects of various smoke releases for military applica-
tions. This capability also is useful in determining visual impacts
on U.S. EPA Class I areas.
Ambient Air Quality
The system is extremely useful for calculating ambient air qual-
ity in complex terrain where traditional straight-line Gaussian
models typically perform unrealistically. The use of the variable-
trajectory calculation capability of the system helps to account
for terrain-induced dispersion. This capability is useful in
transport studies such as odor and toxics impacts.
CONCLUSIONS
CARE gives one the capability to perform a variety of emergen-
cy preparedness functions:
• Risk Assessment
• Hazmat Training
• Real-Time Monitoring
• Emergency Response
CARE is a system that puts truly capable emergency prepared-
ness on one's desk. In the event of an accident that releases a
hazardous material into the air, CARE will tell one where it will
go, who it will effect, which areas are safe as evacuation centers,
how long it will take to disperse and which preplanned response
options would be most effective. CARE gives one the ability to
neutralize the danger, to assess alternative defense strategies and
to keep abreast of changing situations. In short, CARE gives one
the tools needed to prevent an emergency from becoming a
disaster.
REFERENCES
I. Fleischer, M.T., "SPILLS, An Evaporation/Air Dispersion Model
for Chemical Spills on Land," Shell Development Company, Hous-
ton, TX, Dec. 1980.
2. Hogstrom, A.S. and Hogstrom, U.. "A Practical Method for De-
termining Wind Frequency Distributions for the Lowest 200m from
Routine Meteorological Observations." J. Appl. Meteor., No. 17,
1978, 942-954.
3. Mitchell, A.E. Jr. and Timbre, K.O., "Atmospheric Stability Class
from Horizontal Wind Fluctuations," paper presented at the 72nd
Annual Meeting of the Air Pollution Control Association, Cincin-
nati, OH, June 1979.
4. Safendorf, J.F., et a/., "XOQDOQ: Computer Program for the
Meteorological Evaluation of Routine Effluent Releases at Nuclear
Power Stations," NUREG/CR-29I9, Sept. 1982.
5. U.S. EPA, "User's Manual for Single-Source (CRSTER) Model,"
EPA-450/2-77-013, July 1977.
6. Raynor, G.S., el a/., "Recommendations for Meteorological
Measurement Programs and Atmospheric Diffusion Prediction
Methods for Use at Coastal Nuclear Reactor Sites," NUREG/CR-
0936, Oct. 1979.
7. Beebe, R.C. and Sorge, J.M., "Coastal and Overwater Dispersion
Characteristics of a Warm Water Environment," unpublished
paper, Apr. 1975.
8. Patnaik, P.C. and Freeman, B.C., "Improved Simulation of Meso-
scale Meteorology Phase I," Science Applications, Inc., Report
SAI-77-915-LJ, Mar. 1977.
9. Sasaki, Y., "Some Basic Formalisms in Numerical Variational
Analysis," Won., H'ea., Rev., No. 98, 1978, 312-319.
10. Sherman, C.A., "A Mass-Consistent Model for Wind Fields Over
Complex Terrain," J. Appl. Meteor., 1978, 312-319.
38
DETECTION OF RELEASES
-------
The Use of PC Spreadsheet-Based Graphics
to Interpret Contamination at CERCLA/RCRA Sites
George A. Furst, Ph.D.
U.S. Environmental Protection Agency, Region 1
Boston, Massachusetts
ABSTRACT
During site investigations, large amounts of environmental data
are generated in the form of lists of chemical analyses from differ-
ent sampling locations over various periods of time. These tables
of data are difficult to analyze and are seldom fully interpreted
because of the magnitude of the task and the difficulty of visualiz-
ing trends from long lists of numbers. Electronic spreadsheet soft-
ware for personal computers, originally developed for use by the
business community for financial analysis, may be employed to
store site data and to graphically display contamination patterns
at CERCLA/RCRA sites. This software allows the input of
numbers, formulas and words into a matrix of rows and columns
and provides an ideal format for listing contaminant concentra-
tions in the air, soil, surface water and groundwater.
When the above data are entered into a spreadsheet and in-
tegrated with graphics software, it is a simple task to rapidly
visualize contamination distribution at a site over space and time.
The following spreadsheet-based graph types are illustrated: the
line, bar, stacked bar, x-y and pie chart. Each type is well suited
to a particular application. For example, the pie chart is useful for
showing a connection between a source area (such as a lagoon)
and a receptor (such as a downgradient well). The expanded bar
chart can be used to indicate if new sources of contamination are
affecting the downgradient groundwater. Specific applications of
the spreadsheet-based graphics are illustrated using data from
Region 1 CERCLA/RCRA sites.
INTRODUCTION
The advent of the personal computer and its increasing
availability, as well as the increase in user sophistication, has led
to new applications of this powerful instrument. The spreadsheet
has become the basis for many financial as well as scientific ap-
plications. One recent example of the use of the spreadsheet is
that of Olsthoorm.' In this case, the spreadsheet was used to
model groundwater flow in two and three dimensions.
The spreadsheet is used because it is a general purpose, inex-
pensive software product and one need not be a computer
specialist to use it. In the above application, the spreadsheet
replaced three or more specific programs. This paper applies the
spreadsheet analysis to aid in understanding the flow of con-
tamination from an uncontrolled hazardous waste site and from
an active manufacturing site.
Often there is a great deal of information known about a
specific site. This information is available as tables of chemical
data collected while sampling the groundwater, surface water,
soils, sediments and air at the site. Without being able to organize
and visualize this data, one often does not utilize the data to the
full potential. The computer-based electronic spreadsheet, com-
bined with graphics, provides the means to more fully interpret
data.
The spreadsheet, more than any other software program, has
revolutionized the use of the personal computer. The spreadsheet
consists of a matrix of rows and columns in which numbers, for-
mulas and words are input. It is also a very user-friendly software
tool that is limited only by the imagination of the user. Examples
of popular spreadsheet-based software are: Visicalc, Multiplan,
Lotus 1,2,3 and Symphony.
The spreadsheet used in this paper is Symphony, which along
with Lotus 1,2,3 represents a new generation of integrated soft-
ware; word processing and graphics are directly available while
working with the spreadsheet. The Symphony spreadsheet con-
tains 256 columns and 8,192 rows. This capacity allows all known
chemical information from a specific site to be stored on one
spreadsheet. The size of the spreadsheet is demonstrated by the
fact that the above matrix may hold 63,000 individual analyses
consisting of 30 chemical elements or compounds. In total, the
spreadsheet contains over 2 million available data points, limited
only by the computer's Random-Access Memory. The advantage
of storing all the data from one site on one large spreadsheet is the
ability to rapidly compare different sections of data and to con-
duct direct statistical analysis on specific ranges of data. The ap-
plication described in this paper is the ability to focus on sections
of data in the spreadsheet and to make a picture or graph of the
numbers. An important assumption made is that the data are ac-
curate and have passed a quality screening test.
APPLICATION OF SPREADSHEET-BASED
GRAPHICS AT A SUPERFUND SITE
Table 1 is a portion of one spreadsheet of data from a Super-
fund site in southern New Hampshire. It is displayed in the stan-
dard spreadsheet format: i.e., the information is listed in rows
and columns. Note that both numbers and words are used. At
this site, there are a sufficient number of monitoring well sam-
pling locations to define the cross-section of a contaminant plume
adjacent to a water supply. The horizontal axis of the table
represents the distance in feet from the assumed center of the con-
taminant plume to the center of the sampled monitoring well
screen. All samples were taken in the same period of time.
An examination of the data shows that total contamination de-
creases with distance from the center of the plume. Also, the con-
centration of 2 -butanone decreases rapidly, but it is not as ap-
parent that t-l,2-dichloroethene is decreasing much less rapidly
across the plume cross-section.
DETECTION OF RELEASES 39
-------
Table 1
Distribution of Contamination, Plume Center to Water Supply
Distance from Plums Center 0 70 130 175 250 310
Sample number NUS 2-3 NILS 2-2 NIJS 2-1 N'JS -2 C5X 11 NUS 2BR WP-4
Benzene
1,2-Dichlorde thane
1,1,1-Trichloroethene
1,1-Dichloroethane
Chloroform
1,1-Dichloroethene
t-1,2-Dichloroethene
1.2-D ichloroethene
Ethylbenzene
Tetrachloroethene
Toluene
Trichloroethene
Vinyl Chloride
2-Butanone
4-Methyl-2-Pentanone
Total Xylene
Chlorobenzene
Total:
(PPB)
10
51
24
48
760
17
14
410
27
13
1100
69
48
2591
7
41
38
43
7
5
810
5
15
12
350
23
10
520
58
43
1987
7
32
24
37
5
5
590
5
9
10
240
20
10
310
37
24
1365
34
25
520
15
11
36
641
19
35
12
300
5
22
11
404
32
18
5
11
27
93
0.5
5
0.5
6
Note: All concentrations in ppb (or jig'l)
Figure 1 is an x-y plot of the distribution of specific organic
contaminants versus distance from the center of the plume. Using
this plot, one can see the partition of different organic chemicals
in the plume diffusion gradient perpendicular to a groundwater
flow direction. This distribution may be due to differences in
sorption as discussed by Mackay, el al.1 This hypothesis can be
modeled by taking additional data from the spreadsheet (such as
analyses from downgradient wells) and comparing these analyses
to the above theory. Because the data are readily available on the
spreadsheet, examination of contaminant distribution in the
water supply well and the bedrock well is facilitated. These two
analyses are listed in the columns headed by WP-4 and NUS 2BR
on Table 1. The pie chart (Figs. 2 and 3) is jsed because it graph-
CONTAMINANT PUT .E CROSS SECTION
VOLATILE COKT/umMON NEAB WP «U
DISTANCE fDOU PUIME CENTER (IT)
TOtUENE 6 H.2-OPCHUWOETHENE
Figure I
X-Y Diagram of Volatile Contamination in a Groundwater Plume
ically depicts the relative percentages of organics impacting these
different receptors and specifically shows the persistence of
t-1,2-dichloroethene.
APPLICATION OF SPREADSHEET-BASED
GRAPHICS TO AN RCRA SITE
Groundwater monitoring at manufacturing facilities is a critical
aspect of the U.S. EPA RCRA permitting process. Implementa-
tion of the 40 CFR Part 265 and 270 groundwater monitoring se-
quence generates much data related to the permitting process,
either closure or post-closure. If the detection monitoring wells at
a land disposal site detect contamination, the site immediately
goes into assessment monitoring.
The goal of this monitoring is to clearly identify the rate am ex-
tent of hazardous waste constituent migration and to establish the
concentration of individual constituents in the plume. For all of
the above assessment data, the spreadsheet provides an ideal
storage file for later retrieval and interpretation.
The following discussion is an example of the ways in which
one can use the above data once they have been entered into the
electronic spreadsheet. The site utilized for illustration is a
northern New England metal machining and plating company
that has been active since the early 1950s. From 1970 to 1985, an
unlined waste lagoon was used to manage hazardous waste at the
site. Detection monitoring wells were installed in 1981 to assess
the site groundwater flow and potential contaminant migration
irom the waste lagoon. During the early 1980s, about 100,000
gal/day of waste from the manufacturing process entered the
lagoon. The waste stream was composed principally of solvent
wastes from metal degreasing, waste oils from metal machining
and sludge from electroplating operations. In order to assess the
changes in contaminant migration at the site, all data collected
from 1981 to 1985 were entered onto the spreadsheet. Table 2
contains the data collected during the fall of 1982. The following
figures are based upon this table.
40
DETECTION OF RELEASES
-------
MONITORING WELL NUS 2 BEDROCK
TOTAL VOLATILES - 93 PPS
2-Bt/TANOtC (».OX)
TTflCHLOROETHENE (tl.«)
1.1-OCHLOROETHANE (J4.4X)
TETHACHLOROETHENE (S.4*>
1-1.2-OCHLOROETHENE (19.4%)
Figure 2
Pie Diagram of Volatile Distribution in Monitoring Well NUS-2BR
WP WATER SUPPLY
TOTAL VOLATILCS - « ffB
TRICHLOROETHE)* (».3»)
1.1.1-TRCHLOROETMANE (8.3*)
1-I.Z-DICHlOroCTHeNE (83.W)
Figure 3
Volatile Organic Contamination in WP Water Supply
Figure 4 is a stacked bar chart that illustrates the composition
of the oily waste lagoon. This graph is used because it gives a
visual summary of the contaminant profile at different depths in
the waste lagoon in 1982. Figure 5 shows the distribution of con-
taminants at mid-level in the lagoon. Figures 6, 7 and 8 illustrate
the volatile distribution in wells 3, 4 and 7, located 50 ft adjacent
to and 50 and 1,450 ft downgradient from the lagoon. The main
difference in contaminant profiles is the absence of methylene
chloride in the farthest downgradient well MW-7. The other
organic chemicals are present at lower concentrations than in the
lagoon.
MAJOR COMPOSITION OF OILY WASTE LAGOON
SAMPLES Of 10/5/82
35000 -
3OOOO -
25000 -
20000
15000
E7"3 METHYI^NE CHLORIDE
[V~l 1.1.1-TOCHLOROETHANE
IT^ TMCHLOROETHYIENE
rvTl 1.1-DICHLOROETHANE
Figure 4
Volatile Organic Concentrations in Waste Lagoon
The pie diagram is an ideal graph to show the hydrologic con-
nection between source area, lagoon and receptor which in this
case is the downgradient well. At sites where there are multiple
source areas, such as separate lagoons used for waste oil and
plating and degreasing wastes, the pie chart represents a signature
that is useful for indicating which downgradient wells are con-
nected hydrologically with a specific lagoon.
Table 2
Lagoon Composition at Surface, Mid-Depth, Lower and Sludge (10/82)
Monitoring Well Groundwater Composition (10/81)
DATt: 10/5/82
SAMPLE NUMBER Surface
DISTANCE FROM LAGOOM 0
1,1-Dichloroethane 630
1,1-Dichloroethylene
Ethylbenzene
Methylene Chloride
Tetrachloroethylone
Toluene
t-1,2-Dichloroethylene
1,1,1-Tr ichloroethane
Trichlorethylene
Total: (PPB)
10/5/82 1-1/5/8 ^
M id-Dap th lower
0 0
12960
7550
1390
10/5/5'2
sludge
0
13700
10/16/81
MW-1UPG
150
0
22775 43731
164320
34
1.0/1.6/81
MW-2ADJ
150
11
45
1.0/16/1)1
MW-3
50
830
170
10/16/131 10/81
MW-4 MW-7
50 1450
1500
570
110
7500
174
2586
2070
9900
0
3000
2325
30500
121
6350
5370
96900
1240
1180
23100
28200
0
34
12
22
820
340
21
140
1000
270
470
14
1300
33
1300
70
1100
430
950
3591
4617 3230
Note: All concentrations in ppb 0
-------
OILY WASTE LAGOON. MID-DEPTH
IOIAL voutus - 2277s m
HWCHUXlOCTmVCW (I0.2X)
M.I-TWCHlOfWCTHKNC (13.2X)
I.I-OCW-OHOCTHWC (31 !»)
urnmrxc cmonct <«3 M)
Figure 5
Volatile Organic Concentrations at Mid-Depth in Waste Lagoon
MONITORING WELL 4, OOWNGRAOlENT
TOT*. KXATUI - «M4 m
I.I-OCMU*OCTNW*OCT»*M*t (07X)
TTTHMXJCMOCTMtUM (OJX?
Figure 7
Volatile Organic Contamination of Groundwater at MW-4
MONITORING WELL 3. DOWNGRADIENT
i - 3M t ^0
MONITORING WELL 7. OOWNGRA0IENT
I0t«t MX*rU3 - 3241 »»•
mO«.OHOCTHTU>C (
1.1-DICHt.O
-------
SO MONITORING WELL 3
PERIOD SAMPLED 10/81 - 10/85
SO MONITORING WELL 3
CONCENTRATION OF METHYLENE CHLORIDE
1.1 -OICHLOROETHANE
t. 1.1 -TR1CHLOROETHANE
BATE SAMPLES ANALYSED
U7~* 2/U/82 r^J 10/5/82
TRICHLOROETHYUNE
r\7\ 8/11/&J
E53 5/25/84
| 10/85
Figure 9
Volatile Organics Present in MW-3 (10/81 - 10/85)
The spreadsheet also provides the basis for comparing changes
in organic contaminants at a specific well over a number of years.
This comparison is facilitated since data can be moved rapidly
from one section of the spreadsheet to another. Table 3 contains
data from monitoring well 3 which were collected from October
1981 to October 1985. Figure 9 is a bar graph of the major organic
contaminants, which, with the exception of methylene chloride,
are contained in the groundwater. The graph illustrates that the
highest organic contamination at well 3 occurred in late 1982.
Since then, there has been a decrease in the concentration of the
three compounds, as shown in this figure. Figure 10 shows the
concentration of methylene chloride over the same time period.
The highest concentrations of this species were observed in Oc-
tober 1982 and October 1985. The later date coincides with the
final closure of the lagoon and may be related to it. This line chart
is particularly useful for showing changes in contamination over a
period of time.
In summary, spreadsheet-based graphics allow one to visualize
the changes in concentration of organics in the source area,
lagoon and downgradient monitoring wells and changes over a
4-year period in well number 3.
O DATES OF SAMPLING
Figure 10
Concentration of Methylene Chloride at MW-3 (10/81 10/85)
CONCLUSIONS
The spreadsheet is the fundamental way to store data in the
computer. Once data from a hazardous waste site or manufactur-
ing site are entered into the spreadsheet, one can readily combine
data, move data and, when the spreadsheet is combined with
graphics software, one can generate graphs from any portion of
the spreadsheet. In this paper, the spreadsheet has formed the
basis for quickly assessing the contamination at two sites. The
software also has been used by the author to prepare graphs for
overhead projection at a meeting, to prepare tables and graphs
for reports and to forecast when increasing contamination in a
well will cause the water to exceed Safe No Action Response
Levels (SNARLS). With the increasingly available portable com-
puters, the above site-specific information can be taken to the
manufacturing company and used as a resource during discus-
sions regarding contaminant distribution at an RCRA site. As
stated earlier, the application of this new generation of software is
limited only by the imagination of the user.
REFERENCES
1. Olsthoorm, T.N., "The Power of the Electronic Worksheet: Model-
ing Without Special Programming," Groundwater, 23, 1985, 381-390.
2. Mackay, D.M., Roberts, P.V. and Cherry, J.A., "Transport of Or-
ganic Contaminants in Groundwater," Environ. Sci. Technol., 19,
1985, 384-391.
DETECTION OF RELEASES 43
-------
Emergency Response to Toxic Fumes and Contaminated
Groundwater in Karst Topography:
A Case Study
P. Clyde Johnston
Mark J. Rigatti
Roy F. Weston, Inc.
Jacobs Engineering Group, Inc.
Atlanta, Georgia
Fred B. Stroud
U.S. Environmental Protection Agency Region 4
Atlanta, Georgia
ABSTRACT
This paper presents a preliminary report on the on-going con-
trol of contaminated aquifers in karst topography. Over the past
several years, there has been a growing concern over the presence
of volatile hazardous chemicals in the karst formation cave sys-
tems in the Bowling Green, Kentucky area. By April 1985, con-
tamination and related chemical fumes warranted an "emergency
response action" by the U.S. EPA, Region 4. The management
of migration procedures under these conditions are discussed and
summarized. Dye tracing, geophysical studies, speleological car-
tography, exploratory drilling, down hole cameras and sampling
techniques have been utib'zed to locate the primary flow char-
acteristics of the "Lost River" and its tributaries and to identify
the source(s) of contamination and mitigate its effects.
Karst formations present peculiar and distinct management
difficulties. The impact of leaking underground storage tanks
and illegal dumping may be magnified by the motility of spilled
substances. Pollution source determination is especially proble-
matic.
Many methods used in this emergency response action were
relatively experimental with few strong precedents. These
methods are comparatively analyzed for the benefit of future
cleanup activities.
INTRODUCTION
Karst topography, evident to some degree almost anywhere
carbonates are found at or near the surface of the earth and sub-
jected to humidity and water movement, presents unique and
perplexing difficulties to the professional involved in hazardous
waste management. By definition, karstic terrains are character-
ized by closed depressions, i.e., sinkholes or dolines and uvalas,
caves and the prevalence of underground drainage. Although
these conditions can be beneficial, as in the case of high capacity
groundwater accessibility (it is the closest thing in nature to a
fresh water pipe), extreme liabilities also may be present. For ex-
ample, the impact of leaking underground storage tanks and
illegal dumping may be greatly magnified by the motility of lost
substances. While permeability in non-karst areas may be meas-
ured in ft/day, permeability in karst can be measured in miles/hr.
Discrete sources of groundwater recharge are numerous and var-
ied and the velocity of recharge so high that pollution source de-
termination can be especially complex.
APPROXIMATE SCALE IN MILES
1 LOST RIVER UVALA I
CAVE ENTRANCE
3 LOST RIVER RISE
Figure 1
Bowling Green Location
Located on the textbook karstic terrain of south central Ken-
tucky, Bowling Green is the only city of such size (pop. 41,000)
in the United States to rest almost entirely over a cave system.
The Lost River, which is exposed in a uvala in southwest Bowling
Green, flows into the Lost River Cave entrance and has been
mapped for 4 miles. At this point, the river sumps and takes an
undetermined flow path through the St. Genevieve and St. Louis
limestones underlying Bowling Green until it exits north of the
city at the Lost River Rise.
44 CONTAMINATED GROUNDWATER CONTROL
-------
Fixed facility and transportation related spills, deliberate
dumpings and undetected leaks, all associated with oil and haz-
ardous materials, have become daily occurrences in every major
city across the nation. Bowling Green is no exception. Its loca-
tion in a sinkhole plain and over an extensive cave system, makes
the city highly susceptible to groundwater pollution and to the re-
lated problems associated with karst areas.
Many contaminants can immediately affect the watershed
either through unobstructed flow into the cave system or through
rapid percolation through thin cover soils. In Nicholas Craw-
ford's earlier research, he showed that heavy metals and road
salts from urban detritus quickly entered the groundwater sys-
tem,1 but volatile materials pose special problems. As the sub-
stances flow through the subterranean labyrinth, they can volatil-
ize in the cave air space and ultimately find their way to the sur-
face. The control of gases and vapors in the atmosphere usually
is not attempted because natural dispersion dilutes vapors and
gases rapidly. However, fumes rising from cave systems may con-
centrate in areas of concern. If these fumes collect in a basement,
crawl space or simply in or near public areas, a potentially explo-
sive situation exists.
Bowling Green's Problem
The toxic and explosive fume problem in Bowling Green dates
back to 1969, when explosive fumes were detected and believed
to have caused a residential explosion. In 1981, the residents of
five homes in the Riverwood area were evacuated when fumes,
rising into their basements and crawl spaces, reached explosive
levels.
In 1982, benzene and methylene chloride were detected in the
Lost River Cave. Subsequently, the U.S. EPA, Region 4, initiated
an emergency response action in March of 1983. Preliminary in-
vestigations showed alarmingly high levels of volatile chlorinated
hydrocarbons2 flowing into the spring-fed Keith Farm pond,
which discharges into the Lost River drainage system. Dye traces
pinpointed the source of the contamination to be a series of leak-
ing industrial underground storage tanks. The U.S. EPA treated
1.9 million gal of contaminated pond water, and the responsible
party removed the leaking tanks and excavated the contaminated
soils.3
Throughout 1984, toxic and explosive fumes plagued the homes
in the Forest Park and Parker-Bennett School areas. In January
1985, toxic and explosive fumes were detected in the Dishman-
McGinnis Elementary School. To date, fumes of this type have
been investigated in over 100 homes and commercial buildings,
two schools, one church, 18 drainage wells and 24 sinkholes.
After an extensive public health evaluation of the fumes by
the Center for Disease Control, a health advisory was issued on
the grounds that explosive levels existed and that the detectable
fume levels of benzene, toluene and xylene exceeded standards
for "non-occupational settings," At that time, the U.S. EPA
Region 4 Emergency Response Branch initiated a "Superfund"
emergency response to address the problems designated in the
public health advisory. Mitigation efforts were initiated to re-
duce the threat, identify the source or sources of contamination
and define the underground drainage system controlling the
transport of chemicals in the Bowling Green area.
upon the assumption that continuous releases or intermittent
dumpings into the Lost River drainage system, i.e., sinkholes,
solution joints and fractures, have resulted in a chronic flow of
chemicals within the caves.
The first hypothesis is that, once in the cave, the chemicals
volatilize; a volatilization that is accelerated by natural struc-
tures. Waterfalls and riffle areas provide an indigenous turbu-
lence which increases the rate of volatilization, increasing, in
turn, the fume concentration within the cave atmosphere. Dur-
ing high water flow, the air space within the cave is decreased,
forcing the gases to the surface. This process not only occurs dur-
ing high water flow, but also during normal cave breathing due
to changes in barometric pressure.
The second hypothesis is that water-filled passages, forming a
natural underflow dam, trap floating chemicals and restrict them
from flowing downstream. This flow restriction, in turn, causes
low density and volatilize chemicals to become concentrated in
these areas. This frequently occurs during increased river dis-
charge when heavy rainfall causes the river within the cave to rise.
Consequently, the floating chemicals and the fumes are pushed
upstream.
As the water recedes, the floating chemicals adhere to the ceil-
ing and walls forming a "bath tub ring" within the system. This
bath tub ring phenomenon coats the surface of upstream perched
pools with floating chemicals. The rise and fall of the river level
results in a repetitive deposition of chemicals within the cave,
leaving a stratified layer of aliphatic hydrocarbons and mud.
Finally, it is proposed that contaminants have been dumped or
spilled in soils and areas underlain by perched water tables or
water pockets and are flushed into the Lost River Cave system
when meteoric water percolates through the soil or overflows a
perched water table. Perched water tables and water pockets are
associated with the pinnacle nature of the limestone at the lime-
stone/soil interface (Fig. 2).
WATER POCKET
SOIL
(Terra Rossa)
PERCHED WATER
TABLE
(VARIABLE SIZE)
SOLUTION CONDUITS
FLOWING TO LOST RIVER
Figure 2
Generalized Illustration of Pinnacle Nature of Limestone Showing
Possible Perched Water Table and Water Pockets
FUME OCCURRENCE HYPOTHESES
Although not fully documented, there is an apparent direct re-
lationship between increased reports of fumes and wet weather;4
similarly, there have been fewer fume incidents during dry
weather. Thus, the fumes appear increasingly manifest in periods
of heavy precipitation. The hypotheses which follow are based
These hypotheses are not mutually exclusive. Investigative tech-
niques in verifying them and solving the resulting problems asso-
ciated with them have been varied. In this verification, geological
techniques were used in conjunction with sound environmental
studies and common hazardous waste site cleanup practices.
CONTAMINATED GROUNDWATER CONTROL 45
-------
SEARCH FOR THE LOST RIVER
The logical place to start this investigation was to locate the
contaminated medium, i.e., the Lost River. In this search, an in-
tegrated team approach based on multidisciplinary scientific sup-
port was utilized to address the diverse problems associated with
the fumes rising from the cave system.
Excavation of sinkholes and crevasses initiated the physical ex-
ploration of the project. In the search for possible entrances into
cave passages, sinkhole breakdown was removed with a track-
hoe or a backhoe and by hand. Excavations were made adjacent
to homes and buildings at identified channels from which fumes
were reaching the surface. Excavations also were made within
sinkholes along the hypothesized Lost River flow path and where
historical accounts indicated the closing of cave entrances.
Concurrent with these excavations, exploratory drilling was
initiated. However, with few means of locating drilling sites for
cave passages other than topographic features, the production of
exploratory wells at 10-ft intervals appeared to be a costly and
time-consuming method of cave discovery. Clearly, a method was
needed to locate drilling sites more precisely and economically.
This need led to the implementation of gravity survey techniques.
Gravity Survey
Gravity surveys are utilized by, and in fact originated in, the oil
industry, where there is great interest in detecting large under-
ground structures. In microgravity surveys, which include cor-
rections for tidal effects, more sensitive instrumentation and
more precise techniques, it is possible to detect very small and
finely localized variations in geologic structure. In Bowling
Green, it was anticipated that microgravity surveys would de-
tect the presence of void space (or absence of solid rock) along a
given transect.
Using a Lacoste and Romberg Model D Microgal Gravity
Meter, 10-ft stations were measured over thousands of feet of
transects in the study area. The readings were plotted and large
anomalies were investigated by exploratory drilling. It often was
found that anomalous readings represented unusually deep
troughs of soil between pinnacles of limestone or break down
zones of possible cave collapse (Fig. 3). Nevertheless, micro-
gravity surveys proved to be the most reliable means of locating
void space or cave passages when no historic or surface indica-
tions were available.
9
.„;
2000 IBOOcoo 1*001200 ioootoo «oo 400 zoo
Figure 3
Representative Comparison Between Topographic (Solid Line) and
Microgravity (Dotted Line) Surveys Showing Probable Breakdown
Zone (A) and Small Solution Void (B)
Drilling
Exploratory and core drilling, utilizing rotary water and forced
air hammer drill rigs, continued throughout the investigation to
verify the microgravity data and to correlate localized strati-
graphic columns. Many of the exploratory and core wells later
became monitoring wells used to record a wide variety of rele-
vant variables. In certain instances, monitoring wells were placed
by survey and drilled into the stream thalweg of cave passages,
eliminating the need for lengthy and labor intensive cave
sampling and/or monitoring trips.
Standard hydrologic methods, including dye tracing and piezo-
metric surface contouring, were used to locate primary flow char-
acteristics, while air monitoring was used to locate and character-
ize areas of fume contamination. A 1.5-in. downhole camera
with both downhole viewing (Lighted Actual Viewing Attach-
ment, LAVA) and sideviewing (Right Angle Viewing Attach-
ment, RAVA) capabilities also were used to investigate the sub-
terranean environment.
The investigation team occasionally experienced difficulty in
distinguishing between solution voids and breakdown zones and
between true void space and mud-filled cavities. The downhole
camera allowed visual inspection of these features, ensuring cost-
effective and time-saving verification of the presence of cave
passages. Also, the downhole camera confirmed the indications
of microgravity data and historic accounts that a large cave
existed in the vicinity of Creason Street and Robinson Avenue
area.
This discovery prompted drilling of a new caliber. A 30-in.
core barrel drill rig was employed to drill entrance wells into
Robinson's Cave and Creason Cave. These caves were actually
one prior to a 1960s road construction blasting accident which
collapsed 100 ft of road, effectively severing the original cave
passage.
Utilizing a team of experienced spelunkers, detailed speleo-
logical investigations and complete surveys of the caves were in-
itiated. From these explorations it was determined that the Robin-
son's and Creason Cave system represented a paleo-channel of
the Lost River, allowing the realization that the river was not
confined by the resistant Lost River chart,' but that it had broken
through into what Crawford calls a "Karst Sandwich" (Fig. 4).
ST. CENEVIEVE
LIMESTONE
GRANULAR LIMESTONE
(FOSSIL FRAGMENTS t COLITESI
DOLOMITE
Figure 4
Partial Section Showing Generalized Flow of Lost River into "Karst
Sandwich" of Chert and Limestone
46 CONTAMINATED CROUNDWATER CONTROL
-------
Unfortunately, the Lost River proper has yet to be found by
way of Robinson's Cave. The cave does, however, represent a
tributary conduit into the Lost River system contributing con-
taminants from perched water tables and/or soil percolation.
Another cave investigation was prompted by fumes rising to
the surface. In one residential crawl space and in an adjacent
storm sewer and drainage well, fumes reached explosive levels
frequently. The investigation team began hand excavation accom-
panied by continuous air monitoring in the drainage well. During
this apparently minor excavation, fumes peaked at levels above
300 ppm and 100% LEL, prompting a more comprehensive ex-
cavation. Using spark-proof tools, an entrance to a new, hitherto
unknown, cave was discovered. This cave, Napier's Cave, be-
came the focal point of further study and investigation. Physical
exploration of this cave required positive pressure flushing of con-
taminant fumes, utilizing two high volume smoke exhaust fans
blowing into the cave. Exploration revealed floating contam-
inants, wall staining and contaminated perched pools, all indi-
cating contamination transport at high and low water flows.
SAMPLING
The ultimate success of determining locations where investiga-
tions could be efficiently performed, of determining mitigation
procedures and of initiating possible enforcement actions de-
pended on accurate and valid analysis of the chemicals involved
and of the vulnerability of the environment to their effects. This
information was obtained by extensive sampling and monitoring
of soil, air and water matrices characterizing the contaminants
and affected medias. Potential and real effects on the environ-
ment, immediate and long-term risk to public health and to the
investigating team were established and monitored. Several
methods were employed.
Air monitoring was performed with the use of real-time, direct-
reading, portable instrumentation, -photoionization detectors,
combustible gas indicators and oxygen meters. This highly mobile
instrumentation proved very convenient in screening cave atmos-
pheres and monitoring stations over the large study area. How-
ever, instrumentation limitations were revealed. The HNU photo-
ionizer, Model PI 101, gave unreliable readings within the cave
environment and in the monitoring wells. This problem was due
to the cave humidity and subsequent condensation on the ultra
violet light source.
The GasTech, Hydrocarbon Super Surveyor, Model 1314, was
the most practical portable instrumentation for use in cave en-
vironments. It enabled monitoring for the parameters of total
organics, combustible gas and oxygen with one instrument. Its
light weight and size, ease in calibration and relative insensitivity
to moisture made it the preferred instrument for speleological
monitoring.
Once initial contaminant parameters were identified, species-
specific instrumentation provided accurate quantification. Specif-
ically, monitoring was aided by the use of a Photovac, a pro-
grammed photoionization gas chromatograph. The Photovac was
placed within a crawl space in the Parker-Bennett Elementary
School. Scheduled samples were automatically collected and
immediately analyzed before and during building occupation.
Generated chromatographs provided the specific qualitative and
quantitative data utilized for public health considerations.
The U.S. EPA Environmental Response Team's mobile en-
vironmental unit (SCIEX) also was utilized in the study. The com-
puter-coupled tandem mass spectrometer, mounted in a specially
adapted bus, provided real-time measurements of contaminants
in ambient air from the atmosphere, monitoring wells and within
buildings and homes. This versatile field laboratory was not lim-
ited geographically and provided quality assured quantitative
and qualitative analysis in minutes. It proved to be the most effic-
ient and accurate means of screening air samples and identify-
ing compounds.
Tedlar air sample bags were used for capillary soil gas
sampling and occasionally for atmospheric sampling when sub-
sequent qualitative laboratory analysis was indicated. The soil
gas technique was employed at industrial sites to establish a corre-
lation between volatile chemicals associated with those sites and
those identified rising from the Lost River Cave system, and
where fumes were detected in the absence of any surface evi-
dence of escape paths. In addition to real time air monitoring and
sampling, Tenax collector tubes were employed, while subse-
quent desorption and analysis provided time weighted average
concentrations for public health considerations.
Soil, sediments and water samplings completed the charac-
terization of chemicals involved and the assessment of this com-
plex problem. Surface and cave sediment samples were collected
from drainage ditches, cave walls and cave stream beds to iden-
tify contaminants and to establish correlations between industrial
releases of chemicals and cave contamination. Soil samples also
were collected, specifically as a means of documentation during
underground storage tank excavation.
Due to the scarcity of surface streams in the Bowling Green
area, surface water samples were limited to the Lost River Cave
Entrance and Rise. However, groundwater samples were collected
directly from the caves, monitoring wells and potable wells. Cave
stream and monitoring well samples helped to verify contam-
inant flow paths, while potable well samples were useful in pub-
lic health determination and helped to locate contaminated
perched water tables.
MITIGATION
A priority throughout the investigation was mitigation of
immediate and potential threats to health and safety. Prompted
by the 1983 U.S. EPA identification of leaking underground
storage tanks as a source of Lost River contamination, an exten-
sive search was performed to locate all storage tanks in study
areas. Data were compiled to indicate the integrity of the tanks.
Several underground storage tanks were tested and removed or re-
paired if leaks were detected.
Similarly, the discovery of one residence that relied entirely on
wellwater and the subsequent determination that this water was
contaminated, led to an extensive data search and public survey
in an effort to identify all water wells in the study area. Residents
on contaminated potable water supplies were connected into the
city water system.
The desired culmination of mitigation with regard to hazardous
fumes is the ultimate elimination and removal of those fumes and
their sources. However, prior to the identification of the chem-
ical nature of the fumes and their precise sources, the most
immediate and obvious mitigation procedure was the venting of
homes and public buildings. Using exhaust fans in crawl spaces,
basements and cave entrances, fumes were dispersed into the
atmosphere instead of being allowed to concentrate in closed and
inhabited areas.
CONCLUSIONS
As a part of an on-going study, the following conclusions on
particular aspects of the investigation and mitigation of contam-
inated aquifers in karstic terrains are tentative. Based on informa-
tion derived from this study, the body of this paper might pro-
vide guidelines for future investigations of hazardous waste gen-
eration and handling in karst areas.
CONTAMINATED GROUNDWATER CONTROL 47
-------
Ventilation of fumes
Venting provides a convenient short-term means of lowering
or eliminating fume concentrations in living areas and public
buildings. Venting is, of course, a short-term solution to a long-
term problem.
Monitoring
Clear-cut advantages were noted between instruments applied
to various monitoring tasks. The GasTech proved superior in
speleological monitoring due to its relative insensitivity to mois-
ture and multiple function capacity. Single function instruments
used in conjunction with multi-function instruments provided
data checks. The HNU with its increased sensitivity was ideal
for surficial monitoring of organic vapors. Similarily, the ex-
plosimeter's specialized function provided a comparison and
verification of data accuracy.
Microgravity Survey
Not overlooking topographic features and historic information
of an area, microgravity surveys were the most effective tech-
nique utilized in this study for the remote sensing of karstic fea-
tures.
Smoke Testing
Although smoke testing met with little success in this investiga-
tion, we believe that a very dense paniculate and aromatic smoke
with an efficient injection system could be an effective tool in
locating small or hidden cave openings.
Drilling
Although rotary water drilling initially was used as a safety
precaution to prevent spark ignition of cave fumes, it is now
thought that forced air rotary and hammer drilling are acceptable
methods in all but the most dangerous areas; the forced air would
flush fumes away from any spark generation.
Tank Testing
As national attention to the Leaking Underground Storage
Tank Program and hazardous material handling grows, special
attention should be paid to karst areas. Precision testing of
underground storage tanks identified leaking tanks. Their re-
moval or repair possibly may have eliminated the source of con-
tamination and the subsequent fume problem in a Lost River
tributary.
Chemical Assessment
The only substance releases whose consequences might be min-
imized in karstic terrains could be those of acids which rapidly
neutralize in contact with limestone. In all other instances, oil
and hazardous substances releases in karst areas can rapidly
migrate and generally pose more complicated and variable prob-
lems than in non-karst regions.
These studies further concluded that all of the water wells in
the study area, with one exception on its boundry, were contam-
inated with a variety of hazardous substances. Some of these
contaminants were not present in the Lost River; hence it was in-
ferred that these contaminants were trapped in perched water
tables, migrating down well into the phreatic zone.
As general methods and specific techniques of hazardous waste
management continue to develop and improve across the country,
special attention should be paid to environmental influences on
hazardous waste dispersion. Karst areas, in particular, require
specific guidelines and possibly regulative measures governing
hazardous material generators, transporters and emergency re-
sponses.
FOOTNOTES
I. Crawford, N.C., "Sinkhole Flooding Associates with Urban Develop-
ment upon Karst Terrain: Bowling Green, Kentucky," in B.F. Beck,
ed., Sinkholes: Their Geology, Engineering, and Environmental Im-
pact, A.A. Backema, Rotterdam, 1984, 283-292.
2. U.S. EPA, "Superfund Cleanup, Keith Farm Pond, Bowling Green,
Kentucky," Vols. I & II, 1983.
3. Ibid.
4. This section owes much to the formal work of, and informal dis-
cussions with, Nicholas C. Crawford; see footnote I, above.
5. The relative position of the Lost River chart in the stratigraphic col-
umn is subject to some disagreement. The Geological Survey places
the Lost River chart at the base of the St. Genevieve formation; how-
ever, authorities in the field place the Lost River chart in the Horse
cave member of the St. Louis formation. Cf. Palmer, N., A Geologi-
cal Guide lo Mammoth Cave National Park, Zephyrus Press, Tea-
neck, NJ, 1981, 196; Pohl, E.R., "Upper Mississippian Deposits in
South Central Kentucky," Kentucky Acad. of Sciences, Trans., 31,
1970, 1-15. This placement has been further confirmed through per-
sonal communication with Nicholas Crawford.
REFERENCES
1. Beck, B.F., ed.. Sinkholes: Their Geology, Engineering and Environ-
mental Impact, A.A. Backema, Rotterdam, 1984.
2. "Bowling Green Toxics," U.S. EPA Report, Vol. I., 1985.
3. Crawford, N.C., "Toxic and Explosive Fumes Resulting from Con-
taminated Groundwater Flow through Caves Under Bowling Green,
Kentucky," unpublished research proposal submitted tothe City of
Bowling Green, KY, April 1985.
4. Palmer, A.N., A Geological Guide to Mammoth Cave National Park,
Zephyrus Press, Teaneck, NJ, 1981.
5. Pohl, E.R., "Upper Mississippian Deposits in South Central Ken-
tucky," Kentucky Acad. of Sciences, Trans., 31, 1970, 1-15.
6. U.S. EPA, "Superfund Cleanup, Keith Farm Pond, Bowling Green,
Kentucky," Vols. I & II, 1983.
7. Tejada, S., "EPA Goes Underground at Kentucky Superfund Site,"
EPA J. 2. July/Aug. 1985, 26-27.
48 CONTAMINATED GROUNDWATER CONTROL
-------
Use of Low Flow Interdiction Wells to
Control Hydrocarbon Plumes in Groundwater
John H. Sammons, Ph.D.
Ninth U.S. Coast Guard District
Cleveland, Ohio
John M. Armstrong, Ph.D.
The Traverse Group, Inc.
Ann Arbor, Michigan
ABSTRACT
This paper presents a case history and discussion of a situation
where the initial hydrocarbon contamination occurred in 1969, re-
mained undiscovered until 1980 and since has resisted massive
efforts to even adequately define the situation. The paper de-
tails the installation of low-flow wells across the plume to block
flow of contaminants. This interdiction resulted in a serendipi-
tous activation of the naturally occurring soil microbes in the
sediments down-gradient from the well field. The net result of the
interdiction and microbiological activation was to effect a rapid
and dramatic reduction in the hydrocarbon contaminants in the
plume down-gradient from the interdiction well field.
INTRODUCTION
Hydrocarbon contamination of groundwater is rapidly becom-
ing one of the most troublesome and costly problems to be con-
fronted by both the regulators and the regulated community.
Virtually every day, an article is written announcing another
underground water supply that has become contaminated and
that it appears virtually impossible and prohibitively expensive
to contain and completely decontaminate the aquifer.
SITE HISTORY
In July 1942, the United States Navy established an Air Sta-
tion at Traverse City, Michigan, a small and isolated community
located in the northwestern section of the Lower Peninsula
(Fig. 1). The purpose of the Air Station was to conduct highly
classified research and development of pilotless drone aircraft
that could be remotely guided by television to the target from
chase aircraft. This research effort was continued until 1944 when
it was suspended. When the war ended, the Air Station was
turned over to the United States Coast Guard (USCG) to serve as
a major Search and Rescue base for Lakes Superior, Huron and
the upper portion of Lake Michigan.
Coast Guard Operation
Coast Guard operations at the site commenced in 1944 and
have continued until the present with a mix of rotary-wing and
fixed-wing aircraft. A review of operations revealed nothing
extraordinary or unusual had been reported although historical
records of the station are sketchy.
History of Problem
In 1979, during the removal of two fuel farms preparatory to
the installation of a new system, there was some indication that
leaks had occurred. The only soil contamination was in the Jet
Fuel (JP-4) storage area. These soils were removed and disposed
of under the direction of the State of Michigan Department of
Figure 1
Location of East Bay Township and U.S. Coast Air Station in Michigan
Natural Resources (MDNR). This area was located some 1,500 ft
upgradient and to the north of the area that was ultimately impli-
cated as the "geographical origin" of the plume. The Aviation
Fuel (115/145) fuel farm, located immediately adjacent to the
"geographical origin" of the plume, was excavated at the same
time with little indication that any unusual occurrences had taken
place. There was some odor noted in the soil, but laboratory
analyses did not confirm gross contamination, so no soil was re-
moved.
In 1979 and 1980, residents in the Avenue E area of the Pine
Grove Subdivision of East Bay Township complained to the local
health department that their wells were producing discolored
-water, were foaming and had bad tastes and foul odors. The first
CONTAMINATED GROUNDWATER CONTROL 49
-------
residence reporting the problem is located 1,200 ft to the north-
east of the Coast Guard Air Station (CGAS). At that time there
was no explanation for the contaminated wells and, unfortunate-
ly, the health department did not test for any of the possible
organic contaminants. Later in 1980, the MDNR did a limited
hydrogeologic study in the area and concluded that the source of
the contamination was from some unspecified site on the CGAS.'
In May 1982, the government was notified of the findings and be-
gan an attempt to elucidate the problem.
COAST GUARD ACTIONS
The U.S. Geological Survey (USGS) was retained in June of
1982 to do a thorough hydrogeological study of the area to define
the plume of contaminants and to find the source. By April 1983,
the USGS had determined the direction and velocity of ground-
water flow through the area and had tentatively identified the
boundaries of the plume (Figs. 2 to 5). The USGS was not able to
determine the source of the contaminants. They did, however,
conclude that the majority of the contaminants identified were re-
lated to components in fuels with some chlorinated compounds
also present.1
Detailed Studies
In November 1983, the Coast Guard contracted with the Uni-
versity of Michigan (UM) for a scientific study and feasibility
study of the site. Objectives for these studies included a tem-
poral analysis of the plume and a positive identification and loca-
tion of the contaminant origin(s). The final report' included a
description of the time variation of contaminants and adsorptive
characteristics of soils, including side-by-side comparisons of
n n w n to w
WEST ARM GRAND
TRAVERSE BAY
EAST ARM GRAND
TRAVERSE BAY
Principol
study area
EXPLANATION
-roo-LINE OF EQUAL ALTITUDE OF
LAND SURFACE--Interval
20 feet NGVO of 1929
10-WATER-TABLE CONTOUR--
Shows altitude of water table.
Contour interval 10 feel.
»- GROUND -WATER FLOW--Arrow
indicates direction of flow
GROUND-WATER DIVIDE
Base from U S Geological
Survey I 62.SOO quadrangles
Figure 2
General Direction of Groundwatcr Flow
50 CONTAMINATED GROUNDWATER CONTROL
-------
water and soil concentrations found by MDNR, USGS and UM;
a description of the quality assurance program for collection and
analysis of groundwater samples; a preliminary risk assessment
for chemicals found in the groundwater and a description of the
effort to numerically model the behavior of the groundwater
plume, a description of the groundwater sampling program and a
preliminary remediation analysis that identified several cleanup
alternatives.
During this time, the USCG was pursuing an internal investiga-
tion which resulted in the discovery of an Aviation fuel spill inci-
dent. In November or December of 1969, a flange in an under-
ground pipe line under a 115/145 high octane Aviation Gasoline
fueling station failed, resulting in the loss of approximately 2,000
gal of product over a 12-hr period.
The Traverse Group Inc. (TGI), an Ann Arbor-based multi-
disciplinary consulting firm, was contracted by the USCG to do a
detailed feasibility study that included the design, construction
and operation of groundwater treatment plants (involving carbon
adsorption and air stripping techniques) and to continue a sam-
pling program for selected wells to monitor the plume. This part
of the investigation was completed in February 1985." TGI was
then asked to continue the project to include the following tasks:
• Selection of a specific cleanup technology
• Design, construction and operation of an interdiction well
system
• Design and installation of monitoring wells
• Design, construction, operation and evaluation of a full-scale
carbon adsorption system
• Assembly, installation and evaluation of an advanced rotary
air stripping and vapor incineration system
Figure 3
Water Table and Direction of Groundwater Flow on Apr. 5-7, 1983
CONTAMINATED GROUNDWATER CONTROL 51
-------
In February 1985, new hydrocarbon contamination was dis-
covered at the JP-4 fuel farm south of the Hangar/Administra-
tion building (HAB). The four fiberglass underground storage
tanks at the station were tested; three were found to be leaking
and were removed. In March 1985, routine sampling indicated
the presence of high levels of benzene, toluene and lighter hydro-
carbons in monitoring wells along the south boundary fence of
the USCG property upgradient from the JP-4 Fuel Farm. The
origin of this contamination was not then, nor is it now, under-
stood since these wells are now clean. One possible explanation
for the contamination in these wells was a spill of approximately
2,000 gal of Jet-A at the Republic Air Lines fuel farm some 1,800
ft directly upgradient from the south boundary of the USCG
property. What cannot be explained is the high contaminant
levels when the wells were sampled in March 1985, and the current
absence of contaminants in the wells.
Plume Characterization
Much of the investigative effort focused on the location and
characterization of the groundwater plume. All agencies involved
installed and monitored observation wells for this purpose. Due
to temporal variations of the contaminants and differences in
analytical techniques, not all results were in complete agreement;
however, there is sufficient similarity to delineate locations and
concentrations of the plume's major constituents.
\ EXPLANATION
*l\ • «VU iOUtiOi
Figure 4
Water Table and Direction of Groundwater Flow on July 19-20, 1983
52 CONTAMINATED GROUNDWATER CONTROL
-------
LOCATION OF
INTERDICTION WELLS
EXPLANATION
— 590— WATER-TABLE CONTOUR--Showi
olliludc ot limuloled -aw table
Aoril 1383 Contour interval I loot.
NCVO ol 1929
»-GROUND-WATER FLDW--Arrow
indtcotci direction of fto*
AIR STATION BOUNDARY
:•:•:•:• PLUME OF CONTAMINATION
O 500 ICOOFEET
Bole odapfrrd Irom U.S Cootl Guard mop and
mopi bv Gojrdic-Frotcr and Atiocialtf. Inc
IOO 200 METERS
Figure 5
Water Table for Apr. 1983, Simulated by Groundwater Flow Model
The UM study3 identified benzene and toluene as the compon-
ents in the plume presenting the greatest health risk. The largest
concentrations occurred in the vicinity of the Hangar/Adminis-
tration Building (HAB) at the geographical head of the plume
(Figs. 6 and 7) although significant amounts of some compounds
such as benzene were found at some distance downgradient.
Other chemicals also were found in the plume but at smaller con-
centrations and reduced distributions.3
The USGS provided a detailed map of the plume's size and
location (Fig. 1). From the HAB, the plume follows groundwater
flow to the northeast and off the base, passes under an industrial
park and turns slightly north, narrowing as it passes underneath
Parsons Road and widening out again under Avenue E (Fig. 5).
The plume is approximately 4,300 ft long and varies from 180 to
400 ft in width. Its vertical dimension ranges from 25 to 80 ft.
Small concentrations of benzene and toluene have been de-
tected in the water of East Bay. The USGS2 reported maximum
values of 20 /ig/1 benzene and 3.1 /tg/1 toluene aproximately
330 ft from shore. The vast majority of subsequent measurements
by the TGI have been less than those found by the USGS with the
majority being below the detectable limit.
Soil Contamination
Both the UM study and the USGS study reported numerous
measurements of organics in the soils at the Air Station, again
with discrepancies due to analytical, temporal and areal varia-
tions. The UM study found maximum concentrations of 25.4 jtg/1
benzene, 27.6 /jg/1 toluene and 229 /itg/1 xylene. Analyses were
done for seven other hydrocarbons with negative results. Analysis
of soil borings indicates that much of the organic material is re-
tained in the soil in a 6 to 12 in. thick layer in the capillary zone
immediately above the water table. It has been suggested that
this zone is slowly leaking organics into the groundwater over
time and is thus serving as a contaminant source for the plume.3
Proposed Cleanup
After consideration of the various long-term treatment or
cleanup options available, it became clear that the most logical
first step in any remediation action program would be to decrease
or stop the further movement of contaminants off U.S. Coast
Guard property.
This option was judged to have several advantages:
• Reduce any possible increase in risk to human populations that
may have been related to fuel based contaminants present in
the groundwater.
• Promote reductions in contaminant concentrations in the
groundwater either by dilution or possible biodegradation of
fuels by indigenous microbial consortia present in the subsur-
face soil-water system.
• Provide a better opportunity (e.g., more time) to efficiently
select and design appropriate method(s) for dealing with the
contaminants present in the geographical origin of the plume.
CONTAMINATED GROUNDWATER CONTROL 53
-------
K9S.O
I
^
/r- »S20 »SI7
H&«ix
-------
system to withdraw the minimum amount of water yet still cap-
ture the contaminants passing through the aquifer so that no
contaminants, or at least a significantly reduced amount, were
leaving government property.
After reviewing several of the USOS groundwater model runs,
the team selected a well field configuration shown in Figure 8.
Seven wells were located laterally across the plume in the east-
northeast area of the Air Station. These wells were 6 in. auger-
drilled wells with full 10-slot stainless steel screens running from
the top of the aquifer to the clay confining layer at the bottom
(Figs. 5 and 8).
The system initially was constructed with all elements above
ground. In 1985, the interdiction system was modified by install-
ing pitless adapters and providing a heated building to protect the
manifold and control systems from the extreme cold experienced
at this site. The water produced from the interdiction wells is
piped to a carbon treatment system consisting of four 20,000-
Ib carbon reactors. The carbon reactors were specified to reduce
the levels of benzene and toluene in the water to less than 1 /ig/1
as measured using headspace technique on a HP5710A Gas
Chromatograph.
The USGS model gave clear indications that these flow rates
would produce closure of the equipotential lines at the interdic-
tion well line. One difficulty was that the two-dimensional nature
of the model did not yield any prediction on the vertical move-
ment of water at or near the individual well locations. However,
because of the uniform nature of the saturated zone and the rapid
mean field velocity of the aquifer, which was measured at 5 ft/
day, it was believed that vertical movement in a fully screened
well at 15 gal/min would be sufficient to capture contaminants
flowing in the lower areas of the aquifer.
Table 1
Specifications for Groundwater Interdiction Systems
North Interdiction Field
Date Wells Installed:
Wells ID 1-ID6 Feb. 26-Feb .28,1985
WellID-7 Mar. 22, 1985
Date of Water collection
system construction: Mar. 19-Apr. 18,1985
Date Pumping
commenced: Apr. 19, 1985
Well and Screen Data:
North Interdiction
(6 in. diameter)
Table 2
Monitoring Wells for North Interdiction
We) 1 *
Flow cpm
Top of
screen
(below
ground)
dottom
of
screen
uel 1
Dump
[ bottom
of pump
Below ;
ground ',
level) 1
.' ID-1
; is
13
53
43
10-2 i 10-3
; is : is
13 ' 13
53 : 43
43 ; 33
;
10-4
15
13
»3
33
10-5
15
13
43
33
ID-6
15
13
43
33
[0-7 \
25 ;
13 :
58
48 :
Five full screen wells installed—Mar. 20-27, 1985
Downgradient of North Interdiction Field on Mar. 20-Mar. 27, 1985
Four-inch diameter wells extending to clay liner beneath the aquifer
Uell «
Depth below
well screen to
bottom of well
screen { top of
water at 13'
below ground
level)
Vertical loca-
tion of sample
Dumps in we 1 I s .
feet below
ground
Sample pump HI
HZ
«3
*4
»5
rci-i
n'-53'
15'8"
25 '8"
35 '8"
«5'8"
54 '5"
TGI-2
13'-69'
14'10.5"
24'
40'
54'
66'9" ;
roi-3
12'-68'
14 '6"
24'
41'
57'
68'
rci-4
n'-6S'
5'5.5"
8'
I1
4 '
6'
rci-s
13'-54'
14'10"
i?' :
33' :
43' :
53'5.P;
TGI-2 sample well located 50' downgradient from ID-3 on a line parallel to the flow lines of the
plume drawn NE from ID-3. TGI-3 located 50' downgradient from TGI-2 on a line parallel to
the flow lines through ID-3 and TGI-2. TGI-4 located 90' northwest of TGI-3 on a line through
TGI-3 perpendicular to the flow lines. TGI-1 located 180' northwest of TGI-3 on a line through
TGI-3 and TGI-4. TGI-5 located at the end of a line drawn perpendicular to a line through
TGI-3 and TG1-4 starting at a point 90' SE of TGI-3 running 50' NE.
In most situations, contaminants such as benzene and toluene
are found more predominantly in the upper areas of the aquifer.
However, in this situation the original spill had been in the
ground so long that visible product was no longer present and
the product not contained in the interstitial area of the capillary
zone had solubilized and had mixed downgradient so that it was
present throughout the vertical cross-section of the aquifer. Thus,
placement of the pump in the well had to be in conjunction with
a full screen configuration to capture contaminants from all ver-
tical locations.
A second consideration in the design was the monitoring of the
interdiction system to obtain data on the effectiveness of the
blocking of contaminant flow. Since the area of the aquifer
downgradient from the interdiction system had been contam-
inated for several years, the monitoring wells must yield data in
sufficient detail to prove any decrease in contaminant levels after
the "valve" had been closed.
A sampling network of five wells was located downgradient
from the interdiction well system (Fig. 9). The wells were located
outside the zone of influence of the interdiction wells. Sampling
was done using a dedicated "Well Wizard" sample pump system.
Groundwater samples were collected following standard field pro-
cedures for the collection of volatiles and analysis was done with-
in a few hours of collection, always on the same day that the
sample was collected. Five sample pumps were located in each
monitoring well at the top and bottom of the saturated zone
and at three equidistant points (Table 1).
RESULTS
The results downgradient from the interdiction system have
been satisfying to date. Toluene and benzene levels in the down-
gradient monitoring wells were monitored in one of the five wells
(M2/TGI2) on a daily basis for the first 8 weeks and on a bi-
weekly basis thereafter. The other four wells are monitored once
a week.
CONTAMINATED GROUNDWATER CONTROL 55
-------
Since benzene is the more mobile of the two constituents of in-
terest and has the higher solubility, it should be slower than
toluene to respond to the interdiction. In addition, there was
speculation that some biodegradation by selective demethyoxyla-
tion might be seen and be reflected in the benzene and toluene
concentrations found in the downgradient wells. A considerable
amount of data have been collected which allow trend observa-
tions to be made regarding the fate of benzene and toluene in the
downgradient area (Figs. 8 to 10).
Figure 8 shows the variations and trend of benzene concentra-
tions as a water column average by month since the interdiction
field was activated on Apr. 15, 1985. Figure 9 provides the same
information for toluene. These data are from Well M-2 (TGI-2)
located immediately downgradient and outside the zone of influ-
ence of the interdiction well field. This well logically would be
the first monitoring well to provide data on the effectiveness of
interdiction. Although microbiological activity would not be sep-
arated from hydraulic blocking, toluene levels decreased from a
baseline level of 10329 jig/I to less than 10 /ig/1 in approximately
100 days.
Figure 10 shows benzene and toluene levels plotted against
time from April to December 1985. Since this figure shows the
variation of the two compounds, it could substantiate the micro-
biological activity. The decrease in toluene levels, together with
the concomittant rise in benzene between May and July, could
be attributed to the demethoxylation of toluene to benzene. Both
compounds then began a steep downward trend with the toluene
being the first to approach non-detectable levels in August. As
can be seen, the benzene reached its lowest level during October.
The increase in benzene levels since October is attributed to the
appearance of a red slime that plugged pump screens and coated
the inside of the piping systems, resulting in a reduction of the
hydraulic capacity of the system. The groundwater in this aquifer
is high in iron content, and the substance has been tentatively
identified as oxidized iron compounds and complexes, mineral
deposits and biomass. Microbial action is believed to play a signif-
icant role in the dissolution of these solids in the aquifer and
subsequent precipitation within the system. The slime is complete-
ly soluble in an acidic solution, and gamma logging of the wells
does not indicate the presence of any clay layers that could be
contributory. Obviously this situation must be better understood
and controlled or the field will lose its effectiveness as may be in-
dicated by the gradually increasing benzene levels.
The USCG has an active project with the U.S. EPA Robert S.
Kerr Environmental Research Laboratory (RSKERL) in Ada,
Oklahoma to evaluate the soil and water chemistry as well as the
microbial consortia present at the site. The results of these studies
should provide a mechanism to control the slime and elucidate
the biodegradation phenomenon.
While monitoring of the interdiction system continues, there al-
ready is substantial evidence that the desired effect was achieved
in terms of contaminant containment. Toluene has decreased to
virtually non-detectable levels and, while it continues to be vari-
able, benzene concentrations clearly demonstrate a downward
trend. The benzene variability may be due in part to some, all or
a portion of the slime problem discussed earlier. Biodegrada-
tion variability, the effect of recharge and the higher mobility of
benzene in the groundwater system also may contribute to the
variability of the benzene.
CONCLUSIONS
Based on the data collected over the past 9 months, it appears
that, for well defined plumes, low flow interdiction is a viable
alternative to intercept and block contaminant flow.
HELL M-2
Hit a* ui JJ< 4J.I U OT' Kl •>• OK
(.:«
Figure 8
Benzene Concentration as a Function of Time
iiv *• wt JLM juf »4 ari XT «* ore
Figure 9
Toluene Concentration as a Function of Time
ST **» ui JU< all 14* apt XT 4* OCC
Figure 10
Benzene and Toluene Concentrations as a Function of Time
Careful attention must be paid to the mechanical as well as
contaminant systems. The appearance of the biomass is a good
example of this since no homeowners in the area using wells have
reported a problem with slime.
The possibility of triggering increased microbial activity is a
potential factor to consider.
Particular attention must be given to the vertical distribution
of contaminants since withdrawal volumes may be significantly
different for new plumes as compared to old plumes.
Minimum volume requirements as dictated by available process
water disposal options are a major determinant in interdiction
system design.
56
CONTAMINATED GROUNDWATER CONTROL
-------
DISCLAIMER
The opinions or assertions contained herein are the private
ones of the writers and are not to be construed as official or re-
flecting the views of the Commandant or the Coast Guard at
large.
REFERENCES
1. Sibo, K., "Groundwater Investigation of East Bay Township," Un-
published, Mar. 3, 1982.
Twenter, F.R., Cummings, T.R. and Grannemann, N.G., "Ground-
water Contamination in East Bay Township, Michigan," U.S. Geo-
logical Survey Water—Resources Investigations Report 85-4064, 1985.
Rossman, R., Rice, C.P., Hartung, R., Simmons, M.S., Armstrong,
J.M. and Wright, S.J., "Scientific Study and Feasibility Study
Groundwater Contamination at Traverse City, Michigan," Unpub-
lished, The University of Michigan, Ann Arbor, MI, Mar. 1985.
"TGI Final Report, Phase I," Unpublished, Apr. 1985.
CONTAMINATED GROUNDWATER CONTROL 57
-------
Computer Groundwater Restoration Simulation
at a Contaminated Well Field
Shih-Huang Chieh, Ph.D.
Ecology and Environment, Inc.
Buffalo, New York
Jeffrey E. Brandow, P.E.
State Department of Environmental Conservation
Albany, New York
ABSTRACT
The most obvious and direct means of contaminated aquifer
rehabilitation is to remove the contaminated water by pumping
from recovery wells. Factors affecting the technical and economic
feasibility of this procedure include the number of recovery wells,
the rate and duration of pumping and the method used to dispose
of the contaminated water. Since the aquifer rehabilitation is
assumed to be the reverse of the contaminant spreading process,
the theory and models that have been developed for predicting the
spread of a contaminant should provide a basis for predicting
contaminant withdrawal.
The drinking water of western Vestal, New York, is supplied by
Water District 1, which consists of wells 1-1, 1-2 and 1-3. Well
1-1, located on the bank of the Susquehanna River about 400 ft
from the Endicott-Vestal Bridge, is contaminated with volatile
organic compounds. Currently, a remedial investigation/feasibil-
ity study is being conducted at the site. The purpose of the
remedial investigation/feasibility study is to locate the contami-
nant source, to define the extent of the contaminant plume and to
evaluate the remedial alternatives. As a practical tool to assist in
the feasibility study, a two-dimensional finite element model is be-
ing applied to this site. First, the model is calibrated against field
data. Then the model is utilized to simulate the movement of the
contaminant plume under the operation of a recovery well. The
model results will be used to evaluate and refine the remedial
alternatives.
INTRODUCTION
A groundwater contamination problem due to toxic organic
chemicals exists in Water District 1 of Vestal, New York. In early
1980, organic chemicals were detected in a municipal well in
Vestal. As a result of this finding, all wells in Vestal were tested
for synthetic organic compounds. Well 1-1 of Water District 1
was found to contain large quantities of organic compounds. This
triggered an assessment study of the potential of the pollutant to
contaminate other water supply wells in the district. Currently, a
remedial investigation/feasibility study is being conducted at this
contaminated well field site. As a result of the remedial investiga-
tion, the contaminant source areas were identified and the extent
of the contaminant plume was defined.1'2
A feasible alternative to cleanup of the contaminated well field
is to remove the contaminated groundwater by pumping from
recovery wells. Since the aquifer rehabilitation is assumed to be
the reverse of the contaminant spreading process, the theory and
models that have been developed for predicting the spread of a
contaminant provide a basis for predicting contaminant
withdrawal.
A two-dimensional finite element model is applied to this con-
taminated well field site to simulate the transport of the pollutant
under a recovery well is being operated. The model utilized a
quadrilateral element and the Crank-Nicholson method for
calculation of the time marching scheme. The model result can be
used to evaluate the effect of the recovery well and to refine the
remedial alternatives.
DESCRIPTION OF THE SITE
The drinking water for most of western Vestal is supplied by
Water District 1, which consists of well I-I and water supply welk
1-2 and 1-3. Well 1-1, located on the south bank of the Susque-
hanna River about 400 ft from the Endicott-Vestal Bridge, is con-
taminated with volatile organic compounds. Contamination was
detected first in early 1980; since spring 1980. the well has been
pumped directly into the Susquehanna River. At present, well 1-2
is supplying the Vestal drinking water; well 1-3 is serving as a
backup supply. Because the water in well 1-3 is highly corrosive
and well 1-2 has a limited capacity, it is important that well 1-1 be
restored, especially to meet anticipated future peak demands for
Water District 1 and other interconnected districts. Figure 1
shows the site of the contaminated well field and the groundwater
table contours under wells 1-1 and 1-2 while in operation. The
groundwater table contours were developed from field measure-
ments.'
The area immediately surrounding well 1-1 is within the Sus-
quehanna River floodplam and consists of marshland, wooded
areas and commercial and residential land use areas. The river oc-
cupies a bedrock valley filled with glacial deposits ranging in size
from clay to gravel. The bedrock surface, situated between 130
and 160 ft below ground level, is overlain by a productive aquifer
consisting of about 40 ft of permeable sand and gravel. The
aquifer is overlain by 120 ft of poorly permeable clays and silts.
Within these deposits are locally permeable pockets of material
which apparently have acted as contaminant pathways between
the ground surface and the producing aquifer.
The high rate of pumping from this aquifer (approximately 1.5
million gallons daily, including the pumping of well 1-1 to waste)
has induced organic solvent movement toward well 1-1 from the
source area. The source area is an industrial park containing a
variety of industries. Volatile organic levels exceed 10 mg/1 in one
monitoring well located within the area, but no information is
available regarding spills or disposal or organic solvents.
DESCRIPTION OF THE MODEL
The governing equation of pollutant transport in porous media
is the following:
58 CONTAMINATE!} GROUNDWATER CONTROL
-------
Figure 1
Map of Vestal Well Field and Groundwater Table Contours with Wells 1-1 and 1-2 in Operation
3c f 3 / 3c 3c \ 3 / oc 3c\l
— = n\ — \D,,, — + £>xv— +— £v,- — + DVV —
3t lax \ 3x y3y/ 3y\ yx dx yy 3y/J
r 3 3
- — (Vxc) + —
Idx 3y
(1)
in which:
c
R = 1 + esA://z
Ss
k
n
Dxv, D
•yx> yy
V V
'x> 'y
\
= concentration of the pollutant
= retardation factor
= bulk density of the porous medium
= distribution coefficient of the
pollutant
= porosity of the porous medium
= component of dispersion tensor
= Darcian velocity component
= 1st order decay constant
= concentration of the source fluid
= flow rate of the source fluid
Let L denote the operator on c in equation (1), thus:
Lc = 0.
For the finite element method:
(2)
{N}T{c} (3)
where £/V} is the shape function. The residue becomes Lc. The
principle of Galerkin method requires that:
Lc,Wt
= 0 .
(4)
Equation (4) states that the inner product of L? and the
weighting function Wj over the solution domain B vanishes. The
integrations of Equation (4) are now carried out, and the results
are expressed in matrix and vector notation as follows:4
{c}T[Af{c} +(K-A+E) {c} -{p}] =0
in which:
dc
dW
— — •} +nDxy — •
dx 3 x
K -// [n£)x
nD
ay
3
+ nDyy { \l
dxdy
E =// n\R{W}{N}'* dxdy
B
M -ff nR{W}{N}'T dxdy
{p} = Qc*{A'}dxd;y+
3c 3c
xs — + "-D^y VjcC Inx
ax ay '
(3c 3c \ "1
nDyx — + nDyy -- VyC rty { W } dS .
dx 3y / J
(5)
(6)
T
(7)
(8)
(9)
(10)
(11)
The advantage of the finite element method over the finite dif-
ference method is its ability to handle complex boundaries and
normal derivatives. In the time dimension, these advantages are
not present, so the finite difference method will be used for the
time derivative term of the solute transport Equation.5
With finite difference in time, Equation (5) becomes
CONTAMINATED GROUNDWATER CONTROL 59
-------
Table 1
1,1-Dichloroelhane Concentration Measurements at Vestal Well Site
in which ^ = 1 for backward-differences and /i = 0.5 for the
Crank-Nicholson differencing scheme. Along the boundary of the
study area, Dirchlet boundary conditions were specified.
SIMULATION OF CONTAMINANT PLUME
For the preparation of input files for the computer simulation
study, it was necessary to set up a grid system for the discretized
site and determine several model parameters. According to the
site history, production wells 1-1 and 1-2 are constantly in opera-
tion. Therefore, a non-uniform but steady-flow field was as-
sumed from the measured ground water table under wells 1-1 and
1-2 while in operation' as presented in Figure I. As shown in
Figure 2, the Vestal well site is discretized into 103 quadrilateral
elements with 125 nodals.
Among the dissolved organics in groundwater, 1,1-dichloro-
ethane was chosen to model the groundwater solute transport in
the study area. The measurements from the samples at the
monitoring wells are summarized in Table I. The locations of the
monitoring wells are shown in Figure 1. The 1- series wells were
installed during the assessment study. The s- series wells were in-
stalled during the remedial investigation by E & E.
Since no field measurements are available for the dispersion
coefficient, this value was selected based on the literature. The
molecular diffusion is assumed to be small compared to the
hydrodynamic dispersion. Values of 205 ft for longitudinal
dispersivity and 12.9 ft for transverse dispersivity were used.
1,1-dichloroethane could be adsorbed on the solid phase of the
aquifer. However, the retardation effect is believed to be small
and will not seriously affect the qualitative prediction of solute
transport. Therefore, a retardation factor of 1 was used for this
study. The values of model parameters are listed in Table 2.
Well
1.1-dlchloroethane
NO.
S-l
S-2
S-6
S-7
S-ll
1-33
1-34
Node
No.
65
77
103-112
94
68
84
78
Date
4/26/85
4/26/85
4/26/85
4/26/85
4/26/85
4/26/85
4/26/85
Cone.
0
945
58
1.280
158
860
l.JOO
Table 2
Model Parameters Used in Computer Program
Number of elements
Number of nodes
Number of we)It
103 Longitudinal dliperslvlty
125 Transversal dlsperstvlty
7 Retardation factor
Hydraulic conductivity 47 ft/day First order decay constant
205 ft
12.9 ft
1
0
In order to verify the dispersion coefficient chosen for this
study, a simulation of the dispersion of 1,1-dichloroethane was
conducted. Two constant-strength sources located at nodals 84
and 94 were assumed in this computer run. The contaminant
sources were found during the remedial investigation. The simula-
tion time was 6 years. The assumptions were based on our best
knowledge of the history of the site. The computed contaminant
plume and comparison with field measurement are presented in
Figure 3. The results show that the simulated concentration is
very close to the field measurements. Other values of longitudinal
and lateral dispersivity also were selected in simulating the disper-
sion of 1,1-dichloroethane. No satisfactory' results were obtained
from these tests. Therefore, the longitudinal and lateral disper-
sivities selected for the current study represent a rational choice
for this site.
Figure 2
Finite-Element Grid System for the Vestal Well Field
60 CONTAMINATED GROUNDWATER CONTROL
-------
1597 Slmul.i.d Cone.
11300) M.I,
ltd Concin
imHon
».„.„
Figure 3
Computed Contaminant Plume and Comparison
with Field Measurement
Since spring 1980, well 1-1 has been pumped and discharged
directly into the Susquehanna River. A possible remedial action
would be the removal of the contaminant source and would con-
tinue use of well 1-1 as a recovery well to withdraw the pollutant
from the aquifer. Computer simulation runs were conducted to
simulate the movement of the contaminant plume after the
removal of the contaminant source and while well 1-1 was in
operation. Figure 4 presents the simulated concentration at the
end of 5 years. A comparison of Figure 4 with Figure 3 shows that
the 1000-/tg/l iso-concentration contour is moving toward well
1-1, and the iso-concentration area is shrinking. Figures 5, 6 and 7
present the simulated concentration at the end of 10, 15 and 20
years, respectively. It can be seen clearly in these figures that the
contaminant plume is moving toward recovery well 1-1 and that
the high concentration region is shrinking.
Figure 4
Computed Contaminant Plume, 5 Years After
Removal of Contaminant Source
CONTAMINATED GROUNDWATER CONTROL 61
-------
Figure 5
Computed Contaminant Plume, 10 Years After
Removal of Contaminant Source
CONCLUSIONS
The model predicted the movement of the contaminant plume
under the operation of well 1-1 and the removal of the contami-
nant sources. The prediction was made for the time periods of 5,
15 and 20 years. The study provided a basis to evaluate the effec-
tiveness and the cost of using well 1-1 as a recovery well. This
summary paper demonstrates the usefulness of the computer
model in assisting the feasibility study at uncontrolled hazardous
waste sites. Further applications of computer groundwater
restoration simulation can be conducted by varying the pumping
rates, locations and numbers of recovery' wells.
ACKNOWLEDGEMENT
The authors would like to express their appreciation to U.S.
EPA and the New York Slate Department of Environmental
Conservation for their support in this study.
figure 6
Computed Contaminant Plume, 15 Years After
Removal ol Contaminant Source
62 CONTAMINATED GROUNDWATKR CONTROL
-------
Figure 7
Computed Contaminant Plume, 20 Years After
Removal of Contaminant Source
REFERENCES
1. Ecology and Environment, Inc., "Town of Vestal Water District
No. 1 Focused Feasibility Study," May 1985.
2. Chieh, S.H., Cook, D., Hwang, J.C. and Brandow, J., "Computer
Model Study of a Contaminated Well Field—Groundwater Pollution
Source Identification and Contaminant Plume Simulation," Proc. of
the 1985 ASCE Hydraulics Division Specialty Conference, Orlando,
FL, 1985, 330-335.
3. Martin, R.J., Coates, D.R. and Timofeefe, N.P., "Well Field Con-
tamination Investigation for Town of Vestal Water District No. 1,"
Report for New York State Department of Environmental Conserva-
tion, 1983.
4. Yeh, G.T. and Ward, D.A., "FEMWASTE: A Finite-element
Model of Waste Transport Through Saturated-unsaturated Porous
Media," Oak Ridge National Laboratory Report No. 5601, 1981.
5. Hwang, J.C. and Koerner, R.M., "Groundwater Pollution Source
Identification from Limited Monitoring Well Data—Part 1, Theory
and Feasibility," J. of Haz. Mat,, 8, 1983, 105-119.
6. Bear, J., Hydraulics of Groundwater, McGraw-Hill, Inc., New York,
NY, 1979.
7. Gray, W.G. and Hoffman, J.L., "A Numerical Model Study of
Ground-Water Contamination from Price's Landfill, New Jersey-II.
Sensitivity Analysis and Contaminant Plume Simulation," Ground
Water, 21, 1983, 15-21.
8. Finder, G.F., "A Galerkin-Finite Element Simulation of Ground-
water Contamination on Long Island, New York," Water Resources
Res., 9, 1973, 1657-1669.
CONTAMINATED GROUNDWATER CONTROL 63
-------
Enhancement of Site Assessments by
Groundwater Modelling
Joseph R. Kolmer, P.E.
John B. Robertson, P.G.
Roy F. Weston, Inc.
West Chester, Pennsylvania
ABSTRACT
It is a commonly held misconception that extensive ground-
water monitoring and other geologic and groundwater related
data must be available before a groundwater model can be
selected, calibrated and implemented. The primary use of models
in this mode of operation has been to predict long-term ground-
water flow and groundwater quality due to implementation of
engineered groundwater supply systems and/or remedial cleanup
programs.
Advances in computer technology have dramatically revised the
utilization of groundwater models and expanded their utility with
respect to groundwater investigation programs. Groundwater
flow models are now readily adaptable to personal computers. To
use these models in a site investigation, only a very limited
amount of information is needed initially, and assumed data can
be used to supplement this initial information.
Given the limited data, the model is responding primarily to the
general equations of groundwater flow and not to unique site con-
ditions. This uncalibrated flow model, however, provides the
technical skeleton to which additional groundwater monitoring
information can be input and assessed. As additional site data are
added to the model, it is refined and calibrated. During the data
gathering process, the model is used not only to identify data
gaps, but also to determine when sufficient data have been
gathered to quantify site conditions.
The groundwater model is not limited to utilization as a predic-
tive model based upon exhaustive site information. Rather, the
groundwater model is now most properly and cost effectively
used in the groundwater monitoring program to assess data gaps,
quantify site conditions and evaluate engineered remedial
programs.
INTRODUCTION
The purpose of this paper is to discuss and present a method of
site investigation using groundwater computer models to guide
the investigation as well as to indicate when sufficient data have
been collected. The methods used to implement this approach are
neither unique nor technically complex, but they do require
utilization of an integrated geoscience and engineering staff. This
technical integration, plus the general availability of groundwater
models on microcomputers, makes this approach to site investiga-
tion technically feasible and advantageous.
The need for quantitative groundwater resource assessment
tools is growing at an accelerating rate. Water resources investiga-
tions for domestic, agricultural and industrial uses are increasing,
and the investigation work associated with definition of con-
taminated groundwater due to hazardous materials is expanding
daily. These latter investigations are particularly difficult because
they require quantitative analysis methods to define the contami-
nant migration conditions as well as the remedial solutions. The
use of groundwater flow models to quantify the site investigation
work and standardize the basis upon which remedial systems are
comparatively evaluated makes use of the inherent attributes of
these methods. A brief discussion of the current practices in
groundwater modelling is presented below followed by a detailed
discussion of how models can be used in site assessments.
EXISTING METHODS OF
SITE INVESTIGATION
The predominant practice for projects involving groundwater
quality and water resource investigations is illustrated in Figure 1.
The objective or the field investigation work is to obtain sufficient
information to describe site conditions, specifically the geologic
conditions. After these data are gathered, they are assembled and
assessed to determine their adequacy.
Report
Sit.
Groundwater
Figure 1
Silc Investigation Followed by Groundwater Modeling
64 CONTAMINATED GROUNDWATER CONTROL
-------
This data evaluation exercise might reveal the need for addi-
tional information to describe an anomalous or vague site condi-
tion, and a second phase of field investigation would commence.
This interactive process would continue until the objective of
understanding the geology, ground water hydrology and water
quality conditions is accomplished. A report describing the in-
vestigative procedures and the defined site conditions is finally
prepared.
After the field investigation is complete, the engineer or
engineering group receiving the final report develops a set of
remedial action alternatives designed to mitigate any adverse
chemical quality conditions identified during the field investiga-
tion. To conduct a comparative analysis of the identified remedial
alternatives, a groundwater flow model is usually employed. The
flow model is calibrated based upon the findings of the field
report and run (exercised) to satisfy the identified needs of the
engineer. If, during the conduct of the engineering analyses, the
need for additional data is identified, it usually is assumed as part
of the model input. The mathematical operations inherent to the
flow model check these assumed inputs against measured site con-
ditions. Thus, assumptions made at this stage of a site investiga-
tion are not totally unchecked, but they usually do not have the
advantage of being field verified.
After the site model has been developed and calibrated, it is ex-
ercised to simulate the various remedial solutions being con-
sidered for the site. For example, a groundwater barrier wall, such
as a soil bentonite slurry cut-off wall, could be simulated by the
insertion of no-flow boundaries around the perimeter of the site.
Guswa1 noted that implementation of a remedial action plan can
be very expensive, and it is desirable to have the preimplementa-
tion assurances that the proposed plan can be effective. While no
type of modelling provides guarantees that any particular plan
will be effective, they do provide a comparative basis against
which all remedial alternatives can be evaluated.
THE MODEL AS INTEGRATOR
The use of the groundwater model is a good tool to evaluate
various remedial solutions. This use of the model is obviously ad-
vantageous, but additional information can be obtained. That is,
the groundwater model can be used not only to evaluate the facts
gathered after conclusion of the field investigation, but also can
be used to better define what information needs to be gathered
and where it can be found.
Figure 2 illustrates how the groundwater flow model can be
used during the site investigation phase of work and in the subse-
quent analysis of remedial alternatives. Comparison of this figure
to Figure 1 shows that the steps leading to the final report have
been reduced. This reduction primarily centers around integration
of the groundwater modelling work into the analysis of data
obtained during the field investigation exercise.
To initiate use of groundwater modelling during the field in-
vestigation phase of any project, a basic amount of site informa-
tion must be known. General site geology, basic groundwater
flow conditions and surface water conditions should be known to
establish the initial flow parameters and boundary conditions for
the model. Published literature is replete with geologic definition
and groundwater ami surface water factors for a large portion of
the United States. If an assessment of existing information on a
site area is conducted in a fairly exhaustive manner, the general
information needed initially to establish the groundwater model
usually can be accumulated.
One of the primary questions raised during conduct of a field
investigation is how does one know when sufficient data are
available to satisfy the objectives of the project? The groundwater
flow model can provide significant assistance in answering this
question. As indicated in Figure 3, in order to assemble or
Figure 2
Integrated Site Investigation
establish the groundwater model, existing data as well as input
from numerous technical disciplines must be obtained. Questions
as to the area to be covered by the model, the geologic conditions
to be considered and the overall engineering requirements
associated with the model must all be discussed and incorporated.
Thus, at this early stage of the project, the flow model has in ef-
fect served as a catalyst for integration of the technical disciplines
that eventually will be needed to resolve the identified problem.
This integration, properly carried to fruition, will result in the
establishment of a set of objective criteria for the site investiga-
tion and the overall remedial action study. Specific data needed to
satisfy these criteria would be identified for acquisition during the
field investigation.
As field investigation work is conducted, new site information
is developed and can be used with the established groundwater
model. As these data are added, the aquifer characteristics are ad-
justed so that the model representations are consistent with iden-
tified field conditions. As more information is gathered in the site
area, the adjustments that need to be made to the model will
Model
Calibration •
Figure 3
Groundwater Model Development
CONTAMINATED GROUNDWATER CONTROL 65
-------
become relatively minimal. As this point is approached, acquisi-
tion of additional field information is not needed to refine the site
model, and field investigation work is complete. Site conditions
are now defined consistent with the set of objective criteria iden-
tified for the project. Assessment of remedial alternatives can
commence.
The model assembly, identification of additional data needs
and model revision work tasks identified in Figure 3 are in effect
the data analysis and additional data input loop identified in
Figure 2. The exercise of identification of additional data needs
and acquisition and input of these data to the flow model con-
stitute calibration of the flow model. An important distinction be-
tween this approach to use of groundwater modelling and the
approach discussed in the previous subsection is that model
calibration is integral to the field investigation exercise.
Finally, no (or at least very little) assumed data are needed to
calibrate the groundwater model. Thus, the flow model that
ultimately will be used to compare various remedial alternatives
already will have undergone the rigors of development and
calibration placed upon it not only by the simulation of known
site conditions, but also by the technical requirements of the disci-
plines involved in overall problem resolution.
THE COMPUTER MODEL
AND SITE INVESTIGATIONS
The principal concerns surrounding the use of computer model-
ling have been its availability to the project's principal
investigator and the cost associated with development and opera-
tion of these models. Over the past few years, both of these con-
cerns have been significantly diminished by the development of
the desk top personal computer. Thus, the principal investigator
or project manager has a significant library of groundwater
models readily available to him that can be developed and exer-
cised at a relatively inexpensive cost. In fact, the graphics
capability as well as the data base management capabilities of
these machines can reduce the overall project costs compared to
hand manipulation and graphics development for a medium to
large scale site investigation.
In the past, groundwater flow models have been misused by
misrepresenting their capabilities and application. One of the
common problems associated with groundwater modelling is the
tendency to use the most sophisticated model available for prob-
lem resolution. This temptation is difficult for the technical com-
munity to resist, but can be managed if integration of technical
skills is accomplished in the early stages of the project. This in-
tegration of the groundwater modelling expertise with the
pragmatism of the engineer usually results in model development
consistent with the identified objectives for the site study. This
management of technical expertise must be coordinated by the
project's principal investigator.
Another technique that can be used to help prevent over exten-
sion and misrepresentation of groundwater model usage is the
phased implementation of a particular model's capability. For ex-
ample, throughout this paper, only groundwater flow models
have been discussed and solute transport models have not been
strongly considered. This analysis does not say that solute
transport models are not appropriate for the problems discussed
herein; it does, however, say that in the early stages of model
development and utilization in site investigations, the first step
should be a characterization of groundwater flow hydrology.
Once groundwater flow characteristics (especially velocity
parameters) have been identified, solute transport models can be
included. One must remember that the model only need to detailed
enough to satisfy the stated objectives of the project.
The following brief example shows how various data elements
can be obtained and input for site investigation modelling. The
advantage of a phased field investigation program is highlighted
in the example.
Existing Data Review and Model Development
During the review of existing information for a site area, data
concerning the history of operation of the site, the geologic and
geohydrologic conditions of the area, precipitation, surface water
hydrology and so forth, are compiled. The completeness of this
information depends upon the location of the site, the notoriety it
has received, site management and numerous other factors. For
the purposes of this paper, only that available data pertinent to
development of the groundwater model will be discussed. The
lack of consideration of any other data does not mean it is not
needed for a site investigation nor does it mean it may not be per-
tinent to model development at some sites.
Figure 4 presents a map of the example site area and contains a
significant amount of information. The example site is located in
a river floodplain below an upland area. Thus, the aquifer of con-
cern is more than likely alluvial in nature, probably containing a
wide distribution of grain-size materials. In addition, channeliza-
tion of sediments due to alluvial action is very possible. Finally,
the upland area probably forms a recharge boundary for the
alluvial aquifer, and the river probably forms a discharge
boundary.
Figure 4
Example Site Area
There are four existing monitoring wells identified at the cor-
ners of the site. These wells could provide information on ground-
water levels as well as the type of sediments contained in the
saturated zone. Figure 5 presents preliminary geotechnical find-
ings showing a gravel seam through the center of the site flanked
by finer grained sediments. If no aquifer characteristic data is
available, the existing monitoring wells could be used to conduct
abbreviated aquifer tests, such as slug tests. These tests are
66 CONTAMINATED GROUNDWATER CONTROL
-------
described by Cooper, Bredehoeft and Papadopulos2 and Bouwer
and Rice.3 Thus, existing site information as well as readily
obtainable information can provide the needed data.
Upland Area
Figure 5
Preliminary Geotechnical Findings
In some site areas, no existing monitoring wells are available. In
such cases, geologic and hydrogeologic reports concerning the
aquifers of the area as well as the estimated permeabilities and
transmissabilities would have to be utilized. In these cases, it
might be advantageous to postpone initial model development un-
til specific site information could be obtained from installation of
groundwater monitoring wells. Generally speaking, sites being in-
vestigated today have a limited number of monitoring wells or soil
borings available for use in the initial stages of the projects.
FIELD INVESTIGATION
Based upon the findings of the data review, it is often advan-
tageous to conduct the field investigation work in a phased man-
ner. That is, early acquisition of general site area information
(such as aquifer characteristics determined by slug tests and
geophysical data) would be advantageous to more detailed plan-
ning of the later stages of field investigation.
Figure 6 shows a probable set of conditions related to
geophysical findings for the example site area. The geophysical
methods employed might include determination of terrain con-
ductivity by electromagnetic techniques. The results of this type
of investigation usually are mapped so that anomolous conduc-
tivity patterns (inferred to be indicative of contaminant transport
conditions) can be identified. In our example problem, it appears
that the gravel seam through the site area is a preferred avenue of
groundwater movement and contaminant transport. In a pre-
ferred groundwater movement area, it is likely that water levels in
the gravel will be slightly lower than in the immediately adjacent
less permeable zones.
Given the information obtained from the data review as well as
the preliminary work conducted in the initial stages of the site in-
vestigation, a plan of overall investigation effort can be for-
Figure 6
Geophysical Survey Findings
mulated. Figure 7 displays how additional data with respect to
groundwater flow conditions and contaminant levels might be ob-
tained by monitoring well installation. At the same time, the
available data can be used to establish the initial conditions of the
groundwater flow model. The groundwater elevations obtained
from the existing monitoring wells coupled with the assumed
groundwater levels associated with the gravel seam can be used to
predict the findings of the investigative work. The actual data
from each well would be put into the model as each well was in-
stalled and the model would be rerun to refine predictions for the
remaining wells scheduled for installation. Based upon these find-
ings, monitoring well locations would be modified to maximize
data acquisition efforts. The use of the model to guide and refine
the field investigation exercise can result in the elimination of
monitoring wells and the early cessation of the field investigation
exercise.
Simulation of Remedial Work
After the field investigation exercise has been completed and
groundwater flow calibration has been accomplished, the impacts
on groundwater movement due to various remedial scenarios can
be simulated. In the case of the example problem, groundwater
flow barrier walls intersecting the gravel zone, coupled with
groundwater pumping wells, could be simulated. Groundwater
pumping without barrier walls also could be modelled and the
results could be compared to estimate the savings in treatment
costs that could be realized due to the difference in groundwater
pumpage. These simulations, as well as numerous other simula-
tions, can be made on a consistent comparative basis through the
CONTAMINATED GROUNDWATER CONTROL 67
-------
developed groundwater flow model. This consistency of com-
parison is essential to the quantitative evaluation of alternative
remedial measures.
jLLLOU-LLLLi
I ILJLl_U-i_LiJ.
Figure 7
Additional Data Needs Definition
THE OVERALL PICTURE
Not all projects may be appropriate for application of ground-
water flow models as described in this presentation. Certain site
problems may be so small that a conceptual model may be suffi-
cient and a computer model may not be warranted. The ground-
water model is simply a tool that can be used to help organize and
streamline the site investigation work and the process of remedial
system selection.
Whenever a model is being developed during a site investigation
and is intended to be used to compare and predict remedial action
impacts, care must be taken to properly apply historical factors
specific to the site area. For example, long-term historical
precipitation patterns should be considered when designing
groundwater pumpage, treatment and recharge networks.
Anomolously high rainfall conditions could cause an increase in
the volume of water requiring treatment or in other ways impair
operation of the remedial system. Anomolously low rainfall could
result in a general lowering of the groundwater table and cause
flow conditions to be significantly different than those considered
by the predictive model. In either of these cases, if the predictive
model did not consider the long-term fluctuations in precipitation
and the resultant impact on infiltration quantities, the predictive
analyses compiled by computer modelling might not be valid.
Mercer and Faust* recommend that groundwater flow models be
calibrated with known historical site information and that the
period of prediction not exceed twice the period of historical data
used in the calibration.
When evaluating the overall factors of a site investigation and
development of remedial alternatives, a general review of the
work should be made to insure that it is consistent with sound
technical "intuition." One of the worst mistakes that can be
made in groundwater modelling is unchecked acceptance of the
model results. Data cross-checking and integration are advisable
with or without a groundwater computer model, but must
definitely be utilized when employing models to protect against
blind acceptance of computer output data.
A good conceptual understanding of the site conditions and
general influences in the site area is a prerequisite to initiation of
the site investigation work and utilization of groundwater model-
ling. Without such understanding and cross-checking of technical
information, little or no integration of technical disciplines will be
accomplished, the overall program effectiveness will be
significantly reduced and the investigation may, in fact, become a
waste of time and money.
CONCLUSIONS
Until now, groundwater models have not been extensively used
as tools in the conduct of site investigations. Generally, models
have been applied after completion of field investigations to
predict the impact of various remedial alternatives. The prolifera-
tion of sophisticated computer equipment and the adaptation of
groundwater model codes to this equipment have literally brought
the groundwater flow modelling capability to the desk top of each
project engineer and scientist. Thus, groundwater modelling is
now a readily available tool that can be used to identify data col-
lection requirements, evaluate these data, conduct data sensitivity
analyses with respect to groundwater flow impacts and assess the
long-term impacts of various remedial actions on the groundwater
flow conditions.
When employing groundwater flow models, a good conceptual
understanding of the aquifer and other pertinent subsurface con-
ditions must be available as the underlying basis for model
development. The extent to which these conceptual understand-
ings backed by actual observed field conditions can be simulated
by the model is the quantitative measure of the ability of the
model not only to adequately represent existing field conditions,
but also to predict future conditions. The ultimate responsibility
to ensure both proper integration of technical talent in model
development and calibration, plus proper technical usage of the
model in simulating remedial activities, rests with the principal in-
vestigator or overall project manager. Proper integration of
technical talent in model development and usage is key to suc-
cessful site investigation and remedial action alternative analysis.
REFERENCES
I. Guswa, J.H., "Numerical Models, Practical Tools in Groundwater
Investigations," The Weston Way. Fall 1984, 10-13.
2. Cooper, H.H., Bredehoeft, J.D. and Papadopulos, I.S.. "Response
of A Finite-Diameter Well To An Instantaneous Change of Water."
Water Resources Res.. 3, 1967, 263-269.
3. Bouwer, H. and Rice, R.C., "A Slug Test for Determining Hy-
draulic Conductivity of Unconfined Aquifers with Completely or
Partially Penetrating Wells," Water Resources Res.. 12. 1976.423-428.
4. Mercer, J.W. and Faust, C.R., "Groundwater Modelling," National
Water Well Association, 1981.
68 CONTAMINATED GROUNDWA TLR CONTROL
-------
Alternative Treatment Techniques for Removal of Trace
Concentrations of Volatile Organics in Groundwater
Mark E. Wagner
Brian V. Moran
Geraghty & Miller, Inc.
Annapolis, Maryland
ABSTRACT
In response to the detection of volatile organics in wells serving
the process and drinking water needs of an industrial manufactur-
ing facility, a groundwater quality investigation was initiated for
an industrial manufacturer in Maryland. The objectives of the
investigation were to define the extent of groundwater contamina-
tion, to identify probable sources and to develop remedial
measures to rehabilitate the affected aquifer while ensuring a
potable water supply.
Field studies confirmed the sources of volatile organics to be
burial trenches located 350 ft southeast of one of the affected
wells. A hydraulic-control recovery system was recommended,
using the affected well as a pumping center.
Several treatment techniques were identified which could treat
the low-levels of tetrachloroethylene and trichloroethylene. The
alternatives included aerated lagoon volatilization, spray irriga-
tion/land application and packed-tower air stripping. Cost com-
parisons and treatment efficiencies for each process were
evaluated. Air stripping was selected as the preferred alternative
based on efficiency, cost and reuse of the treated water to satisfy
the facility's water needs.
INTRODUCTION
An industrial manufacturer in Maryland depended on wells
on the plant property for process cooling and drinking water.
When low levels of tetrachloroethylene (PCE) and trichloroethy-
lene (TCE) were detected in several of the supply wells, the facility
installed individual granular activated carbon (GAC) cartridges at
various points in the plant. In response to increasing levels of
volatile organics found in the water supply, a groundwater quality
investigation was initiated to determine the probable sources of
the contamination and the degree of aquifer degradation, and to
recommend remedial measures to rehabilitate the aquifer while
ensuring a continuous potable water supply for the plant.
GEOLOGY/HYDROGEOLOGY
The plant site is located within the eastern division of the Pied-
mont physiographic province. The region is characterized by
moderate relief, gentle slopes and rounded hills. Locally, the
bedrock geology consists of folded biotite and chloritic-albite
schist of the Wissahickon formation. The Wissahickon is overlain
by saprolitic material maintaining parental bedding-plane struc-
tures. The saprolite consists of silty clay varying in thickness from
PROPERIY_L|NE.
X STUDY AREA _. :^°'
WOODED AREA
TREATMENT
PLANT
EXPLANATION
300FEET
H
I WATER SUPPLY WELL
MANUFACTURING
BUILDING
Figure 1
Site Plan of the Groundwater Investigation
CONTAMINATED GROUNDWATER CONTROL 69
-------
SURFRCE MRGNETIC SURVEY (Gammas+50K)
Figure 2
Geophysical Survey Results Using Magnetics Instrumentation
"' SURFRCE CONDUCTIVITY SURVEY (mmhos'm)
Figure 3
Geophysical Survey Results Using Conductivity Instrumentation
\
"/ .N
/•*.
?»«
-
\
f~. Scare* *!•••
.„ .5"Ts~r
\ S^b\NS-^f $
100 200
F..I
•SS
\
\
3 ^
Figure 4
Well Locations and Source Areas of Contamination
70 CONTAMINATED GROUNDWATtR CONTROL
-------
as*
Figure 5
Geologic Cross-Section and Groundwater Flow Patterns
25 to 125 ft. Groundwater exists in both bedrock and the overly-
ing unconsolidated deposits predominantly in fractures of
bedding-plane microfractures.1 The secondary permeability
results in a wide areal variation in well yields dependent on prox-
imity of the well to fracture zones. The facility and surrounding
local populace depend on the Wissahickon formation for water
supply. Wells in the area typically penetrate the bedrock from 50
to 450 ft, with well yields varying from 5 to 250 gal/min.
FACILITY DESCRIPTION
The manufacturing facility is located on top of a hill with a
small stream located along the northwestern and western property
boundaries. This stream locally represents the ground water
divide. On-site water-supply wells for the plant are located along
the northwest property line. At the southwest corner of the facil-
ity, a lagoon collects and treats stormwater runoff from the boun-
dary swaleway and treated wastewater effluent from the plant
(Fig. 1). Water-supply well 7 (WS-7), located at the northwest
property boundary, was found to contain the highest concentra-
tions of volatile organic compounds (VOCs) with 1300 /ig/1 tetra-
chloroethylene.
SOURCE INVESTIGATION
Facility personnel tentatively identified possible VOC source
areas as waste pits previously excavated in a clearing 350 ft
southeast of Well WS-7. Wastes, which were suspected of being
present in these pits, included process waste drums, construction
debris and defective products. A surface geophysical investigation
was conducted in this area using a proton precession mag-
netometer and terrain electromagnetic conductance instrumenta-
FRACTURED ROCK
MOVEABLE PACKER /PUMP STRIN6
INFLATABLE PACKER
INFLATABLE PACKER
Figure 6
Straddle Packer Apparatus Used for Stressing Wells
tion to locate buried metallic objects. Both surveys were run on
common transects with 20-ft grid spacings. The proton precession
method delineated three anomalous areas of low and high signals.
Terrain conductance delineated three areas of anomalous signals
located between the proton precession highs and lows (Figs. 2
and 3).
WATER QUALITY INVESTIGATION
Subsequent to the surface geophysical work, numerous nested
well clusters were installed around the major anomalies to: (1) de-
fine horizontal and vertical flow patterns and (2) determine local
groundwater quality. Seventeen monitor wells were installed at
discrete water-bearing intervals in the saprolite material. Wells
were designated by series P, W or S, representing the first water-
bearing zone (20 ft), intermediate water table (40 ft) and bedrock/
overburden interface (60 ft), respectively. The well locations are
shown in Figure 4.
Based on water-level measurements from the installed monitor
wells, groundwater flow is directed to the west toward the small
stream. In addition, three significant seeps were discovered
downgradient of the area under investigation. These seeps are
believed to represent surface discharge of groundwater from the
first water-bearing zone. Vertically, a strong downward com-
ponent of groundwater flow also exists. Groundwater samples
from the wells and seeps were collected and analyzed for
suspected VOCs. Based on the sampling, Anomaly B was iden-
tified as the major contributor of VOCs to the groundwater
regime (Fig. 5). Wells located between Anomaly B and Well WS-7
exhibited tetrachloroethylene concentrations equal to levels
discovered in the water-supply well.
AQUIFER TESTS
In an effort to better define the groundwater regime in the
vicinity of Well WS-7, several aquifer test programs were in-
itiated. A downhole pneumatic straddle packer/pump assembly
was used in Well WS-7 to stress discrete intervals of the openhole
well (Fig. 6). Short-term pumping was performed at various inter-
CONTAMINATED GROUNDWATER CONTROL 71
-------
Table 1
Lagoon Disposal Cosl Estimates
EquIpment
Llni-
-------
Table 2
Slow Infiltration (Spray Irrigation) Cost Estimates
•a
-. I
' U)
a
TOTAL
CAPITAL
COSTS
4J
10
0
-------
Table 3
Stripping Tower Cost Estimates
S t r i p|i l nq Towe i (with b 1 owo r )
P t p i nq
French Drat n
s ?s,ooo
$ 10,000
S 3,000
$ 15,000
$ 19,000
C4M, T«Bpa
Mean* C CD
Means CCD
G*« I
Means CCD
CoetB Un Iquc to
Stripping IUC)
Total Capital
Costa (TCC)
Subcontracting Fees
Eng 1 nee r ing and De» ign
Start-up/Shake Down
C
TOTAL
CWITU,
COSTS
Operating Labor
Electricity for Pumps
$ 7,600 10% o( TCC
S $, 100 9% of TCC
S 3,300 M of TCC
$ 32,700
5108,200
$ 11,000 2 nan-hri/day 1-30
S 3,000 6.25 h.p. * 1-30
Be/kilowatt
Electricity for Blower
3 h.p. •
ec/kllo«att
Monitoring
Not Included
Administration
S 1,000
Maintenance Contingency Reserve
S 1,000
TOTAL
OPERATING
COSTS
evaluated with and without a liner for the swaleway. Mass balance
calculations were made, treating the lagoon as a continuous reac-
tor with first-order kinetics for volatilization of the PCE and
TCE. Evaluation of flow conditions in the lagoon at both max-
imum and average flow rates showed that there would be suf-
ficient retention time for volatilization of PCE to concentrations
below 10 /ig/1 throughout the year.
Total capital costs were estimated to be $41,000 for the unlined
swaleway option and $79,000 for the lined swaleway. Operational
costs would be about $16,000 yearly. Cost estimates for both op-
tions are presented in Table 1.
Spray Irrigation
Treatment of the volatile organics in the groundwater also was
evaluated using a spray irrigation method. With this option,
groundwater would be pumped through a network of sprayers
located throughout part of the wooded land located above the
swaleway. It had been estimated that approximately two acres of
land would be required to allow infiltration of the 100 gal/min of
water. Approximately 60% of the PCE and TCE would volatilize
during spraying. Water containing residual PCE and TCE would
enter the swaleway and flow to the lagoon; however, we deter-
mined that some could infiltrate to deeper sediments.
Since spray irrigation for this site represented a combination of
slow rate and overland flow methods (for water to the swale), the
method was evaluated using both design procedures. Both design
procedures confirmed that sufficient land was available at the site
for this treatment option. Volatilization rates of the contaminants
were evaluated using a mass transfer approach as in the lagoon
option; volatilization rates compared favorably to values in the
literature for similar studies by others.'
Estimated capital costs for the spray irrigation/recovery system
were $108,200; ytarly operational costs would be approximately
$15,000 (Table 2). One problem with this option was that the
spray irrigation system would not be operational during the
months of January and February due to cold weather conditions.
An additional cost consideration for spray irrigation would be the
need for a pilot study to determine actual PCE removal rates and
the ability of the soils to accept the sprayed fluids.
Air Stripping and Reuse of Groundwater
The last option considered was pumping and treating the
groundwater using a packed column tower air stripper. Using
mass transfer calculations, we calculated a 99<% removal of the
volatile organics (PCE and TCE). This option would involve
pumping the water to be treated to the top of the aeration tower
and allowing it to flow downward against a countercurrent flow
of blown air. The system would be designed so that the treated
water could achieve volatile organic concentrations as low as
1 /ig/1 which, in turn, could be used as the water supply for the
74 CONTAMINATED GROUNDWATER CONTROL.
-------
Table 4
Summary of Alternatives, Costs and Evaluation
1 .
Option
Lagoon Disposal
wo/Liner
Total
Capital
Costs
57,730
Annual
Operating
Costs
could be
capital
Comments
desirable on
costs basis
2. Spray Irrigation
Air Stripping/Reuse
S 88,600 S 15,000
$108,200 S 18,000
(S 5,060)
Cold weather operation
problems makes this option
less attractive
Option would allow for
S13,000/yr. savings by
eliminating in-plant GAC
system.
plant. Locating the tower adjacent to the plant buildings would
make it convenient to use excess steam, if available, to ensure
operation of the treatment system throughout the winter.
Total capital costs for the completed stripping-tower/recovery
system were estimated to be about $108,200; yearly operational
costs would be $18,000. The cost estimiates are presented in
Table 3. These costs assumed that the volatile organic air emis-
sions would be permissible and that no carbon recovery system
would be required.
CONCLUSIONS
The three treatment options were evaluated for effectiveness
and cost (Table 4). Assuming that the only contaminants of con-
cern are trichlorethylene and tetrachloroethylene, both the strip-
ping tower and the lagoon systems can effectively reduce VOC
concentrations in the groundwater to acceptable levels; the spray
irrigation option would remove approximately 60% of the
volatiles. The capital costs of the lagoon system would either be as
expensive as the stripping tower or 50% less expensive, depending
upon whether or not lining the swaleway would be required.
However, considering the operating cost of the in-plant activated-
carbon filters currently used for drinking water (which would still
be necessary if the lagoon system or the spray irrigation system
were used), the $18,000 per year operating cost of the stripping
tower would be offset by a $13,000 savings by eliminating the ex-
isting GAC system currently installed. This option would also
allow reuse of the treated groundwater in the plant, since the con-
taminants could be reduced to a "safe" level.
The air stripping option was recommended to the client and will
be constructed for treatment and reuse of the contaminated
groundwater. The lagoon treatment system, although competitive
from a capital cost perspective, was eliminated from further con-
sideration because it did not eliminate the high costs of the
presently installed in-plant activated carbon system nor did it pro-
vide the water-supply advantages that resulted from the packed
tower air stripping treatment system.
REFERENCES
1. Meyer, G. and Beall, R.M., "The Water Resources of Carroll and
Frederick Counties, State of Maryland," Department of Geology,
Mines and Water Resources, Bull. 22, 1958, 355.
2. Jenkins, T.F., Leggett, D.C., Martel, C.J., and Hare, H.E., "Over-
land Flow: Removal of Toxic Volatile Organics," Cold Regions Re-
search and Engineering Laboratory, Special Report 81-1, Feb. 1981.
CONTAMINATED GROUNDWATER CONTROL 75
-------
Cost-Effective Soil Sampling Strategies to Determine
Amount of Soils Requiring Remediation
Gregory J. Gensheimer, Ph.D.
William A. Tucker, Ph.D.
Steven A. Denahan
Environmental Science and Engineering, Inc.
Gainesville, Florida
INTRODUCTION
Designing a sampling strategy to determine the extent of con-
tamination in uncontrolled hazardous waste sites is not as simple
as putting out some transects or an evenly spaced grid. Similarly,
contouring the data should not be considered a simple exercise in
running software. The spatial distribution of the data plus the
statistical distribution will determine the type of sampling scheme
and the type of contouring (interpolation scheme) that will repre-
sent field conditions with greatest certainty.
Sampling schemes and contouring programs are most easily
designed and applied when the spatial distribution of con-
taminants is continuous across the field and when the statistical
distribution tends toward normality or lognormality. If these two
conditions are met, the certainty about the boundaries separating
clean from contaminated soils will be highest. This high level of
certainty also can be achieved by using a minimal sampling
strategy.
Such ideal conditions of spatial and statistical distributions are
rarely met on hazardous waste sites. Hazardous waste sites
typically are very heterogenous. The statistical distributions most
often are dominated by outliers or exist as multiple populations,
and the spatial distributions usually are discontinuous across the
landscape. Such non-ideal conditions reduce the efficacy of sim-
ple sampling schemes or simple contouring programs. Non-ideal
conditions also increase the chances of non-optimal remediation
costs; costs that are much higher than necessary considering the
"true" extent of contamination.
A sampling strategy designed to overcome the non-ideal condi-
tions will be expensive to accomplish, but the extra expense of
sampling can be compensated by reducing the amount of material
removed during remediation. An example of sampling scheme/
remediation optimization is described in detail later in this paper.
Designing the appropriate soil sampling strategy and choosing
the methods of interpolation depends on a variety of factors in-
volving the spatial and statistical distributions of the data set as
well as the expected remedial measures. These factors include: (1)
the method of disposal, spatial scale of disposal areas, site history
and previous sampling results; (2) chemical and physical
characteristics of the contaminants and site including mobility of
contaminants, soil organic carbon, water table and groundwater
flow direction; (3) present spatial extent of the contamination tak-
ing into account the homogenizing effects of the environment; (4)
appropriate sampling, analysis, methods and their costs consider-
ing site conditions, depth of contaminant penetration into the
subsurface and anticipated compounds; and (5) most likely
remedial alternatives and their costs.
Extrinsic vs Intrinsic Variability
The firsi three factors above can be broken into two major
groups relative to the variability contributed by each factor. Ex-
trinsic variability can be thought of as the variability applied to
the system by non-natural forces (humans).
Methods of disposal, spatial scale and distribution of the
disposal areas are all factors directly imposed onto the system by
humans. There is no way to predict the locations or size of
disposal areas from viewing the natural system. A series of aerial
photos may indicate the presence of roads, process buildings,
burning grounds or landfills, the location of which can be used in
designing the appropriate sampling scheme. Human actions set
up the heterogeneity for natural factors to homogenize.
On the other hand, intrinsic variability can be thought of as the
variability applied to the system by the natural environment. In-
trinsic factors such as temperature, precipitation and physico-
chemical characteristics of the media act in conjunction with the
characteristics of the contaminant to homogenize the extrinsic
variability.
When extrinsic factors are more evident than intrinsic factors,
the spatial and statistical distributions tend to be non-ideal (i.e.,
very heterogenous). As the effects of intrinsic factors become
more apparent, the distributions tend to be closer to ideal. This
certainly is not always the case as evidenced by the extreme
variability of contaminant distribution in naturally occurring
fractured bedrock.
For an example of the intrinsic/extrinsic effects of a hazardous
waste site, consider a drum storage area with some leaky drums
interspersed through the area. At time zero, numerous small areas
of heavily contaminated soil would exist in an area of
predominantly clean soils. A gridded sampling scheme would
probably miss most of the contamination, but the statistical
distribution would be dominated by a few extreme outliers. The
extrinsic factors dominate in this situation.
If the same site is studied 10-20 years later, the contaminants
will have begun migrating vertically and horizontally through the
soil. The homogenizing effects of the intrinsic factors are evident.
The same gridded sampling scheme will identify more of the con-
taminated areas, especially at depth. The statistical distribution
will not be dominated by the extreme outliers as before. The
statistical distribution will tend toward lognormality (more ideal),
exhibiting some outliers but not as extreme as at time zero. The
spatial distribution will be more continuous than at time zero.
Factors 4 and 5 consider the economics of the total remedial ef-
fort including sampling and analysis costs as well as cleanup costs.
The importance of designing an optimum sampling strategy is
76 CONTAMINATED SOIL TREATMENT
-------
minimized if the costs of the expected remedial alternatives are
minimal. If one of these feasible remedial alternatives is to ex-
cavate, the number of samples to be taken can be optimized by
considering the estimated volumes of soils needed to be removed
considering sampling density.
OPTIMIZATION OF SAMPLING DENSITY
When the pattern of contamination is relatively simple (a con-
tiguous contaminated area with little or no contamination outside
that area) an optimal sampling density can be defined to minimize
total remediation costs, including both the investigation and
cleanup phases. This concept is illustrated best by considering
some extreme examples. Assume that 100 m2 of soil are con-
taminated. Next assume that samples were taken to define this
contamination: one sample at its center and then samples on a
uniform rectangular grid with a sampling interval of 100 m2 for a
total of 9 samples as in Figure 1. As has been discussed, soil con-
tamination is extremely heterogeneous, and interpolation between
data points is risky at best. It is prudent, based on the data
available from this hypothetical site, to remediate all soils within
the region defined by the clean samples, approximately 40,000 m2.
X SAMPLING LOCATIONS
CONSERVATIVE ESTIMATE OF
CONTAMINATED AREA
ACTUAL
CONTAMINATED AREA
I—-
Figure 1
Example of Inappropriate Sampling Density
On the other hand, one might have started by taking a sample
every meter along a similar rectangular grid. Since phased sam-
pling frequently is not feasible, this approach would have re-
quired 40,000 samples to define the contamination distribution
over the same area. Even if we had been clairvoyant, or had some
independent information on the soil contaminant distribution,
sampling at a density of 1 m would have required 121 samples to
define a contaminated area of only 10 m2.
Obviously these are ridiculous extremes, but they clarify an im-
portant problem. How does one define an appropriate sampling
density? If we sample intensively, sampling and analysis costs
could exceed the cleanup costs. If the sampling grid is too coarse,
we may not accurately define the contaminant distribution. Based
on a reasonable requirement to ensure the public health and safe-
ty, remediation of large quantities of clean soils may be indicated,
wasting valuable resources.
This problem is resolved as a simple linear optimization prob-
lem where the objective is to minimize total remediation costs, in-
cluding the wasteful remediation of clean soils and the cost of
sampling and analysis. The cost of remediating contaminated
soils must occur in any case and can be ignored in the analysis.
The cost of wasteful remediation is calculated as follows. It can
be assumed that two rings of sample locations will define the con-
taminated perimeter. The perimeter of most simple shapes is 3 to
4 times the width (a circle is 3.14 times; a square is 4 times). It can
be assumed approximately that the perimeter, P, is given by P =
3.6 W, where W is a length scale, such as the width. The area
where wasteful remediation can occur is the distance between
sampling points, L, times the perimeter of the contaminated area.
The sample density, D, is the number of samples per unit area.
Thus, L = D - '/2. Combining these definitions, the area in which
wasteful remediation is likely to occur is AR = 3.6 WD- '/*. The
cost of wasted remediation is given by a unit cost, R$, the
remedial cost per unit area times the area of wasteful remediation.
The sampling and analysis costs will be simply the S&A costs
per sample, SA$, times the number of samples, N. The number of
samples is given approximately by N = DW2 (sample density
times area). The total costs thus are given by:
$ = SA$ (W2) D + R$ (3.6 W)D -
(1)
The solution of the optimization problem is derived by differ-
entiating with respect to sample density and setting the derivative
to zero. This will define the sample density at which total costs are
minimized. The result is given by:
D=
^_T
SA$J
(2)
Consider the following example:
SA$ = $1,000 per sample
R$ = $900/m2
W = 75 m
Then D = 0.069 sample/m2.
The sampling interval, L = D -l/z = 4 m, and one sample every
4 m would optimize overall cost. Approximately 400 samples
would be indicated with a sampling program cost of $400,000.
Contrast this with the cost of wasted remediation, estimated in
this example to be $810,000, even though a relatively dense sam-
pling grid is employed. If the sampling interval were 20 m, the
sampling costs would be reduced to $15,000, but the cost of
wasted remediation would increase to $4,000,000. Total costs
would be approximately $3,000,000 higher than the optimal
response.
By inspection, Equation 2 indicates that a lower sampling den-
sity (greater spacing between samples) is optimal if the con-
taminated area is large (large W) or if remedial costs are less than
in the example. If remediation is inexpensive, then overdoing it is
less costly than sampling so there is less need to define the con-
taminated area precisely.
GEOSTATISTICS
Geostatistics are best applied to normal populations that are
spatially continuous. As normality and continuity criteria
deteriorate, geostatistics become less optimum in ability to inter-
polate. Somewhere between ideal and non-ideal data sets, inter-
polation schemes such as method of triangles, inverse distance
weighting and kriging become useless.
Kriging is a valuable interpolator because, unlike other geosta-
tistical methods, kriging produces an error value for each inter-
polated value. By paying close attention to the errors, the user can
CONTAMINATED SOIL TREATMENT 77
-------
determine when kriging becomes a liability in determining extent
of contamination. There is no easy way to tell when other inter-
polators fail to produce accurate estimates.
The Semi-Variogram (SV) is the tool within kriging that iden-
tifies the expected ability of this interpolator to work. A combina-
tion of SV characteristics and goodness-of-fit values helps iden-
tify the usefulness of kriging before interpolation even begins.
Goodness-of-fit values are typically called jackknifing errors' and
have been the standard method of determining the usefulness of
the kriging system and its SV.
Characteristics of the modeled SV are as valuable or more
valuable than jackknifing errors as indicators of the ability of
kriging to produce optimum estimates. The important SV
characteristics include the nugget value (y-intercept), sill value
and shape (value of gamma where the SV levels out) and noise.
The ideal SV has a nugget value of zero, a well defined flat sill and
minimal noise. The noise can be defined as the variability of the
SV data points about the fitted model.
Pitfalls of Using Geostatistics
User knowledge of SV characteristics and the Cumulative Den-
sity Function (CDF) is vital to be confident that the kriging
algorithm actually is producing optimum estimates.
Using jackknifing errors alone may be misleading. Certain
hazardous waste site characteristics tend to result in problems if
kriging is to be used as the interpolator. The problems are
evidenced by SV or CDF characteristics.
Waste sites consisting of numerous small waste spills inter-
spersed through generally uncontaminated soils most likely will
result in a statistical population dominated by outliers. The CDF
would be heavily skewed. The effects of the outliers also will be
noticeable as significant noise and/or a significant nugget value
on the SV. Sometimes the SV is so noisy that a model cannot be
fit to it.
Log transformation is the simplest remedy for this situation
but, depending on the extremity of the outliers, it does not always
work. The transformation will usually smooth the SV, enough to
be modeled, but rarely affects the nugget value.1 Transformation
removes the extreme variance caused by the outliers but does not
remove the baseline variance between the majority of points. In
one case, log transformation reduced the noise but resulted in a
pure nugget (horizontal SV) indicating that the sampling scheme
was not of sufficient density to identify spatial variability.
Kriging this system will tend to greatly overestimate the extent
of contamination. The peak true contamination values will be re-
duced while the true extent of contamination will be overesti-
mated. The interpolation scheme will not generate any new peaks
in the field but will incorrectly increase the extent of low level con-
tamination especially around the measured outliers. This becomes
significant when the cleanup criterion is near background.
Another waste site characteristic that causes problems is one
that contains several lagoons or larger areas of contamination in-
terspersed through relatively uncontaminated soil. A grid system
of points covering this site most likely would show a bimodal
distribution, one peak of near background values and another
peak of high values. The SV may show a poorly defined sill or
possibly even two sills representing the effects of the two popula-
tions of values. If the true SV model cannot be well estimated,
one of the populations will be overemphasized causing a severe
over or under estimate of the true extent of contamination. One
solution to this problem is to work with the two populations in-
dividually to calculate individual SVs and then use both SVs as a
comparison for interpolation.
Waste sites characterized by basin environments are conducive
to kriging. In a basin, the extrinsic variability is decreased by the
intrinsic factors involved with seasonal flooding. Each time the
system floods, the contaminants disperse somewhat creating a
more continuous spatial distribution and decreasing the extreme
outliers to create a more lognormal population. These spatial and
statistical distributions are more ideal than those found on other
types of hazardous waste sites.
Case Histories
The following example shows how information about extrinsic
and intrinsic factors can be used to optimize the utility of kriging.
The site is a basin environment, so the values are dominated by in-
trinsic variability. Boundaries of expected contamination were
identified from previous investigations (extrinsic variability) and
used to help locate the sampling grid (Fig. 2). The results from
this first phase are not available yet, so we've described one of the
probable outcomes. We assume that the eastern third of the site
was uncontaminated while the samples up to the western boun-
dary were contaminated. Our confidence along the western half
boundary is bolstered by the previous information on the location
of the perimeter (extrinsic factors). There is less need of sampling
along this boundary since it is more certain. The dashed line, P2,
represents the line of proposed phase 2 sample locations based
only on seeking a precision of 35% in estimation of contaminated
soil areas. This line would encompass 1.35 times the area cir-
cumscribed by P|. This precision of estimation of contaminated
soil lies in the requirement under recently issued CERCLA
Guidance' that feasibility study cost estimates be accurate to
within + 50% or - 30% of the actual costs. Where the perimeter
extends beyond line Pi, an additional line of samples is proposed
(P3) which again would encompass an area 1.35 times larger than
the previous area.
L-PNA
SE I
EXPLANATION
=, BOOING LOCATIONS FOUND 1
u TO QE UNCONTAMINATEO
. BORING LOCATIONS FCLNO I
TO DC CONTAMINATED j
• PHASE II BORING LOCATIONS
P, IDENTIFIED PERIMETER
P, MINIUM EXTENT OF CONTAMINATION
^ I JS • AREA CIRCUMSCRIBED 8» P
Pj I 33 • AREA CIRCUMSCRIBED 8Y P
TS\ CONFIDENCE INTERVAL
— 90% CONFIDENCE INTERVAL
Figure 2
Sampling Scheme for Hazardous Waste in a Basin Environment
78 CONTAMINATED SOU. TREATMENT
-------
We have less confidence in the boundary along the eastern half
of the site since it is completely estimated. A parameter, C, has
been used to help locate subsequent sample locations.3
Parameter C is defined as follows:
C = [(Criterion) - (Est + (Conf)(KE)]/Standard Deviation (3)
Where:
Criterion = Cleanup criterion
Est = Interpolated value
Conf = Confidence interval (1.28 for 90% confidence)
KE = Kriging error associated with interpolated point.
The 75 and 90% confidence intervals about the boundary of
minimum extent of contamination were drawn using parameter
C. It is obvious that many more samples are required to identify
this boundary than the western boundary due to increased uncer-
tainty in this area.
Parameter C is unique since it combines the cleanup criterion
with the kriging error to arrive at the most likely Phase 2 sampling
locations. Areas that are significantly above or below the criterion
are eliminated from further consideration even if uncertainty is
large in those areas. The use of kriging error alone to indicate
future sampling areas discriminates based only on uncertainty
rather than boundary locations which can be misleading.
Kriging was used in conjunction with a phased sampling
scheme on another site and reported by Tucker et al.3 Parameter
C was used here to identify confidence intervals for the whole site
since the expected extent of contamination was not well known on
this site. This site was not a basin area as described previously, so
it had less than ideal spatial and statistical distributions. Log
transformation of the data eliminated SV noise resulting in a
smoother curve. The uncertainty of boundary locations was still
significant however, and caused a large portion of the site (1/3 of
the area) to be considered for Phase 2 sampling. This level of
uncertainty was caused mostly by a few outliers in the data set and
resulted in the need of a significant Phase 2 effort.
CONCLUSIONS
With the advent of user-friendly computer contouring pack-
ages, there will be an increased tendency to misuse kriging or
overemphasize the positive benefits of kriging when used on inap-
propriate data. Kriging along with other geostatistic methods
should not be used as standard procedure at all hazardous waste
sites. The data must be examined for statistical and spatial
distribution patterns before initiating geostatistics. Geostatistics
become more valuable as the spatial distribution tends to be con-
tinuous and the statistical distribution tends to be normal or
lognormal. Both patterns are relatively rare on hazardous waste
sites since extrinsic factors usually dominate intrinsic factors in
controlling site variability. Extrinsic factors tend to result in
discontinuous spatial data exhibiting statistical outliers.
Kriging is more useful than other interpolators since it produces
an error term with each estimated value. This error can be used to
plan a second phase of sampling using Parameter C when the data
from Phase 1 sampling are less than ideal. This error also may in-
dicate that kriging or other interpolators are not useful on that
particular site.
Finally, sampling schemes can be designed to optimize the total
costs of remediation (sampling costs + remediation costs) by op-
timizing the relationship between the number of samples taken
and analyzed versus the costs of soil remediation. By examining
this relationship mathematically, it appears that costs of extra
sampling can be easily overcome by reducing the final costs of soil
remediation.
REFERENCES
1. Gambolati, G. and Volpi, G., "Groundwater Contour Mapping in
Venice by Stochastic Interpolators. 1. Theory," Water Resources
Res., 155, 1979,281-290.
2. Gensheimer, G.J., Flaig, E.G. and Jessup, R.E., "Effects of Log
Transformation on Semi-Variogram Modeling Using Kriging Error,"
Agronomy Abstracts, 1985, Chicago, IL.
3. Tucker, W.A., Gensheimer, G.J. and Dickinson, R.F., "Coping
with Uncertainty in Evaluating Alternative Remedial Actions," Proc.
Fifth National Conference on Management of Uncontrolled Haz-
ardous Waste Sites, Nov. 1984, Washington, DC, 306-312.
4. U.S. EPA, Memorandum from William Hedeman, Director of Office
of Emergency and Remedial Response, and Gene Lucero, Director of
Office of Waste Programs Enforcement, to Waste Management Di-
vision Directors, Regions 1-10; Subject: Preparation of Records of
Decision for Fund Financed and Responsible Party Remedial Ac-
tions, Mar. 1984.
CONTAMINATED SOIL TREATMENT 79
-------
Land Treatment of Wood Preserving Wastes
John R. Ryan
Remediation Technologies
Ft. Collins, Colorado
John Smith
Koppers Inc.
Pittsburgh, Pennsylvania
ABSTRACT
Numerous wood preserving plants currently are evaluating dis-
posal alternatives for sludges and contaminated soils generated
at these plants. Land treatment is a potentially environmentally
sound and cost-effective means of treating these wastes on-site.
Results of four pilot-scale studies completed at wood preserving
plants in California, Montana, Minnesota and the northeastern
United States are summarized. The land treatment studies in-
volved a detailed evaluation of the degradation, adsorption and
toxicity reduction of the hazardous constituents present in these
wastes.
INTRODUCTION
This paper presents a description of on-going industry spon-
sored studies concerning the land treatability of sludges and con-
taminated soils resulting from wood preserving operations. The
objectives of this paper are to:
• Identify the types of field and laboratory studies being con-
ducted in this area
• Identify the methods being used in the treatability studies
• Present preliminary results of the on-going studies
Background
Of the estimated 400 wood treating plants in the United States,
over 80% use creosote and/or pentachlorophenol as wood pre-
servatives. These plants may generate sludges and contaminated
soils containing the above preservatives as a result of spills, leaks
and settled material from wastewater impoundments. Many of
the wood preserving plants are regulated under RCRA or are on
the National Priority List and must identify alternative means
for disposal of sludges and contaminated soils.
This paper briefly discusses four current laboratory and field
studies concerning the land treatability of wood preserving
wastes. Table 1 summarizes the location at each study and the
types of wastes involved. All the studies have been designed to
meet the objectives of a land treatment demonstration under
RCRA as defined in 40 CFR 270.20. These objectives are to:
• Accurately simulate proposed full scale conditions including
soils, climate and design and operating conditions
• Demonstrate that the wastes can be degraded, transformed and
immobilized within the treatment zone of the land treatment
facility
The four studies described in this paper will provide valuable
data in determining whether land treatment of wood preserving
wastes can successfully meet RCRA standards. It should be em-
phasized that additional studies concerning the land treatability
of wood preserving wastes have been completed. These include
but are not limited to studies by Sims', Sims and Overcash2, Um-
fleet el al.\ Umfleet,4 Edgehill and Finn,' Baker and Mayfield'
and McGiniss.' Laboratory studies also are being conducted
under the sponsorship of the U.S. EPA Robert S. Ken Labora-
tory in Ada, Oklahoma.
CASESTUDY A
The objective of this study was to develop design and operat-
ing criteria for full scale on-site treatment of creosote-contam-
inated soils at a wood treatment plant in Minnesota. The study
included both bench scale and pilot scale evaluation of several
performance, operating, and design parameters. These param-
eters include:
Soil characteristics
Climate
Treatment supplements
Reduction of organics, phenolics and PAH compounds
Toxicity reduction
Effect of initial loading rate
Effect of reapplication
Soil Characteristics and Climate
The study soil is a fine sand which comprises the upper 2 ft of
the bottom of a RCRA impoundment previously used for the
storage of wastewater from a tie treatment plant. The soil is con-
taminated with creosote constituents which primarily consist of
PAH and phenolic compounds. Total benzene extractable hydro-
carbons in the contaminated soil range from approximately 8 to
12% by weight.
The soil has a natural pH of approximately 6.5 and, therefore,
no pH adjustments were necessary prior to initiating the studies.
Specific conductance values and metal contents of the soil are
generally within the range of values reported for natural soils.
The carbon to nitrogen (C/N) ratio of the contaminated soil is
high, and nutrient additions were necessary to reduce the C/N
ratio to a range which promotes microbial growth.
Because the natural soils are fine sands and extremely perme-
able, it was decided that the full scale system would include a liner
and leachate collection system to prevent leachate break through.
To replicate the proposed full scale conditions, the pilot studies
consisted of five lined, 50 ft1 test plots with leachate collection.
Table 2 shows the experimental conditions for each of the test
plots.
The field studies were conducted from July 25 through Oct.
30, 1984. Average ambient temperatures generally exceeded 50° F
during this period. July and August were exceedingly dry months,
80 CONTAMINATED SOIL TREATMENT
-------
Table 1
Summary of Land Treatment Studies
Fj
3C 11 i ty
A
B
Loca tion
Minnesota
C<
s 1 i f or nia
Type
of
Study
L,F
L,F
Type(s) of waste
Creos
soil
:ote
and
Creoso te
con tamina ted
sludge
and pentachlorophenol
Principle
Con taminants
PAH, T. Phenol
PAH, clorinated phenol
ics
contaminated soi 1
C
D
Montana
North East
F
L
Creosote
soil
Cr eoj
con tf
and
>ote
imi n;
contamina ted
sludge
and pentachlorophenol
ated soil
PAH, T. Phenol
PAH, chlorinated phenolic s
L— Laboratory Study
F— Field Study
and the field plots required weekly watering to maintain them
near field capacity. October was a very wet month, and the plots
were near saturation during this period. The active degradation
period appears to extend through the month of October at this
site.
Treatment Supplements
The studies were designed to maintain soil conditions which
promote the degradation of hydrocarbons. These conditions
include:
• Maintain a pH of 6.0 to 7.0 in the soil treatment zone
• Maintain soil carbon to nitrogen ratios of 25 to 1
• Maintain soil moisture near field capacity
In addition, the studies evaluated the effect of seeding the soil
with commercially available microbes adapted to hydrocarbon.
The soil pH was within the desired range at startup and no lime
additions were necessary. Soil pH decreased below 6.0 in most of
the plots after 2 months of operation. Subsequent additions of
1.0 to 1.5 tons/acre of agricultural grade limestone raised the pH
of the test plots to the desired range and no subsequent drops in
pH were observed during the test period.
The equivalent of 10-20 tons/acre of 10:10:10 fertilizer (10 Ib
nitrogen, 10 Ib phosphorous and 10 Ib potassium per 100 Ib fertil-
izer) were added to all the plots at the beginning of the study
based on the estimated carbon content of each plot. The fertil-
izer additions successfully reduced the carbon to nitrogen ratios
of each plot below 25:1. Residual ammonia levels following the
fertilizer application were fairly high, however, creating a con-
cern about possible ammonia toxicity problems.
All the test plots were watered based on the results of daily rain-
fall and pan evaporation data as well as daily soil tensiometer
reading. Soil irrigation was necessary on a weekly basis during
most of the study. The monitoring and irrigation program suc-
cessfully maintained the soil near field capacity. No significant re-
ductions in organics were observed in the plot which was not
watered. These data illustrate that maintaining the soil near field
capacities is a critical operational parameter for obtaining suc-
cessful degradation. Field capacity of the fine sands appears to be
approximately 10% moisture by weight.
No significant advantages in organic degradation can be attrib-
uted to seeding the plots with the adapted microbial culture used
in the study. It is hypothesized that an active natural soil micro-
bial population exists in the contaminated soil. Due to the length
of time the soil contamination existed, the existing soil microbes
are well adapted to the contaminants present in the soil.
Reduction of Organics, Phenolics and PAH
Reductions of benzene extractable hydrocarbons were fairly
similar between all the field plots. Similar reduction trends were
observed in the laboratory reactors, however, at lower rates. Per-
cent removal efficiencies and first order kinetic rate constants
were fairly similar for all the field plots. Average removals for all
field plots over the 4 months were approximately 40% with a cor-
responding first order kinetic constant (K) of 0.004.
The greatest mass of benzene extractable removals was asso-
ciated with the highest initial loading rates. Total removals over
the 4-month test period on the basis of pounds of organics re-
moved/ft3 of soil/degradation month are presented in Table 3.
Table 2
Experimental Specifications for Test Plots
Test
Plot
1
2
3
t,
5
Soil Initial Oil1
Specifications Content (%)
Con t aminated
Hydrobac CL
Contaminated
Hydrobac CL
Contaminated
no microbial
Contaminated
Hydrobac CL
Visibly non-
Hydrobac CL
soil and clean sand; 10
soil and clean sand: 5
soil and clean sand, 5
seed
soil and peat moss, 5
contaminated soil, 3
Fertilizer Cultivation
(Ib) Frequency
50 Weekly
24 Weekly
25 Weekly
25 Weekly
15 Weekly
'Benzene Extractable Hydrocarbons.
•10-10-10 Fertilizer.
CONTAMINATED SOIL TREATMENT 81
-------
Table 3
Total Benzene Extractable Removals In Field Plots
Extractable Content
Plot Number (percent by weight)
1 8.83
2 4.00
3 4.06
4 10.59
5 1.73
Total B
ExtracLnbl
( Ibs/f t of
0
0
0
0
0
e n z e ne
e Removals
soil/month)
.94
. 56
. 34
.91
.20
These removal rates can be compared to data from petroleum
refinery land treatment facilities.' Typical removal rates at well
operated refinery facilities range from 0.1 to 1.0 Ib of oil re-
moved/ft5 of soil degradation month with the lower end corres-
ponding to a "low" initial concentration (such as in Plot 5) and
the high end corresponding to "high" initial concentrations such
as Plots 1 and 4. The data, therefore, are in good agreement with
removals documented at refinery land treatment facilities.
Phenol removals did not exhibit the same linear relationship
over time as the benzene extractable contents. Phenol removals
were observed over the first three months followed by a market
increase in the phenol content of the soil in the fourth month.
The trends observed may be partially a result of analytical vari-
ability but also may be affected by the breakdown of complex
organic compounds releasing phenolic constituents.
First order rate kinetic constants for PAH were fairly similar
between the field and laboratory studies with the exception of 4 +
ring compounds in test runs 1 and 4. The laboratory tests had
higher first order rate constants than the field plots for these com-
pounds.
Table 4 compares the first order rate constants for benzene
extractables and PAH compounds between the laboratory and
field studies which had an initial benzene extractable content of
4 to 5% and the laboratory and field studies which had an initial
benzene extractable content of 8 to 10%.
The median kinetic values (of the 4 to 5% studies and the 8 to
10% studies) are approximately equal for all the parameters. The
most notable variance is for the 4 + ring PAH compounds. The 4
to 5% initial loading rates resulted in slightly higher kinetic rates
for these compounds as compared to the 8 to 10% initial loading
rate. The kinetic rates appear independent of the initial loading
rate within the range of loading rates tested.
Table 4
Comparison of the Range and Median
Kinetic Values at Two Initial Loading Rales
4 to 5 percent
Range
(K. day-')
Median
to 10 percent
Range Median
(K, day-')
Ben zene
Extractables
2 Ring
3 Ring
4 + Ring
PAH
PAH
PAH
Total PAH
0,
0,
0.
0.
0,
.001-0.
,021-0.
.014-0.
.002-0.
.008-0.
007
024
017
007
Oil
0.
0.
0.
0.
0
,003
,023
.016
.004
.009
0,
0.
0,
0.
0.
.002-0.
.01 3-0.
,013-0.
,000-0.
006-0.
004
033
023
005
013
0.
0.
0.
0.
0.
003
023
016
001
008
Toxicity Reduction
A battery of toxicity assays was completed on 3 test plots (1,2
and 5) at the completion of the 4-month study as well as on creo-
sote sludge and contaminated soils which approximately repre-
sented the start up conditions in plots 1 and 2 at the beginning of
the study. The battery of assays included:
• Microtox
• Ames mammalian mulagenicity
• 96-hour static acute fish bioassay
The battery of assays resulted in the following observations:
• Test plots 2 and 5 were non-toxic after 4 months
• Test plot 1 showed intermediate toxicity after 4 months
• Both contaminated soils and the creosote sludge were toxic to
all the bioassays
The microbial assays (microtox and Ames test) on the contam-
inated soils would have suggested that the initial concentrations
in test plots 1 and 2 would have resulted in a toxic response to the
soil microorganisms and therefore no significant degradation
would occur. The results from these plots, however, have shown
significant degradation. This indicates that an acclimated micro-
bial population existed in the contaminated soil prior to the initia-
tion of the studies. The creosote constituents (within the range of
initial concentrations tested) do not result in toxic effects to the
acclimated soil microorganisms used in the toxicity tests.
From an environmental toxicity standpoint, however, the toxic-
ity tests show that the toxicity of the contaminated soil has been
decreased through the treatment process. Complete toxicity re-
duction appears to fall between 2.5 and 5% benzene extractable
content. Complete toxicity reduction was achieved after 4 months
in plot 2.
Effect of Initial Loading Rate
First order rate constants were fairly similar between all the
loading rates. The intermediate loading rate (4 to 5% benzene
extractionables) may demonstrate a slightly higher removal of
high molecular weight PAH compounds. Toxicity reduction also
appears to occur at a faster rate for a 4 to 5% initial loading rate
than at the higher initial loading rates.
The higher loading rates, however, showed the greatest mass
removals and a clear detoxification trend. The studies suggest
that all the loading rates tested are feasible. The selection of an
initial loading rate should balance additional land area require-
ments against time requirements for completing the treatment
process. Moderate loading rates (5%) will result in a faster de-
toxification whereas higher loading rates will decrease land area
requirements.
Effect of Reapplication
Greater kinetic rates were observed after waste reapplication
to a treated soil. The results suggest that average loading of 1 Ib of
benzene extractable/ft' of soil/degration month to a treated soil
can be treated effectively when the application does not exceed
3.0 Ib/ft.1 The first application should occur after the initial
application has been shown to be successfully treated.
Operating and Design Criteria
The studies have successfully developed operating and design
criteria for a full-scale system at this site. These criteria include:
• Treatment period can be extended through October
• Soil moisture should be maintained near field capacity
• Soil pH should be maintained between 6.0 to 7.0
• Soil carbon:nitrogen ratios should be maintained at 25:1
82 CONTAMINATED §OIL TREATMENT
-------
• Fertilizer applications should be completed in small frequent
doses
• Initial benzene extractable hydrocarbon contents of 5 to 10%
are feasible
• Bioassays can be used to determine treatment effectiveness
• Waste reapplication should occur after initial soil concentra-
tions have been effectively degraded
• Waste reapplication rates of 2 to 3 Ib of benzene extractables/
ft3 of soil/2 degradation months can be effectively degraded
The studies completed have shown that the visibly contam-
inated soil at the site can be effectively degraded and detoxified.
The data indicate that the organic content of the visibly contam-
inated soil present in the upper two feet of the RCRA impound-
ment can be degraded to levels similar to the organic content of
the soils beneath the visibly contaminated soils. The data also
indicate that the treatment is effective in detoxifying the waste.
CASE STUDY B
This study involves a field pilot program to evaluate the feas-
ibility of land treatment of contaminated soils from a wood pre-
serving plant in California. Principal hazardous constituents in-
clude polynuclear aromatic hydrocarbons (PAH), pentachloro-
phenol, dioxins and furans. The studies were initiated in the sum-
mer of 1985 and may be continued through the fall of 1986.
Treatment Medium Soils
The test plot soils are located east of a spray irrigation field
which is proposed as the full-scale treatment area. The general
characteristics of the surface soils at the spray field and the test
plots nearby are characterized as clayey silts which grade to silty
clays with depth. Depth to groundwater in the area is generally
30 to 40 ft beneath the ground surface. Specific soil character-
istics of the test plots based on core samples to greater than 7 ft
show the soils to be relatively homogeneous with depth. The
soils through the test plots are characteristic of a brown silty
clay, tight to very tight, with small gravel clasts to 1 in.
Field Plot Experiments
Three experimental field plots were constructed at the facility
near the spray irrigation field. The plots were graded to a 1%
slope and earth berms were constructed between the plots. Run-
off is completely collected at the low slope position of each plot.
The collection areas for the plots are sized to collect the run-off
from the 25-year, 24-hour storm (i.e.: approximately 4.5 in).
Also included in the plot design was the installation of soil-
pore liquid samplers. The samplers are used to collect water from
the unsaturated zone for the purpose of determining if hazardous
constituents are migrating out of the treatment zone. The plots
were fitted with three types of lysimeters per plot: a commer-
cially available vacuum lysimeter, a fabricated glass block lysi-
meter and a gravity flow trench lysimeter. During the course of
the study, the vacuum and glass block lysimeters have proven to
be severely limited in extracting water for analysis. The trench
lysimeter has worked effectively in collecting macropore flow
through the treatment zone which is the most critical factor in
determining if hazardous constituents are moving through the
treatment zone.
The plot design also included the installation of tensiometers.
Tensiometers are important in monitoring soil suction or the
force that determines which way moisture will move in soil. For
the study, tensiometers are used to measure moisture in the zone
of incorporation and at 18 in. to determine irrigation scheduling
and at 5 ft to monitor soil moisture movement through the treat-
ment zone of the three plots.
Startup
After the plots were constructed and the treatment area was
readied, the contaminated soil was spread and mixed in place
with native soil to achieve an approximate benzene extractable
content of 2% in plot 1 and 3000 ppm of pentachlorphenol in
plot 2. Sewage sludge was applied at a rate of 20 tons/acre to
Plots 1 and 2 to increase the organic matter of the contaminated
soil as well as reduce the carbon:nitrogen:phosphorous ratio to
a range to promote microbial growth. Also, agricultural lime was
spread as a step to raise soil pH to 6.0 to 7.0. The final step of
startup activities was the mixing of contaminated soil, fertilizer
and lime with a tractor mounted rototiller to a depth of 6 to 8 in.
Table 5 summarizes the operational activities conducted dur-
ing the pilot scale studies. These activities included test plot culti-
vation, irrigation, sampling and on-site climatological and sub-
surface monitoring.
Table 5
List of Operational Activities
Activity
Test Plot Cultivation
Lime Additions
Fertilizer Additions
Moisture Additions
On site monitoring
-soil temperature
-pan e va poration
-ra in fa 11
-soi 1 suction
-soil moisture
Freq uenc y
Weekly
As necessary to maintain pH 6-7 '
As necessary
As necessary
Daily
Daily
Daily
Weekly
Weekly
'Determined from monitoring data.
Soil and Water Sampling
Soil samples were collected from Plots 1 and 2 at 2-week inter-
vals for the first 6 weeks of the study and then monthly until the
end of the study. Each plot was subdivided into three subplots for
sampling purposes. The subplots were each 12 ft by 16 ft with
1-ft grid spacings for a total of 192 grid squares per plot. Four
subsamples were collected in each subplot during each sampling
event. These four subsamples were then composited into one sam-
ple for each subplot. Each subsample was collected from one of
the grid squares. The four grid squares sampled in any one event
were selected from a random number table. Each grid square was
sampled only once during the course of the study. The random-
ized sampling was recommended as part of quality assurance pro-
cedure to account for field variability.
Also, during the study, nine grab samples were taken from one
subplot during one sampling event as well as the normal com-
posite sampling. This sampling was done to compare composite
sampling versus grab sampling.
Soil core sampling was conducted prior to waste application,
60 days and 90 days after waste application. Cores were taken in
each subplot including the control plot. The cores were collected
every 1 ft to a depth of 7.5 ft.
Two water samples were taken during the study. Runoff sam-
ples were collected from each plot in September, and trench
lysimeter samples were collected from each plot in December. All
samples were shipped in coolers packed with ice under standard
chain-of-custody procedures. Sampling equipment was decon-
taminated with deionized water and methanol after each sam-
pling round.
CONTAMINATED SOIL TREATMENT 83
-------
Table 6
Range of PAH Kinetic Constants
Subplot
1A
IB
1C
2A
2B
2C
2 Ring
0.020
0.014
0.035
0.019
(1)
(1)
3 Ring
0.200
0.014
0.029
0.027
0.020
0.018
4 Ring
0.006
0.007
0.027
0.017
0.012
0.001
Tola 1
0.010
0.009
0.028
0.019
0.013
0.004
(1) not delected at startup.
Organic Reduction Rates
Complete analytical results for the study were not available
during the preparation of this paper. Preliminary results, how-
ever, indicate that benzene extractable removals in each of the
subplots ranged from 34-66% for the 4-month time period with
a corresponding first order kinetic constant (K) ranging from
0.003 to 0.009.
Table 6 summarizes the first order kinetic constants computed
from the preliminary analytical results for various ring classes of
PAH compounds. The following compounds are listed under the
noted ring classes:
2 Rings 3 Rings 4+ Rings
Naphthalene Fluorene Fluoranthene
Acenaphthylene Phenanthrene Pyrene
Acenaphthene Anthracene Benzo (a) anthracene
Benzo (a) pyrene
Benzo (b&k) flouranlhene
Chrysene
Benzo (ghi) perylene
Indeno (1,2,3-cd) pyrene
Dibenzo (a,h)anthracene
Considerable variability was observed in the pentachloraphenol
(PCP) data. Table 7 summarizes the measured values over the
course of the study and the corresponding calculated removal
rates. The PCP data are non-conclusive and continuations of
these studies into 1986 are necessary to draw any treatability
conclusions.
As a byproduct of PCP production, furans and dioxins may be
present at variable concentrations. Tetra and 2378 furans and
dioxins were not detected at startup in Plot 1 and were below
5 ppb and 1 ppb for the Tetra and 2378 furans and dioxins re-
spectively in Plot 2. The remaining furans and dioxins (Penla,
Hexa, Hepta) show removals after 4 months between 75 and
95%. The octa furans and dioxins have the least percentage re-
movals with values ranging from 0-81%.
In summary, significant removals of furans and dioxins were
observed over the 4 month period, and initial startup concentra-
tions were relatively low in comparison to PCP and PAH startup
concentrations.
The studies may be extended through 1986 to evaluate long-
term performance of the plots and evaluate the immobilization
of the contaminants.
CASE STUDY C
This study is an on-going RCRA land treatment demonstra-
tion for a proposed full scale land treatment facility in Montana.
The wastes being evaluated are creosote contaminated soils and
sludges resulting from the closure of a RCRA surface impound-
ment. The proposed full scale land treatment facility will be used
to treat and detoxify the wastes removed from the impoundment
at closure. Design parameters for the field studies were developed
based on results of Case Study A.
Soil and Climate
The treatment area soils consist of loamy silts and silty sands.
The caution exchange capacity of the surface soil is approximate-
ly 20 milliequivalents/100 g of soil and the organic matter content
of the surface soil is approximately 1.5%.
The climate of the area is semi-arid. Annual evapotranspira-
tion exceeds annual precipitation and there is limited recharge to
the groundwater resulting from precipitation and run-off.
Monthly ambient temperatures exceed 50° F approximately 6
months of the year.
Description of Study
The soil and climatic conditions at the site are considered high-
ly favorable for attenuation of the creosote contaminants. There-
fore, unlike Case Study A, this system is not lined. The results
Table 7
Summary of Pentachlorophenol Removals in Plots 1 and 2
Concentration (ppm)
at Time (t) Daya
Percent of Removal
at Ti«e(l) Days(l)
Da to
Date
7/12/85 8/16/85 8/27/85 10/10/85 11/11/85 8/16/85 8/27/85 10/10/85 11/11/85
Plot t-0 t-35 t-46 t.90 1-122 t-35 t-46 t-90 1-122
PIA
PIB
PIC
P2A
P2B
P2C
120
130
160
6800
1400
1700
1 10
95
69
5300
5200
4200
260
1 20
150
8200
5700
5600
54
44
48
1500
2300
1900
210
1 10
1 1
310
1000
480
8
27
57
22
0
0
0
8
6
0
0
0
55
66
70
78
0
0
0
1 5
93
95
29
72
(I) Percent Removal at Time (C(O)-C'(T)/C(O))IOO.
84 CONTAMINATED SOIL TREATMENT
-------
of Case Study A, however, were used to select the design param-
eters for the study including waste application rate and treat-
ment supplements.
Waste Application Rate
Waste application rates are based on gross organic loading as
defined by benzene extractable hydrocarbons. The loading were
selected on the basis of results from Case Study A and involved
initial loadings of 2% and 5% benzene extractables. The 2% plot
includes subsequent bi-monthly applications of 2% benzene ex-
tractables to evaluate the effect of waste re-applications. The 5%
plot does not include waste re-application in order to evaluate the
effect of one batch loading.
Treatment Supplements
To minimize the need for replicate plots and minimize the
analytical costs of the study, certain management practices were
selected as "good management practices" for the study and were
applied consistently to all the plots. These "good management
practices" are anticipated to optimize treatment (based on liter-
ature data) and include:
• Maintain surface soil pH between 6.0 and 7.0
• Maintain a soil carbon:nitrogen ratio of 25:1
• Maintain soil moisture near field capacity through irrigation
• Apply manure or organic amendments to provide nutrients
and increase the organic carbon content of the surface soil
• Till the soil once every 2 weeks
• Prevent storm water run-on from entering the plots and capture
run-off
Construction of the Plots
The field plots were constructed in June 1985. One plot is a
control where no wastes have been applied but all the "good
management practices" are implemented. The other two plots are
for the two waste loading rates being evaluated.
Each plot is 12 ft x 50 ft with a defined treatment depth of 5 ft.
The plots are separated by earthen berms which prevent run-on
from entering the site. Each plot is graded to a 1 % slope, and a
lined run-off collection sump capable of handling two 25-year,
24-hr storm events is present at the toe of the slope. The entire
study area is fenced:
Three types of soil-pore water collection devices were installed
in each plot. These include:
• A vacuum lysimeter capable of placing a suction on the sur-
rounding soil
• A pan lysimeter constructed from a hollow glass block approx-
imately 1 ft2
• A trench lysimeter consisting of a 10-ft long 6-in. diameter
PVC drain, placed at the bottom of the treatment zone
In addition soil tensiometers were placed at 3 ft and 5 ft. These
tensiometers in conjunction with a hand tensiometer used to take
soil suction readings at the surface are used to evaluate the
hydraulic flux through the soil.
Monitoring
Monitoring at the site includes monthly sampling of the zone of
incorporation to evaluate the degradation of specific organic con-
stituents and bi-annual sampling of soil cores, soil pore water and
groundwater to evaluate contaminant immobilization.
Table 8 presents the analytical parameters evaluated in each
medium. The principal hazardous constituents being evaluated
are polynuclear aramatic hydrocarbons PAH. These constituents
are monitored on a monthly basis along with indicator para-
meters such as benzene extractable hydrocarbons and total phen-
Table 8
Analytical Parameters, Site C
Parameter/Method
Medium
Z
C
Z
Z
2,0,1
Z,C,L
L
Z,C
C.L
Z,L
Z
-Benzene Extractables/Soxhlet extraction
-Metals/ICAP
-Total kjeldahl nitrogen/automated titrimetric
-Ammonia nitrogen/automated titrimetric
-pH/combination electrode
-Conductivity/conductivity bridge
-Total organic carbon/combustion or oxidation
-PAH/Method 8270
-Base neutrals/Acid extractables
-Microtox/Toxicity Analyzer System
-Fish bioassay
Z zone of incorporation soil samples.
C = core soil samples
L lysimeter water samples.
ols. In addition, nutrients, soil electrical conductivity and soil
pH are monitored on a monthly basis to evaluate the need for
treatment supplements. Soil moisture is checked on a weekly basis
to evaluate the need for supplemental irrigation. The relative
toxicity of the surface soil is evaluated through monthly micro-
tox analyses.
Bi-annual soil core sampling includes collection of soil cores in
1 ft increments down to the bottom of the treatment zone (5 ft).
The soil cores, groundwater and lysimeter samples are analyzed
for indicator parameters and for a complete scan of base neutral
and acid extractable organics using GC/MS techniques. Inorgan-
ic metals also are included in the analyses using ICAP techniques.
All sampling and analyses include triplicates from each plot to
determine statistical variability. The startup of the study was in
July 1985, and it is scheduled to be completed in 1986. Insuffic-
ient analytical data were available at the time of this writing to
evaluate degradation trends. However, the soil core, lysimeter
and groundwater analyses have not detected any migration of
hazardous constituents from the treatment zone.
CASE STUDY D
This project involves a detailed laboratory study using creo-
sote and pentachlorphenol contaminated soils from an aban-
doned wood preserving plant in the northeastern United States.
The study was initiated in December 1985 and is scheduled to run
for 8 months.
The laboratory study program evaluating land treatment is
divided into the following three areas:
• Evaluation of leachate movement through soil columns and
contaminant volatilization
• Examination of contaminant degradation rates in soil pans
• Development of soil/water partition coefficient for PAH and
pentachlorophenol associated with the site soil
Soil Column Evaluation
The studies outlined below are designed to evaluate the rate
and extent of treatment, including biodegradation and retarda-
tion. A mass balance of selected constituents will be determined
by taking into consideration the initial loading rate, the concen-
tration of constituents in the leachate for the duration of the 8-
month experiment and the concentration of constituents in the
soil core at the termination of the experiment.
The soil treatment process will be evaluated using 4-in. diam-
eter, 5-ft long glass column reactors packed with clean soils to
represent actual site conditions. The columns contain the same
CONTAMINATED SOIL TREATMENT 85
-------
soil profile materials that exist at the site. Native soils were
collected in 6-in. intervals and packed into the columns in the
same order as obtained in the field to obtain a representative
soil column profile.
Before any contaminated material was applied to the soil col-
umns, chloride tracer studies, using sodium chloride, were per-
formed. This work was completed to determine the proper hy-
draulic loading rate once contaminated material is applied.
Actual contaminated soil from the site was mixed into the
upper 6 in. of the clean soil in a porportion to give approximate-
ly 5% benzene extractables. Common commercial fertilizer and
manure also were added and the soil pH was adjusted to and
maintained between 6-7 by lime additions. The columns are to be
operated for approximately 8 months with only one initial waste
application. Watering of the columns is to be done with its pH
adjusted to approximately pH 4. The pH adjustment is done to
simulate the existing pH of rain water for the area. In addition
to routine watering, the columns are to be watered with the equiv-
alent of a 25-year, 24-hr storm event in January and May during
the study. The soil is to be tilled every other week. Volatile emis-
sions are to be addressed as part of the column work.
For QA/QC purposes, two duplicate columns containing the
waste material are operated. A third column containing only
clean soils serves as a control and is operated in the same man-
ner as the columns with contaminated material. All three columns
are to be sampled periodically for analyses.
Soil Pan Reactor Evaluation
In conjunction with the operation of the soil columns, a soil
pan study phase is to run for approximately 8 months with
degradation rates of specific compounds addressed. Four soil pan
reactors are operated. The soil pans are constructed of aluminum,
have a surface area of approximately 1.5 ftj and contain approx-
imately 6 to 8 in. of soil. The soil is a mixture of contaminated
material and clean surface soil from the site in a proportion to
give a concentration of approximately 5% benzene extractables.
Fertilizer and manure have been added to give a C:N:P:ratio of
approximately 50:2:1, respectively. Manure was applied corres-
ponding to a loading rate approximately 10-20 dry tons per acre.
This is the same soil misture that is to be used in the column re-
actors. The soil pH is maintained within a range of 6-7 by lime
addition as needed, and the soil moisture is maintained within
50% of field capacity by laboratory tap water adjusted to a pH of
4. During the course of the experiment, hand tilling is to be done
every other week.
For QA/QC purposes, three duplicate pans containing the soil
waste material lime, fertilizer, manure are to be set up. Two of
the pans are to be maintained in a moist environment by water
applications as needed. The third pan is not to be watered and is
to serve as a contaminated soil control. A fourth pan, contain-
ing only clean soil, is to be operated in the same manner as the
contaminated soil pans.
Soil/Water Partition Coefficient Development
Soil/water partition coefficients are to be developed for penta-
chlorophenol and PAHs as these compounds specifically relate to
the contaminated site soils. For this work, different solution
pHs are to be evaluated to examine the effect of pH on the leach-
abilities of PAHs and pentachlorophenol.
This information is needed to develop a better theoretical
understanding between pentachlorophenol and PAH attenua-
tion by the soil matrix.
CONCLUSIONS
Land treatment is a potentially technically effective and eco-
nomic means of treating and disposing of wood preserving wastes
from sites currently regulated under RCRA or CERCLA. Rigor-
ous 2-year studies are currently under way to determine appro-
priate engineering design features for these systems. The data
from those studies will provide a sound scientific basis for eval-
uating the feasibility of this technology at other wood preserv-
ing sites.
REFERENCES
1. Sims, R.C., "Land Treatment of Polynuclear Aromatic Com-
pounds." Ph.D. dissertation. Depi. Biol. Agr. Eng.. No. Carolina
State Univ., Raleigh, NC, 1982.
2. Sims, R.C. and Overcash, M.R., "Fate of Polynuclear Aromatic
Compounds in Soil-Plant Systems," Residue Reviews 88. 1983, 1-68.
3. Umfleet, el al.. "Reclamation of PAH Contaminated Soils." Pre-
sentation made at Environmental Engineering 1984. Specialty Con-
ference, ASCE, Los Angeles, CA. June 1984.
4. Umfleet, D.A., "Land Treatment of PAH Compounds," Unpub-
lished M.S. thesis. Department of Civil and Environmental Engi-
neering, Utah State University, Logan, UT, 1985.
5. Edgehill and Finn, "Microbial Treatment of Soil to Remove Penta-
chlorophenol. " Appl. Environ, .\ficrobiol. -IS. 1981, 1120-1125.
6. Baker, M.D. and Mayfield, C.I.. Water. Air and Soil Pollul., 13,
1980,411.
7. McGinnis, G.D., "Biological and Photochemical degradation of
pentachlorophenol and creosote," Mississippi Forest Products Lab-
oratory. Mississippi Slate, Ml, 1985.
8. American Petroleum Institute. "Land Treatment Practices in the
Petroleum Industry." API. Washington, DC, 1983.
86 CONTAMINATED SOIL. TREATMENT
-------
Method for Determining Acceptable Levels
of Residual Soil Contamination
William A. Tucker
Carolyn Poppell
Environmental Science and Engineering, Inc.
Gainesville, Florida
INTRODUCTION
With the exception of PCBs, regulatory standards or estab-
lished numerical criteria that define acceptable levels of soil con-
tamination at hazardous waste sites do not exist. However,
numerical criteria are readily available for air and water. In con-
trast to air and water, soils are not fluid and are more readily sub-
ject to property rights. People experience little direct exposure to
contaminated soils. For example, adults inhale approximately
17,000 I/day and ingest roughly 2 I/day of water, but they
generally are believed to accidentally ingest less than 0.00005 1 of
soil per day. For these and other reasons, Congress has not
enacted the "Clean Soils Act," and soil quality standards have
not been implemented.
The extent of human exposure to residual contamination in
soils can be extremely variable, depending on such factors as
climatic factors and soil properties. The most important variable
is land use. Contaminated soils in a backyard playground and
subsistence garden (common in many residential areas) pose a
much greater health risk than in a paved shopping mall develop-
ment.
Consequently, it is reasonable to determine acceptable levels of
residual soil contamination on a site-specific basis, considering
the probable (or possible) future uses of the land during the
period that it will remain contaminated. Many feasibility studies
for remediation of contaminated sites and remedial actions im-
plemented at sites reflect this concept. On the other hand, im-
plications of this concept have been addressed only implicitly or
qualitatively in many Remedial Investigation/Feasibility Studies
(RI/FS) conducted to date. Systematic and quantitative im-
plementation of this concept can be accomplished with multi-
media exposure and risk assessment procedures that are readily
available and well documented.
The purpose of this paper is to present these procedures in the
context of the endangerment assessment process as they currently
are implemented by the authors. The methods used by the authors
are similar to the PPLV method developed by Dr. David
Rosenblatt and co-workers of the U.S. Army Medical Bioengi-
neering Research and Development Laboratory; the interested
reader should consult Rosenblatt and Dacre7, Rosenblatt, et a!.'
and Small9 for further discussion of their procedures. Following
the presentation of the method, a few examples will be presented
to illustrate some features of its implementation.
APPROACH
Establishment of acceptable residual soils concentrations often
will be an essential element of the endangerment assessment pro-
cess. Endangerment assessment links the remedial investigation to
the feasibility study in two ways:
• The process results in determination that an endangerment or
public health risk exists. If the determination is positive, it
virtually rules out the no-action alternative and demonstrates
the need for the feasibility study.
• The endangerment assessment should clearly identify the en-
vironmental quality criteria that would be required for the
CERCLA alternative; that is criteria which, if met, would sub-
stantially eliminate the endangerment.
The establishment of acceptable levels of residual soils con-
tamination contributes to both objectives of the endangerment
assessment. Endangerment may be suspected if the "endanger-
ment criteria" are exceeded. The role of "endangerment criteria"
in the RI/FS process is presented in Figure 1.
The recommended approach for determining acceptable levels
of residual soils contamination is presented in Figure 2. Results of
the remedial investigation will identify the contaminants of con-
cern for the site. Inputs from the endangerment assessment team
will be useful at an early stage in the remedial investigation to
refine the list of contaminants of concern to focus RI resources on
the "riskiest" contaminants. Further refinement and screening
undoubtedly will be required as the RI data are acquired and re-
viewed. In some circumstances a formal hazard ranking system
may be useful to screen contaminants, and applications of a for-
mal screening model may be a component of the uppermost
"box" of Figure 2.
Once contaminants of concern have been identified, investiga-
tions proceed along two parallel and simultaneous paths which in-
terface at intermediate points in the process as well as in the in-
tegrative final steps. The two paths are contaminant fate (down
the right in Fig. 2) and contaminant toxic effects (down the left).
Effects
First, the available toxic effects data for the contaminants of
concern are reviewed. It is important to identify any particularly
sensitive subpopulation (e.g., children, women of childbearirig
age, aquatic species or age classes). Considering this information
on sensitive subpopulations and site-specific factors (e.g., in-
dustrial site where children would not be exposed) permits the
identification of populations at risk. Populations at risk are those
potentially exposed populations who also are sensitive to the con-
tamination. The most sensitive population at risk normally will be
considered as the basis of the endangerment criteria. However, if
the number of individuals in the most sensitive population at risk
is likely to be very small, it may in some instances be reasonable to
base the criteria on another sensitive population while allowing
for special or auxiliary protection of the special, rare subpopula-
tion.
CONTAMINATED SOIL TREATMENT 87
-------
REMEDIAL INVESTIGATION
COLLECT AND ASSESS SITE
SPECIFIC DATA
DEVELOP PRELIMINARY
ENOANOEFIMCNT CRITERIA
NO FS NEEDED
NO ACTION ALTERNATIVE
ACCEPTABLE
Figure I
Role of "Endangcrmenl Criteria" in RI/FS Process
The next step in the effects side of the process is to determine
acceptable exposure levels for the sensitive population at risk.
Determination of an acceptable exposure level is often the most
controversial step in the process depicted in Figure 2. Much of the
effort, techniques and procedures frequently referred to as carcin-
ogenic or quantitative risk assessment are almost exclusively
devoted to this element of the process. Later in this paper, it will
be shown that other aspects of the process also introduce substan-
tial uncertainty and variability in estimates of acceptable residuals
levels. For carcinogenic contaminants, this step also introduces
the controversial issue of acceptable risk: recent policy guidance
from the U.S. EPA indicates that a 10-* lifetime individual
cancer risk level should be evaluated as a cleanup goal at
CERCLA sites, consistent with its guidance for groundwater
Alternative Concentration Limits, a site-specific standard setting
mechanism under RCRA. Alternative cleanup standards resulting
in lifetime individual cancer risks in the range of 10 ~8 to 10-4
also should be considered.
Land Use Scenario
As illustrated in Figure 2, a critical requirement of the soil
criteria setting procedure recommended here is the need to define
an exposure scenario. As used in this discussion, an exposure
scenario is a characterization of the behavioral patterns of the
population at risk that could result in, or reduce, exposure to the
contaminants at the site. The critical element of the exposure
scenario will be the land use scenario used in the analysis.
Depending upon the degree of conservatism that may seem
warranted to affected parties, and the persistence of site-related
contaminants, the exposure scenario may be based on: (1) the ex-
isting land use, (2) probable future land use over a specified
future period, based on existing or predicted patterns of land
development and (3) the use (of all possible future uses) that leads
to the greatest potential for exposure. The choice among these
alternative scenarios is not strictly a technical decision, but the
technical implications of selecting among these alternatives can be
addressed by a technical specialist. Ultimately, the selection of a
land use scenario for the analysis and the selection of an "accep-
table risk" are political/regulatory decisions, not technical ones.
It will be shown later that these decisions can affect the ultimate
soils criteria by factors of a thousand or more.
Environmental Fate
Parallel to and simultaneous with the review of toxic effects
data, the environmental physical chemical properties of the con-
RCVIEW ENVIRONMENTAL
BEHAVIOR OF COMPOUNDS:
ENVIRONMENTAL SETTING
IDENTIFY DOMINANT
PATHWAYS OF MIGRATION
FROM SITE
DETERMINE QUANTITATIVE RELATIONSHIPS
BETWEEN SOIL CONCENTRATION AND HUMAN
EXPOSURE ALONG EACH PATHWAY
EXPOSURE VIA DRINKING WATER • EXPOSURE VIA
CONSUMPTION OF FISH • _< ACCEPTABLE DOSE
ACCEPTABLE SOIL CONCENTRATION
Figure 2
Determination of Acceptable Soil Concentrations
by Exposure and Risk Assessment
taminant are reviewed in the context of the environmental setting
of the contamination/soils to evaluate the site-specific en-
vironmental behavior of the contaminants.
Dominant pathways of migration from the site are identified by
screening level analyses and calculation. Screening level analysis
procedures are diverse, and flexibility must be maintained in their
application in order to derive the most information possible from
the available data, while conserving "level of effort" resources.
Identification of the dominant migration pathways usually can be
deduced on the basis of the judgment of senior environmental
scientists familiar with the site. Occasionally, formal screening
level calculations as represented, for example, by methods of En-
field, el at.' or Tucker, el a/.10, can be applied at this step.
Integrating the information from the exposure scenario with
the dominant migration pathways yields the dominant pathways
of exposure. These are the pathways, from soil to an exposed
member of the population at risk, that are capable of carrying
CONTAMINATED SOIL TREATMENT
-------
and delivering a significant dose of contamination. An example of
an exposure pathway would be "surface run-off erodes surficial
soils and carries both dissolved and adsorbed soil contaminants to
a nearby estuary; fish and shellfish bioconcentrate the con-
taminants from the water, and the organisms are harvested for
human consumption."
Once the dominant exposure pathways are identified, each is
expressed as a mathematical equation relating the dose (exposure)
via the pathway to the soil contaminant level. These equations are
representations of actual processes and are designed to express the
relationship conservatively (inferring a higher-than-expected ex-
posure for a given soil concentraton) to favor protection of public
health. The most important pathways should be evaluated with
the greatest degree of sophistication to make their exposure
estimates more reliable.
Finally, the exposures via all of the individual pathways are
added together, and the sum (total exposure) must be less than the
acceptable exposure. This step links the exposure assessment with
the toxic effects evaluation to produce the risk assessment-based
criterion. Each pathway is defined as a factor times the soil con-
centration, and the acceptable exposure and the factors are
known. Consequently the equation may be inverted to derive the
acceptable soil concentration.
APPLICATIONS OF THE PROCEDURE
In most applications of this procedure conducted by the
authors, the following pathways of exposure have required con-
sideration:
• Incidental direct soil ingestion
• Dermal absorption
• Uptake by plants in the human food chain
• Leaching by infiltrating rainwater leading to contamination of
shallow water table aquifer
• Groundwater or surface run-off discharge to surface water
bodies supporting a fishery
• Resuspension of contaminated soil particles to the atmosphere
by wind erosion or mechanical entrainment associated with
earth moving activities
Consequently, the assumptions related to exposure via the pro-
cesses and the uncertainty associated with procedures used to
quantify these processes, become critical issues in determining ac-
ceptable levels of residual soil contamination.
Importance of Land Use Scenario
The land use issue can be investigated with reference to these
processes by contrasting the assumptions involved for a residen-
tial land use scenario vs. an alternative land use. Consider a
hypothetical example: rural land now used for lumber production
has been determined to be contaminated. Assume it has been
estimated that the contaminants may persist in soils for 50 to 200
years. If the existing land use is maintained for that period, then
the exposure potential is very low. Workers involved in lumbering
will be exposed to soil contamination via dust inhalation, inciden-
tal direct soil ingestion and dermal absorption. Small9 recom-
mended the use of 10 mg/m3 as a reference exposure level for this
situation. Since levels higher than 15 mg/m3 are not permitted in
the workplace, this is a useful conservative estimate. Actual levels
may be much lower. Depending on the importance of this
pathway relative to others, a more realistic estimate considering
level of activity, soils and climatic factors may be warranted. Ex-
posure to airborne dust also depends on the pulmonary ventila-
tion rate. Data are presented in Anderson, et al.' and Small.*
Adults probably ingest very little soil. Soil ingestion occurs as a
result of contamination of food and food wrappers, dirt on hands
(especially prior to eating) and ingestion of inhaled particles
trapped in the upper respiratory tracts. Definitive data on typical
or extreme rates of soil ingestion among adults are not available.
Small' recommends the use of 100 mg/day, an estimate consistent
with Ford and Gurba" recommendations. Although there is very
little basis for this estimate, it is commonly used. General support
for the estimate relates to studies of juvenile soil ingestion that in-
dicate ingestion rates from 100 to 1,000 mg/day.2 It is generally
accepted that most adults ingest less soil than most children.
A basis for estimating dermal absorption exposure to con-
taminated soils was presented recently by Layton, et al.6 Their
analysis indicates that it is conservative to assume that adults are
exposed to the contamination contained in 43 mg of soil per day
via this exposure pathway. Alternative approaches to estimating
dermal absorption are available, but few address the details of ab-
sorption from contaminated soils, so the Layton, et al.6 method is
a useful basis for screening level analysis.
The information presented above may be combined to estimate
the total dose experienced by the individual as follows:
Incidental
Ingestion
Airborne
Inhalation Dust
Dermal
Absorption
Dose = [0.1 g/day + 17 mVday (0.01 g/m3 + 0.043 g/day]Csoil (1)
Expressed in terms of the soil concentration:
DoseO»g/day) = 0.31 g/day x Csoilf4g/g
(2)
To determine the acceptable soil concentration, ACS, we sim-
ply require that the actual dose be less than or equal to the accep-
table dose:
Dose = 0.31 g/day x'ACS S Acceptable Dose
implying ACS < Acceptable Dose
0.31 g/day
(3)
(4)
To complete the hypothetical example, insert an acceptable
dose for a 70 kg adult of 2,000 /ig/day. Then the acceptable soil
concentration is 6,450 /tg/g.
Figure 3
Characteristic Elements of a Residential Land Use Scenario
CONTAMINATED SOIL TREATMENT 89
-------
The recommended residual contamination level would have to
be much lower if it were anticipated that the land would be
developed residentially. Figure 3 illustrates some of the exposure
pathways operating under a residential land use. It must be
assumed that children may spend most of their time playing on
contaminated ground. Subsurface contamination can be exposed
during normal residential landscaping which also may resuspend
dust into the air. Families may derive a significant fraction of
their diet from a backyard garden. The residential water supply
may be derived from shallow individual wells, depending on local
hydrogeologic conditions and the availability of community water
supplies. These factors can lead to substantially higher exposure,
especially for children.
Assume that the population at risk is children ages 1 to 3 having
a typical body weight of 12.5 kg. Their acceptable daily dose for
the (same) contaminant would be 350 /ig/day. Based on the data
presented by Binder, el at..1 it seems likely that the average child
ingests 100 to 200 mg of soil per day and a value of 150 mg/day
will be used in this sample calculation. Kolbye, et al.' provide
estimates that a 2-year-old ingests 0.4 I/day of water and 125 g of
potatoes and other vegetables per day (20 g/day, dry weight);
while Layton, el a/." estimate that a child is exposed to the con-
tamination found in 41 mg of soil per day, by the dermal absorp-
tion route. These exposure factors will account for the major ex-
posure pathways to children in a residential setting. To relate
these exposure factors back to the soil concentration, it is
necessary to establish the relationship between soil concentrations
and: (1) plant concentrations (plant bioconcentration factor) and
(2) groundwater/drinking water concentrations.
The best method for estimating plant bioconcentration factors
will depend on the combination of contaminant and plant(s) and
the data available for that combination. Assume that Small's"
guidance has been used to indicate that plant concentration on a
dry weight basis will be approximately five times soil concentra-
tion.
Groundwater/drinking water concentrations may or may not
be related to soil concentrations. However, if the shallow water
table aquifer yields water of acceptable quality and quantity for
residential potable use, then it usually would be assumed to be the
source of drinking water. Further, and for the sake of simplicity
in this example, if the contaminated land is situated at a ground-
water divide, then the quality of water in the wells will be similar
to the quality of the soil solution in equilibrium with the con-
taminated soils. Then the concentration in drinking water will be:
mg/1
SOil
(5)
where: Kj is the adsorption coefficient. Assume that Kj = 125
for the contaminant in site soils. Combining this information for
the various pathways, the total dose for the 1-3 year old children
comprising the population at risk can be calculated in proportion
to the soil concentration.
Soil Drinking Garden Dermal
ingestion water vegetables absorption
Dosc= CSO,| (0.15 g/day + 0.4 I/day li'** + 2(1 g/day • 5 • 0.041 g day) (6)
kd
Expressed in terms of the soil concentration:
Dose O^g/day) = 103 Csoi, (/tg/g)
97% of the dose (100/103) is calculated to come from the garden
vegetable pathway.
To calculate an acceptable soil concentration (ACS), the dose
must be less than the acceptable dose.
Dose = 105 g/day Csoi, < 350 ^g/day
ACS<3.4Mg/g
Comparing this value with the ACS for the lumber production
land uses scenario (ACS S 6,450/ig/g), it is seen that the land use
scenario assumption can affect the recommended soil criterion by
more than three orders of magnitude. This degree of sensitivity to
assumptions is apparently at least as important as the uncertainty
regarding toxic effects and quantitative carcinogenic risk assess-
ment procedures.
Refinement of Exposure Factors/
Reduction of Uncertainties
The results of any predictive method, such as the one presented
here, always will be uncertain. In multimedia exposure and risk
assessment, the number of variables and uncertainty in each
results in large uncertainties. A traditional response to uncertain-
ties in environmental assessment has been a reliance on conser-
vative estimates. The true situation is unknown, but estimates and
decisions arc based on extremely adverse assumptions.
The simple predictive methods presented in the examples above
can be refined by acquisition of more data or application of more
sophisticated predictive methods. As the estimates become more
reliable, the degree of conservatism can be reduced. This process
usually will result in a less stringent, yet more firmly based,
recommendation.
As an example, consider the basis for estimating inhalation ex-
posure for the lumber production scenario. It was assumed that
air concentration of dust that had been stirred up from con-
taminated areas could be as high as 10 mg/m'. The basis for that
estimate was conservative: if levels were higher, the workers
would wear respirators. This estimate could be refined either by
sophisticated wind erosion/fugitive dust modeling or by measur-
ing air concentrations under the conditions of the exposure
scenario. Assuming that fixed air monitors were placed in a
similar, though uncontaminated area, and that personnel
monitors also were employed in a 1-month study of worker ex-
posure to nuisance dust, a more realistic dust concentration of 0.4
mg/m-1 was determined. Applying this finding, the previous
analysis can be refined. The dose is now estimated to be:
Soil
ingestton inhalation
Dose - CSO,| (0 I g da> * 17 m.Y das •
= 0.15 mg da> C'ioil I eg g)
Airborne Dermat
dim absorption
0 IXKM g m» * 0.043 mg day) (7)
and the acceptable soil concentration would be:
ACS=
0.15
= 13.400
Such refinements almost invariably should result in a relaxation
in the recommended criterion, as long as uncertainties always are
resolved with a conservative assumption.
CONCLUSIONS
A method for determining safe levels of residual soils con-
tamination as a component of the endangerment assessment pro-
cess for hazardous waste sites has been presented. The method
presented is based closely on the PPLV approach developed by
Dr. David H. Rosenblatt and co-workers at the U.S. Army
Medical Bioengineering Research and Development Laboratory,
Ft. Detrick, MD."* The method has been restated here, using
slightly different terminology, to illustrate its relationship to the
endangerment assessment process.
Development of a hypothetical example illustrates the sensitiv-
ity of the recommended criteria to the exposure assumptions
adopted, particularly those relating to the future land use of the
site. The land use scenario adopted for the analysis affected the
recommended acceptable soils concentration by more than three
90
CONTAMINATED SOIL TREATMENT
-------
orders of magnitude in the example problem. It is possible to con-
struct alternative scenarios that affect the criteria by even greater
amounts. Thus, uncertainty regarding future land use and the
assumptions adopted in the face of that uncertainty can influence
the endangerment assessment, the evaluation of the no-action
alternative and the costs of remedial action just as severely as
uncertainties regarding the "toxic effects/quantitative risk assess-
ment" aspects of the problem. The exposure assumptions are
readily understood and thus readily debated by all concerned par-
ties, and so they can and should receive a thorough airing during
the RI/FS. Decisions regarding appropriate future land uses for
the site are quintessentially political, rather than technical deci-
sions. Affected parties should be consulted, and these decisions
should be made by regulatory authorities and/or affected parties.
The second major conclusion of the paper is based on an
analysis in which one of the criteria was refined by acquisition of
additional data. This allowed a more realistic analysis and the
discarding of an excessively conservative assumption. More
sophisticated theoretical analysis frequently can achieve a similar
result. To develop criteria in the face of great uncertainties, it fre-
quently is prudent to adopt conservative assumptions. This af-
fords the decision-maker the certainty that the criterion level is
safe and that concentrations below the criterion do not pose an
unacceptable risk. The converse, however, is untrue: levels above
the derived criteria are not necessarily unsafe. The refinement of
estimates presented bears this out: with a conservative assump-
tion, applied in lieu of site-specific data, it was determined that
soils having 6,450 jtg/g of contaminant "X" were "safe" and
presented no endangerment. Upon acquisition of additional data,
it became apparent that soils as high as 13,400 /ig/g can be left in
place. This does not contradict the previous analysis. Before ac-
quisition of the new data, the analysis indicated that 6,450 jtg/g
was a safe level. The new analysis confirms this conclusion.
Generally, the public and their representatives interpret results of
this type to mean that soil contamination at 6,500 /ig/g, or any
level above 6,450 /ug/g, is harmful. Drawing this interpretation
from the conservative analysis would be inappropriate and should
be avoided.
REFERENCES
1. Anderson, E., Browne, N., Duletsky, S. and Warn, T., "Develop-
ment of Statistical Distributions or Ranges of Standard Factors"
used in Exposure Assessments, U.S. EPA Contract No. 68-02-3510,
1984.
2. Binder, S., Sokal, D. and Maughan, D., "Estimating the Amount of
Soil Ingested by Young Children Through Tracer Elements," Un-
published Manuscript. Div. of Env. Hazards and Health Effects,
Center for Env. Health, Centers for Disease Control Public Health
Service, U.S. Dept. of Health and Human Services, Atlanta, GA.
3. Enfield, C.G., Carsel, R.F., Cohen, S.Z., Phan, T. and Walters,
P.M., "Approximating Pollutant Transport to Ground Water,"
Ground Water, 20, 1982, 711-22.
4. Ford, K.L. and Gurba, P., "Health Risk Assessments for Contam-
inated Soils," Proc. of the Fifth National Conference on Manage-
ment of Uncontrolled Hazardous Waste Sites, Washington, DC,
Nov. 1984, 230-731.
5. Kolbye, A.C., Mahaffey, K.R., Fiorino, J.A., Corneluissen, P.C.
and Jelinek, C.F., "Food Exposures to Lead," Environ. Health
Perspectives, May 1974, 65-74.
6. Layton, D.W., Hall, C.H., McKone, T.E., Nelson, M.A. and
Ricker, Y.E., "Demilitarization of Conventional Ordnance: Priori-
ties for Data Base Assessments of Environmental Contaminants,"
UCRL-53620 Draft prepared for USAMBRDL, Ft. Detrick, MD,
1985.
7. Rosenblatt, D.H. and Dacre, J.C., "Preliminary Pollutant Limit
Values for Human Health Effects," Environ. Sci. Techno/., 14,
1980, 778-84.
8. Rosenblatt, D.H., Dacre, J.C. and Cogley, D.R., "An Environ-
mental Fate Model Leading to Preliminary Pollutant Limit Values
for Human Health Effects" in R.A. Conway (ed.), Environmental
Risk Analysis for Chemicals, Van Nostrand Reinhold, New York,
NY, 1982, 475-505.
9. Small, M.J., "The Preliminary Pollutant Limit Value Approach:
Procedures and Data Base," U.S. Army Medical Bioengineering Re-
search and Development Laboratory, Ft. Detrick, MD, Tech. Re-
port 8210, 1984.
10. Tucker, W.A., Dose, E.V., Gensheimer, G.J., Hall, R.E., Pollman,
C.D. and Powell, D.H., "Evaluation of Critical Parameters Af-
fecting Contaminant Migration through Soils," Report AMXTH-
TE-TR-85030, prepared by Environmental Science and Engineering,
Inc. for U.S. Army Toxic and Hazardous Materials Agency, 1985.
CONTAMINATED SOIL TREATMENT 91
-------
Objective Quantification of Sampling Adequacy
and Soil Contaminant Levels Around
Point Sources Using Geostatistics
Jeffrey C. Myers
Geostat Systems International
Golden, Colorado
ABSTRACT
This paper describes the events and conclusions resulting from
a study focused on two lead smelters within the Dallas city limits.
These smelters were implicated as the sources of elevated con-
centrations of lead in the surrounding soils. The presence of high
lead is of considerable concern to nearby inhabitants.
This report focuses on the application of geostatistical tech-
niques to sampling around point sources. The application of geo-
statistical methods to smelter contamination had not been prev-
iously attempted. Many of the questions concerning applicability
of geostatistics, geometric and spatial structure of the contam-
inant plume, modeling of the area, sampling strategies and gen-
eralization to future sites have been resolved. Geostatistical tech-
niques have proven to be very useful in the resolution of these
questions.
INTRODUCTION
Increased environmental awareness dictates that virtually any
type of industrial "smokestack," either existing or planned, is
subject to public scrutiny. Coal-burning power plants and steel
mills are being blamed for acid rain problems. Other heavy metals
such as uranium that are emitted from coal-powered plants also
are causing concern. These emissions pose a threat to nearby
plant and animal life. Insights gained from modeling heavy metal
airborne contamination hopefully can improve models in these
other situations or provide a starting point for understanding
these unique problems.
The results and conclusions of the Dallas Lead Project have
provided analytical tools and valuable insights into the nature of
contamination at other locations. This information then can be
applied to future modeling projects.
This paper discusses the techniques applied to the modeling of
the Dallas project in detail as well as the insights that have been
gained by the authors. The U.S. EPA commissioned this study
as an experiment to determine applicability of geostatistical tech-
niques in an attempt to develop more objective procedures. The
following conclusions are reached in the paper:
• Geostatistical methods are very objective and effective in
modeling contamination around smelters. Two separate
smelters in the Dallas area have been studied. Both sites have
produced excellent results and nearly identical geostatistical
models indicating that other sites may follow a similar pattern.
• Geostatistical modeling has many advantages over traditional
estimation techniques. Its Best Linear Unbiased Estimator
(BLUE) property provides more objective answers to questions
that are difficult to provide otherwise. Results of studies such
as this one influence decisions in cleanup phases that affect
human health and safety and should be made based on the
best possible data. The results also will have significant legal
ramifications and are subject to close scrutiny in courts of law.
Thus the need for objectivity and reproducibility becomes in-
creasingly important as issues enter the courtroom.
• Modeling with regionalized variable techniques permits the
earth scientist to isolate different phenomena acting simul-
taneously within the same geographic area. Different patterns
of dispersion from the smelter can be recognized and isolated
from other sources. Trivial contributions from secondary
sources (e.g., auto emissions) can be identified. In this way
liability can be correctly assessed.
• Objective methods to determine sampling sufficiency easily can
be defined. Sufficient grid sizes to model and assess the area
with an acceptable level of confidence can be obtained with
minimum sampling. If too many samples already have been
taken, sampling can be reduced to determine sufficient levels
for obtaining both variograms and desired accuracy as a start-
ing point for future projects. The technique known as the com-
position of extension variances has been applied and discussed
with respect to overall precision.
The results can be applied to the design of future smelters;
they also provide a method of limiting liability in environmental
impact situations by isolating sources; finally, the results can be
used by the environmental and legal professions to further the
cause of public health and safety within a framework of cost con-
sciousness and objectiveness.
BACKGROUND: THE DALLAS LEAD
PROJECT
During the last quarter of 1980, a sampling team from the
University of Texas at Arlington, under contract with the U.S.
EPA's regional office in Dallas, Texas, collected samples of soil,
house dust, paint and tap water from a number of private resi-
dences and schools. These sample locations were situated near
and in the general vicinity of two lead smelters and one battery
plant in Dallas, Texas. In addition to the two smelter sites and
the battery plant, a fourth site identified as the Reference area
(REF) also was selected as a control area for sampling. The pur-
pose of this initial phase of the investigation was to determine
whether additional effort was warranted, i.e., did a problem
exist?
All four sampling sites were located within the Dallas city limits
and all except the REF site were geographically located in the
city's western section. Figure 1 shows the location of the two lead
smelters: RSR Corporation (RSR) and the Dixie Metal Company
(DMC). General Battery Corporation (GBC) is located near the
intersection of Bengal and Amelia Streets. The GBC site was not
analyzed in the geostatistical portion of the study, however. The
92
CONTAMINATED SOIL TREATMENT
-------
Figure 1
Location of the DMC & RSR Smelters and the Reference Area
REF site was selected for its similarity to the RSR, DMC and
GBC sites in automotive emissions, population and ethnic back-
ground.
SAMPLING
Data collected by the University of Texas at Arlington have
shown some soils in the immediate vicinity of the Dixie Metal
Company (DMC) and the RSR Corporation (RSR) to have lead
levels that exceed 1,000 ppm, believed to be the critical danger
level for exposure. Due to the presumed behavior of airborne
lead from smelting facilities such as these, the sampling area was
confined to a 2-mile radius around each site. To better evaluate
the degree of contamination of these two industrially exposed
areas, a site known as the Reference area (REF) was selected.
The sampling scheme, except for the total number of samples
collected in the reference area, was identical to that used at the
DMC and RSR sites. In addition, a 2-mile area surrounding a
third site, the General Battery Corporation (GBC), was sampled
using the same scheme.
SAMPLING AND GRID DESIGN
The information collected in sampling phase 1 (20 data points)
was used to construct a variogram for each of the three study
areas. The calculated and recommended sampling scheme to
identify the geographical extent and level of soil lead contamina-
tion for phase 2 of the project was as follows:
• The dimensions of the square grid should be 750 ft x 750 ft.
• The total area covered by the grid should be contingent upon
the minimum lead concentrations of concern.
• The grid should be oriented along the axis of the plume.
• The grid used in the REF area should be identical to that used
in the DMC and RSR areas.
• Soils within a radius of 100 ft of each grid intersection should
be sampled using procedures identical to those described in
phase 1.
• Six samples should be collected for each site to be composited
into a single sample.
• Duplicate samples should be collected at 5% of the grid loca-
tions.
• Samples should be separated into three "splits" at 5% of the
grid locations and sent to different laboratories.
During the summer of 1980, the U.S. EPA collected more than
3,000 individual core samples that were composited into more
than 500 grid samples throughout the three study areas. These
data are the foundation of the subsequent geostatistical analysis
and mapping portion of the project.
STRUCTURAL ANALYSIS
Variograms of the Ln-transformed soil sample values were
computed in each of the three areas. In each case, an average
variogram and four directional variograms were produced. The
four directional variograms have been calculated along the east-
west (0°), north-south (90°), northeast-southwest (45°) and
northwest-southeast (-45°) directions. For each directional vario-
gram, an angular tolerance window of 22.5 ° (total width of win-
dow) around the given direction was allowed. A distance lag sep-
aration of 800 ft also was used. For the average variogram, which
considers all points, the angular tolerance was 180°.
The results of the calculations were very good. Salient features
of the variograms include the following:
• All the variograms show some kind of continuity between
holes. They all tend to increase with distance until they reach
a sill value.
• Continuity in the DMC and RSR areas shows very similar pat-
terns when compared to the reference area. Modeled range
and sill values bear this out.
• As expected, the REF area shows a much lower sill and a lesser
degree of continuity.
VARIABLE LEAD
DM-LAS LEAD DHC
LOGARITHMIC
Figure 2
DMC Variograms of Ln-Transformed Lead Values for Phase 2
CONTAMINATED SOIL TREATMENT 93
-------
t i-coo Q I s i
LOGARITHnIC
Figure 3
RSR Variograms of Ln-Transformed Lead Values for Phase 2
Isotropic spherical models have been fitted to the experimental
average variogram curves. In all cases, they are the sum of a small
nugget effect (calculated from the duplicate samples) and one or
two spherical structures of varying ranges. The curves of the fitted
models are shown as a solid line in Figures 2 to 4.
^ LCAO - K' I
Finally, one anticipated problem in the calculation of the vario-
grams did not arise. This "non-problem" was the possible pres-
ence of a drift or a trend. A drift is defined in geostatistical
terms to be a systematic increase or decrease in the data along
given directions. This drift causes problems in the variogram.
Modeling such a structure is difficult and also causes problems in
the kriging process. Special forms of kriging (universal, general-
ized IRF) have been created just to handle this problem. Although
the possibility of a drift seemed reasonable, one did not arise.
MODELING THE AREAS: KRIGING
LEAD VALUES
The soil sample data were kriged in each of the three areas
using their respective variograms. The kriged area has been made
large enough to include three reference locations of geographic
interest. These locations have been used for map orientation.
The contour maps resulting from the kriged grid are shown in
Figures 5 to 7. The round symbol in the center represents the lead
smelter. The contour lines show the estimated soil lead concentra-
tion in ppm. In the DMC area (Fig. 5), a highly visible concen-
tration or plume has developed around and to the north and east
of the source. The large number of concentric contours encircling
the smelter shows a steep gradient of rapid change in a short dis-
tance between a low of 200 ppm outside and a high of 3,000 in-
side.
The effects of flooding by the Trinity River are well defined
(to the north) as a distinct area of low lead concentration which
can be seen along the flood plain of the river. On either side of
the river, the concentrations are noticeably higher. Several pos-
sible anomalous sample values also occur to the north of the river.
In at least one case a sample whose concentration was 10,400 ppm
contamination was found later to have been exposed to automo-
tive contamination.
In the RSR area, the smelter is located within the major con-
taminant plume that disperses rapidly east-west and more slowly
north-south (Fig. 6). The southern end of the plume also contains
SCO 2700 SJOO 7900 »»OO
I23OO
12300
T80O-
SKX>-
2700-
Figure 4
REF Variograms of Ln-Transformed Lead Values for Phase 2
»002700^00 TioO 9900
Figure 5
Isomap of Estimated Lead Values in the DMC Area
94 CONTAMINATED SOIL TREATMENT
-------
29OO S3OO 7TOO
1500 3OOO 45OO 60OO 75OO 90OO IOSOO
H20OO
IOIOO
Figure 6
Isomap of Estimated Lead Values in the RSR Area
a possibly anomalous value that has the effect of extending the
plume in this direction. This finding suggests that the true dis-
persion nature of the plume is primarily north and slightly west
from the smelter.
As expected, the REF area shows no systematic areas of lead
concentration (Fig. 7). Local highs and lows fluctuate randomly
and at considerably reduced levels overall from those seen in the
DMC and RSR sites. There is no evidence to suggest that this is
anything other than a background lead level map.
Accompanying each isopleth map is a second contour surface
that contains isovalues for the kriging standard deviation (Figs.
8 to 10). The values shown are the logarithmic kriging deviation
multiplied by a factor of ten (to facilitate plotting). The "swiss
cheese" effect noted here is common for cases such as this where
the sample grid spacing is a relatively large percentage of the
range of the variogram. Because the quality of estimation is solely
a function of geometry and the variogram, estimates are better as
they approach a sample point.
37OO 6IOO 85OO \O9OO
1300 3700 6100 6500 I09OO
Figure 7
Isomap of Estimated Lead Values in the REF Area
I05OO-
9000 -
6000
3OOO -
I50O 3OOO 450O 6OOO 75OO 9OOO K55OO
Figure 8
Isomap of the Estimation Variance in the DMC Area
IOIOO "TOO
RD.a
THE MKTRft TRACK
IOIOO
-6900
370O
500
IOIOO II7OO
Figure 9
Isomap of the Estimation Variance in the RSR Area
/? °
y * iLLiNoia AVC a
2IOO
6500 IO9OO
•ucxtrr AVE
-8500
JNNTV»LE IT.
-5300
"2IOO
.-IIOO
85OO IO9OO
Figure 10
Isomap of the Estimation Variance in the REF Area
CONTAMINATED SOIL TREATMENT 95
-------
SAMPLING SENSITIVITY: REDUCED
SAMPLE SIZES
The results of this first mapping stage proved that kriging and
geostatistical techniques apply very well to soil data and point
source contamination. The second stage of this study addressed
the concerns of the U.S. EPA and others as to the adequacy of
sampling: have enough samples been taken to assure reasonable
level of certainty? Also, what changes might be made in future
studies of similar phenomena? These questions will not be ad-
dressed.
The techniques used to answer these questions will be the analy-
sis of extension variance and the use of reduced data sets to de-
termine the effects on the variogram. The results of the prelim-
inary geostatistical analysis provided models of the spatial con-
tinuity structure in each area through their respective variogram
models. These models will now be assumed to be the "true"
underlying spatial structure. The data sets of 206 and 208 samples
in the DMC and RSR sites respectively will be taken as the "uni-
verse" of data.
The procedure, then, is to reduce the level of sampling in the
areas of study and to recompute the variograms using a smaller
data set. Multiple data sets will be produced and analyzed at each
level of reduced sampling to minimize any possible selection bias.
The new parameters of the variograms then will be compared to
the "true" values to see how accurately the subset can predict the
"real" structure. This information from small data sets can be
used to design future two-phase sampling schemes.
The two smelter areas have been divided into equal sized grid
cells from which one sample per cell has been taken at random.
This random selection has been done by computer to eliminate
human bias. Since the "true" variograms are isotropic (equal
zone of influence in all directions), the sampling grids were de-
signed to be isotropic (square) so that this relationship can be
honored.
From the variograms, it is possible to quantify the variances
of polygonal (block) extension errors and then combine them,
since they are the variances of independent variables. This has
been used as a check on total precision.
Small sample subsets were selected in both the DMC and RSR
areas on grids of 11 x 11, 7 x 7, 5 x 5 and 4x4. These represent
sample sizes of approximately 100, 50, 25 and 16 respectively.
The 11x11 grid was selected as the first grid size with the idea
that subsequent sizes would be enlarged or reduced as needed de-
pending upon the outcome of this first attempt. Ten sample sub-
sets were taken using the same ten random seeds in both smelter
areas. The ten new data sets in each area then were variogram-
med, their sample locations were plotted and their histograms
were plotted.
The 20 variograms then were modeled. The results showed re-
markable similarity to the original data sets both visually and in
variogram parameter comparison. The results of the 11 x 11 grid
data were so successful that the data set was again halved by us-
ing a 7 x 7 grid. Excellent variograms were produced although
more variation was noticed in the model parameters. The RSR
models tended to have a lower sill than the true one, but the
ranges were generally the same as for the true model. The DMC
models showed on the average the same sill but generally showed
a shorter range. Because the object of this exercise was to de-
termine sampling grid sizes from small initial data sets, the range
was the most important parameter. For this grid size, both areas
provided accurate ranges with the DMC models being a little
more conservative.
Again the data set was halved by going to a 5 x 5 grid. At this
level of sampling, the directional variograms had deteriorated to
such a point that they were valueless. A lack of data was the
cause. However, the average variogram still could be used to
advantage as good structures still were found. Sill variance
showed a much greater fluctuation, with a number of low-valued
variances where sampling failed to pick up high values near the
smelters. Ranges generally were shorter than the "true" ranges.
This corresponded to the shorter range found in phase 1 of the
sampling where only 20 samples were taken. A few very long
range values did appear, but they resulted mainly from an attempt
to fit a good model at short distances where the variance rose
quickly for a time and then slowly crept toward the sill.
The final level of sampling was produced on a 4 x 4 grid. Again
directional variograms were of no value. The sill values in the
RSR area showed a large fluctuation, but the ranges were quite
stable and not too far from the "true" parameters. In the DMC
area, sills were generally near or above the "true" parameters
and the ranges were at or below those expected. Overall, this grid
produced about the same quality estimates as the 5x5 grid and
showed slightly better stability of ranges.
DMC AREA
"sooo
2
t-
m
til
000
0 497 H 14 19
NUMBER OF GRID CELLS
Figure 11
Global Precision vs. Sampling Density in the DMC Area
RSR AREA
I
2000.
497 II 14 19
NUMBER OFGRIDCELLS
Figure 12
Global Precision vs. Sampling Density in the RSR Area
96 CONTAMINATED SOIL TREATMENT
-------
GLOBAL PRECISION
An estimate of global precision provides a measure of the cer-
tainty with which the global mean is known. This value is anal-
ogous to the "standard error of the mean" in classical statistics.
In the geostatistical sense, obtaining a high value for precision of
the global mean suggests that local estimates also will remain
quite accurate. This condition is not strictly true in all local cases
but offers a starting point for decision making.
The precision for symmetric grids of sides 4 through 15 have
been calculated and have been plotted in Figures 11 and 12. The
14 x 14 and 15 x 15 intervals represent the precision when con-
sidering a full sample set of 206 or 208. The graphs show that in
terms of global estimates the sampling has reached the point of
diminishing returns. Both figures indicate that further sampling is
unnecessary. What also is critical here is that variograms have
been produced for even the smallest of data sets.
CONCLUSIONS
Geostatistical techniques have proven to be an extremely valu-
able tool in the analysis of pollutant contamination. A strongly
identifiable plume of lead is associated with smelter source. No
such structure is found in the REF area. The plume is marked by
a central core of concentrated high values, generally within 1 mile
of the source, with a continual decrease in lead values to the
borders of the areas. It appears certain that these two smelters
were indeed the source of the contamination and that automo-
bile emissions are of negligible contribution.
The smelter areas have been adequately sampled with the 800-ft
grid. The range of the variograms was five to seven times this
sampling interval and proved quite sufficient to model the spatial
continuity of the lead values.
Global estimates of precision for the second stage sampling
campaign are very conservative and indicate that additional
sampling would have been of little value. Variogram parameters
remain surprisingly constant down to extremely small levels of
sampling.
Two-stage sampling, though already in regular use in the en-
vironmental industry, can be greatly enhanced by the use of geo-
statistical technique. Optimal sampling geometries can be devel-
oped for a variety of situations and constraints. As a result, cost
benefit curves can be produced to select development drilling
grids. These analyses will provide a measure of confidence and
objectivity.
CONTAMINATED SOIL TREATMENT 97
-------
Innovative Application of Chemical Engineering
Technologies for Hazardous Waste Treatment
Robert D. Allan
Michael L. Foster
IT Corporation
Knoxville, Tennessee
ABSTRACT
Regulatory and public pressure are creating an increasing de-
mand for more innovative methods of hazardous waste manage-
ment. In response to this need, some proven technologies are
being adapted to provide alternatives to the more classical remed-
iation techniques in certain applications. This paper reviews the
site-specific parameters that need to be considered before select-
ing one of these alternatives. The operational details of the can-
didate technologies are discussed, and specific examples of appli-
cations are given.
INTRODUCTION
Application of the RCRA and CERCLA statutes often creates
the need for corrective action, particularly with regard to ground-
water treatment and surface impoundment closure. Of particular
interest to industries are the on-site treatment alternatives due to
the cost and liability factors of land disposal and the lack of con-
venient incineration options.
This paper concentrates on proven treatment solutions avail-
able to meet the needs of corrective action situations. Particular
emphasis is placed on those technologies offering the opportun-
ity for cost-effective, on-site treatment or waste volume reduc-
tion. Included is a review of technologies utilized in mobile
and/or in situ applications.
PROBLEM DEFINITION
Unfortunately, no single treatment technology is going to be
universally applicable for all corrective action situations. A
thorough evaluation of the site-specific parameters involved is
necessary in order to choose the most technically suitable and
cost-effective treatment technology for the job. Because major
corrective action programs, especially those involving aquifer
restorations, are likely to take several years to accomplish, the
investment of time and effort to properly evaluate the situation
initially will more than pay for itself in the long run.
Among the major site-specific parameters that should be de-
fined are those discussed below.
Contaminant Composition and
Concentration
An accurate analysis of the site contamination is important.
The contamination could be a single component from an under-
ground solvent storage tank leak or a wide variety of compon-
ents from a disposal site leachate. The level of contamination
could range from low >jg/l to several hundred mg/1. The treat-
ment costs of technologies such as carbon adsorption and oxida-
tion are dependent on concentration and type of compound,
while costs for technologies such as air stripping are more closely
related to the treatment rate and desired removal efficiency.
Desired Cleanup Criteria and
Effluent Disposition
Stringent cleanup criteria (e.g., drinking water standards for
aqueous streams) will eliminate some technologies from consid-
eration immediately, but it must be recognized that different
levels of cleanup are likely to be required for many remedial pro-
grams. The ultimate disposition of the effluent for aqueous treat-
ment will have significant bearing on the required effluent qual-
ity. If the treated stream can be discharged to the sewer for
further treatment by a municipal waste facility (POTW), effluent
guidelines are likely to be less strict than if the stream is to be re-
charged to an aquifer or discharged directly to a surface water
source under an NPDES permit.
Volume to be Treated
Treatment volumes for remedial programs can vary widely. A
small volume may be treated best by a technology with low capital
costs and high operating costs, while economic remediation of a
larger magnitude is more likely to be accomplished utilizing a
technology with low operating costs. An evaluation of treatment
options should include consideration of leasing equipment to
treat small volume sources.
Fate of Treatment Byproducts
Local requirements for disposition of treatment byproducts
will have a major impact on the economics and viability of vari-
ous treatment technologies. Air stripping can be an extremely
cost-effective technology if the organic-laden air can be emitted
directly to the environment; but strict air emissions standards may
require that air stripping be followed by a vapor-phase adsorp-
tion system, significantly increasing treatment costs. Similar con-
siderations exist with the generation of potentially toxic by-
products from chemical oxidation processes and the possible need
for nutrient removal (phosphorus and nitrogen) from the effluent
of biological oxidation processes. The economics of activated
carbon adsorption (especially on a throwaway carbon basis)
can be affected by the accessibility of an approved disposal site
for the contaminated carbon.
Utilities Availability and Cost
The selection of certain technologies will be influenced by the
availability of utilities such as steam at the treatment site. Like-
wise, some processes such as UV-catalyzed oxidation can require
a substantial source of electrical power. The energy-intensive
processes are likely to be more attractive in the southern part of
the country where electrical costs are approximately $0.05/kWh
than on the West Coast where costs can approach $0.15/kWh.
A clear definition of the goals of a remedial action program
along with an understanding of the above site-specific parameters
98
ON-SITE TREATMENT
-------
and the technical limitations of the technologies under considera-
tion will allow selection of the most cost-effective treatment
option.
TECHNICAL EVALUATION
A number of technologies are potentially applicable to remed-
iation programs. Some are best suited for specific applications
and others may be eliminated from widespread consideration be-
cause of the inability to meet stringent discharge standards or
other technical limitations. In this section, the technologies with
the greatest potential in remedial actions are identified and spe-
cific applications are discussed.
Incineration
Incineration, or thermal destruction, is an effective way to de-
stroy organic liquids and sludges and organics in a solid matrix.
It can be used for a relatively broad waste profile depending on
the type of incineration system utilized. Incineration is defined
generally as a controlled, high-temperature, oxidation reaction.
The end products of the incineration reaction are generally car-
bon dioxide and water from hydrocarbon wastes; but industrial
and hazardous wastes also can generate hydrogen chloride, sulfur
dioxide and less desirable combustion byproducts depending on
the nature of the wastes, the operating features of the specific in-
cineration system and the compatibility of the waste profile and
incineration system.
The technology of high-temperature incineration is in wide-
spread use, both at private and commercial installations. It often
is considered in remedial action alternative evaluations, but the
waste volume must be substantial to justify the expense of permit-
ting and building an on-site unit. For more limited volumes, com-
mercial incineration facilities offer the advantage of destruction
of the wastes involved.
Thermal Desorption
Thermal desorption removes organic chemicals from soil and
solids by heating the soils to temperatures sufficient to convert the
organics to vapors and then holding the soil at that elevated temp-
erature long enough for the removal process to occur. Desorbed
organic and water vapors can be recovered by solvent scrubbing
or condensed and collected in an accumulator where the water
and organics are separated into their respective phases. Water
can be treated with activated carbon before discharge, while the
organics phase can be sent to a commercial incinerator or treated
chemically or photolytically.
Thermal desorption is effective for any organic contaminant,
with more volatile compounds such as solvents, being more read-
ily removed. The types of compounds for which this technology
is expected to be applicable range from chlorinated solvents and
gasoline/fuel components to phenols, aromatic hydrocarbons,
PCBs and dioxin. The concentration of the contaminant does
not have a significant influence on the applicability. Although
the soil characteristics will impact the operating rate and process
design requirements, thermal desorption is amenable to a wide
range of soil/site conditions.
IT has developed and demonstrated thermal desorption in
bench scale and pilot scale equipment. Laboratory testing has
been performed on a variety of soil types containing such chem-
icals as Agent Orange (containing chlorinated dioxins), penta-
chlorophenol and creosote materials (PNAs). Pilot scale work,
conducted as part of the Air Force's Environmental Restora-
tion Program, will be reported at an upcoming American Chem-
ical Society meeting.1 Briefly, a pilot plant was assembled to treat
up to 100 Ib/hr of soil contaminated with 2,3,7,8 TCDD from
Agent Orange. The desorber was operated at 460 to 560 °C. The
vaporized contaminants were captured by a solvent scrubber and
destroyed in a UV photolysis reactor. Soil analysis showed a
greater than 99% reduction in the TCDD concentration and resid-
ual soil concentrations of less than 1 ppb.
The thermal desorption process offers several advantages, as
compared to thermal oxidation, for decontamination of organ-
ically contaminated soil. These advantages include lower temper-
ature operation resulting in less energy consumption and simpler
design criteria, relative ease of regulatory approval due to classif-
ication as a physical/chemical treatment process and the ability
to easily skid or truck mount the equipment for mobile treat-
ment.
UV Photolysis
Ultraviolet light can catalyze normal chemical oxidations by
creating a supply of "free radical" species. IT has applied this
technology to the destruction of dioxin at the Syntex Agribusi-
ness plant in Springfield, Missouri.2 The dioxin destroyed was in
a distillation residue which was a byproduct of previous pro-
duction of hexachlorophene by the Northeast Pharmaceutical
Company. The process involved extraction of dioxin from the
residue using a common solvent and then destruction of the
dioxin in reactors equipped with 10-kW, high intensity ultra-
violet lamps. Approximately 13 Ib of dioxin were destroyed using
this process.
Chemical/Biological Oxidation
Oxidation technologies use chemical or biological means to
completely oxidize organic contaminants to carbon dioxide and
water, or partially oxidize them to non-toxic intermediates. Most
of the oxidation options are best suited to specific applications.
Chemical oxidation accomplishes detoxification of waste
streams through the addition of chemical oxidation agents. Near-
ly all chemical oxidation processes use either chlorine or oxygen
as the oxidizing agent.
Chlorine oxidation is of questionable value in most remedial
applications because of the difficulty in assessing the impact of
potentially toxic chlorinated byproducts. Oxygen itself has very
limited effectiveness but may be of use in the form of ozone or
hydrogen peroxide.
As an example of an application of chemical oxidation, IT has
utilized hydrogen peroxide to successfully treat a sizeable spill of
formaldehyde. The spill resulted from a leak in an underground
pipeline and had contaminated both the subsurface soil and the
groundwater. Because the spill contamination was below a ship-
ping and receiving area, common cleanup techniques involving
excavation would have caused an expensive disruption of plant
operation. Instead the formaldehyde was chemically oxidized in
situ by injecting a hydrogen peroxide solution into the contam-
inated zones. Formaldehyde contamination levels in the soil, orig-
inally as high as 8% (80,000 ppm), were reduced to 15 ppm or less
by this method.
Biological oxidation uses active microorganisms to biodegrade
organics to acceptable forms. The two major forms of biological
treatment are aerobic (which produces carbon dioxide and water)
and anaerobic (which produces carbon dioxide and methane).
Biological treatment is getting increased attention as a remedial
alternative because of its potential for in situ treatment. Bio-
reclamation basically is the use of indigenous soil bacteria to de-
grade organic contaminants. Nutrients, such as oxygen, and spe-
cific biological cultures can be added to enhance the degradation.
Vandalism of a rail car parked on a railroad siding in Ukiah,
California resulted in an opportunity for the application of bio-
logical oxidation to an emergency situation. Twenty thousand
gallons of a strong formaldehyde solution drained from the rail
ON-SITE TREATMENT 99
-------
car into and through the rail-track ballast and into the associated
drainage system. Initial efforts focused on containing and isolat-
ing the spill.
Once the spill had been controlled, the more difficult task of
decontamination of the affected area was addressed. The rail-
way company was vary concerned that normal excavation pro-
cedures would disrupt rail service on the adjacent mainline track.
Thus, IT developed an alternative approach which initially in-
volved treatment of the area with alkaline hydrogen peroxide.
This chemical oxidation process was successful in reducing the
formaldehyde concentrations to a level where biological oxida-
tion could be used.
The biological process made use of a portable aeration tank, a
spray system and the railroad ballast itself. Liquid was pumped
from a sump dug next to the ballast, passed through the aeration
tank and sprayed back onto the ballast. The system was inocu-
lated with a specially cultured microorganism that degraded
formaldehyde but was free of pathogens. The ballast, being com-
posed of coarse rock, supported the growth of the biological med-
ium in a manner similar to a trickling filter. Nutrients were added
to the system as needed. Over a 25-day period, the system suc-
cessfully reduced the formaldehyde concentration from several
hundred ppm to less than 1 ppm.'
Carbon Adsorption
The use of activated carbon adsorption for aquifer restora-
tion programs has drawn widespread attention. Several literature
sources point to its ability to achieve exceptionally good efflu-
ent quality, and the U.S. EPA has endorsed it as the preferred
treatment method for meeting drinking water standards. This
section discusses three basic ways in which carbon can be used:
• Throwaway carbon
• Thermal regeneration
• Nondestructive regeneration
Throwaway Carbon
One way to consistently ensure good effluent quality is to use
activated carbon absorption on a once-through carbon basis.
Virgin carbon is capable of removing a broad range of organic
contaminants to low ug/1 levels. A once-through carbon adsorp-
tion system is easy to operate, requires a minimum of operator
attention and its capital cost requirements are relatively low.
Unfortunately, carbon replacement costs associated with once-
through carbon adsorption systems are very high. The large treat-
ment volumes and/or high concentrations usually associated with
remedial programs result in a high carbon consumption rate. In
addition, hazardous substances, when loaded onto activated car-
bon, make the carbon a hazardous waste, requiring disposal in
an approved hazardous waste facility.
Thermal Regeneration
The most common regeneration technique for activated car-
bon is thermal oxidation, usually accomplished in a multiple
hearth, fluidized-bed or rotary-kiln furnace. A thermal regenera-
tion unit can be built at the treatment site, but the level of carbon
consumption associated with most remedial programs usually
makes it more economical to utilize an off-site thermal regenera-
tion service.
The advantages of activated carbon adsorption with a thermal
regeneration service are low capital requirement and ease of oper-
ation, as with throwaway carbon, but the disadvantages of
thermal regeneration are numerous. Using a thermal regenera-
tion service basically amounts to "renting" the carbon. How-
ever, in many cases, the cost per pound of carbon for thermal
regeneration is only slightly less than that for purchasing virgin
carbon. Additionally, carbon losses associated with thermal re--
generation are typically 5 to 10% per cycle due to handling losses
and carbon that is destroyed during the regeneration process.
However, thermally regenerated carbon can be expected to have
different performance characteristics and reduced removal effic-
iency as compared to virgin carbon.
A major problem associated with thermal regeneration is that
many hazardous substances cannot be thermally regenerated for
one reason or another. Because the hazardous substances are de-
sorbed from the carbon and destroyed in an afterburner during
thermal regeneration, air permits for thermal regeneration sys-
tems often contain restrictions on the types of materials that may
be accepted. Consequently, thermal regenerators often refuse to
deal with carbon that has been used to adsorb toxic materials
such as PCBs.
Nondestructive Regeneration
There are three basic ways that granular activated carbon can
be nondestructively regenerated:
• Using pH shift for weak acids or bases
• Using steam for volatile organics
• Using a solvent for a wider variety of organics
Steam regeneration, which would be applicable in many remed-
ial situations, is accomplished by passing steam through a spent
adsorber to a condenser and then to a decanter where the con-
densate and immiscible organics are separated. Carbon losses
associated with nondestructive regeneration are significantly less
than those associated with thermal regeneration. The carbon is
not physically altered by the nondestructive regeneration process,
and if the carbon adsorbers are made of the proper material, the
regeneration can be accomplished in situ, eliminating carbon
handling losses.
The process is relatively easy to operate and the nondestruc-
tive nature of the process allows organic recovery, if desired.
Even if recovery is not desired, the disposal requirements are re-
duced from several thousand pounds of contaminated carbon to a
few gallons of organic material.
IT has evaluated the applicability of carbon adsorption systems
with nondestructive regeneration in a variety of remedial situa-
tions. A recent feasibility study compared air stripping, air strip-
ping with vapor-phase adsorption, adsorption using virgin car-
bon, adsorption using steam-regenerated carbon, steam strip-
ping and chemical oxidation using UV peroxide for an aquifer
restoration. The critical design parameters were: a contaminated
groundwater concentration of 30 mg/1 or halocarbons includ-
ing EDB (1,2-dibromoethane); a maximum flow of 200 gal/min
and an average flow of 100 gal/min; and a mandated effluent
concentration of ^1 ug/1 of each halocarbon present.'
Details on the result of this evaluation can be found in the
referenced paper. Based on these design parameters, the costs of
air stripping and steam-regenerated carbon adsorption were
basically equal and fell in the range of $300-400/day of system
operation. Subsequent laboratory work, using samples of the
actual groundwater, verified that carbon adsorption with steam
regeneration could achieve acceptable effluent quality and main-
tain a stable working adsorption capacity.
Laboratory evaluation of the air stripping option was post-
poned until the issue of exhausting the vaporized organics to
the atmosphere could be resolved. This issue is frequently a re-
striction in the application of air stripping.
Air Stripping
Air stripping is a contaminant removal technique based on con-
centration differentials between a liquid phase and a contacting
gas phase. As air is contacted with a wastewater stream in a strip-
ping tower, the concentration differential drives the organic con-
100 ON-SITE TREATMENT
-------
taminant from the liquid to the gas phase. The major advantage
of air stripping is its low overall treatment cost. Both capital and
operating costs requirements are low compared to most other
technologies.
The key to air shipping's low overall treatment cost is the
assumption that it can stand alone as a treatment technology. In
many cases, however, air emission standards will require that air
stripping be used in conjunction with a vapor-phase adsorption
unit, significantly affecting the cost-effectiveness of the technol-
ogy.
Air stripping is most applicable to compounds of a volatile
nature with relatively low solubility in water. Chlorinated organ-
ics such as trichloroethylene, 1,1,1 trichloroethane and aromatic
compounds such as benzene and toluene are good examples of
chemicals that can be successfully removed by air stripping. The
overall applicability of the technology can be expanded to com-
pounds with lower Henry's law constants by preheating the water
prior to its entering the stripping column.
IT has evaluated air stripping a number of times in treatability
evaluations for a wide variety of clients. In a study conducted for
the American Petroleum Institute, IT engineers conducted side-
by-side, bench-scale evaluations of air stripping and carbon ad-
sorption with steam regeneration to remove dissolved gasoline
components from groundwater.5 Air stripping effectively re-
moved the aromatic gasoline components from groundwater even
at the normal groundwater temperature of SOT. However, as
would be expected, air stripping was ineffective in removing more
soluble components of gasoline such as tert-butyl alcohol.
MOBILE TREATMENT
As the expense and regulatory limitations of waste disposal be-
come increasingly important factors, the innovative option of
mobile, on-site treatment has become an important remedial
alternative. Many of the technologies described above can be, and
have been, adapted to mobile applications.
The U.S. EPA and a limited number of commercial firms cur-
rently operate mobile incineration systems. A number of aque-
ous treatment technologies have been mobilized for on-site re-
mediation including activated carbon, air stripping and ion ex-
change. In addition, mobile equipment is available to dewater
sludges and reduce the volume of material that ultimately must be
sent to disposal or destruction facilities. These options include
mobile centrifuge and mobile belt filter press equipment.
CONCLUSIONS
A number of technologies exist and have been proven in re-
medial action situations. Experienced, knowledgeable personnel
in the hazardous waste field are able to efficiently evaluate
options for corrective action programs, but, in order to do so
properly, they must evaluate the site-specific parameters involved
and fully understand the constraints, options and goals of the
program.
REFERENCES
1. Helsel, R., Alperin, E., Geisler, T., Groen, A., Fox, R., Stoddart,
T. and Williams, H., "Technology Demonstration of a Desorption/
UV Photolysis Process for Decontaminating Soils Containing Herbi-
cide Orange," To be presented before the Division of Environmental
Chemistry, American Chemical Society, April 1986.
2. "Destroying Dioxin: A Unique Approach," Waste Age, Oct. 1980,
60-63.
3. Sikes, D.J., McCulloch, M.N. and Blackburn, J.W., "The Contain-
ment and Mitigation of a Formaldehyde Rail Car Spill Using Novel
Chemical and Biological In Situ Treatment Techniques," Proc. of
the Hazardous Materials Spill Conference, Nashville, TN, Apr. 1984,
38-44.
4. Parmeli, C.S., Allan, R.D. and Mehran, M., "Steam-Regenerated
Activated Carbon: An Emission-Free, Cost-Effective Groundwater
Treatment Process," Paper presented at American Institute of Chem-
ical Engineers Annual Meeting, Chicago, IL, Nov. 1985.
5. Parmeli, C.S. and Allan, R.D., "Treatment Technology for Removal
of Dissolved Gasoline Components from Groundwater," Third Na-
tional Symp. and Exposition on Aquifer Restoration and Ground-
water Monitoring, Columbus, OH, May 1983.
ON-SITE TREATMENT 101
-------
Field Studies of In Situ Extraction
and Soil-Based Microbial Treatment
of an Industrial Sludge Lagoon
David S. Kosson
Erik A. Dienemann
Robert C. Ahlert, Ph.D., P.E.
Rutgers, The State University
Department of Chemical and Biochemical Engineering
Piscataway, New Jersey
ABSTRACT
Over a period in excess of 10 years, several industrial sludges
were disposed of by landfilling in a surface impoundment. The
resulting lagoon contains more than 30,000 yd3 of sludge.
Leachate from the sludges has impacted the local groundwater,
thus requiring remediation. Laboratory studies indicated that in
situ extraction with aqueous sodium hydroxide is a viable method
for removal of organic materials from the sludges. Treatment of
recovered extract was accomplished through use of a soil-based
microbial treatment system. Sequential aerobic and anaerobic
biodegradation in the laboratory reduced TOC by greater than
99.5%.
A field, pilot-scale treatment system has been designed, con-
structed and operated to demonstrate feasibility of the proposed
renovation program. Field results indicate that exhaustive extrac-
tion of the sludges is possible. Up to a 50-fold increase in the rate of
removal of contaminants over natural processes can be achieved.
Biodegradation treatment efficiencies in excess of 95°'o have been
demonstrated.
INTRODUCTION
Over a decade or more, several sludges were disposed of by
landfilling in a surface impoundment. During this period of
operation, the composition and rate of deposition of sludges
varied. The resulting lagoon contains more than 30,000 yd1 of
sludge. The principal sludges in the lagoon are primary (lime neu-
tralized) and secondary (biological) sludges from treatment of ef-
fluent from diverse chemical manufacturing operations. The
sludges range from solid to gelatinous in physical state and are
layered in the lagoon. Leachate from the sludges has impacted the
local groundwater.
Cleanup of the lagoon is viewed a,s two interrelated problems.
The first problem is the removal of contaminants from the lagoon
without major excavation. The second problem is treatment of
the stream containing the stripped contaminants, including both
organic and inorganic species. To achieve these goals, a process
corsisting of in situ extraction of the sludges followed by on-sile
treatment of recovered extract has been developed.
Laboratory results have indicated that exhaustive leaching of
the sludges with an alkaline aqueous extractant (sodium hydrox-
ide) is possible.1 Natural leachate and the sludge extracts can be
treated effectively for total organic carbon (TOC) removal
through use of a soil-based aerobic/anaerobic microbial treat-
ment process employing an acclimated mixed microbial popula-
tion in a soil matrix.2 Aerobic biodegradation occurs near the soil
surface where oxygen is available. Anaerobic biodegradation oc-
curs at greater depths. Conversion of organic solutes to carbon
dioxide and methane is the ultimate result. Treatment efficiencies
in excess of 99% were attained.'
Preliminary design estimates for implementation of these pro-
cesses indicated that renovation of the lagoon could be ac-
celerated to a period of less than 5 years with substantial cost
reductions compared to traditional treatment technology. Thus, a
field pilot plant was designed, constructed and operated for
demonstration of the process.
PILOT-PLANT DESIGN
The pilot system consisted of several sequential process steps
(Fig. 1). The first process step is the extraction of sludges present
in a representative section of the lagoon. Sodium hydroxide solu-
tion is mixed batch-wise in a 200 gal process tank (Tank 1). This
solution is applied to sludges present in the Extraction Bed. The
solution either can be applied to the surface of the Extraction Bed
through perforated pipe or injected into the sludges through six
well-points. Extract is recovered from the Extraction Bed through
Wells 1 and 2. Extract removed from each well is pumped through
a basket strainer and cartridge filter and into a 1,000 gal storage
tank (Tank 2).
The second process step is adjustment of pH, dilution if
necessary and addition of nutrients to extract stored in Tank 2.
This process occurs as a continuous process in Tank 3 which is a
200 gal process tank, divided into four equal sections by two
overflow baffles and one underflow baffle. Extract is pumped
from Tank 2 to the first chamber of Tank 3. Carbon dioxide is
bubbled through the extract to adjust the pH to between 7.0 and
7.5. Extract passes under an underflow baffle to the second
chamber of Tank 3. In the second chamber, floe created by the
recarbonation process is allowed to settle. Clarified extract passes
over an overflow baffle into the third chamber of Tank 3. In the
third chamber, dilution with recycle or potable water and addi-
tion of nutrients occurs. Finally, extract underflows into the
fourth chamber from which it overflows and is applied to the sur-
face of the Treatment Bed.
The third process step is treatment of the modified extract (ef-
fluent from Tank 3) through an aerobic/anaerobic soil-based
microbial treatment process. The treatment process consists of a
soil bed in which an aerobic microbial population is maintained in
the upper region and an anaerobic microbial population is main-
tained in the lower region.1 Extract applied to the surface of the
Treatment Bed percolates through the soil bed where it is
biodegraded. Effluent from the Treatment Bed is recovered
through Well 3. Recovered effluent is pumped to a 500 gal storage
tank (Tank 4) from which it can be recycled onto the Treatment
Bed, recycled to Tank 1 or discharged.
The entire process was designed to be sufficiently flexible to
compensate for fluctuations in either the extraction or treatment
processes as well as weather conditions. Ranges of expected condi-
tions and flow rates for each process step are presented in Table 1.
102 ON-SITE TRLATMhNT
-------
* BLEED
-------
WEU.2
WCU.3
- n
»*
0^
•T
1.
—
.OT771
"
. ,
\A
SUU06E
^
VWEL
OS
or^fr VCNTdWD
rxwii *
IS
ii l
L"
^
SOIL
ri
T.6AS VEKT
/O'PVC)
i J
1
r
,
vj
--HJ-Y-"
•r^se
^-— GRA
u-^w
^
^1C ^•r°00
DETAIL A
JIIIUHVilllllllllll [&&
KOTO.** LAC«» ..J[J?T^~
T
1
i: O
Figure 2
Extraction and Treatment Bed Construction
placed in the center. An additional 2 ft of sand were backfilled
around the well. Two 1-in. methane vents were installed on top of
the sand, and the remainder of the bed was backfilled with 5 ft of
soil mixture.
The soil mixture used to backfill the Treatment Bed consisted
of local top soil, sand and granular activated carbon IS: 1:1 by
volume, respectively. The soil mixture was mixed on-site by a
bucket loader prior to backfilling.
Modified extract is applied to the surface of the Treatment Bed
through two 0.5-in. perforated PVC pipes. Two water level sen-
sors are located on the surface of the bed to prevent flooding. Ef-
fluent from the process is collected through Well 3 which contains
a stainless steep deep well pump, a low water level sensor and a
high water level sensor. Water collected from Well 3 is pumped
directly to Tank 4.
Process Area Design
The Process Area consists of a 30 x 35-ft compound located
orr top of shale fill approximately 100 ft east of the Extraction and
Treatment Beds at an elevation of approximately 10 ft above the
surface of the lagoon (Figs. 3 and 4). Tanks 1, 2, 3 and 4, a
storage shed for CO2 cylinders and nutrient solution, electrical
power supply and controls, a wash basin and safety shower are
located within the compound.
Interconnections between tanks and the Extraction and Treat-
ment Beds consist of PVC solvent-cemented pipe. All piping,
where possible, is buried in shallow trenches approximately 1 ft
deep and backfilled with top soil. The surface of the compound is
covered with roadstone.
Process Controls
The process controls consist of four operationally independent
controllers. Three of the controls are identical units, i.e., one
pump controller each for Wells I, 2 and 3. The well pump con-
trols are a high water level sensor and a low water level sensor
(float switches) in each well. Liquid rising to the level of the high
water level sensor activates the pump in that well. When the liquid
level falls below the level of the low water level sensor, the pump
is deactivated. The actual operating time of each pump is ac-
cumulated on a timer located on the control panel.
104 ON-SITE TREATMENT
-------
TOP OF BERM
PROCESS
8
oo
(VI
•II
•TO UTILITIES
(POWER AND WATER)
"i-WELL
TO DISCHARGE
EXISTING ROAD
LAGOON SURFACE
GRADE AND LEVEL WITHIN PERIMETER
CHAIN LINK FENCE
I' DEEP TRENCH FOR PIPING
(TO BE BACKFILLED)
T
3
BED
BED
TREATMENT
BED
EXTRACTION
BED
Figure 3
Pilot Plant Layout
The fourth controller controls operation of pumps, valves and
sensors associated with Tanks 2, 3 and 4 and sensors on the sur-
face of the Treatment Bed. Tank 1 is operated manually, only.
The primary function of the fourth controller is to regulate ap-
plication of influent to the Treatment Bed. A master 24 hr/7 day
timer regulates operation cycles for application. When set on
automatic, the controller will activate whichever of the Tank 2, 3
and 4 valve and pump combinations are set on automatic during
the preset cycle. These combinations include:
• The actuation valve and pump on Tank 2 (extract feed to Tank
3 or Tank 1)
• The actuation valve and pump on Tank 4 (recycle to the third
chamber of Tank 3, Tank 1 or discharge)
• The valve controlling potable water for dilution in the third
chamber of Tank 3
• The pump for addition of nutrients to the third chamber of
Tank3
• The valve controlling recarbonation in the first chamber of
TankS
• The valve controlling floe removal from Tank 3
LINK /=jr/vcr
r—
n
•v i
v » i . .e
*
T*
-n
A*
r~*
T
CAT*1
ti-L£ D
\. Sfi^S ^
r«e ivcr* fo «
(SHALE)
Figure 4
South View of Pilot Plant
SLUD6£^
TREATMENT
AMD EXTAAcr/0/J
BEOS
PILIN&~
ON-SITE TREATMENT 105
-------
The system will be shut off when set on automatic when:
• A sensor on the surface of the Treatment Bed indicates
flooding
• A sensor in Tank 2 or 4 indicates the tank is empty, if the valve/
pump combination for that tank is set on automatic
• An application cycle is completed
In addition, each pump and valve combination can be activated
or deactivated manually. The system includes accumulating
timers to monitor operating time for each pump in the system. A
main power disconnect is provided for this controller, also.
PILOT-PLANT CONSTRUCTION
Construction of the pilot-plant began in May 1985. Brush was
cleared from the surface of the lagoon in the area proposed for
placement of the Extraction and Treatment Beds. Approximately
1-2 ft of cover were scraped from the lagoon surface. The
resulting surface was leveled. In addition, an area on top of shale
fill on the eastern portion of the lagoon was leveled for placement
of the Process Area.
Wells 1 and 2 were installed on May 24 and 25, 1985. Three bor-
ings were conducted before an appropriate location for placement
of the Extraction Bed was found. It was considered important to
locate the Extraction Bed in a section of the lagoon that was filled
continuously with sludges rather than in an area containing semi-
continuous lenses of shale fill. Sheet piling was installed around
Wells 1 and 2 between June 10 and 13, 1985. Construction and
water batching of the pilot-plant was completed on July 15, 1985.
PILOT-PLANT OPERATION
The pilot-plant has been inspected either daily or every second
day during operation. Field crews consist of at least two trained
workers. Routine maintenance and inspection of the pilot-plant
includes:
• Recording of elapsed time on the control timers for each pump
• Recording of the liquid levels in Tanks 2 and 4
• Visual inspection of the Extraction and Treatment Beds
• Measurement of the water level elevation in Wells 1, 2 and 3
• Mixing NaOH solution in Tank 1 and feeding the Extraction
Bed, if required
• Adjusting feed stream controls and timers for Tank 3
• Sampling Wells 1, 2 and 3, and Tanks 2, 3 and 4 at times re-
quired
II
^-d
^ 3
So
9 -
8 -
.7 -
6 -
5 -
4 -
.3 -
.2 -
. I
I
09-
o.a -
0.7 -
0.6 -
o.s -
0.4 -
0.3 -
0.2 -
0 I -
0
10 GO
TIME (DAYS)
Routine maintenance and inspection usually requires 1 to 2 hours
of site work per day.
Extraction Bed Operation
Recovery of leachate from Wells 1 and 2 began on July 17,
1985. Application of potable water to the surface of the Extrac-
tion Bed began on July 23. The purpose of the potable water addi-
tion was to test the application system from Tank I and the
hydraulic response of the Extraction Bed. It was indicated that
the apparent hydraulic response of the bed was greater than
several days.
On July 31, influent to the Extraction Bed was changed to
0.005 N NaOH. Throughout August, low infiltration and
recovery rates of extractant were observed (Figs. 5 and 6). Pond-
ing frequently occurred at the sludge surface. In response to this
problem, six well-points were installed in the Extraction Bed to ef-
fect subsurface injection of extractant. Installation of the well-
points was completed on August 30 (Day 40). Subsequently,
sodium hydroxide solution was injected through the well-points
and applied to the surface of the Extraction Bed. Substantial in-
creases in extract recovery were observed (Fig. 6). On Sept. 19
(Day 60), influent to the bed was increased in strength to 0.1
N NaOH. This change was made in an attempt to increase TOC
concentration and quantity of extract recovered. Recent data in-
dicate the permeability of the Extraction Bed is increasing.
100
90 -
BO -
70 -
CO -
50 -
tO -
30 -
20 -
10 -
40 60
r/.ur (DATS)
Figure 5
Extraction Bed Influent Volume
Figure 6
Extraction Bed Effluent Volume
Treatment Bed Operation
Addition of potable water to the Treatment Bed surface began
on July 15. This application continued until field status of the soil
system was established. On July 19, influent to the Treatment Bed
was changed to a nutrient solution containing dextrose as the
primary carbon source (Table 3). Inoculation with an activated
sludge culture occurred on July 24.
Table 3
Makeup of Nutrient Solution Applied Beginning July 19
50 g/1 Dextrose
8 g/1 NH4C1
45 g/1 Ca(NO3)2 • 4H2O
0.3 g/1 K2HP04
This solution is diluted approximately 100:1 in Tank 3
106 ON-SITE TREATMENT
-------
11
10. S -
to -
9.5 -
9 -
8.5 -
8 -
7.5 -
TIME (DAYS)
D WELL I + ITJiY.i 2
Figure 7
Extraction Bed Well 1 and 2 (pH)
Operation with nutrient solution as influent continued until
adequate quantities of extract for treatment had been ac-
cumulated in Tank 2. Treatment of extract began on Sept. 13.
The target TOC concentration for influent to the Treatment Bed
(effluent from Tank 3) was 1,000 mg/1. The nutrient solution
composition during this phase of operation included 45 g/1
Ca(NO3)2 and 0.3 g/1 K2HPO4. This solution was diluted by a
factor of approximately 100:1 in Tank 3. The target influent rate
to the Treatment Bed during this phase of operation was 1
in./day. The typical application cycle was 4 hr/day.
On Oct. 17, the flows to Tank 3 were adjusted to the target
Treatment Bed influent TOC of 2,000 mg/1. The target influent
rate of application was reduced to 0.5 in./day. These changes
were made so that operation at a similar mass loading with a dif-
ferent hydraulic flux and influent concentration could be exam-
ined. Target flow and concentration were difficult to control
because of variability in water pressure.
RESULTS
Extraction Bed
Influent and effluent volumes for the Extraction Bed are
presented in Figures 5 and 6. Initially, influent volumes exceeded
the volume recovered through Wells 1 and 2 by considerable
->^
If
Si
14 -
13 -
12 -
11
10 -
a -
8 -
7 -
6 -
5 -
4 -
3 -
9 _
+
+
% + + + * ++ +
9
D
D +
+ •*• +
D + ++ ++
DmD ++ + 0°° Q+ &
DaDfiQD P a Da++
D
— . -i 1 r- 1 1 1 1 1 1 1
amounts. These differences are attributed to several factors. In-
itially, the Extraction Bed was unsaturated and a large volume of
liquid was required to saturate the sludges. Leading to the dif-
ference between the two flows were losses through the joints of
the sheet piling. Although the area for leakage was considered
small, a significant hydraulic gradient may have existed across the
joints. Losses through the Extraction Bed bottom also accounted
for more liquid loss.
Recently, increases in the volume of extract recovered from
Wells 1 and 2 have been observed. These increases correlate close-
ly to the installation of the well-points for subsurface injection of
extractant. Additional increases have been observed, most likely
the result of increased sludge permeability occurring from the ad-
dition of more concentrated sodium hydroxide solution. Current
influent application rates are approximately 0.5 in./day; current
effluent recovery rates are approximately 0.3 in./day. Both in-
fluent application rates and effluent recovery rates are increasing.
During the interval between Days 60 and 100, extract recovery
was approximately 60%.
Effluent pH, TOC and TDS for the Extraction Bed are
presented in Figures 7, 8 and 9. Initially, all three quantities were
high. The TOC and TDS of extract recovered from Well 2 were
approximately 11,000 mg/1 and 28,000 mg/1, respectively; the pH
was 10. However, all three parameters decreased between Days 40
and 60. On Day 60, the TOC and TDS of extract recovered from
Well 2 were 5,200 and 11,000, respectively; the pH was 9.2. Initial
high pH and TDS values are attributed to lime previously applied
to the sludges in the Extraction Bed. Decreases in these
parameters between Days 40 and 60 are considered to be the result
of the lime dissolution and removal from the sludges.
32
30
28
2B
24
22
-------
•"V
Cfmy/2;
tnLsandsi
g£
*
t.
1
^
s
1
a»
^•5
p E
UENT I
(T/icm»,
fe
^
15 -
14 -
13 -
12 -
II -
10-
9 -
B -
7 -
e -
5
3
B.S -
g .
7.5 -
7 -
S.S -
s -
5.5 -
z -
.9 -
.8 -
.7 -
6 -
S -
.4 -
.3 -
2 -
j
t
0.9 -
0,8 -
0.7 -
06-
0.5 -
0.4 -
0.3 -
0.2 -
O.I -
0 -
I
*
+
*** '*
+
D
«m +
D* 3 +
0 * *
_ +
D ^ + ft * *
D ri
cP* o of
" OKLD^
1 1 1 1 1 1 1 1 T— T —I 1 1 '
8 8.4 8.8 3.2 9.6 10 10.4 108
pH
D ITELL i » "ELL 2
Figure 10
Extraction Bed - Well 1 and 1
DO
0 D D^ D
^^ i — " — i 1 1 1 1 1 1
' 20 40 SO tO 100
TlUf fDArsi
Figure 12
Nutrient Tank Tank 3 (to Treatment Bed)
Figure 15
Treatment Bed - Well 3 (Effluent TOC)
attributed to development of the microbial population and
modification of the target application rate to 0.5 in./day. Signifi-
cant rainfall occurred during this period, also.
Effluent TOC for the Treatment Bed is presented in Figure 15.
The peak observed between Days 70 and 90 is the response ex-
pected during development and acclimation of the microbial
108 ON-SITE TREATMENT
-------
p,
EFFLUENT
7.8 -
7.B -
7.4 -
7.2 -
7 -
C.8 -
6.B -
B.4 -
6.2 -
6 -
s.a -
S.B -
5.4 -
S.2 -
S -
Rfu
D
O DO
00° D ° °
one ° D D
D QDnD n
O
O D
so ao 100
TlilE (DAYS)
Figure 16
Treatment Bed Well 3 (Effluent pH)
population. This peak was accompanied by a decrease and subse-
quent increase in effluent pH (Fig. 16). This variation in pH also
was an expected response of the system. Both phenomena were
observed in previous laboratory column experiments.
Throughout operation, process effluent pH was between 6 and
8 and effluent TOC and TDS were less than 50 mg/1 and 900
mg/1.
CONCLUSIONS
A pilot-plant to demonstrate a state-of-the-art process for
cleanup of an industrial sludge lagoon has been designed, con-
structed and operated successfully for a period of over 100 days.
The cleanup process consists of two coupled treatment steps. The
first step is removal of contaminants from the sludges through in
situ extraction with aqueous sodium hydroxide. Extractant is in-
jected into the sludges through well-points or applied to the sur-
face through a perforated pipe distribution network. Extract is
recovered by means of two wells screened near the bottom of the
sludge deposits. Operation of this system indicates that up to a
15-fold increase in removal efficiency, over natural processes, can
be obtained through increased extract concentration. An addi-
tional 4-fold increase in the rate of renovation is obtained through
increased hydraulic flux through the sludges.
Extract recovered from the sludges is treated for removal of
TOC through a soil-based sequential aerobic/anaerobic microbial
treatment system. Treatment occurs on-site, immediately adja-
cent to the extraction process. A diverse, mixed microbial popula-
tion is developed in the soil system. Neutralized extract is applied
to the surface of the Treatment Bed and allowed to percolate
through the soil system. Aerobic and anaerobic microbial pro-
cesses metabolize organic contaminants to carbon dioxide and
methane. Treatment efficiencies in excess of 95% have been
demonstrated.
ACKNOWLEDGEMENTS
The authors would like to acknowledge several people at
Rutgers who have performed principal roles in the design, con-
struction and daily operation of the pilot-plant. These people are
Nick Bosko, Betsey Brush, Jim Claffey, Irene Legiec, John
Magee, David Robertson, Judy Svarczkopf and Valerie Ver-
miglio. In addition, the authors would like to acknowledge Joe
Evankow and Terry Burke for design and fabrication of the con-
trol panel.
REFERENCES
1. Kosson, D.S., Ahlert, R.C., Boyer, J.D., Dienemann E.A. and Ma-
gee II, J.F., "Development and Application of On-Site Treatment
Technologies for Sludge Filled Lagoons," Proc. International Con-
ference on New Frontiers for Hazardous Waste Management, EPA/
600/9-85/025, Sept. 1985, 118-127.
2. Kosson, D.S. and Ahlert, R.C., "In-situ and On-site Biodegradation
of Industrial Landfill Leachate," Environ. Prog. 3, 1984, 176-183.
3. Dienemann, E.A., Magee II, J.F., Kosson, D.S. and Ahlert, R.C.,
"Rapid Renovation of a Sludge Lagoon," paper presented at the
AIChE Annual Meeting, Chicago, IL, Nov. 1985.
ON-SITE TREATMENT 109
-------
Cleanup of Contaminated Soils and Groundwater
Using Biological Techniques
Paul E. Flathman
Jason A. Caplan, Ph.D.
O.H. Materials Co.
Biotechnology Division
Findlay, Ohio
ABSTRACT
On-site biological cleanup following spills of biodegradable
hazardous organic compounds in lagoon, soil and groundwater
environments is a cost-effective technique when proper engineer-
ing controls are applied. Biodegrdation of hazardous organic con-
taminants by microorganisms minimizes liability by converting
toxic reactants into harmless end products.
The two case histories presented in this paper describe:
• Bench-scale evaluation of the potential for biological cleanup
in the spill site matrix
• Field implementation of biological cleanup systems
O.H. Materials Co. has performed biological cleanups of spilled
substances since 1978 when a railroad incident resulted in spillage
of acrylonitrile. Subsequent biological environmental restoration
projects have included additional acrylonitrile spills and cleanup
of other materials such as methylene chloride, methylethylketone,
crude oil, petroleum hydrocarbons, butylcellosolve, ethylacry-
late, n-butylacrylate, toluene, styrene, acetone, isopropanol,
tetrahydrofuran and various phenolics.
Cost-effectiveness, minimal disturbance to existing operations,
on-site destruction of spilled contaminants and permanence of the
solution are several of the advantages identified for implementing
biodegradation as a technique for spill cleanup and environmen-
tal restoration.
TREATMENT OF METHYLENE CHLORIDE-
CONTAMINATED GROUNDWATER
Many current and emerging technologies are available for the
on-site removal of groundwater contaminants. For cost-effective
cleanup following spills of hazardous organic materials, combina-
tions of treatment alternatives often are employed.
The case history presented here describes the physical and bio-
logical treatment of methylene chloride-contaminated ground-
water following rupture of an underground pipeline. After 2
months of field operation, air stripping techniques provided an
estimated 97% reduction in the concentration of methylene
chloride in the groundwater. In a downgradient monitoring well
located within 20 ft of the pipeline break, the methylene chloride
concentration was reduced from 9,300 mg/l to 300 mg/ by the
end of the second month.
Since it became increasingly difficult to remove the residual
methylene chloride from the groundwater by physical treatment
methods, biological techniques were initiated. Biological treat-
ment, using adapted microbial strains, was implemented by the
end of the third month and achieved an estimated 97% reduction of
the residual methylene chloride in groundwater. After 4 months of
field operation, greater than a 99.9% reduction in the concentra-
tion of methylene chloride was attained.
On Aug. 16, 1983, it was discovered that a buried methylene
chloride line, located in proximity to a water main, had ruptured
and an undetermined amount of methylene chloride had leaked
into the soil and groundwater. Following discovery of the methy-
lene chloride leak, OHM was summoned to conduct an emergen-
cy response site investigation. Free product was collected first and
staged in vessels for eventual on-site treatment. Monitoring wells
that were installed established that the contamination was within
a confining clay layer at the 20-ft level and did not reach the
aquifer located at 100 ft. Monitoring wells were utilized to iden-
tify the contaminant plume, and pumping techniques were em-
ployed to contain the spilled methylene chloride.
Emergency response to the spill involved only containment. An
investigative phase was initiated immediately to examine various
alternatives for environmentally restoring the site once the spill
was contained. Several alternatives were considered, including:
excavation and disposal; physical containment; and biological
techniques. After reviewing ihe alternatives, it was determined
that a combination of the alternatives would be the most
beneficial for site remediation.
Following removal of 160 yd-' of highly contaminated soil to a
Class A secure landfill, positive placement and suction lift tech-
niques were implemented. Air stripping was used to remove
methylene chloride from the recovered groundwater. When it
became increasingly difficult to remove methylene chloride from
the groundwater, biological techniques utilizing adapted micro-
organisms were employed.
Physical Treatment
Air stripping was selected as the preferred treatment technology
for removing methylene chloride from the recovered ground-
water. This decision was based on several important factors. The
first factor was OHM's documentation of the strippability of this
compound in past emergency and remedial cleanup operations.
Second, the regulatory agencies did not require the implementa-
tion of vapor phase scrubbing units to cleanse the air emissions
from aeration devices. Finally, unlike granular activated carbon,
the use of air stripping is generally maintenance-free, and
regeneration and/or disposal of contaminated media is not
necessary.
Bench-Scale Evaluation
Characterization and bench-scale testing of recovered ground-
water was performed prior to final selection of the packed column
air stripping treatment alternative. Among the wastestream
parameters characterized were:
• Influent concentration of volatile organic(s)
• Presence of phased product(s)
• Hardness and alkalinity
110 ON-SITE TREATMENT
-------
FEED
FROM
VACUUM — —
RECOVERY
UNIT
1 SAND I
1 FILTER I
EQUALIZATION
VESSEL
1
t\ HEAT
UEXCHANGER
CONDENSATE
AIR STRIPPER
TO PLANT
WASTE
-+• WATER
TREATMENT
FACILITY
SOLVENT
RECOVERY
Figure 1
Physical Treatment System Schematic for Methylene Chloride-Contaminated Groundwater
• Suspended solids concentration
• Scale-forming tendency at influent pH
Characterizing these parameters determined if pretreatment was
necessary prior to implementation of the primary treatment pro-
cess. A particular primary treatment process may be cost-
effective, but the pretreatment costs may render the process cost-
prohibitive.
Methylene chloride was obtained when the groundwater was in-
itially pumped from several monitoring wells. The recovered
groundwater often had a murky brown color with visible par-
ticulate matter that slowly settled with the heavier-than-water
methylene chloride.
Following the separation of methylene chloride (by phase
separation) from recovered groundwater in quart jar samples, the
supernatant was analyzed for methylene chloride concentration.
The methylene chloride concentration approached the accepted
limit of solubility at ambient temperature.
Dilling et al.1 calculated the volatilization rate of methylene
chloride for concentrations ranging from several hundred to
several thousand mg/1 from the liquid-air interface in open
laboratory vessels. OHM assumed that similar volatilization rates
of methylene chloride would occur in process settling chambers
following gravity separation. OHM also assumed that the action
of system pumps and surface winds blowing over water surfaces
into open vessels would result in even greater rates of volatiliza-
tion. Evaluations of the remedial alternatives were based upon
those assumptions.
Field Implementation
Figure 1 is a schematic of the treatment system designed to treat
groundwater at a rate of 10 to 15 gal/min. The recovered ground-
water was pumped through a mixed-media filter bed to remove
sand and other particulate matter. Three types of porous media
were used in this filter: (1) anthracite, (2) silica sand and (3) pea
gravel, with the sand placed above the gravel and the anthracite
above the sand. This media configuration provided a large pore
size near the filter surface to capture particulates with a relatively
low head loss. Smaller sized particles were trapped in the lower
sand layer.
Groundwater leaving the mixed-media filter bed was piped to
an equalization vessel having a residence time of 14 hr. This vessel
was basically a rectangular holding tank which allowed the denser
methylene chloride to separate from the water by gravity. The
vessel was equipped with sludge collection equipment to remove
accumulated solids and float controls in the last chamber to main-
tain a constant flow to the next treatment step.
Liquid from the equalization vessel was then pumped through a
skid-mounted shell and tube heat exchanger to raise the
temperature of the water from 10°C to more than 40 °C. The
heated water was then pumped to the top of the air stripping col-
umn for removal of soluble methylene chloride.
Data collected during the project demonstrated that methylene
chloride, as pure product, settled in the equalization vessel
together with fines and other particulate matter that were not
trapped by the mixed-media filter. While the total volume of the
REMOVAL EFFIENCY
—O CALCULATED
• OBSERVED
40 45 50
COLUMN OPERATING TEMPERATURE
CC)
Figure 2
Removal Efficiencies of Methylene Chloride from Groundwater by Air Stripping at a Function of Column Operating Temperature
ON-SITE TREATMENT 111
-------
pure product collected in this manner did not exceed more than a
few gallons, the vessel operated efficiently. Methylene chloride
concentrations in the water entering the air stripping column rare-
ly exceeded 150 mg/1. The mixed-media filter removed virtually
all paniculate matter in the water, and backwashing of the filter
was infrequently required. The tower operating temperature was
maintained between 27 and 60 °C.
Figure 2 is a summary of actual removal efficiencies obtained
during the 3 month operation of the air stripper. The same plot
contains predicted data. The predicted removal efficiency was
calculated from the mass transfer model of Onda et a/.2-M At op-
erating temperatures less than 38 °C, removal efficiencies agreed
with those predicted.
However, significant deviations from expected values were
observed at the higher operating temperatures. One explanation
for those deviations is the non-adiabatic nature of the air strip-
ping system. The column was constructed of mild steel and was
not insulated since the system was meeting the discharge criteria
of 20 mg/1 methylene chloride. As ambient temperatures became
progressively colder, column heat losses were greater. Those
observations suggested that substantial deviations from the
predicted removal efficiency would occur as the average operating
temperature of the air stripper was increased.
The method used for predicting mass transfer rates of methy-
lene chloride during the design phase of the project was adequate
for estimating removal efficiencies at elevated temperatures. This
same procedure has been implemented by OHM in numerous air
stripping operations where mobilization of available equipment
was used for emergency situations. Unlike other volatile organics,
equilibrium and mass transfer data for methylene chloride have
not been documented for aeration applications at elevated
temperatures.
Results for the removal of methylene chloride from monitoring
well B-5 are presented in Figure 3. This well was located 20 ft
downgradient from the ruptured methylene chloride line. An ex-
ponential decay curve was used to quantify the removal rate
(decline in concentration) of methylene chloride from the ground-
water surrounding the well. Holding other variables constant, the
rate of decrease was assumed to be a function of methylene
chloride concentration, i.e.
dC
— = -kC (!)
where,
C = the concentration of methylene chloride remaining
(mg/1)
t = time (days)
k = the rate constant (day-')
The curve generated was fitted to the following first-order
equation:
C = C0e-i« (2)
where,
C0 = methylene chloride concentration at time zero (mg/1)
The first-order rate constant, k, was determined by linear
regression using least squares, and the first order equation was
converted to:
InC = InC0 - kt (3)
A 50% reduction in methylene chloride concentration was
observed within the first 11 days of physical treatment. After 2
months of field operation, air stripping techniques provided an
estimated 97% reduction in the concentration of methylene
chloride in the groundwater being extracted. Since it became in-
creasingly more difficult to remove the residual methylene
chloride from the groundwater, biological techniques were con-
sidered for implementation.
Biological Treatment
Bench-Scale Evaluation
Prior to initiation of a methylene chloride biodegradation
feasibility study, the following screening analyses were performed
on 17 selected water samples collected from the site: methylene
chloride, ammonia-nitrogen, nitrate-nitrogen, orthophosphate-
phosphorus, pH and aerobic heterotrophic bacterial plate counts.
Results of the screening analyses indicated that the site matrix
samples contained viable microbial populations; i.e., the site
matrix did not appear toxic to microbial growth (geometric mean
= 2.0 x 104 CFU/ml, n = 17). The analyses also indicated that
inorganic nitrogen and phosphorus nutrient additions would be
necessary for effective biological treatment of the contaminated
groundwater. Ammonia-nitrogen, nitrate-nitrogen and ortho-
phosphate were not detected in any of the groundwater samples
analyzed. The average pH for all 17 wells was 7.7, an acceptable
value for microbial growth.'
The biodegradation study was performed using the electrolytic
respirometer (Exidyne Instrumentation Technologies, Inc., Col-
orado Springs, CO). For this study, both the fate of methylene
chloride and chloride release, resulting from the biodegradation
of methylene chloride, were used as the test parameters. Chloride
concentration was determined using an Orion model 96-17B com-
bination chloride electrode.
PHYSICAL TREATMENT
DAYS 0-Si
UETHYLEMC CHLORIDE
r* m 0.7t
M C • » 11000 - 00«It
INJECTION OP
BIOLOGICAL MEDIUM
INTO GROUND WATCH
INITIATED
20 40
too ito
Figure 3
Physical Removal of Methylene Chloride
from Monitoring Well B-5
112 ON-SITE TREATMENT
-------
An injection pool sample collected in November 1983, was used
for the electrolytic respirometer study. Sludge collected from the
plant's industrial wastewater treatment basin was used to inocu-
late the respirometer vessels. According to reports, the sludge in
the wastewater treatment basin had been continuously exposed to
low levels of methylene chloride for years and was, therefore,
thought to contain populations of microorganisms acclimated to
methylene chloride. The electrolytic respirometer vessels were
prepared as follows:
Treatment
Vessel
Number
U
3,4
Inoculum
Source
S
S
Test
Matrix
IP
IP
Poisons
X
Poisons added were KCN and NaN3
S = wastewater treatment basin sludge
IP = injection pool
Methylene chloride was added to all respirometer vessels at a
concentration of 100 mg/1. This concentration approximated
average methylene chloride concentrations found in the ground-
water at the site. Vessels 1 and 2 were replicates that demonstrated
the feasibility of using the natural microbial flora. Any nonbio-
logical loss of methylene chloride from the test environment
would be quantified with the abiotic control (i.e., vessels 3 and 4).
Respirometer vessels were prepared using procedures described by
Young and Baumann34 with diammonium phosphate and sodium
dihydrogen phosphate added to each culture vessel. Nitrification
was not suppressed.
Figure 4
Biodegradation of Methylene Chloride with
Chloride Release in a Microbial Reactor
On a periodic basis, 40 ml aliquots were removed from
respirometer vessels and analyzed for methylene chloride and
chloride concentrations. Throughout the study, aliquots were
periodically removed from culture vessels and analyzed for pH
and for ammonia-nitrogen, nitrate-nitrogen and orthophosphate
concentrations. When necessary, pH and inorganic nitrogen and
orthophosphate concentrations were adjusted to ensure that the
chemical environment remained optimal for microbial growth.
Methylene chloride biodegradation and chloride release data
for respirometer vessel 1 are presented in Figure 4. Results for
vessel 2 were similar. Neither loss of methylene chloride nor
chloride release was observed in the abiotic control vessels.
The theoretical chloride release for mineralization of methylene
chloride was calculated from the following relationship:
CH2C12
+ 2HC1
The theoretical chloride release is 0.835 mg chloride per mg of
methylene chloride following complete oxidation.
Methylene chloride was biodegraded stoichiometrically to
chloride in respirometer vessels 1 and 2 (Fig. 4). Using the follow-
ing relationship, it was determined that 130 and 120% of the
theoretical chloride release for methylene chloride biodegradation
was met in respirometer vessels 1 and 2.
The results indicated that methylene chloride was completely
oxidized to carbon dioxide, water and the chloride ion. The lack
of a significant lag period indicated the presence of adapted,
naturally occurring microbes that could biodegrade methylene
chloride.
Field Implementation
Use of the underground recovery and treatment system for
aquifer restoration has been described previously.n'15' 25>27> 28 A
schematic of the biological treatment system is shown in Figure 5.
A modified activated sludge system was the preferred method
for aboveground treatment. In addition to providing efficient
biological treatment, the activated sludge system provided a sup-
ply of adapted microorganisms for inoculation of the ground-
water through the injection system. The injection system was used
to inoculate the groundwater with naturally occurring microbes
capable of biodegrading methylene chloride and to provide a
suitable chemical environment necessary to support microbial
growth.
NUTRIENTS
»H
ADJUSTMENT
TO
UNDERGROUND
INJECTION
SYSTEM
Figure 5
Biological Treatment System Schematic for
Oxidation of Methylene Chloride in Groundwater
ON-SITE TREATMENT 113
-------
I!
BIOREACTOR
BIOREACTOR INFLUENT
* BIOREACTOR DRAINED INTO INJECTION
POOL WHILE BEING REFILLED
INJECTION OF CONTINUOUS
BIOLOGICAL MEDIUM BIOLOGICAL
MTO GROUND WATER TREATMENT
HT1ATED INITIATED
Figure 6
Chloride Release in the Bioreactor Resulting from the
Biodegradation of Methylene Chloride in Groundwater
The recovery system was used to withdraw contaminated
groundwater for aboveground treatment. The supernatant from
the treatment system was reinjected into the subsurface environ-
ment, creating a closed-loop system. Biodegradation of
methylene chloride took place in the ground as well as above the
ground in the biological treatment system.
On day 82 of field operation, the bioreactor was prepared for
batch treatment of methylene chloride-contaminated ground-
water to provide a large quantity of adapted microorganisms for
injection. To maintain a methylene chloride concentration in the
bioreactor capable of supporting microbial growth, air containing
methylene chloride vapor was bubbled into the bioreactor. Be-
tween days 82 and 87, an 80.4% increase in the chloride concen-
tration was observed as a result of chloride release from the
biodegradation of methylene chloride. Between days 87 and 90,
the bioreactor was drained into the injection pool while being
refilled. Injection of methylene chloride adapted microorganisms
into the groundwater environment began on day 89 (Fig. 6).
Batch biological treatment of methylene chloride-contaminated
groundwater again was performed in the bioreactor between days
90 and 92. A 110% increase in chloride concentration was ob-
served as a result of methylene chloride biodegradation. The
bioreactor again was drained into the injection pool while being
refilled. On day 95, continuous biological treatment was initiated.
Over the 10-day operational phase, bioreactor effluent chloride
concentration averaged 131 ±65% greater than influent concen-
tration. Therefore, the bioreactor contained a population of
microorganisms adapted to methylene chloride.
Removal of methylene chloride from the groundwater was
rapid. A 50% reduction in methylene chloride concentration was
observed in monitoring well B-5 (Fig. 7) within 8 days after the
commencement of biological treatment. The rate of decrease was
assumed to be first-order;" i.e., holding other variables constant,
the rate of decrease was assumed to be a function of methylene
chloride concentration.
With the onset of winter, biological treatment was temporarily
suspended on day 123 (Jan. 2, 1984). At that time, a 91% reduc-
tion in methylene chloride concentration had been achieved since
the injection of adapted microorganisms into the groundwater.
The theoretical chloride release for biodegradation of 186 mg/1
methylene chloride (i.e., from 192 to 6 mg/1 methylene chloride)
is 155 mg/1. The observed chloride release was 156 mg/1; i.e.,
chloride concentration in monitoring well B-5 increased from 175
mg/1 on day 80 to 331 mg/1 on day 123. Therefore, we concluded
that biological treatment was responsible for the removal of
methylene chloride from the groundwater.
TREATMENT OF ETHYLENE
GLYCOL-CONTAMINATED GROUNDWATER
At the Naval Air Engineering Center in Lakehurst, New Jersey,
biodegradation techniques were used successfully to treat ethylene
glycol-contaminated groundwater following the loss of an
estimated 4,000 gal of cooling water from a lined surface storage
lagoon. The cooling water was estimated to contain 25%
(vol/vol) ethylene glycol.
The problem developed on Jan. 5, 1982, following a storage
lagoon liner break. A subsequent investigative program con-
firmed soil contamination around the lagoon and identified a
180-ft long by 45-ft wide contaminant plume extending to the
east. At the start of the project, the average ethylene glycol con-
centration in the groundwater was 1,440 mg/1. Approximately 85
to 93% of the ethylene glycol was removed from the groundwater
within the first 26 days of biological treatment. By the completion
of the project, ethylene glycol was reduced to below the limits of
detection (<,50 mg/1) in all production wells at the site.
I ,0
INJECTION OF >IO(.OaiCAl
UEOAJUINTO OROUHD
WATER INITIATED
METHYLCNE CHLORIDE
•* • 0.71
In C B In 100000 - 0.0851
•o »o 100 no no iJo MO i»o
DAYS
Figure 7
Biodegradation of Methylene Chloride with Chloride Release
Based on Samples Taken from Monitoring Well B-5
114 ON-SITE TREATMENT
-------
Biofeasibility Evaluation
The biofeasibility study assessed the biodegradation potential
of ethylene glycol in the ground as well as the presence of a toxic
or inhibitory environment to microbial growth. In addition, tech-
niques to increase the biodegradation rate in the ground also were
evaluated. Such techniques commonly include the use of surfac-
tants, cosubstrates, primary substrates, vitamins, trace elements
and selected microbial strains. The results of the biofeasibility
evaluation form the basis for subsequent field implementation, as
discussed in this case history.
Ethylene glycol can be used as a carbon and energy source for
aerobic migroorganisms.6'10.17^1 The aerobic metabolism of ethy-
lene glycol is relatively common, and the pathways of its
metabolism are known.4,5,18,19,26,31-33
Anaerobic metabolism of ethylene glycol has been reported by
Dwyer and Tiedje.8 Using a sewage sludge inoculum under
methanogenic conditions, ethylene glycol was converted to
ethanol, acetate and methane. The ethanol produced was further
oxidized to acetate with methane as the final end product. It has
been shown that clostridium glycolicum fermentation of ethylene
glycol yields equimolar amounts of acetate and ethanol.16
Prior to initiation of the feasibility study, screening analyses, as
previously described for the first case history, were performed on
representative soil and groundwater samples collected from the
site. Ethylene glycol concentration was determined by direct
aqueous injection into a gas chromatograph.
Screening analyses indicated that the groundwater contained a
viable microbial population; i.e., the environment did not appear
toxic to microbial growth. Aerobic heterotrophic bacterial
population densities ranged from 102 to 106 colony forming units
(CFU)/ml. Bacterial population densities of soil samples collected
within the spill area varied from 102 to 106 CFU/g oven-dry soil.
Screening also indicated that both pH adjustment and inorganic
nitrogen and phosphorus nutrient additions would be necessary
for biological treatment of ethylene glycol. The average pH of the
13 samples was 4.5.
• TOTAL BOD
n = 40
r2 = 0.85B
y = 1740 - 4480 •"'
ETHYLENE OLVCOL
n = 6
r2= 0.813
In y = In 44100 - 0.7»»t
The biodegradation study was performed using the electrolytic
respirometer with the composite groundwater sample. All samples
were prepared in a liquid nutrient medium containing basal salts
and yeast extract.
Treatment
Vessel
Number
1
2
3
4
Test
Matrix
OW**
GW
LGWf
GW
Glucose
(1000 mg/1)
—
X
X
Poisons*
—
....
X
*Poisons added were HgCl2, KCN, NaN3
**GW = Composite groundwater sample
tLGW = Laboratory grade water
Results from vessel 1 would measure the rate of biodegradation
of ethylene glycol by the natural microbial flora. Comparison of
oxygen uptake rates in vessels 2 and 3 would demonstrate the
presence of significant quantities of substances toxic or inhibitory
to microbial growth in the composite groundwater sample. Any
nonbiological loss of ethylene glycol would be quantified with the
abiotic control (i.e., vessel 4).
Ethylene glycol biodegradation data for respirometer vessel 1 is
presented in Figure 8. A loss of ethylene glycol was not observed
in the abiotic control (vessel 4). The theoretical oxygen demand
(ThOD) for aerobic mineralization of ethylene glycol was
calculated from the following relationship:
2 (CH2OH)2 + 5 O2 > 4 CO2 + 6 H2O
The ThOD is 1.29 mg O2 per mg of ethylene glycol. With an in-
itial concentration of 1440 mg/1 ethylene glycol, 1860 mg O2
would be required for complete oxidation. The ultimate BOD
(BODu) was determined by fitting oxygen uptake data to a
modified crescent curve:30
BOD = a + be-kt (4)
where,
BOD = the amount of BOD expressed or exerted at time t
(mg/1)
a = BODu, the ultimate amount of oxygen uptake to be
expressed (mg/1)
b = a lag period parameter (mg/1). (Note: b = a if the fitted
crescent curve intersects the origin)
k = the rate constant (hr-i)
t = time (hr)
The ultimate BOD is defined as the total amount of oxygen re-
quired to biodegrade the immediately available organic matter
present in a sample.24
Using the following relationship, it was determined that 94% of
the theoretical oxygen demand for ethylene glycol biodegradation
was met:
BODu
ThOD
x 100
(5)
Figure 8
Biodegradation of Ethylene Glycol in Composite
Groundwater Sample (Vessel 1)
Thus, the results indicate that ethylene glycol was completely ox-
idized to carbon dioxide and water without the accumulation of
incomplete oxidation products. With a BODu of 1700 mg/1, an
estimated minimum of 170 mg/1 NH3-N and 17 mg/1 PO,}-?
would be required in the groundwater to prevent nitrogen and
phosphorus from limiting microbial growth during biological
treatment."
ON-SITE TREATMENT 115
-------
1000
too
BOO
700
-»-• VESSEL,
COMPOSITE OROUNO WATER SAMPLE
4- BASAL SALT! 4- OLUCOSE
O.B8S
Ts In SB.4 + O.OSB7I
• BASAL SALTS » OLUCOSE
t MICROaiAL INOCULUM
n 7
rf = O.BBS
In y = hi BS.1 4- 0.04781
30
HOURS
Figure 9
Oxygen Uptake Data for the Composite Groundwater Sample
Containing Glucose and for a Glucose Control
The lack of a significant lag period in both oxygen uptake and
ethylene glycol biodegradation by the natural microbial flora in
vessel 1 indicated the presence of adapted microbes that could
biodegrade ethylene glycol. That Finding was significant because
it indicated that in situ biodegradation of ethylene glycol already
was occurring and that the management approach should be to in-
crease the natural biodegradation rate.
The groundwater appeared to be slightly stimulatory to
microbial growth, as evidenced by the slightly greater rate of oxy-
gen uptake in the composite groundwater sample compared to the
glucose-basal salts control (Fig. 9). At the termination of the
study, initial oxygen uptake data in the toxicity/inhibition control
(respirometer vessels 2 and 3) were compared on a semi-log plot.
A comparison was made of the slopes of the lines-of-best-fit
through the linear portion of the natural log transformed oxygen
uptake data. In the linear portion of the semi-log plot, the rate of
oxygen uptake in the glucose-containing composite groundwater
sample was 1.2 times greater than in the glucose-basal salts con-
trol. Thus, evidence for an environment toxic or inhibitory to
microbial growth was not found.
The feasibility study results indicated that biodegradation tech-
niques were a viable option for reducing ethylene glycol concen-
trations in the groundwater.
Field Implementation
The biological treatment program at the site was divided into a
14-day operational phase, a 3-month monitoring phase and a
9-month maintenance program. The operational phase was de-
signed primarily to provide maximum recovery, treatment and en-
hanced bacterial growth in the groundwater both within the spill
area and within the contaminant plume (Fig. 10). The monitoring
program was designed to assess the ethylene glycol degradation
rate in the groundwater following nitrogen and phosphorous
nutrient addition, pH adjustment and enhanced bacterial growth.
The maintenance program was designed to provide an environ-
ment suitable for the continued biodegradation of any residual
ethylene glycol remaining in the groundwater environment.
STORAGE LAQOOr^
0
\
jO.
V" ,
\
o
• 0
CONTAMINANT
PLUUE.
OnOUNO WATCH
'LOW
O RECOVERY WELL
O (OIL SORINO
• BOIL IORINO / MONITOR WELL
Figure 10
Schematic of Ethylene Glycol Storage Lagoon
and Contaminant Groundwater Plume
Ethylene glycol contamination at the site was divided into two
zones. The first zone was the unsaturated zone between the sur-
face and the water table where ethylene glycol had been retained
through capillary action. The highest contamination level
detected in that zone was 4,900 mg/1 ethylene glycol. Surface con-
tamination also was indicated following analysis of shallow
samples (0 to 2 ft) collected adjacent to the lagoon.
The second zone of contamination was the groundwater.
Groundwater samples within the spill area had ethylene glycol
concentrations as high as 2,100 mg/1. Monitoring wells were in-
stalled, and subsequent water analyses indicated significant
groundwater contamination. A downgradient contaminant plume
was estimated to be 180 ft long by 45 ft wide.
A two-phased approach was implemented to deal with both soil
and groundwater contamination. Using injection and recovery
wells, initial efforts addressed the highly contaminated soils
underlying the storage lagoon. The second phase concentrated on
groundwater cleanup. Its goals were (1) to prevent further migra-
tion of contaminated groundwater and (2) to treat any con-
tamination released during treatment of the unsaturated zone.
The injection system was used to adjust groundwater pH as
well as provide the inorganic nitrogen and phosphorus necessary
to support microbial growth. The recovery system withdrew con-
taminated groundwater for above ground treatment in an ac-
tivated sludge treatment system (Fig. 11). Supernatant from the
treatment system then was reinjected into the subsurface environ-
ment, creating a closed-loop system.
pH ADJUSTMENT
PREMIX I NUTRIENT ADDITION
1 TANK '
RECOVERY—.
WELLS «
WASTE SLUDGE
FOR SURFACE
APPLICATION
Figure 11
Flow Diagram for the Activated Sludge Treatment System
of Ethylene Glycol-Contaminatcd Groundwater
116 ON-SITE TREATMENT
-------
ETHYLENE GLYCOL
O BELOW DETECTION LIMIT
MICROBIAL POPULATION
DH
Figure 12
Ethylene Glycol Concentration, Bacterial Population Density, and pH as a Function of Time for a Contaminant Plume Production Well
Based on the information obtained during the site investiga-
tion, five recovery wells were installed to recover contaminated
groundwater (Fig. 10). Three wells located near the lagoon pro-
vided zones of attraction for treated water injected in that area.
The remaining two recovery wells were positioned east of the
lagoon to recover contaminated water from the contaminant
plume and aid in the distribution of treated water from the plume
injection system.
Following pH adjustment and nutrient addition, supernatant
from the biological treatment system was recharged through a
three-phase injection system. The lagoon injection system was
employed to flush contaminated soil and thereby transmit con-
taminated water to the three recovery wells located in the vicinity
of the lagoon. The plume injection system was similar to the
lagoon system in both construction and operation. The primary
functions of the plume injection system were to enhance bacterial
growth through pH adjustment and nutrient addition and to
create a gradient from the fringe of the plume toward the two
recovery wells located at the center of the contaminant plume.
The third injection phase was implemented through surface ap-
plication. Surface application was used primarily in the lagoon
area to flush the unsaturated zone and enhance bacterial growth
in the contaminated soil.
The activated sludge biological treatment system was designed
to reduce ethylene glycol concentration through microbial bioxi-
dation using the indigenous microbial flora. A second role of the
bioxidation reactor was to provide adapted microorganisms for
the three-phase injection system (Fig. 11). The biofeasibility study
had demonstrated that the natural microbial flora were adapted
to biodegrading ethylene glycol and that the management ap-
proach for the project should be to increase the natural
biodegradation rate using microbes indigenous to the ground-
water.
Throughout the operational phase, treatment system and
recovery well samples were analyzed for aerobic heterotrophic
bacteria, pH, dissolved oxygen, inorganic nitrogen, phosphorus
and ethylene glycol. Based on the results of those analyses, en-
vironmental parameters were modified to maintain an effective
ETHYLENE QLYCOL
O BELOW DETECTION LIMIT
MICROBIAL POPULATION
pH
Figure 13
Ethylene Glycol Concentration, Bacterial Population Density, and pH as a Function of Time for a Downgradient Spill Area Production Well
ON-SITE TREATMENT 117
-------
rate of biological treatment. During the monitoring phase, those
same parameters were quantified on a monthly basis from
recovery well samples.
In the initial phases of treatment, ethylene glycol concentra-
tions in two plume recovery wells were reduced from 690 and 420
mg/1 to below limits of detection (LOD = 50 mg/1) within 26 days
of treatment. With the exception of one monitoring phase data
point (day 40), ethylene glycol concentration remained below
detection limits. As shown in Figure 12, groundwater pH was
maintained within an optimal range for bacterial growth (i.e., pH
6 to 7). Bacterial population density was returning to background
levels (i.e., 105 CFU/ml) following a high of 10« CFU/ml.
In the two downgradient recovery wells adjacent to the storage
lagoon, the groundwater concentration of ethylene glycol was
reduced by more than 85% and 92%, respectively, within the first
26 days of treatment (Fig. 13). Initial concentrations were 3,400
and 1,100 mg/1, respectively. In the upgradient recovery well ad-
jacent to the lagoon, the ethylene glycol concentration in the
groundwater increased from 100 mg/1 on day 0 to 880 mg/1 by
day 13. However, monitoring data showed a continued decrease
in ethylene glycol concentration after day 13. The increase in
ethylene glycol concentration prior to day 13 reflected the
project's aggressive operational phase, whereby, the injection
system flushed pockets of ethylene glycol from the unsaturated
zone into the groundwater.
The maintenance program focused on the removal of those re-
maining pockets of contamination, particularly in the lagoon
area. As part of that program, both lime and diammonium phos-
phate were applied to the soil surface in the lagoon and contami-
nant plume areas. Lime increased the pH of the soil and ground-
water to a favorable range for microbial growth. Diammonium
phosphate, a readily available source of nitrogen and
phosphorus, supported the growth of an increased microbial
population.
By the completion of the project, ethylene glycol was reduced
to below the limits of detection in all production wells at the site.
CONCLUSIONS
The two case histories presented demonstrate the application of
biological techniques for environmental restoration of areas con-
taminated with spilled organic materials.
The combination of physical and biological techniques effec-
tively removed methylene chloride from the groundwater at a
relatively rapid rate. If it had been necessary to remove the con-
taminated soil from the site, transportation and disposal (T&D)
costs for 14,000 yd3 of contaminated soil would have been at least
$1,050,000. The estimate for traditional treatment was based on a
transportation cost of $3/mile in 20-yd3 sized trucks to a hazar-
dous waste disposal facility 100 miles from the treatment site. The
disposal cost for the contaminated soil was $60/yd3.
By successfully decontaminating the soil and allowing ii to re-
main on-site, a two-fold cost savings was achieved and all future
liability was reduced substantially. This cost estimate for T&D
was conservative since removal of soil from an area increases void
volume for transportation and disposal often by a factor of 1.2 to
1.3. Moreover, hazardous materials cannot be shipped off-site in
bulk, but most be packaged into drums. Additional chemical
analyses generally are required to characterize the waste prior to
off-site transportation and disposal. Inclusion of these additional
factors that must be considered for off-site disposal resulted in an
even greater cost benefit for in situ cleanup.
Biological techniques, as described in the second case history,
were also effectively used to remove ethylene glycol from the
groundwater at a relatively rapid rate. The flexibility of the injec-
tion/recovery system in maintaining an environment conducive to
biodegradation while flushing ethylene glycol from the ground-
water was a key factor in the removal of scattered pockets of con-
tamination. If it had been necessary to remove the soil from both
the spill area and the contaminant plume to a hazardous waste
disposal facility 500 miles from the treatment site, the T&D cost
would have been more than 16 times the cost for on-site treat-
ment.
Biodegradation as a method for spill cleanup and environmen-
tal restoration is considered to be a promising technology. Land
treatment techniques have been engineered and are accepted as an
economical and environmentally sound means of destruction for
many types of industrial wastes. With regard to the cleanup of
contaminated soil and groundwater, excavation has been a com-
mon method for remediation. However, combinations of physical
and biological techniques now are gaining increasing acceptance as
a practical, cost-effective alternative for environmental restoration.
REFERENCES
I. Adas, R.M. and Bariha, R.. Microbial Ecology: Fundamentals and
Applications, Addison-Weslcy Publishing Company, Reading, MA,
1981.
2. Bolles, W.L. and Fair. J.R., "Improved Mass-Transfer Model En-
hances Packed-Column Design," Chem. Eng.. 84(14), 1982, 109-116.
3. Bordner, R., Winter, J. and Scarpino, P., Eds. "Microbiological
Methods for Monitoring the Environment-Water and Wastes,"
EPA 600/8-78-017. Environmental Monitoring and Support Labora-
tory. U.S. EPA, Cincinnati, OH, 1978.
4. Caskey, W.H. and Taber, W.A., "Oxidation of Ethylene Glycol by
a Salt-Requiring Bacterium," Appl. Environ. Microbiol.. 42, 1981,
180-183.
5. Child, J. and Willetts, A., "Microbial Metabolism of Aliphatic
Glycols. Bacterial Metabolism of Ethylene Glycol." Biochem.
Biophs. Ada. 538, 1978, 316-327.
6. Claus, D. and Hempel, W.. "Specific Substrates for Isolation and
Differentiation of Azotobacler vinelandii," Arch. Mikrobiol, 73,
1970, 90-96.
7. Dilling, W.L., Tefertiller, N.B. and Kallos. G.J., "Evaporation
Rates of Methylene Chloride. Chloroform, 1,1,1-Trichloroethane,
Trichlorocthylene, Tetrachloroethylene, and Other Chlorinated
Compounds in Dilute Aqueous Solutions," Environ. Sci. Technol.,
9, 1975, 833-837.
8. Dwyer, D.F. and Tiedje. J.M., "Degradation of Ethylene Glycol
and Polyethylene Glycols by Methanogenic Consortia," Appl. En-
viron. Microbiol., 46, 1983, 185-190.
9. Anon., "Current Developments, Hazardous Waste," Environ.
Reporter. May 6. 1983. 11-12.
10. Fincher, E.L. and Payne, W.J., "Bacterial Utilization of Ether Gly-
cols," Appl. Microbiol.. 10, 1962, 542-547.
11. Flathman. P.E., McCloskey. M.J., Vondrick. J.J. and Pimlett,
D.W., "In Situ Physical Biological Treatment of Methylene Chlor-
ide (Dichloromethane) Contaminated Ground Water," Proc. of the
Fifth National Symposium on Aquifer Restoration and Ground
Water Monitoring, Columbus. OH, May 1985, 571-597.
12. Flathman, P.E. and Caplan, J.A., "Biological Cleanup of Chemical
Spills," Proc. of HAZMACOM 85. Oakland, CA, Apr. 1985, 323-
345.
13. Flalhman, P.E. and Githens, G.D., "In Situ Biological Treatment
of Isopropanol, Acetone, and Tetrahydrofuran in the Soil 'Ground-
water Environment," E.K. Nyer, Ed., Ground Water Treatment
Technology. Van Nostrand Rcinhold Company, New York. NY,
1985, 173-185.
14. Flalhman, P.E.. Quince, J.R. and Bottomley, L.S.. "Biological
Treatment of Ethylene Glycol-Contaminated Groundwater at Naval
Air Engineering Center, Lakehursl, New Jersey," Proc. of the Fourth
National Symposium on Aquifer Restoration and Ground Water
Monitoring. Columbus, OH, May 1984. 111-119.
118 ON-SITE TREATMENT
-------
15- Flathman, P.E., Studabaker, W.C., Oithens, G.D. and Muller,
B.W., "Biological Spill Cleanup," Proc. of the Technical Seminar
on Chemical Spills, Toronto, Ontario, Oct. 1983, 117-130.
16. Gaston, L.W. and Stadtman, E.R., "Fermentation of Ethylene Gly-
col by Clostridium glycolicum sp. /. Bacterial., 85, 1963, 356-362.
17. Gonzalez, C.F., Taber, W.A. and Zeitoun, M.A., "Biodegrada-
tion of Ethylene Glycol by a Salt-Requiring Bacterium," Appl.
Microbiol., 24, 1972, 911-919.
18. Harada, T. and Nagashima, Y., "Utilization of Aklylether Com-
pounds by Soil Bacteria," J. Ferment. Techno!., 53, 1975, 218-222.
19. Jones, N. and Watson, G.K., "Ethylene Glycol and Polyethylene
Glycol Catabolism by a Sewage Bacterium," Biochem. Soc. Trans.,
4, 1976, 891-892.
20. Kemmer, F.N., Ed., Nalco Water Handbook, McGraw-Hill Book
Company, New York, NY, 1979.
21. Klecka, G.M., "Fate and Effects of Methylene Chloride in Acti-
vated Sludge," Appl. Environ. Microbiol., 44, 1982, 701-707.
22. Kobayashi, H and Rittman, B.E., "Microbial Removal of Haz-
ardous Organic Compounds," Environ. Sci. Tech., 16, 1982, 170A-
183A.
23. Larson, R.J., "Role of Biodegradation Kinetics in Predicting En-
vironmental Fate," A.W. Maki, K.L. Dickson, J. Cairns, Jr., Eds.,
Biotransformation and Fate of Chemicals in the Aquatic Environ-
ment, American Society for Microbiology, Washington, DC, 1980
67-86.
24. Mitchell, R., Introduction to Environmental Microbiology, Pren-
tice-Hall, Inc., Englewood Cliffs, NJ, 1974.
25. Ohneck, R.J. and Gardner, G.L., "Restoration of an Aquifer Con-
taminated by an Accidental Spill of Organic Chemicals," Ground
Water Monitoring Review, 2(4), 1982, 50-53.
26. Pearce, B.A. and Heydeman, M.T., "Metabolism of Di(Ethylene
Glycol) [2-(2'-Hydroxyethoxy) Ethanol] and Other Short Poly (Ethy-
lene Glycol)s by Gram Negative Bacteria," J. Gen. Microbiol, 118,
1980, 21-27.
27. Quince, J.R. and Gardner, G.L., "Recovery and Treatment of Con-
taminated Ground Water: Part I," Ground Water Monitoring Re-
view, 2(3), 1982, 18-22.
28. Quince, J.R. and Gardner, G.L., "Recovery and Treatment of Con-
taminated Ground Water: Part II," Ground Water Monitoring Re-
view, 2(4), 1982, 18-25.
29. Roberts, P.V., Hopkins, G.D., Munz, C. and Riojas, A.H., "Evalu-
ating Two-Resistance Models for Air Stripping of Volatile Organic
Contaminants in a Countercurrent, Packed Column," Environ.
Sci. Tech., 19, 1985, 164-173.
30. Shammas, N.C., "Modified Crescent Curve Fitting, Program Num-
ber 341C," Users' Library, Hewlett-Packard Company, Corvallis,
OR, 1983.
31. Thelu, J., Medina, L. and Pelmont, J., "Oxidation of Polyoxy-
ethylene Oligomers by an Inducible Enzyme from Pseudomonas
P400," FEMSLett., 8, 1980, 187-190.
32. Wiegant, W.M. and DeBont, J.A.M., "A New Route for Ethylene
Glycol Metabolism in Mycobacterium E44," /. Gen. Microbiol.,
120, 1980, 325-331.
33. Willetts, A., "Bacterial Metabolism of Ethylene Glycol. Biochem,"
Biophys. Acta, 677, 1981, 194-199.
34. Young, J.C. and Baumann, E.R., "The Electrolytic Respirometer—
I Factors Affecting Oxygen Uptake Measurements," Water Res.,
10, 1976, 1031-1040.
ACKNOWLEDGEMENTS
The authors wish to acknowledge the contribution of their col-
league, Lucy S. Bottomley, P.E., Environmental Engineer,
Public Works Department, at Naval Air Engineering Center,
Lakehurst, New Jersey. At O.H. Materials Co., we specifically
thank Robert H. Panning and Robert J. Ohneck for their support
throughout these projects.
ON-SITE TREATMENT 119
-------
Physical/Chemical Removal of Organic Micropollutants
from RO Concentrated Contaminated Groundwater
L. Simovic
Environment Canada
Environmental Protection Service
Wastewater Technology Centre
Burlington, Ontario
J.P. Jones
University of Sherbrooke
Sherbrooke, Quebec
I.C. McClymont
ICAN Consultants
Burlington, Ontario
ABSTRACT
The reverse osmosis (RO) concentrate of contaminated ground-
water from the Gloucester Landfill site near Ottawa, Ontario, was
treated by air stripping (AS), ozonation (O3) and ultraviolet (UV)
irradiation for the removal of organic micropollutants.
Fourteen compounds, predominantly volatiles, were monitored
during this 6-week study. The concentration ranges of some of the
compounds in the reverse osmosis concentrate were: trichloro-
ethylene 2-280 /ig/1; benzene 28-1,350 pg/1; chloroform
800-40,100 pg/l; diethyl ether 136-2,300 /ig/1 and acetone
750-36,000 jig/1.
Based on the results of this study, the air stripping process
alone is considered to be adequate for treating the RO-concen-
trated contaminated groundwater at the Gloucester Landfill site.
Residual concentrations of four of the six compounds with pro-
posed groundwater quality objectives met or exceeded those
levels.
The removal efficiency achieved by all three unit processes
combined, however, was better than that obtained by AS alone.
For the degree of additional improvement that might be achieved,
it is doubtful that the addition of Os/UV to the AS process would
be economically justified.
INTRODUCTION
From 1969 to 1980, federal government laboratories in Ottawa,
Ontario, disposed of their hazardous wastes (mainly organic
solvents) in a Special Waste Compound (SWC) at the nearby
Gloucester Municipal Landfill.1 In 1981 contaminants including
diethyl ether, acetone and some chlorinated solvents were
detected in the groundwater migrating from the landfill site. This
contamination was a potential danger to the drinking water
source of nearby residents. Subsequently, Environment Canada
and Transport Canada undertook a number of studies to define
the extent of the problem and identify possible remedial
measures.'
In 1984 Environment Canada's Wastewater Technology Centre
(WTC), in cooperation with the Environmental Emergencies
Technology Division, carried out a pilot scale study on the
removal of toxic organics from the contaminated groundwater
within the SWC. To define one remedial action, the objectives of
this study were: (1) to evaluate the effectiveness of selected reverse
osmosis (RO) membranes for removing toxic organics from the
contaminated groundwater and (2) to evaluate the effectiveness of
air stripping, ozonation and ultraviolet (UV) irradiation for
treating the concentrate from the RO system.
The results related to the RO tests have been presented
elsewhere.' This paper presents the results for the removal of
organics from the RO concentrate by air stripping, ozonation and
ultraviolet irradiation. For each treatment process, the results are
discussed with reference to specific toxic organic compounds
which were present in the groundwater at concentrations higher
than the proposed groundwater quality objectives.4
METHODS AND MATERIALS
A trailer facility outfitted for this project was set up at the
Gloucester site beside test well 36 W. This well, located within the
SWC, was found to have the highest concentrations of organic
micropollutants.' A schematic of the pilot plant used during this
study is shown in Figure 1.
The air stripping unit in the pilot plant consisted of two 1.22 m
by 10 cm (I.D.) plexiglass columns connected in series. The col-
umns were packed with 1.3 cm Intalox* saddles. A Whispair
1707Jt blower was used to force air through the stripping columns.
The ozonation experiments were conducted using a bubble col-
umn 91 cm by 10 cm (I.D.). Ozone was generated from oxygen us-
ing a Welsbach T-816J ozonator. The Oj concentrations in the inlet
gas and the off-gas were measured using a standard iodometric
method.' Ozone in the liquid was measured using the indigo
method.'
The UV irradiation chamber used was a Trojan System 2000.**
It consisted of four 9 W low pressure mercury lamps. The total UV
output of all the lamps in the reactor was determined by ferriox-
alate actinometry'and was found to be 4.1 x 10 ~4 einsteins/min.
The flow of liquid through the reactor was maintained during the
tests to give a theoretical retention time (RT) of 22 min.
Air Stripping
During air stripping experiments, the following three variables
were studied: air flow rate, the packing height and the water flow
rate. A full 23 factorial experimental design was used to allow for
the determination of the relative significance of the effect of these
variables (Table 1). The air and water flow rate levels were chosen
to bracket a wide range below the conditions at which the col-
umns would become flooded. From the calculated stripping fac-
tor for the volatile compounds present in the concentrate, the
values for the packing height were chosen to give a sufficient
number of transfer units for ? 99% removal of these compounds.
• Norton Chemical Company, Akron, OH
t Roots Blower Operation, Dresser Industries, Inc., Cornerville, IN
t Polymetrics Inc., Sunnyvale, CA
"Trojan Technologies Inc., London, Ont.
120 ON-SITE TREATMENT
-------
AIR
STRIPPING
COLUMN
+
RESERVOIR
EXHAUST
^
' t
PUMP
00
CONCENTRATED
GROUNCVATER
Figure 1
Process Schematic for Groundwater Treatment System
Table 1
Air Stripping Process Variable
Levels
Variables
Air flow rate (1/min)
Packing height (m)
Water flow rate (1/min)
15
1.22
0.5
300
2.44
2.0
Ozonation/Ultraviolet Irradiation
The effects of ozonation, UV irradiation, air stripping and
water flow rates on the removal of organic micropollutants were
observed using a full 24 factorial experimental design (Table 2).
Table 2
AS/Os/UV Experiment Variables
Levels
Variables
Air stripping
Os dose (mgO3/l of gas)
UV irradiation
(einsteins/min)
Water flow rate (1/min)
OFF
0
0
0.5
ON*
65
4.1 x 10-4
2.0
•Air now rate = 300 1/min; packing height 1.22 m
Sampling and Methods of Analysis
The locations where samples of pilot plant influent and treated
effluent were collected during the tests are shown in Figure 1. For
each experiment, two sets of samples were collected: one set was
collected at the beginning of the test and another set after 2 hr of
continuous operation of the unit process(es) being tested. After
collection, the samples were stored at 4°C in glass vials with
teflon-lined septums. Sample analyses were completed within 10
days of collection of a given set of samples.
The organic compounds, predominantly volatiles, were analyzed
by the purge and trap method using a Hewlett Packard model
5830A Gas Chromatograph coupled to a Flame lonization Detec-
tor (FID). The following pairs of compounds coeluted during the
analysis and could not be quantified individually: (1) 1,1-di-
chloroethane (RT = 13.6) and tetrahydrofuran (THF) (RT =
13.7; (2) 1,1,1-trichloroethane (RT = 17.4) and 1,4-dioxane (RT
= 17.4).
Water Characteristics
During this 6-week study the concentrations of various com-
pounds in the RO concentrate varied widely. These data are
presented in Table 3, together with the proposed groundwater
quality objectives for these compounds.
STUDY RESULTS
Forty-seven tests were conducted during this study. Of these
tests, 27 used AS, 7 used O3 and 3 used UV alone to remove the
organic micropollutants from the concentrate. In 11 tests, O3 was
ON-S1TE TREATMENT 121
-------
Table 3
The Influent Concentration for Some of the Organic Mlcropollulanls
Organic Contaminant
Dichloromethane
Acetone
1,1 -Dichloroethylene
Chloroform
Diethyl Ether
1,2-Dichloroethane
Trichloroethylene
Benzene
NA — Nol Available
Groundwaler
RO Concen- Quality Objectives
Irate 0*g/l) (/'g/l)
9- 1,670
750-36,000
2- 1,531
800-40,100
136- 2,300
32- 653
2- 280
28- 1,350
150
NA
0.3
30
NA
10
30
10
followed by UV irradiation. Air stripping was followed by UV in
2 tests, by Oj without UV in 5 tests and by O3 in the presence of
UV in 8 tests. Not all of the samples collected during these tests
could be analyzed for all the parameters, so some of the data
points required to complete the factorial analysis of the data are
missing. A stepwise regression using backwards elimination
routine was, therefore, applied for the analysis of the data.
Although the concentration of organic compounds in the RO
concentrate varied greatly (see Table 3), the removal efficiency of
the various processes tested was not affected by this variability in
the concentration. Variations in the removal efficiencies were due
to the experimental conditions applied.
Air Stripping
Table 4 summarizes the results for air stripping experiments.
Table 4
Percentage Removal and Residual Concentration Ranges
Achieved During AS Treatment of Groundwaler
Contaminant
Dichloromethane
Acetone
1,1 -Dichloroethylene
1 , 1 -Dichloroethane/THF
Chloroform
Diethyl Ether
1,2-Dichloromethane
1,1,1 -Trichloroethane/
1 ,4-Dioxane
Trichloroethylene
Benzene
Range of
Removal
Efficiency
CVo)
57- > 99
32- 99
43->99
25- 99
31- >99
25- >99
24- 99
2->99
41->99
25->99
Residual
Concentration
Oig/D
2.3- 38.2
67.4-9,340.4
0.5- 130.2
6.4- 667.9
33.6-2,775.7
0.7-1,294.0
1.6- 228.6
12.5- 813.7
0.5- 63.4
1.2- 78.2
In addition to the chemicals listed in Table 4, the removal effi-
ciencies of four other compounds, listed in Table 5, were
estimated. The estimates were based on the percent reduction in
the peak areas for the compounds observed during the gas
chromatographic analysis of the RO concentrate and the air strip-
ping samples.
Analysis of the data from the AS experiments showed that
water flow rate had the greatest effect on the removal of most
compounds. The best removals were obtained in the experiments
where the water flow rate was low (0.5 1/min). Packing height and
air flow rate did not have a substantial effect on the efficiency of
removal within the range of values used in these experiments.
Under the best conditions (water flow rate of 0.5 1/min), four
of the compounds were removed to below their groundwater
quality objective concentrations (see Table 6).
Table 5
Percent Reduction Range Estimated from the Peak Areas
Chemical
Dichloropropane
Bromoform
Toluene
Chlorobenzene
Estimated
Removal 0
21->99
15-799
7- 98
5- >99
Table 6
Effluent Concentrations Achieved with Air Stripping
Compounds
Dichloromethane
1 ,2-Dichloroethane
Trichloroethylene
Benzene
Ground water
Quality
Objectives
Oig/l)
150
10
30
10
Residual
Concentration
99%, having residual concentrations of 67.4 /tg/1
and 0.7 /»g/l, respectively. Since there are no proposed ground-
water quality objectives for these compounds, it is not clear
whether the residuals obtained through AS alone would be con-
sidered acceptable. If they were not acceptable, then AS would
have to be augmented by some other process(es).
Ozonation
Results from direct ozonation of the RO concentrate are shown
in Table 7.
The removal efficiencies for ozonation alone were generally
either similar to or lower than those achieved by AS alone. The
reason for lower removal during ozonation was probably due to
the lower gas flow rate through the column. In AS, the gas flow
rate was in the ranges of 15 to 300 1/min; for ozonation, the max-
imum gas flow rate was 2 1/min.
In one of the tests using Oj alone, the ozonation column was
operated at a gas flow rate of 2 1/min but the ozone generator was
turned off. From the results of this test, it appears that approx-
imately 30% of the removal observed was due to the stripping ef-
fect of the gas and 70% was due to the chemical oxidation of the
organics.
With ozonation of air-stripped RO concentrate (Table 7), the
removal of the organic micropollutants generally improved, par-
ticularly for acetone and diethyl ether. It is not clear whether this
additional removal could be sufficient to justify the use of ozona-
tion in conjunction with air stripping.
Ultraviolet Irradiation
Ultraviolet irradiation was used alone, as well as in combina-
tion with O3 or AS. For most compounds, the removal was not as
122 ON-SITE TREATMENT
-------
good as that achieved by AS alone. For this reason, the data from
these tests are not reported. The reader is referred to an earlier
data report.8 However, since the best removal efficiency for most
of the organic compounds was achieved when all three unit pro-
cesses were applied, the removals and residuals from these tests
are presented in Table 8.
Table 7
Average Removals of Organic Chemicals from Groundwater after
Ozonation and Air Stripping/Ozonation
Contaminant
Dichloromethane
Acetone
1 , 1-Dichloroethylene
1 , 1-Dichloroethane/THF
Chloroform
Diethyl Ether
1 ,2-Dichloroethane
1 , 1 , 1-Trichloroethane/
1 ,4-Dioxane
Dichloropropane (1,2- or 1,3-)
Trichloroethylene
Benzene
Bromoform
Toluene
Chlorobenzene
Removal
After AS
(%)
93
57
73
61
90
84
86
69
57
92
91
83
77
60
Removal
After Oa
W
37
28
56
67
33
74
39
53
99
93
73
82
53
64
Removal
After AS/
03 (%)
84
64
76
83
70
95
76
81
100
99
99
95
78
83
Table 8
Percentage Removal and Residual Concentration Ranges
Achieved During AS/Os/UV Treatment of Groundwater
Contaminant
Dichloromethane
Acetone
1 , 1-Dichloroethylene
Chloroform
Diethyl Ether
1,2-Dichloroethane
Trichloroethylene
Benzene
Range of
Removal
Efficiency (%)
89- 99
38- 95
29->99
87- > 99
84- > 99
63- > 99
99- > 99
85- > 99
Range of Residual
Concentrations
Oig/i)
7.1-45.5
36.5-1,933.8
0.5-45.4
6.2-2,020.0
0.7-95.0
1.6-56.5
0.5-3.5
0.3-13.5
Generally, the range of removal efficiencies obtained by the
combination, AS/O3/UV, was higher than AS alone (Table 4).
The residuals were either similar or lower than the ones obtained
by AS alone.
Based on these results, AS alone appears to be adequate to
remove volatile organic compounds from the RO concentrate.
However, all three unit processes combined provided marginally
better removal efficiencies. It is doubtful, though, that the degree
of improvement would justify the increase in cost that would oc-
cur by including O3 and AS units in the treatment train.
CONCLUSIONS
The conclusions drawn from this study follow below.
Two pairs of compounds (1,1-dichloroethane/tetrahydrofuran
and l,l,l-trichloroethane/l,4-dioxane) coeluted during the gas
chromatography analysis and could not be quantified individually.
The results of data analysis by stepwise regression showed that
during air stripping experiments the water flow rate had the
greatest effect on the removal of most organic compounds. The
best removals were obtained at a water flow rate of 0.5 1/min.
During air stripping experiments, four compounds
(dichloromethane, 1,2-dichloroethane, trichloroethylene and
benzene) were removed to below the proposed groundwater quali-
ty objectives.
In the absence of groundwater quality objectives for acetone
and diethyl ether, it is not clear whether the effluent concentra-
tions obtained by air stripping (67.4 and 0.7 /xg/1) would be con-
sidered as acceptable residual values. If they were not acceptable,
O3/UV in conjunction with air stripping could be considered.
Removal efficiencies observed after O3 or UV irradiation alone
were generally either similar to or lower than the results obtained
by AS alone. Combinations of AS/O3, AS/UV and O3/UV unit
processes compared to AS alone did not offer significant im-
provements in the removal efficiencies in most of the tests.
The removal efficiency achieved by all three unit processes
combined was better than that obtained by AS alone. For the
degree of improvement that can be achieved, it is not clear that
the addition of O3/UV to the AS process would be economically
justified.
The data obtained in this study can be used to predict process
performance for those situations where the contaminated ground-
water has large variations in the micropollutants concentration
range.
REFERENCES
1. Jackson, R.E., Patterson, R.J., Graham, B.W., Bahr, J., Belanger,
D., Lockwood, J. and Priddle, M., "Contaminant Hydrogeology of
Toxic Organic Chemicals at a Disposal Site, Gloucester, Ontario,"
NHRI Paper No. 23, IWD Scientific Series No. 141, Ottawa, Ont.,
5985.
2. GEC, WESA, Canviro, "Gloucester Landfill Waste Site Engineering
Study—Problem Definition and Remedial Alternatives," submitted
to Transport Canada as project # L10137-1174 by A.J. Graham En-
gineering Consultants Ltd., Water and Earth Science Associated Ltd.,
and Canviro Consultants Ltd., Sept. 1984.
3. Whittaker, H., Adams, C.I., Salo, S.A. and Morgan, A., "Reverse
Osmosis at the Gloucester Landfill," Proc. of the Technical Seminar
on Chemical Spills, Environment Canada, Toronto, Ont., Feb. 1985.
4. Canviro Consultants Ltd., "Treatment of Organic Contaminants in
Landfill Leachate: Final Report," submitted to Wastewater Technol-
ogy Centre, EPS, Burlington, Ont., May 1984.
5. APHA, AWWA, WPCF, Standard Methods for the Examination of
Water and Waste-waters, New York, NY, 15th ed., 1980.
6. Bader, H. and Hoigne, J., "Determination of Ozone in Water by the
Indigo Method; A Submitted Standard Method," Ozone: Science
andEng., 4, 1982, 169-176.
7. Hatchard, C.G. and Parker, C.A., "A New Sensitive Chemical Acti-
nometer II. Potassium Ferrioxalate as a Standard Chemical Acti-
nometer," Proc. of the Royal Soc. of London, Series A., 235, 1956,
518.
8. Simovic, L. and Jones, J.P., "Removal of Volatile Organic from RO
Concentrated Contaminated Groundwater—Gloucester Site," Un-
published manuscript, Environment Canada, Wastewater Technology
Centre, Burlington, Ont., 1985.
ON-SITE TREATMENT 123
-------
State-of-the-Art Technologies of Removal, Isolation
and Alteration of Organic Contaminants Underground
Walter W. Loo
George N. Butter
McKesson Environmental Services
Pleasanton, California
ABSTRACT
The rising cost and increasing long-term liability of off-site
treatment and disposal have caused a demand for on-site in situ
treatment of organic contaminants located underground. There
are generally three types of in situ treatment: removal, isolation
and alteration.
The purpose of this paper is to provide the reader with a de-
scription of available state-of-the-art, in situ treatment alterna-
tives for the treatment of various organic contaminants in both
the vadose and saturated zones in the subsurface. Guided format
tables are developed for the reader to search for the application
that may work in an individual situation while considering regu-
latory, economic, risk and liability consequences.
INTRODUCTION
The treatment and disposal of hazardous waste and material
off-site has become increasingly difficult due to closure of an al-
ready limited number of qualified waste handling and disposal
facilities, tightening of regulatory standards, escalating transpor-
tation and disposal costs and risks of long-term liability. While
the technology of in situ treatment is advancing at a rapid pace,
there is a need to evaluate periodically the applicability of various
technologies to specific on-site contamination problems and
underground conditions.
The purpose of this paper is to provide the reader with a de-
scription of available in situ treatment alternatives for applica-
tion on various organic contaminants in both the vadose and
saturated zones in the subsurface. However, the best technology
will work only if it is in agreement with regulatory remediation
standards and permits, economic feasibility and long-term risks
and liability as shown in Figure 1.
The following sections discuss the three possible types of in situ
treatment alternatives: removal, isolation and alteration. These
alternatives will encompass a broad combined spectrum of science
and engineering areas:
Chemistry
Physics
Geology
Geohydrology
Biology
Environmental Engineering
IN SITU REMOVAL
An initial requirement for on-site remediation always is the
removal of the source and control of the contaminated plume,
whether it is in the soil (vadose zone) or in the groundwater
(saturated zone) or both. The in situ removal process may in-
volve primary, secondary and tertiary methods.
In-iltu
Treatment
Technical
Alternatives
Regula tory
Remedia tion
Standards &
Permits
Environaenlal
Risks & Liability
Considerations
Economic
Feasibility
Figure I
The primary methods include the following:
• Vadose zone—air vacuum venting
• Saturated zone—pumping of recovery well
The air vacuum venting works well with volatile organic com-
pounds located underground, in particular, permeable strata in
the vadose zone where the soil can serve as a migration pathway.
In general, the venting process will require an air emission permit
which may be prohibitive in some areas. If organic emissions to
the atmosphere are not allowed, an oil stripper (OS) or an acti-
vated carbon adsorption unit is required. If the vent stream is
steady and loaded with high concentration volatile organics, it
may be advisable to recover the product by either an oil strip/de-
sorption or a regenerative activated carbon process (Fig. 2).
The secondary method of removal involves both injection and
recovery wells. Usually, injection wells for the vadose zone should
be discouraged because the injected air/water stream may
"push" the volatile organics toward undesirable off-site loca-
tions.
The tertiary method of removal involves temperature enhance-
ment and addition of surfactants to increase the release rate and
mobility of the organic compound movement through the porous
media toward ultimate recovery. The addition of surfactants may
create new problems in recovery and surface treatment if not
properly applied. Temperature enhancement is generally a desir-
124 ON-SITE TREATMfiHT
-------
Recovery
Phase
Emission
Phase
Atmosphere
Capture/Disposal
Phase
Recycling
Phase
Figure 2
Generic Flow Diagram of In Situ Removal Process
able process (if duration of treatment can be prolonged) by a
passive solar heating system. The system works as underground
space heating under steady circulation. The system cost and main-
tenance are usually very attractive.
For the saturated zone, the application is similar to the vadose
zone with the exception of depth of the contaminated zone and
liquid handling. \
The following design criteria are important for recovery well
field applications:
• The recovery rate must exceed the recharge/safe yield of the
geohydrologic system
• Drilling and well installation design must now allow vertical
migration of contamination from shallow depth
• Pumping of recovery wells must begin with shallow wells and
move to deeper wells to avoid pressure induced vertical migra-
tion of contamination
• It is not advisable to establish pumping wells in low perme-
ability material (generally less than 1 X 10~4cm/sec)
• Well screen design must be designed specifically for "sinkers"
or "floaters" organic compounds
Unfortunately there will never be 100% removal of the con-
taminants. Some portion of the contamination will remain affixed
to the fine grain matrix in the subsurface material.
IN SITU ISOLATION
There are specific geohydrologic conditions which will not
allow in situ removal treatment. Also, less volatile organic com-
pounds may not be easily removed by even secondary or tertiary
treatment processes. The utilization of hydraulic isolation may be
the solution in these situations. Hydraulic isolation can be used as
a follow-up to in situ treatment after removal of a major portion
of the contamination.
There are three different in situ hydraulic isolation methods:
surface cover, subsurface barriers and sorption processes.
The surface cover method generally involves placement of a
clay or concrete cover over a defined contaminated area to pre-
vent percolation of surface water into the contaminated zone to
contact the contaminants below the surface.
Subsurface barriers include a number of isolation techniques:
Slurry trench walls
French drain
Sheet pile
Interceptor ditch
Curtain grouting
The first four types of hydraulic barriers are best for hydraulic
isolation of low permeability media and shallow depths (20 to 40
ft maximum). Curtain grouting involves injection of impervious
material into permeable strata to block off potential migration
pathways in the subsurface environment.
Sorption in situ treatment is a relatively new technology in-
volving replacement of the permeable strata in a known contam-
inated area with a specially formulated fluid with both absorption
and adsorption properties. In the absorption process, the material
will expand up to 30 times its original volume while absorbing the
contaminant and blocking off or filling in the pore spaces. Due to
the microscopic size of the material, the large surface area of con-
tact with contaminants will make an effective sealant to the per-
meable pathways when contaminants are released. The formu-
lated fluid will work for both polar and non-polar organic com-
pounds. Since this treatment process is still in the experimental
stage, it may not be very cost-effective at the present time.
A shallow interceptor ditch may work in relatively low perme-
able saturated material. The ditch may act and perform like a line
of hundreds of wells. Sometimes a combination of the in situ
hydraulic isolation techniques may work for sites with multiple
organic compounds of different physical and chemical properties
when other treatment processes do not provide a total solution.
IN SITU ALTERATION
In situ alteration treatment involves the breakdown of haz-
ardous organic compounds into harmless compounds. It has
been demonstrated that biodegradation of non-halogenated
organic compounds to harmless compounds can be effected in
surface aerobic treatment of sludge and waste water. The same
result can be attained in subsurface conditions which control the
growth of aerobic bacteria. The parameters controlling bacterial
growth and metabolism are:
Pressure
Temperature
pH
Salinity
Dissolved oxygen
Nutrients
Biotoxins
Radiation
Mixing
The transport mechanism in the vadose and saturated zones is
the major problem that hinders subsurface treatment. If the con-
taminated fluid can be pumped to the surface, it is better to per-
form surface treatment. The application of the in situ alteration
process should not be a primary means of treatment.
The in situ alteration treatment of halogenated hydrocarbons is
still in the experimental stage. The recent advances are well docu-
mented in the references found at the end of this paper. The pro-
cess of breakdown of halogenated hydrocarbons depends on the
successful control of methanogenic conditions (addition of meth-
ane mixed with air to the contaminated soil and water). When
aerobic bacteria are exposed to methanogenic conditions, selec-
tive species will degrade the halogenated organic compounds into
carbon dioxide and chlorides which are harmless compounds.
While the researchers are developing a better understanding of the
mechanism of degradation, forseeable problems of field applica-
tion are as follows:
• Aerobic bacteria only occur in shallow depths
• A specially formulated carrier fluid must be coupled with meth-
ane and other nutrient when introduced underground
• Introduction of methane gas underground may create haz-
ardous conditions (i.e., explosion) if not properly contained
ON-SITE TREATMENT 125
-------
Table 1
Practical Application of In Situ Treatment of Volatile Organic*
Table 2
Practical Application of In Situ Treatment of Acid/Base Neutral
Extraclables, Pesticides and Herbicides
In- si tu Trea traenl
Techniques
Remova 1
• Flushing
tempe ra ture
Isolation
• Subsur face
barriers
• SorplLon
Alt a tto
. Aerobic Process
NA Not applicable
(a) Air stripper
(b) Oil stripper
Vadosc Zone
Less
difficult
NA
ef f ec tlve
Yes
Yes
NA
(b) or (c)
Yes, need (a)
(b) or (c)
(b) or (c)
Yei
Yes
Vol. n««d Id)
Saturated Zone
Leu
Yes, but
difficult
NA
Yet
NA
Tachn Ique a
Permeable
Renova 1
Yti , need * Pu» ptng/e vacua 1 1 on
{a) (b) or (c) to create sink
Yet. need • Flushing
(a) (b) or (c)
(a) (b) or (c) temperature
enh«nc*aen t
I tola I i on
Yt i • Surface cover
bo r r I e r t
Yet, need (d) • Sorptlon
A 1 tera t Ion
( • ) Oil s tripper
Vadote Zone
Lea*
NA
NA
NA
Yei
Yet
VA
NA
N««d (c)
Long dura t Ion
Yes
Yet
Yet, r.«ed (c)
Saturated Zone
Ul«
NA
NA
NA
Yet
NA
Pcraeablc
Need (a) or
(b)
Need («) or
(b) and (c)
Lont duration
Yea
Yea, can be
expcnilva
£«(,er .o»jntal for halo-enaled organic co« pound*. Need
(c) Detail geohydrologlc definition
The understanding of the anaerobic degradation of halogen-
ated organic compounds is in its infancy, but it has great poten-
tial. Because of the low cost of bacteriological processes, it is
worthwhile to have engineers perform opportunistic trial and
error experiments on both laboratory bench and field pilot tests
using this technique. Once demonstrated, the cost saving over
other in situ treatment alternatives may be substantial.
At the present state of knowledge of biodegradation, there is
immense room for improvement and development of original re-
search and application. Hopefully, the economic advantage of
innovative in situ technology will push the alteration techniques
to a major breakthrough in the near future.
PRACTICAL APPLICATIONS
The above mentioned in situ treatment techniques may work
individually or in combination for specific contamination cases.
However, there is no general formula to produce an optimal
remediation solution for a case-specific problem.
An attempt is made to provide a general guidance for prac-
tical application of the specific in situ treatment techniques on
organic compounds in the vadose and saturated zones in Tables
1 and 2. These tables will not provide a total solution for a spe-
cific site. However, they will provide an individual with an oppor-
tunity to evaluate the "cans," "cannots" and effectiveness of the
full spectrum of the available alternatives.
These in situ treatment techniques are only one aspect of the
total remediation plan. The engineer will still have to consider
regulatory standards and permits, economic feasibility and long-
term risks and liability for a total remediation program.
REFERENCES
1. Bouwer, E.J. and McCarty, P.L.. "Transformation of 1- and 2- Car-
bon Halogenated Aliphatic Organic Compounds Under Methano-
gcnic Conditions," Appl. and Environ. Microbiol., 45, 1983, 1286-
1294.
2. Haber, C.L. el al., "Methylotrophic Bacteria: Biochemical Diver-
sity and Genetics," Science 221, 1983, 1147-1153.
3. Parsons, F., el al., "Transformations of Tetrachloroethene and Tri-
chloroelhene in Microcosms and Groundwater," JAWWA, 76,
1984,56-59.
4. Wilson, J.T. and McNabb, J.F.. "Biological Transformation of Or-
ganic Pollutants in Groundwater," American Geophysical Union EOS
64, 1983,505-506.
5. Wilson, J.T. and Wilson, B.H., "Biotransformation of Trichloro-
ethylene in Soil," Appl. and Environ. Microbiol.. 49, 1983, 242-243.
126 ON-SITE TREATMENT
-------
Assessment of Volatile Organic Emissions from a
Petroleum Refinery Land Treatment Site
Robert G. Wetherold, Ph.D.
Bart M. Eklund
Radian Corporation
Austin, Texas
Benjamin L. Blaney, Ph.D.
U.S. Environmental Protection Agency
Hazardous Waste Environmental Research Laboratory
Cincinnati, Ohio
Susan A. Thorneloe
U.S. Environmental Protection Agency
Office of Air Quality Planning and Standards
Durham, North Carolina
ABSTRACT
A field assessment was performed to measure the emissions of
volatile organics from a petroleum refinery land treatment site.
As part of this study, the emissions of total volatile organics from
surface-applied and subsurface-injected oily sludge were meas-
ured over a 5-week period. The effect of soil tilling on the emis-
sions also was monitored.
Volatile organics emission rates were measured using the emis-
sion isolation flex chamber method. Soil samples were collected
during the test periods to determine soil properties, oil levels and
microbe count. Soil surface and ambient temperatures, both in-
side and outside the flux chambers, were measured throughout
the test periods.
INTRODUCTION
The U.S. EPA currently is developing background information
on air pollutants at hazardous waste treatment, storage and dis-
posal facilities (TSDFs) to provide supporting data for setting
standards as necessary under Section 3004 of RCRA as amended
in 1984. This regulatory development is the responsibility of the
Office of Air Quality Planning and Standards (OAQPS). The
U.S. EPA Hazardous Waste Environmental Research Labora-
tory-Cincinnati (HWERL) has the responsibility of providing
technical support to OAQPS in the area of atmospheric emis-
sions determination from hazardous waste management.
In order to assess emissions from hazardous waste TSDFs, a
field study was performed at a land treatment facility located at a
large West Coast petroleum refinery. This study was intended to:
(1) measure the total volatile organics emission rates from sur-
face- and subsurface-applied oily refinery sludge and (2) to de-
termine how emissions of volatile organics varied with time.
Based on these measurements the total emissions over a 5-week
period were estimated.
Little work has been performed to characterize volatile organic
emissions from large full-scale land treatment sites. The study
approach and selected results are summarized in this paper. A
more detailed description of the work and its limitations in in-
cluded in a previous U.S. EPA report.1 The emissions of indi-
vidual measured compounds are provided in that U.S. EPA
report.
DESCRIPTION OF THE TEST SITE
The test site was located at a large West Coast refinery. The
land treatment area covers approximately 42,000 m2 (10 acres).
Three test plots, each approximately 420 m2 (0.1 acre) in area,
were located side-by-side at one corner of the facility. The area
containing the test plot has been in land treatment service for
several years.
During normal operation, sludge from refinery operations is
collected and stored in a fixed-roof tank at the land treatment
area. The sludge is injected subsurface using an 8,0001 (50-barrel)
capacity injector vehicle equipped with four separate injector
tines. The injector vehicle consists of a tractor and an 8,000 1
(50-barrel) capacity tank trailer. The tines are attached to the tank
trailer. The waste is injected 15 to 28 cm below the surface and
immediately disked by a disk unit pulled behind the injector ve-
hicle. The area is cultivated with a disk and/or rototiller several
times per week until the next sludge application.
According to the operators of this facility, the annual applica-
tion rate at this site is about 5.4 - 9.1 x 106 kg/year, and the in-
dividual application rates are typically about 16 1/m2 (400 bar-
rels/acre). The sludge typically consists of 50 to 75% DAF/API
separator float, 20 to 30% API separator sludge and about 5%
miscellaneous oily waste. The soil is irrigated periodically to
maintain proper soil moisture content.
EXPERIMENTAL APPROACH
The experimental design, sampling techniques, analytical
methods and test procedures are summarized below.
Experimental Design
The experimental design of this program was developed to ob-
tain data of sufficient accuracy to fulfill the major objectives of
the study. The primary emphasis of the program was on obtaining
volatile organics emission rate measurements. The test design was
a synthesis of two sampling strategies for obtaining emission
measurements: totally randomized sampling over the test plots
and semi-continuous sampling at a single location in each plot.
To provide an overall estimate of the volatile organics emission
rate for a total test plot area, a number of measurements at ran-
domly selected points must be performed. One method of ac-
ON-SITE TREATMENT 127
-------
complishing this is to divide the area into discrete segments and
then randomly select segments to be tested. On the other hand,
the best strategy for examining the change in emission rate with
time is the repeated sampling of a limited number of points over a
period of time.
Three test plots were used in this study. Sludge was surface-
applied to one plot (Plot A) and subsurface-injected at another
plot (Plot C). In between these two plots was a third plot (Plot B)
with no sludge applied. Plot B served as a baseline or control plot.
Each plot was approximately 420 m2 (0. 1 acre) in size, and each
was divided into 21 segments, as shown in Figure 1 , to provide for
randomized sampling.
Subturlac* Background Surlact
Application Plot Plol Application Plot
Cl
C«
C7
CIO
C1J
cie
C1I
a
CS
c<
C11
Cl«
C17
C20
Cl
C*
O
C12
CIS
C1I
C2I
SI
84
B7
BIO
B13
Bit
B1I
Bl
BS
se
Bit
B14
Bir
BJO
U
M
B9
B12
811
B21
A1
A4
A7
A10
A16
Alt
A]
AJ
At
An
AI7
A»
A)
At
At
A12
Alt
A2I
1
(
t
»
Table 1
Typical Dally Sampling Schedule
Flux Charter
frMUno
ApprmlMt* Grid Cm F/C £MU?IM Cnll«rfpd
"TIB* Point* 10 ID" Canlitw Soil Syrlng* Couwnt*
TEXT PLOT A DATE, 11/09/84 APPL 1 CAT ION METHOD i SURFACE
OBOO 002 003 - - X OC
0843 A07e 002 003 - X X
0930 603 002 003 X XX
1013 A03 002 003 - - X
1100 AI9C 002 003 X X X
IMJ AOI 002 003 X
1230 002 003 - - - TIM
r 1400 A07e 002 003 X - X
1449 Mr 002 003 X - X
1330 B14 002 003 X X
TEST PLOT A DATEi 11/09/84 APPLICATION METHOD! SUBSURFACE
0600 001 002 - - X OC
0845 C1IC 001 002 X X X
0930 CI3 001 002 X
1015 CI4 001 002 - - X
5 1100 CIS* 001 002 - - XX
E 1145 BIS 001 002 - - . X
: 1230 001 002 - - - Till
1315 C13e 001 002 X X
1400 Cll' 001 001 X
1449 C04 001 002 - X X
1530 C06 001 002 X X X
1615 BOS 001 002 X
i Prefix A. B, C refers lo plot A (surface applied). pk>l B (background or control plot), and
Eacn Plol . O3 lq m.l.fl (4.450 K). II.)
Eacn Grid > 20.1 M m.l«r« (217 M IIJ
b. F/C « flux chamber
C. Control points
Figure 1
Test Plot Grid Assignment for Assess; ;ent of Emissions
from Land Treatment of Petroleum Wastes
Sampling schedules for each plot were developed for each
separate day of sampling. Each sampling day was divided into
two parts. During each half day, one sampling crew made
measurements at the surface-applied plot, and the other crew per-
formed similar measurements at the subsurface injected plot.
Each crew performed emission measurements at two separate
control points, at two randomly selected points in their assigned
plot and at one randomly selected point in the control plot. Ran-
dom number tables were used to select the order of sampling, the
crew used on each plot and the equipment set used by each crew.
The sampling order for composite soil samples, soil cores, gas
samples and their associated duplicates within a given plot were
also randomly selected. A typical daily sampling schedule is
shown in Table 1.
Sampling Procedures
The main sampling technique employed at this site involved the
direct measurement of emissions using the emission isolation flux
chamber. The enclosure emission measurement approach has
been used by others to measure emission fluxes of sulfur and
volatile organic species.2-iA
A diagram of the flux chamber is shown in Figure 2. The flux
chamber was placed on the emitting surface. Clean, dry air was
passed through the chamber at a controlled and measured rate.
The concentration of the specie(s) of interest was measured at the
outlet of the chamber. The emission rate of the measured specie(s)
from the enclosed surface was then calculated from the flow rate
and concentration in the gas.
Figure 2
Cutaway Side View of Emission Isolation Flux Chamber
and Sampling Apparatus
The flux chamber encloses a surface area of 0.13 m1 and has a
volume of 0.031 m'. The sweep air flow rate was 0.003 - 0.005
mVmin. Air samples were collected from the outlet gas line using
both gas-tight syringes (for on-site analyses) and evacuated
stainless steel canisters (for off-site analyses).
Liquid grab samples of the sludge and of the service water used
to irrigate the soil were collected. Sludge from two tank trucks
was transferred to the application vehicle (tractor) and applied to
the soil. Each tank truck held two application tractor loads. After
the first load was transferred from the tank truck to the applica-
tion tractor, 4-6 sample containers were filled with sludge from
the tractor. The samples were collected consecutively, while the
tractor tank was full, from a gravity-fed line. While two tractor
128 ON-SITE TREATMENT
-------
loads were required to empty each tank truck, samples were col-
lected after only one of the two transfers from each tank truck
load. One service water sample was collected from the plant ser-
vice water line to determine the concentration of organics in this
water which the facility applies as part of its site management
technique.
Two types of solid samples were collected: soil composites and
undisturbed soil cores. Both types of samples were collected im-
mediately adjacent to points where flux chamber measurements
were made. Soil composites were used in the determination of oil
and moisture contents. These samples were collected using a hand
auger and boring to a depth of 30 to 40 cm. The undisturbed soil
samples were used for determining the physical properties of the
soils. They were obtained by driving a brass sleeve approximately
20 cm into the ground, removing the sleeve and capping both
ends.
Type K thermocouples were used to monitor the air and soil
temperatures both inside and outside each flux chamber. Two
separate thermocouples were used to measure soil surface
temperatures inside and outside the flux chamber. Each was plac-
ed in the ground to a depth of 1.3 to 1.9 cm.
Analytical Procedures
The on-site analyses were limited to gas-phase analyses of the
air samples collected in gas-tight syringes from the outlet of the
flux chambers. A Byron Instruments Model 401 Total Hydrocar-
bon (THC) Analyzer was used to determine the concentrations of
total hydrocarbons (THC), CO2 and methane in the effluent air
samples from the flux chamber.
The off-site analyses included the chemical speciation of the
flux chamber air samples collected in stainless steel canisters and
of the liquid sludge and water samples collected at the land treat-
ment site. The oil, moisture and microbe levels in the soil also
were determined, as were the physical properties of the soil
samples.
Chemical speciation of the air and liquid samples for C2 and
C10 hydrocarbons was performed with a Varian 3700 Gas
Chromatograph. The chemical species were cryogenically concen-
trated to increase the sensitivity of the analyses. Liquid samples
were analyzed using a purge-and-trap technique modified to
utilize the cryogenic trap. For both air and liquid samples, total
volatile organics concentrations were obtained as the sum of the
species concentrations.
The oil, water and solids contents of the sludge samples were
determined by solubilizing the oil and water in the waste with
tetrahydrofuran (THF). This liquid phase was then separated
from the insoluble solids. Residual solids were weighed, and the
amount of water in the THF was determined by Karl Fischer titra-
tion. The oil content was calculated by difference. The oil and
grease content of the soil samples was determined using U.S. EPA
Method 413.1. Standard methods were used to determine the bulk
density, particle density, total porosity, moisture content and par-
ticle size distribution.
Sludge Application
The sludge was applied as evenly as possible to the surface-
applied Plot A (SA plot) and to the subsurface-injected Plot C
(SSA plot). The application tractor started injecting sludge at grid
point (GP) C19 (Fig. 1) and moved across Plot C from GP C19
through GP C21. The injector then was lifted above the soil, and
the tractor proceeded across Plot B (control plot) from GP B19
across GP B21. No sludge was applied to Plot B. Sludge then was
allowed to flow from the raised injector tines onto the surface of
Plot A as the tractor passed across it from GP A19 through GP
A21. The tractor then made a U-turn and returned across Plots A,
B and C while applying sludge. Several application passes were
made to complete the application to the entire area of Plots A and
C. Three complete tractor loads and part of another were re-
quired. The average waste loading was 1.4 x 104 kg/plot. All
three plots were tilled with a disk tiller pulled behind the tractor
during the sludge application.
Emission sampling with the flux chambers was started im-
mediately after application and continued for 4 days. Emission
sampling was performed during two other 4-day periods in the 3rd
and 5th weeks following application. Each plot was tilled 2-3
times per week during the 5-week test period. The plots also were
irrigated with water once between the first and second sampling
periods and three times between the second and third sampling
periods. The complete tilling and irrigation schedule for the test
period is summarized in Table 2.
Table 2
Tilling Schedule for Land Treatment Facility During Test Study
Date
10-10-84
10-12-84
10-15-84
10-16-84
10-18-84
10-19-84
10-24-84
10-26-84
10-29-84
10-30-84
10-31-84
11-01-84
11-02-84
11-05-84
11-09-84
11-12-84
Till Ing
Time
0950-1030
1224-1258
0930-1100
1300-1400
0800-0930
1300-1430
1222-1235
1218-1233
0815-0930
0900-1100
1000-1100
0900-1100
0900-1030
1300-1430
1225-1245
1047-1110
Hours from Start a
0
50.5
120.5
147.5
191
220
338.5
386.5
455
480
504.5
528
552
628
722.5
793
Activity
Apply waste and till
Till
Till and Irrigate
Till
Till
Till
Till
TIM
TIM
TIM and Irrigate
TIM
Till and Irrigate
TIM
Till and Irrigate
Till
TIM
a. Rounded off to nearest 1/2 hour.
SAMPLING AND ANALYSES RESULTS
The results pf the sample analyses for the gas, liquid and soil
samples are summarized below.
Gas Samples
The results of the on-site Byron THC analyses agreed quite well
with the off-site Varian 3700 analyses. The results were correlated
with a correlation coefficient of 0.977. The Varian GC generally
gave slightly higher concentrations, with an average difference of
13.8% between the Byron and Varian instruments.
No methane was detected in any of the on-site samples analyzed
with the Bryon 401, which is capable of detecting methane at
levels of 1 ppm.
The CC>2 concentrations in the on-site gas samples were found
to range from about 50 ppm to almost 500 ppm. Most values fell
between 100 and 300 ppm. Background CO2 levels, as determined
by sealing the bottom of the flux chambers and passing air
through them, were in the range of 10 to 200 ppm.
ON-SITE TREATMENT 129
-------
Liquid Samples
The sludge samples were analyzed several times to define the
sludge nonhomogeneity and obtain an average value for the level
of volatile organics. The analyses were performed over a period of
several months (during which the samples were kept under re-
frigeration). The measured volatile organics content of the
sludge, as determined with the purge-and-trap technique, ranged
from 4,250 to 16,600 /ig/g. Most volatile organics concentrations
were between 4,000 and 6,000 /ig/g, with the mean value being
7,800 jig/g. The variability was believed to be at least partially at-
tributable to the difficulty of obtaining a representative sample in
small quantities for analysis. The sludge volatile organic com-
ponent was made up principally of paraffins (average concen-
tration = 3,600 pg/g), olefins (1,200 ug/g) and aromalics
(2,900 jag/g). The remainder of the volatiles were halogenated
and oxygenated hydrocarbons and unidentified organics. From
the characterization tests, the sludge was found to consist of
3.9 to 4.5% solids, 57.5 to 59.1 % water and 37.0 to 38.0% oil.
The sample of service water used to irrigate the soil was also
analyzed for volatile organics to determine whether this water
could be a significant source of volatile organics (this sample was
kept under refrigeration and analyzed approximately 3 months
after collection). The volatile organics content was found to be
0.089 mg/1. The total amount of volatile organics applied with the
water over the duration of the test was estimated to be about 0.03
kg/plot. This amount is equivalent to less than 0.5% of the
volatile organics applied in the sludge.
Soil Samples
The mean values for the physical properties determined for the
soil samples are summarized in Table 3. Also included are the
mean moisture content and oil/grease levels for the collected
samples. The results are summarized for two of the three sam-
pling periods.
Table 3
Mean Soil Properties
Test
Keek
1
3
Particle
Density
^g/cm )
2.50
2.54
Moisture
Content
lot J>
7.40
7.89
Bulk
Density Porosity,
(g/ciT5) (J)
1.21 52
1.42 44
Oil and
Grease
(mg/g)
64
56
STATISTICAL ANALYSES OF THE DATA
A statistical analysis was performed primarily to investigate the
effects of time, temperature inside the chamber (soil and air),
plot samples, location within each plot and chamber shading on
emission rates. The frequency distribution of the emission rate
data was highly skewed and a log normal distribution was used
to model the frequency distribution of emission rates. To de-
termine whether measurable differences existed among the emis-
sion rates from the test plots, an analysis of covariance was per-
formed. The analysis of covariance technique eliminates the vari-
ation in emission rates due to other variables such as time,
temperature (inside the chamber) and chamber surface shading.
In this way, the average emission rates for the test plots can be
directly compared. A discussion of the analysis of covariance
can be found in Brownlee.'
The statistical analysis indicated that the emission rates were
significantly affected by test plot and elapsed time. The results of
the statistical analysis are summarized in Table 4. Soil or air
temperatures within the chamber were marginally significant.
None of the other variables were found to have a significant af-
fect on the emission rates.
Table 4
Components of Variability
Source of Variability
Variance
Percent of
Component Total Variance
Percent of Total
Minus Temporal
Variance
Temporal (day-to-day) 58.3 63.0
Air Temperature
In Chamber 3.3 3.6
Plot 28.1 30.4
9.6
82.2
Sanpl Ing Location
(•Ithln the plot)
Surface Chamber
Shading
Sampl Ing/ Analytical
Total
1.4
0.42
1.0
92.5
1.5
0.4
1.1
100
4.1
1.2
7.9
100
a Variance components arc equal to standard dotations squared and ihus have units which are
the squares of that used for the log-emivsion rate% f>ig m--sccp
b Relative contribution of parameters to the total tanance
The oil and moisture contents of the soil samples were exam-
ined to determine whether there was any significant change with
time in these parameters. There appeared to be little change in the
oil content of the soil over the duration of the test. The moisture
content of the soil appeared to decrease by about 14 to 40% over
the test period. Both changes were so small that no correlation
with emission rates could be determined.
MEASURED VOLATILE ORGANICS
EMISSION RATES
The volatile organics emission rates were examined with respect
to elapsed time since application, ambient conditions at the time
of sampling and tilling activities.
A = Surface
3 = Background
C = Subsurface
01234 5678 9 10 11 12
Sampling Days
f = tilling episode
Figure 3
Measured Emission Rates as a Function of Time Elapsed
Since Sludge Application
130 ON-SITE TREATMENT
-------
Table 5
Average Measured Emission Rates by Plot by Half-Day
Table 6
Cumulative Measured Emissions
Fmlsslon Rate
Sampl Ing
Day
0
1
2
3
4
5
6
7
8
9
10
11
12
Date
10-09-84
10-10-84
10-10-84
10-11-84
10-11-84
10-12-84
10-12-84
10-13-84
10-13-84
10-23-84
10-23-84
10-24-84
10-24-84
10-25-84
10-25-84
10-26-84
10-26-84
11-08-84
11-08-84
11-09-84
11-09-84
11-10-84
11-10-84
11-12-84
11-12-84
A Plot
(Surface)
1.88
249
142
22.5
32.6
48.2
146
33.4
46.3
16.6
15.0
8.04
49.1
6.41
13.5
21.8
71.1
2.56
7.25
1.24
12.34
2.49
6.4
2.97
10.5
(uo/nT-s)
B Plot C Plot
(Background) (Subsurface) Comments
1.48
24.7
10.0
4.09
5.11
4.09
9.93
8.18
8.86
2.71
6.13
2.73
10.0
2.04
5.80
1.32
2.59
1.22
1.39
1.43
1.19
4.42
1.95
176
52.8
13.8 early start
26.5
120
295 till
71.2
67.4
12.4
15.1
10.1
91.3 till
12.1 early start
22.0
13.0
59.3 till
2.93
5.76
3.94
16.2 till
11.5
7.13
4.06
16.2 till
Each Plot Area = 423 m2
Effect of Elapsed Time from Application
on Emission Rates
The measured emission rates for the three plots are shown as
functions of elapsed time in Figure 3. The average measured emis-
sion rates are summarized for each half-day of sampling. These
rates are tabulated in Table 5. The first point (Day 0) was
measured before sludge application. The tabulated values and the
figure show the approximately exponential decline in emission
rates with time elapsed from application (and, to some extent,
elapsed time from tilling episodes). The emission rates from the
control Plot B remained substantially below those of the test plots
throughout the duration of the testing.
The cumulative measured volatile organics emissions are sum-
marized in Table 6 and shown graphically in Figure 4. An approx-
imately exponential decline in the emission rates is evident.
A = Surface Applied
B = Background Plot
C = Subsurface Injected
3456789
Sampling Days
10 11 12
Fml<;>:lon Rate (kn)
Sampl Ing
Day
1
2
3
4
5
6
7
8
9
10
11
12
A Plot
(Surface)
1.327
2.084
2.204
2.378
2.634
3.413
3.591
3.837
3.926
4.006
4.049
4.310
4.345
4.417
4.533
4.912
4.925
4.964
4.971
5.036
5.050
5.084
5.100
5.156
B Plot
(Background)
0.1317
0.1850
0.2068
0.2340
0.2558
0.3088
0.3524
0.3996
0.4140
0.4467
0.4613
0.5146
0.5254
0.5563
0.5634
0.5772
0.5837
0.5911
0.5987
0.6051
0.6286
C Plot
(Subsurface) Comments
0.938
1.219
1.293 early start
1.434
2.074
3.646 till
4.026
4.385
4.451
4.531
4.585
5.072 till
5.136 early start
5.254
5.323
5.639 till
5.655
5.685'
5.706
5.793 till
5.854
5.892
5.914
6.000 till
Each Plot Area = 423 m2
Effect of Tilling on Emission Rates
Tilling of the soil had a significant effect on the emission rate.
The effect can be seen clearly in Figure 3 as spikes in the emission
rates immediately after tilling. The percentage increase in average
emission rates by half-day immediately following tilling episodes
is shown for each plot in Table 7. The five tilling episodes for
which emission measurements were made were distributed among
the three sampling periods.
Table 7
Increase in Emission Rate Due to Tilling
Observed
Tilling
Episode
1
2
3
4
5
Percent
A Plot
(Surface)
303
611
326
995
354
Increase In Emission Rate
B Plot
(Background)
243
366
0*
0°
371
C Plot
(Subsurface)
246
904
456
411
399
Figure 4
Cumulative Measured Emissions as a Function of Sampling Day
(includes only measured rates for days on which sampling was performed)
a. Incomplete data set
ESTIMATED EMISSION RATE FOR
THE 5-WEEK TEST PERIOD
The emission rates and cumulative emissions for the entire
5-week test period were estimated from the measured rates. The
following assumptions were made to estimate the total emissions:
ON-SITE TREATMENT 131
-------
• Tilling had no effect on emission rates beyond 6 hours after
tilling occurred
• Each day had 12 hours of cooler temperatures (night)
The emission rates were estimated for three distinct times during
each sampling period. These were: (1) nighttime, (2) non-tilling
daytime and (3) tilling daytime. The nighttime emissions for each
sampling period were estimated from the emission rate data ob-
tained during the early morning hours of that period. The tilling
and non-tilling day emissions for each sampling period were de-
rived from the appropriate emission rate measurements. Emis-
sions for the interval between sampling periods were obtained by
interpolation among the three emission rates for the three sam-
pling periods. Summing all the night-time, non-tilling daytime and
tilling daytime emissions, the total emissions for each of the plots
for the 5-week period were estimated to be:
Plot A (surface applied): 33.3 kg
Plot B (control): 5.2 kg
Plot C (subsurface inj.): 39.0 kg
The estimated total emissions per plot per week are shown in
Figure 5.
£.\J
» 18
= o IK
eo* ID
1 Jf 14
sl 12
Q) 10 _
Estimated We
Organic Emi
to *. a> CD a
o
A = Surface Applied
„ B = Background Plot
A
77
I
I
//
B.
_/
^
^
§
|
^
C = Subsurface Injected
\
*y
s
^
N
\
ll
I/A/' \~ fVyfTPxJ f77t7~T^
12345
Week
Figure 5
Total Estimated Weekly Emissions from Each Plot
CONCLUSIONS
The results of the detailed sampling and analysis of total
volatile organic emissions from this land treatment facility in-
dicate that emissions from both surface-applied and subsurface-
injected waste varied in a similar, approximately exponentially
decaying manner over the 5-week period of this test. This result is
in agreement with the predictions of the Thibodeaux-Hwang
emissions model for surface-applied plots.'
Also, the effect of tilling is similar for both application
methods. Tilling resulted in a 2- to 10-fold increase in the emission
rates. A large portion of the 5-week cumulative emissions from
both plots is due to emissions which occurred within 4 hours after
tilling was performed. The magnitude of emissions which occur-
red immediately after tilling decreases from tilling event to tilling
event approximately in proportion to the decrease in emissions
during non-tilling periods.
Based on a statistical analysis of the emissions, it was found
that the emission rates were significantly affected by the elapsed
time. Soil and air temperatures within the chamber were mar-
ginally significant. None of the other variables that were mon-
itored were found to have a significant effect on the emission
rates.
NOTICE
The information in this document has been funded wholly or
in part by the U.S. EPA under contract to Radian Corporation.
It has been subject to the Agency's peer and administrative re-
view, and it has been approved for publication. Mention of trade-
names or commercial products does not constitute an endorse-
ment or recommendation for use.
REFERENCES
1. Radian Corporation, "Field Assessment of Volatile Organic Emis-
sions and Their Control at a Land Treatment Facility," Draft Re-
port, DCN 85-222-078-15-05, U.S. EPA Contract No. 68-02-3850,
Oct. 1985.
2. Hill, F.B., Aneja, V.P. and Felder, R.M., "A Technique for Meas-
urements of Biogenic Sulfur Emission Fluxes," J. Environ. Sci.
Health, AIB(3), 1978, 199-225.
3. Adams, D.F., Pack, M.R., Bamesberger, W.L. and Sherrard, A.E.,
"Measurement of Biogenic Sulfur-Containing Gas Emissions from
Soils and Vegetation," Proc. of 71sl Annual APCA Meeting, Hous-
ton, TX 1978.
4. Schmidt, C.E., Balfour, W.D. and Cox, R.D., "Sampling Tech-
niques for Emissions Measurements at Hazardous Waste Sites,"
Proc. of 3rd National Conference on Management of Uncontrolled
Waste Sites, Washington, DC, 1982, 334.
5. Brownlee, K.A., Statistical Theory and Methodology in Science and
Engineering, John Wiley & Sons, Inc., New York, NY, 1965.
6. Thibodeaux, L.J. and Hwang, S.T., "A Model for Volatile Chemical
Emissions to Air from Landfarming of Oily Waste," paper presented
at the Annual Meeting of the American Institute of Chemical En-
gineers, New Orleans, LA, Nov. 1981.
132 ON-S1TE TREATMENT
-------
Technology for Remediation of
Groundwater Contamination
David V. Nakles, Ph.D.
James E. Bratina
ERT, Inc.
Pittsburgh, Pennsylvania
ABSTRACT
The AquaDetox* technology is a high-efficiency stripping tech-
nology for the removal of organics from water. The technology
was developed by the Dow Chemical Company for internal use
through considerable process research, and only recently has the
technology become available for general use outside of Dow via a
licensing program.
Two types of AquaDetox* units are available, air or steam
AquaDetox*. Air AquaDetox* is capable of efficient removal of
chlorinated solvents and other pollutants considered easily strip-
pable from water. Steam AquaDetox* is used for the removal of
low-volatility organic materials from water including many com-
pounds not considered strippable by any prior process
technology.
The technology has a wide range of applications including
removal of 92 of the 111 organic priority pollutants as well as
other organic materials. It is applicable to the removal of trace
quantities of organics in water intended for drinking water pur-
poses with the capability of producing waters with organic con-
centrations from /ig/1 down to non-detectable levels.
This paper describes the AquaDetox* technology, describes its
potential applications and presents economic and operating data
from several operating commercial units.
INTRODUCTION
Extensive groundwater contamination from both active and in-
active sites has been identified across the United States. It has
been stated that more than 700 synthetic organic chemicals have
been discovered in groundwater resulting from over 30 different
types of sources of contamination. The House Public Works Sub-
committee of Investigations and Oversight has estimated from an
inquiry last year that 4,000 city drinking wells were affected by
hazardous waste seepage into groundwater. Hence, the subcom-
mittee has assessed groundwater contamination as potentially one
of the most expensive problems confronting Congress and the
country.
ERT has been involved extensively in the assessment of ground-
water contamination at a variety of active and inactive sites.
These efforts have focused largely on determining the nature and
extent of contamination, modeling the migration of contaminants
beyond the site boundaries, assessing risks associated with this
off-site migration and preparing conceptual approaches to
remediating the contamination as a means to reduce risk to accep-
table levels. To address the implementation of cost-effective
•Trademark and Service Mark of the Dow Chemical Company
remediation strategies, ERT has licensed a high efficiency
air/steam stripper known as the AquaDetox™ technology from
Dow Chemical.
This paper discusses the current regulatory climate governing
groundwater remediation, describes the AquaDetox technology
(including selected performance/economic case studies) and pro-
vides an overview of ERT's approach to offering it to the
industry.
Regulatory Overview
In response to the growing concern for groundwater quality,
the U.S. EPA has developed a national groundwater protection
plan which is derived from eight different laws passed by Con-
gress. This plan divides groundwater into three classifications
based on its use or potential use. There are specific groundwaters
that are irreplaceable drinking water sources and are particularly
vulnerable to contamination; groundwaters that are currently
used or available for drinking water; and groundwaters that are
not potential sources of drinking water because they already are
contaminated.
Under the various statutes, drinking water aquifers will be pro-
tected using RCRA (and its 1984 amendments) to discourage
pollution from landfills, the Superfund law to select areas for
urgent cleanup and TSCA to develop additional restrictions on
the use and storage of potentially threatening chemicals over these
areas. Other laws to be used by the U.S. EPA and state agencies
to control groundwater pollution are: the Clean Water Act under
its construction grants provisions; the Uranium Mining and Mill
Tailings Reclamation and Control Act (radioactive contamina-
tion); and the Federal Insecticide, Fungicide and Rodenticide Act
(pesticide contamination). Last, the National Environmental
Policy Act's sole source aquifer program and the Safe Drinking
Water Act's provisions for controlling underground injection will
also be used.
With regard to the Safe Drinking Water Act, current versions
of the Senate and House both refer to Phase I and Phase II lists of
contaminants which the U.S. EPA published in 1982 and 1983.
The Senate version of the Act requires the U.S. EPA to set stan-
dards for 85 contaminants with nine regulations due within a
year, another 40 within 2 years and the remainder within 3 years.
In January 1988, the U.S. EPA will be directed to publish a
new list of 25 priority pollutants with a year to propose recom-
mended and maximum contaminant levels and an additional year
to promulgate final standards. This process is to repeat on a
3-year cycle. The House version of the Act addresses 64 con-
taminants and dictates either the promulgation of regulations or a
ON-SITE TREATMENT 133
-------
determination that no rational basis exists to believe the contami-
nant may have adverse health effects. This action must be com-
pleted on 14 volatile organic compounds within a year, and deci-
sions on the remaining contaminants are due within 3 years.
Beginning in 1988, the bill demands annual priority lists for possi-
ble regulation, calling for standards within 3 years of publication.
Under either version of the Act, the U.S. EPA must propose
Filtration and disinfection as treatment for drinking water. The
target maximum contaminant levels will be set as close to health-
based recommended maximum levels as possible, considering cost
and available technology. In both cases, granular activated car-
bon (GAC) is suggested, though not mandated, as the "feasible"
technology. Generally speaking, however, GAC is a very expen-
sive treatment technology, and trade organizations such as the
Synthetic Organic Chemical Manufacturer's Association, as part
of a Safe Drinking Water Coalition, have stated that GAC is too
expensive and impractical. It is this same thinking which led to the
development of the AquaDetox stripping technology by the Dow
Chemical Company and ERT's participation as a Licensee of that
technology.
The AquaDetox™ Technology
The technology was developed by the Dow Chemical Company
for internal use through considerable process research and only
recently has it become available for general use outside of Dow
via a licensing program. AquaDetox achieves high stripping effi-
ciency at lower operating and capital costs when compared to
other technologies or unit operations. In many cases, removal of
organic contaminants, including many which have boiling points
in excess of 200°C and are typically considered "not strippable,"
can be achieved by a single, continuous stripping operation.
The process has a wide range of applications including removal
of 92 of the 111 organic priority pollutants as well as other
organic materials. It is applicable to the removal of trace quan-
tities of organics in water intended for drinking purposes with the
capability of producing waters with organic concentrations from
jig/1 down to non-detectable levels. The ability to achieve these
low levels of organics can eliminate the need for a polishing step
using GAC.
Two types of AquaDetox units are available as shown in
Figures 1 and 2. The air unit is capable of extremely efficient
removal of chlorinated solvents such as perchloroethylene, tri-
chloroethylene, 1,1,1-trichloroethane, methylene chloride and
other pollutants considered easily strippable from water. The
steam AquaDetox is used for the removal of low volatility organic
mateials from water including many compounds not considered
strippable by any prior process technology. For example, it has ef-
fectively removed chlorinated phenols (i.e., pentachlorophenol)
and chlorinated benzenes. The steam unit also removes easily
strippable pollutants from water when vent gas emissions from the
air process are unacceptable.
VENT TO ATMOSPHERE
RAW WATER -
AQUADETOX
UNIT
«. TREATED WATER
AIR •
Primary Advantages
The benefits afforded by the application of AquaDetox units
are many. First and foremost, it is a fully developed, proven
technology. The Dow Chemical Company has 12 units operating
at its Michigan Division treating a range of process wastewaters at
flow rates of 10 to 3,000 gal/min. In addition, four field applica-
tions of AquaDetox for outside firms have been installed and are
operating on contaminated groundwater.
AQUEOUS CONDENSATE
RAW WATER —
1
CONCENTRATE
ORGANIC
D
UNIT
OVERHEAD
— » TREATED WATER
STEAM
Figure 2
Steam AquaDetox Unit
The continuous nature of the technology offers simplicity of
operation as compared to batch processes and has low
maintenance requirements and low operating costs (as will be
demonstrated in the following section). The technology provides
maximum flexibility for the sizing and location of units for both
permanent and mobile installations. Finally, its ability to achieve
high-efficiency removal (99 to 9.99%) of organics to achieve at
least /ig/1 levels makes it a viable alternative to GAC, which is ex-
pensive, introduces solids handling equipment and poses disposal
and/or regeneration problems.
PERFORMANCE/ECONOMIC CASE STUDIES
Current applications of the AquaDetox technology have
demonstrated its ability to cost-effectively remove organics from
water. Three case studies emphasizing groundwater remediation
are briefly presented to demonstrate this point.
Case Study No. 1
An air AquaDetox was applied to groundwater contaminated
with 1,1,1-trichloroethane, methylene chloride, trichloroethylene
and perchloroethylene. The system was designed to treat a 10
gal/min stream and achieved extremely high removal efficiencies
for all pollutants as shown in Table 1.
The estimated capital cost for this installation was $80,000 with
an operating cost (including utilities, labor, maintenance, taxes
and depreciation) of $2.82/1000 gal (or $0.0028/gal). This com-
pared to an estimated GAC operating cost of $20.42/1000 gal for
the same level of pollutant removal.
Table 1
Performance Data: Air AquaDetox
Case Study No. 1
Pollutant
Pollutant levels, ppm
Contaminated
Croundwater Treated Effluent
Figure 1
Air AquaDetox Unit
1,1,l-Trichloroethane
Methylene Chloride
Trichloroethvlene
Perchloroethvlenc
220
180
0.08
non-detectable
non-detectable
non-detectable
134 ON-SITE TREATMENT
-------
Case Study No. 2
A steam AquaDetox system was installed to treat groundwater
contaminated with 6,000 mg/1 of a wide variety of organics shown
in Table 2. Typical pollutant removals from this 10 gal/min
stream exceeded 99% with several in excess of 99.99%.
Estimated capital cost for the unit was $225,000 with an
operating cost of $10.00/1000 gal (or $0.010/gal). Estimated
GAC operating costs to achieve these levels of pollutant removal
were $113 to $135/1000 gal of water.
Table 2
Performance Data: Steam AquaDetox
Case Study No. 2
Pollutant levels , ppm
Contaminated Treated Percent
Pollutant Groundwater Effluent Removal
Methvlene choride 18
1,1,1-trichloroethane 434
laopropanol 5674
Acetone 772
HER 252
2-butanol 399
MIBK 130
2-hexanol 21
0.0007 99.
0.0019 99.
0.191 99.
2.7 99.
0.01 99.
0.014 99.
0.002 99.
0.059 99.
996
999
996
6
996
996
998
7
Case Study No. 3
An air AquaDetox system was utilized to treat groundwater
contaminated with trace quantities of toluene (22,000 /ig/1) and
benzene (150 /tg/1). The technology achieved an effluent quality
with respect to both pollutants of less than 2 pg/1, the analytical
detection limits.
This removal was achieved at an estimated operating cost of
$1.35/1000 gal versus an estimated cost for carbon adsorption of
$16.00/1000 gal.
OTHER CONSIDERATIONS
One of the primary considerations when evaluating the use of
stripper technologies for groundwater contamination is the accep-
tability of the vent gas discharge. In general, this is evaluated on a
case-by-case basis at the state level, but it is estimated that less
than half-a-dozen states have any specific legislation at this time.
The crux of the matter centers on a cost-benefit analysis where the
trade-offs are between the additional cost to control vent gas
emissions (or avoid them entirely) using GAC and the health risk
to these uncontrolled emissions.
Since a number of industries already emit volatile organic car-
bon at legal levels much higher than those under consideration for
the groundwater cleanup situation, the key question becomes
whether a reduction of the already low health risk, especially with
the rapid dispersion of pollutants, is worth the price of control-
ling air emissions from groundwater cleanup. To put this issue in-
to perspective, it has been determined in many instances that the
concentrations of pollutants in the vent gas prior to their disper-
sion in the atmosphere do not exceed OSHA workplace standards
for the individual pollutants. Clearly, each application will have
to be reviewed individually with consideration given to the relative
health risk associated with the lack of control of such emissions.
In instances where vent gas control is dictated, the use of GAC
to remove the pollutants from the gas may be considered. Such a
combination of stripping and GAC has been applied on a Super-
fund site in Michigan and has proven cost-competitive with direct
adsorption of the organics from the water. This competitiveness
has been enhanced by the incorporation of humidity control prior
to the carbon beds since the adsorptive capacity of GAC is re-
duced dramatically when it gets wet.
CONCLUSION
As a process wastewater treatment technology, AquaDetox can
be used alone or in combination with GAC or biological treat-
ment systems to produce a more cost-effective or reliable
operating system.
ON-SITE TREATMENT 135
-------
Anaerobic Biological Treatment
of Sanitary Landfill Leachate
A.K. Mureebe, P.E.
D.A. Busch
P.T. Chen, Ph.D.
Wehran Engineering
Middletown, New York
ABSTRACT
The anaerobic treatment of sanitary landfill leachate was in-
vestigated in bench-scale studies at ambient room temperature.
The test runs were conducted in continuous operation for more
than 6 months using the leachate collected directly from the
leachate collection sump installed at a lined landfill site. No pre-
adjustments of pH or other chemical or physical features were
made. The leachate used in the study showed highly variable
characteristics with a COD of 2,000 to 35,000 mg/1, BOD5 of
1,800 to 24,000 mg/1 and ammonia nitrogen of 50 to 500 mg/1.
An anaerobic fixed-flim reactor, with a volume of 22.24 1 and
with a controlled rate of sludge recycling, was operated with
upflow feeding at organic loading between 0.23 and 2.48 kg
COD/mVday and hydraulic detention times of 6.9 to 10.3 days.
The test resulted in filtered COD removals from 92 to 94%, un-
filtered BOD5 removals of 90-94% and ammonia removal from
40 to 50%. Despite the wide range of loading conditions, the
system showed a high level of stability in terms of COD removal
efficiency which hovered about 93%. With influent COD of 1,700
to 17,000 mg/1, the effluent leaving the system was in the range of
100 to 1,300 mg/1, depending on the organic loading rate. The ef-
fluent can be further purified by an aerobic biotreatment unit to
reduce the COD and BOD5 levels below 250 and 35 mg/1, respec-
tively.
INTRODUCTION
The landfill is located off Route 17M east of the Wallkill River
in the Town of Goshen, Orange County, New York (Fig. 1). The
site was designed to protect surface and groundwater from
leachate contamination by utilizing the following technologies:
• Low permeability on-site glacial till material used as soil liner
to minimize the impact of leachate on adjacent surface and
groundwater
• A sloped landfill bottom to direct leachate toward a collection
system
• A leachate collection piping system to intercept and collect the
leachate
• A collection well with level-controlled pumps that transfer the
leachate collected in the landfill to a geomembrane-lined pond,
thus maintaining the hydraulic head in the landfill at a low level
The leachate management issues currently facing the landfill
are the establishment of a feasible and economically sound
disposal scheme that will meet the needs of the current operation
and future closure of the landfill, and the acquisition of permits
from state and/or local agencies for disposal of the leachate
generated.
Figure 1
Location Map — Turi Landfill
Wehran Engineering has been retained to conduct a study that
will provide a viable system for leachate treatment and to furnish
the necessary framework for implementation.
This paper includes a detailed investigation of all relevant
features and data pertaining to leachate treatment. Specifically,
the scope included the following:
• Analysis of leachate characteristics to determine the quality and
quantity of leachate generated by the site
• Performance of bench-scale treatability tests to confirm the
technical feasibility of on-site treatment options
• Development of engineering design criteria
Quantity and Characteristics of Leachate
An extensive monitoring, sampling and analysis program was
conducted for a period of 1 year at Turi Landfill to determine the
quantity and quality of leachate generated from the operational
cell of the landfill. Weekly and monthly samples were collected
during this period and analyzed in the laboratory.
The weekly samples were analyzed for the conventional pollu-
tant parameters of chemical oxygen demand (COD), 5-day bio-
136 TREATMENT OF HAZARDOUS WASTES
-------
chemical oxygen demand (BOD5), total solids (TS), ash, total
Kjeldahl nitrogen (TKN), ammonia nitrogen (NH3-N), hard-
ness, conductivity and pH.
The monthly samples were analyzed for iron (Fe), zinc (Zn),
copper (Cu), cadmium (Cd), lead (Pb), manganese (Mn) and total
phosphate (T-PO4).
Leachate Flow
The flow of leachate generated in the active landfill operation is
shown in Table 1. The actual monthly average daily leachate flow,
measured in the operating landfill area for the period between
April 1984 and March 1985, ranged from 100 to 18,000 gal/day,
with an annual average of 6,300 gal/day. The variation in flow
probably is due to the absorptive characteristics of the solid
waste.
Characteristics of Leachate
Starting in November 1984, sampling and analysis of the con-
ventional pollutant parameters described above were performed
as part of the study program. In conjunction with this sampling
and analysis, leachate flows were recorded weekly.
With the exception of pH, the pattern of changes in the concen-
trations of these pollutant parameters appears to be in general
correlation with the changes in leachate flows. The concentration
changes are inversely proportional to the leachate flow rates
Table 2
Leachate Characteristics - Turi Landfill
Parameter
BOD5
COD
Am monia Nitrogen
Organic Nitrogen
Total Solids
Ash
Hardness
Conductivity
pH
Range
180- 36,900 mg/1
200 -45,800 mg/1
5 - 640 m g/1
1 - 140 mg/1
1,005-25,330 mg/1
805- 13,580 mg/1
345 - 9,460 mg/1
900 - 15,500 u mhos/cm
5.30 - 6.00
Arithmetic
Average
12,042 mg/1
19,527 mg/1
340 mg/1
46 mg/1
14,367 mg/1
7,037 mg/1
5,647 mg/1
9,360 u mhos/cm
5.55
The monthly results for the metals and total phosphate analyses
are summarized in Table 3. These concentrations are comparable
to reported ranges for sanitary landfills.4 All concentrations
observed in this study are below reported threshold limits for
biological assimilation in a wastewater treatment facility.''4'8
Table 3
Monthly Analysis of the Concentration
of Metallic Elements and Phosphates
wmcn may DC aunouiea 10 me percoiaiion or ramiau. i ne exier-
nal water percolation in the landfill will thus act as a dilutant in
carrying leachables out of the landfill bed.
Table 1
Leachate Generation Rates
Measured Values Estimated Values (5)
Ppt(l) Perc (2) Ratio Flow (4) Ppt Perc Ratio Flow
Month (inch) (inch) (3) (gpd) (inch) (inch) (3) (qpd)
April 84 4.65 1.080 0.23 12,840 3.86 2.12 0.55 25,135
May 10.03 — -- — 4.22 1.48 0.35 17,548
June 3.95 0.270 0.07 3,212 3.86 1.35 0.35 16,006
July 6.27 0.430 0.07 4,920 4.01 1.41 0.35 16,717
Aug. 2.61 0.290 0.11 3,367 3.39 1.19 0.35 14,109
Sept. 1.18 -- - - 3.01 1.06 0.35 12,567
Oct. 2.88 — — — 3.03 1.67 0.55 19,800
Nov. 2.58 0.008 0.03 92 2.88 1.59 0.55 18,852
Dec. 2.99 0.540 0.18 6,139 2.46 2.09 0.85 24,780
Jan. 85 1.06 0.127 0.12 1,463 2.46 1.23 0.50 14,584
Feb. 1.99 1.430 0.72 18,192 2.68 1.34 0.50 15,888
March 2.21 0.516 0.23 5,916 3.88 3.30 0.85 39,127
Annual 42.4 4.7+ — -- 39.74 19.83
Average .19 6,239 .50 19,584
Monthly
Notes:
(1) Precipitation (ppt) from climatological data, Middletown Station, NY, NOAA.
(2) Calculated percolation (perc) through cover soil and underlying solid wastes based on actually
measured leachate flow within 13.1 acres of operating landfill.
(3) Ratio of perc to ppt.
(4) Actual leachate flow measured.
(5) Calculated leachate flow, based on the estimated values of percolation for 13.1 -acre area.
Metallic
Elements 10/25/84 12/20/84 1/24/85 2/27/85 4/4/85
Fe 770.00 420.00 1,100.00 1,140.00 840.00
Cu 0.05 0.08 0.05 0.05 .35
Pb 0.05 0.05 0.16 0.05 .14
Mn 190.00 41.00 310.00 200.00 200.00
In 11.00 14.00 33.00 6.30 11.00
Cd 0.01 0.01 0.01 0.01 .01
T-P04 8.40 1.90 3.40 2.50 6.50
Note: All units are in mg/1
The concentrations of priority pollutants which were present in
the leachate at levels above analytical detection limits are given in
Table 4. All organic chemicals on the list are reported to be effec-
tively removed by conventional biological treatment processes at
both POTWs and industrial wastewater treatment facilities.'
Leachate Treatability Evaluation
A review of the literature reveals that leachate may be treated
by aerobic1 '2'3'4'6'9 or anaerobic biological processes7'8, physical
methods and chemical means.4-9 Existing reports discuss mainly
laboratory-scale studies.
Bench-Scale Treatability Tests
In the selection of systems for a treatability study, considera-
tion was given to the processes available at local POTWs, in-
dustrial wastewater treatment facilities and for the purchase of
packaged treatment systems for possible on-site use. The follow-
ing process techniques were chosen for the bench-scale treatability
studies:
• Aerobic bio-oxidation system utilizing local POTW activated
sludge for the evaluation of maximum ratio of leachate to
The range and average concentrations for the conventional
pollutant parameters during the study's test period are given in
Table 2. The ranges are comparable with reported ranges for
sanitary landfills published in technical literature.4'7
sewage flow
Anaerobic biological system for treating raw leachate
Aerobic treatment for polishing the effluent of the anaerobic
system
Physical/chemical treatment for removing solids or COD as a
pre-treatment or post-treatment process
TREATMENT OF HAZARDOUS WASTES 137
-------
Table 4
Concentrations of Detectable Priority Pollutants
in Turi Landfill Leachatc
Parameter
Benzene
Chloroethane
Dichlorodifluoro methane
1,1-Oichloroethane
1,2-Dichloroethane
1,2- Dichloro propane
Ethylbenzene
Methylene Chloride
Tetrachloroethylene
Toluene
1,2-Trans-Dichloroethylene
1,1,1-Trichloroethane
Trichloroethylene
Vinyl Chloride
Diethyl Phthalate
Isophorone
Naphthalene
Arsenic
Cadmium
Chromium
Copper
Lead
Nickel
Selenium
Zinc
Total Cyanide
Total Phenolics
Reported
Concentration
(mg/1)
0.025
0.062
0.089
0.301
0.081
0.015
0.051
2.600
0.079
1.000
0.228
0.111
0.130
0.010
0.076
0.067
0.010
0.018
0.015
0.240
0.027
0.090
0.390
0.007
12.000
0.100
4.000
The bench-scale treatability tests were performed using the
following equipment:
• For anaerobic biological system: A fixed film upflow reactor,
6 in. diameter x 5 ft high with hydraulic detention time of
7 days, was used in a plug flow mode with recycle (Fig. 2).
• For aerobic bio-oxidation system: A 4-1 reactor with air dif-
fusers was used in a completely mixed mode. Sludge was
wasted directly from the unit or recycled from a settling unit
to maintain the desired MLVSS under aeration (Fig. 3).
• For physical/chemical test: A multiple stirrer, Phipps-Bird
stirring apparatus was used to test the effect of coagulation/
precipitation with lime, caustic and/or alum.
POTW ACTIVATED SLUDGE
TOLERANCE STUDY
In this test, activated sludge seed from the Harriman Waste-
water Treatment Facility was aerated in a bio-oxidation unit. The
system was run to a steady state MLVSS with a sewage and
nutrient feed prior to the addition of raw leachate. Once the
bench-scale activated sludge system achieved this steady state, raw
leachate was added in ratios of raw leachate to sewage of 1 to 5%.
The results of this study are given in Table 5. At dosage ratios
of raw leachate to raw sewage greater than 2%, the aerobic treat-
ment system is adversely impacted. The system failed to maintain
a steady state MLVSS at a 5% dosage, and higher order organ-
isms were observed. At the high dosage levels, increased phos-
phate addition on the order of 1,000 mg/1 was necessary for ade-
quate cell development.
r "
^-U
x
\ /
A
H
^
i — >>
^
i
"^x j
e ai 3 H <|
,4..
H ,
'
r-
k.
rOAS/LIOa/SOL
SEPARATOR
I •'» i a' H
r
L
Oil
METER
1
i
f
-| 1
10 1
^
Ln
i
T
AAS
ANALYZER
PUMPIN8 RATE
CONTROL
PANEL
'Q~\
I
RECYCLE PUMP
FEED FEED 31UDOE EFFLUENT
PUMP TANK TANK TANK
Figure 2
Bench-Scale Anaerobic Fixed Film Biological Treatment System
EFFLUENT
TRAP BOTTLE
Figure 3
Aerobic Bio-Oxidation Treatment Unit
Table S
POTW Tolerance S(ud> of Aerobic
Bio-Oxidation System Treatment of Leachate
Raw Leichatc,
Raw Sewage
•Ai 5°'o ihc system coult) not aliam u steady stale Ml VSS
Average COD
Removal
81
70
54
37
12
ANAEROBIC TREATMENT
The parameters monitored in studying the efficiency of the
anaerobic biological treatment system were COD, BOD5 and
TVA. The total volatile acid removal (TVA) is an important step
in methane fermentation and is therefore included. The results for
the removal of these parameters are summarized in Table 6.
In evaluating the anaerobic treatment of landfill leachate, the
removal of COD and BOD are the most important parameters to
study. As can be seen from the data in Table 6. the efficiency of
138 TREATMENT OF HAZARDOUS WASTES
-------
Table 6
Anaerobic Test Results of Leachate Treatment
Run
1
2
3
4
5
6
Influent
--
4,500
8,775
11,250
11,625
8,700
BOD<;(mq/l)
Effluent
—
390
855
960
1,200
540
X Removal
—
91.3
90.3
91.5
89.7
93.8
Influent
1,670
6,480
13,430
16,990
16,990
9,875
COO (rag/1)
Effluent
100
355
935
950
1,275
545
X Removal
94.0
94.5
93.0
94.4
92.5
94.5
TVA
Influent
1,730
6,520
12,300
14,485
14,160
8,060
(rag acetate/1)
Effluent
405
660
1,545
925
1,710
205
% Removal
76.6
89.9
87.4
93.6
87.9
97.5
•TVA is total volatile acids.
the removal of these two parameters is quite constant. For Run
No. 1, which had an organic loading rate of 0.23 kg COD/m3
• day, the COD removal was 94%, while with an organic loading
rate of more than 10 times that in Run No. 1, the COD removal
rate (Run No. 5) was 92.5%. The system showed an ability to
operate efficiently under a wide range of COD loading condi-
tions, which is an important characteristic in the treatment of
landfill leachate. In conjunction with the COD removal, the study
of two other system paramters is important: methane and sludge
production.
Methane Production
One of the major advantages of the anaerobic process is that
organic material in the wastewater is converted to methane gas.
The methane gas evolution rates and composition for the
anaerobic treatability study are summarized in Table 7.
Table 7
Gas Production Data for Anaerobic Treatment of Leachate
Run
1
2
3
4
5
6
Gas Evolution
Rate
(ml/hr)
125
500
745
965
1,350
745
Methane
(%)
67
71
64
70
67
—
Methane Rate
(I/day)
(at STP)
1.84
7.81
10.48
14.85
19.89
10.97
As can be seen, methane production rates determined from ex-
perimental removal of organics are higher than theoretical rates.
This may be a result of average daily influent conditions actually
being higher than those detected in the samples analyzed; this
would mean that the organics removal rate is higher than the rate
used in the calculations. An important finding of this portion of
the study is that the methane production rate was directly propor-
tional to the organics loading rate.
Solids Production
Another advantage of the anaerobic system is the low solids
production rate. Sludge production from anaerobic biological
treatment systems typically ranges from 0.04 to 0.10 g VSS/g
COD removed. The sludge production rates as a function of
organics removal are summarized in Table 9. The amounts of
solids produced were estimated from the increase in volatile solids
concentrations in the circulation stream.
Table 9
Anaerobic Sludge Production in Treatment of Landfill Leachate
Run (g/day)
2.82
2.81
1.67
6.67
COD
Removed
(g/day)
18.52
26.99
34.64
50.92
TVS/ COD
(g solid/g
COD
removed)
.15
.10
.05
.13
To determine the efficiency of the methane fermentation, the
experimental gas production rate can be compared to the
theoretical value which is 0.35 1 CH4/g of organics removed. The
methane production rates are shown as a function of the removal
of BOD, COD and TVA in Table 8.
Table 8
Methane Production Rates for Anaerobic Treatment
of Landfill Leachate
Gas Rate as a Function of Organic Removal
Average
.11
Run
1
2
3
4
5
6
Liters CH4/
g COD removed
0.39
0.42
0.39
0.43
0.39
0.36
Liters CH4/
g BOD removed
0.53
0.61
0.67
0.59
0.41
Liters CH4/
g TVA removed
0.46
0.44
0.45
0.51
0.49
0.43
The sludge production of the anaerobic digestion of the
leachate was slightly higher than normal anaerobic sludge produc-
tion, yet the amount of sludge was still significantly lower than
the amount of sludge produced by a conventional activated sludge
process.
AEROBIC TREATMENT OF
ANAEROBIC EFFLUENT
The aerobic treatment of the effluent from the anaerobic
system was conducted in a fill and draw manner. Anaerobic ef-
fluent was added to activated sludge from the Harriman Sewage
Treatment Plant, and the system was aerated. Samples were
drawn from the system at 12, 24 and 48 hrs for analysis. The
characteristics of these samples were compared to the anaerobic
effluent that was added initially. The test was conducted with
COD and BOD values of 1,300 and 200 mg/1. The MLSS, MLVS
and SVI values were 8,100, 3,950 and 135, respectively.
TREATMENT OF HAZARDOUS WASTES 139
-------
Table 10
Results of Aerobic Treatment of Anaerobic System Effluent
Hour
0
12
24
48
COD
1,300
400
260
235
COD X
Removed
69.2
80.0
81.9
BOD
1,200
100
70
35
BOD X
Removed
__
91.7
94.2
97.1
NH3-N
135
110
45
5
NH3-NX
Removed
__
18.5
66.7
96.3
COO/
BOO^
1.1
4.0
3.7
6.7
The removal of the conventional pollutants (COD, BOD and
NH3-N) was monitored to determine the efficiency of the pro-
cess. The results of these tests are summarized in Table 10.
Removal of BOD and COD
In discussing the removal of organics by the aerobic system, an
important parameter to consider is the ratio COD/BOD5.
The general trend is an increase of the COD/BOD ratio with in-
creased aeration. This increase is due to a rise in the non-
biodegradable portion of the leachate over time. The test for
chemical oxygen demand will detect some organics which are non-
biodegradable and thus will not be measured by the BOD test. As
the aeration time increases, the biodegradable organics will be
removed by the system microorganisms, and the non-biodegrad-
able organics will represent an increasing proportion of the re-
maining organics.
Based on the results of this test, we concluded that a removal
rate of 80% of the COD and 95% of the BOD is attainable by the
conventional aerobic treatment systems.
Removal of Ammonia
The aerobic system removes ammonia by the process of
nitrification; the ammonia is oxidized to nitrate. The bacteria
responsible for nitrification have a growth rate slower than that of
the bacteria responsible for the removal of organics. Thus, it is ex-
pected that a longer aeration time will be required for the removal
of NH3-N than that for BOD. As can be seen from the test
results, the removal of BOD was as high as 90% in only 12 hr,
while an aeration time of 48 hr was needed to remove 96% of the
ammonia.
average sludge production rate of 0.68 g VSS per g COD re-
moved. This increased sludge production is attributed to the ex-
tended aeration time and the lack of sludge wasting in the test to
maintain a steady state MLVSS. The results do indicate that
under typical conditions for aerobic biological treatment, the
sludge production from treatment of the leachate would be within
the expected range for such systems.
CHEMICAL COAGULATION/PRECIPITATION
The effect of chemical coagulation and precipitation on COD
and suspended solids removal also was evaluated in this study.
The test conditions for the study are summarized in Table 12. In
this test, the samples were dosed with the coagulating chemicals,
rapidly mixed using a Phipps-Bird stirring apparatus at 50
revs/min for 30 sec, flocculated at 15 revs/min for 20 min and
then allowed to settle for 30 min. The supernatants then were
sampled and analyzed for COD and TSS.
Table 12
Chemical Coagulation/Precipitation
Treatment Operating Conditions
Waste Stream
Raw Leachate
Anaerobic Effluent
Aerobic Effluent
Ca(OH)2
(•g/1)
1. 800-4200
2. 2800
3. --
1. 1800
2. --
3. --
4. --
Cnemical Coagulant Dosages
NaOH
(•9/1)
1200
Alum
0-333
0-333
0-333
0-333
0-333
0-2000
0-1455
0-1500
0-2000
The results for the chemical coagulation/precipitation
treatability studies on the raw leachate, anaerobic filter effluent
and aerobic polishing unit effluent are presented in Tables 13, 14
and 15, respectively.
Solids Production
The production of sludge is an important parameter in the design
of the aerobic process. Table 11 contains a summary of the solids
production as a function of the amount of organics removed.
The reported range of sludge production for activated sludge is
0.25 to 04 g of VSS formed per gram of COD removed. The
results of the bench-scale aerobic polishing unit indicate an
Table 11
Aerobic Sludge Production for
Treatment of Anaerobic System Effluent
Time
(hrs)
0
12
24
48
Average
COO
(in 9/1)
1,300
400
260
235
MLVSS
(rcg/1)
1,610
1,860
2,030
3,060
COO
Removed
(mg/1)
__
900
1,040
1,065
MLVSS
Increase
(mg/1)
._
250
420
1,450
MLVSS/
COD
--
.28
.40
1.36
.68
Table 13
Results of Chemical Coagulation/Precipitation
Treatment of Raw Leachate
Chemical Coagulant
Dosages (ng/1)
Run
1
2
3
Ca(OHj>
0
830
1,670
2,500
3.330
4,170
0
2,800
2,800
2,800
2,800
2,800
__
.-
—
—
—
--
NaOH
„
_.
..
.-
,_
--
__
-.
..
._
.. .
--
0
1.200
1,200
1,200
1,200
1,200
Alum
..
._
„
..
__
--
0
0
33
83
167
333
0
0
33
83
167
333
COD
(•q/1)
31.740
32.105
32,230
32.350
31,130
31,660
31,335
30.805
31,700
30,600
31,660
30,805
26,110
22,010
22,560
22,440
23.665
--
Parameters
Percent
Removed
..
..
1.9
0.3
..
1.7
_.
2.4
.,
1.7
..
15.7
13.6
14.1
9.4
--
TSS
(•g/il
3.700
1,400
900
—
5
600
3,780
2,045
2,585
3,050
1.540
1,495
4,915
3.340
4,535
3,560
1,765
955
Percent
Removed
„
62.2
75.7
—
99.9
83.8
_.
45.9
31.6
19.3
59.3
60.5
„
32.0
7.7
27.6
64.1
80.6
140 TREATMENT OF HAZARDOUS WASTES
-------
Table 14
Results of Chemical Coagulation/Precipitation Treatment
of Anaerobic Filter Effluent
Table 16
Operating Conditions and Performance of
Bench-Scale Anaerobic Filter
Chemical
Coagulant
Dosages (mg/1)
Run
1
2
3
4
Ca(OHt>
0
1,800
1,800
1,800
1,800
1,800
__
—
—
—
—
—
~
—
—
—
—
Alum
0
0
33
83
167
333
0
33
83
167
250
333
0
83
167
333
0
1,000
1,500
2,000
COD
(mg/1)
485
600
600
620
550
625
610
620
575
560
555
530
980
955
955
880
685
635
565
585
P ara m eters
Percent
Removed
—
—
—
--
—
__
—
5.7
8.2
9.0
13.1
--
2.6
2.6
10.2
—
7.3
17.5
14.6
TSS
--
45
5
35
0
35
120
80
40
70
45
50
39
12
19
37
45
Percent
Removed
—
—
__
—
--
33.3
66.7
41.7
62.5
58.3
..
69.2
51.3
5.1
—
Operating Conditions Removals
Hydraulic
Oet. Time
(days)
Run 1 7.7
Run 2 7.3
Run 3 10.3
Run 4 10.3
Run 5 6.8
9 V5/
kg COO/ g COD rem.
m 3 day "ercent Solids
Org. Loads COD rem.
0.216 94.0
0.419 94.6
1.302 93.7
2.100 94.0
2.380 92.0
Prod.
-
0.150
0.100
0.040
0.112
Products
LCH4/gCOD
rem.
0.478
0.458
0.418
0.497
0.434
LCH4/g
VFArem.
0.760
0.672
0.504
0.596
0.656
* liters of methane produced/g COD removed
Summary
Table 17
of Performance of the
Four Treatment Systems
Performance
8
—
82.2
Process
COD
(X)
in Removal of
NH3-N
(%)
Table 15
Results of Chemical Coagulation/Precipitation Treatment
of Aerobic Polishing Unit Effluent
Run
Chemical
Coagulant
Alum
0
545
910
1,455
0
500
1,000
1,500
0
1,500
2,000
Parameters
COD
235
220
205
180
320
310
270
245
485
360
330
Percent
Removed
„
6.4
12.8
23.4
__
3.1
15.6
23.4
__
25.8
32.0
TSS
..
--
—
--
70
50
20
20
240
2
--
Percent
Removed
..
—
—
--
28.6
71.4
71.4
__
99.2
--
Removal of COD
Raw Leachate
Of the three tests run, only the addition of caustic and alum
showed any significant reduction in COD, and this was less than
20%. The addition of lime only, and lime and alum showed no
appreciable decrease in COD from pH elevation or coagulation.
The results of the third run indicate that the best COD removal
was obtained with caustic alone, and that alum was ineffective as
a coagulating agent for COD removal.
The results obtained in this study indicate that chemical
coagulation and precipitation processes do not appear to be ade-
quate as a pretreatment process for disposal of the leachate either
off-site at a POTW or through an on-site treatment system.
Anaerobic Filter Effluent
The results of the chemical coagulation and precipitation of the
anaerobic filter effluent indicate that large dosages of alum are re-
quired to obtain moderate removal efficiencies. The results of the
second and third runs indicate that a dosage of 333 mg/1 of alum
results in a 10-13% removal of COD. By increasing the dosage to
Anaerobic system
Aerobic system
Alum coagulation
Integrated efficiency
94 49
80 95
25 --_
99.1 97.5
1,500 mg/1 in Run No. 4, a removal of only 17.5% is attained.
These results indicate that an increase in the alum dosage of 500%
results in only a 5% increase in COD removal. This process is,
therefore, not considered effective for reducing residual COD
from the anaerobic filter effluent.
Aerobic Polishing Unit Effluent
The results presented in Table 15 indicate that very high levels
of alum (1,500-2,000 mg/1) will remove only 25-30% of the COD
in the aerobic polishing unit's effluent. This process, therefore, is
not considered effective for removing any residual COD in the
aerobic polishing unit, should such removal be essential to meet
discharge standards.
Removal of Suspended Solids
Raw Leachate
The test results presented in Table 13 indicate that effective
suspended solids removal in the raw leachate is achieved only with
very high dosages of lime (3,330-4,170 mg/1). Such high dosages
generate proportionally high sludge volumes which would require
further processing for proper disposal.
Based on these results, we concluded that this process is not ef-
fective as a pretreatment process for disposal of the leachate
either off-site at a POTW or through an on-site treatment system.
Anaerobic Filter Effluent
The results presented in Table 14 indicate that effective removal
of suspended solids in the anaerobic filter effluent is achieved
only with high levels of alum (2,000 mg/1). At lower dosages
(33-83 mg/1), moderate removals were attained, indicating that
this process may be viable for solids removal in conjunction with
a clarification system, should subsequent biological or
physical/chemical treatment for residual organics and ammonia-
nitrogen require pretreatment for solids removal.
TREATMENT OF HAZARDOUS WASTES 141
-------
Aerobic Polishing Unit Effluent
The results presented in Table 15 indicate lhat relatively high
levels of alum (1,000 mg/1) effectively reduce the suspended solids
level in the aerobic polishing unit's effluent to acceptable
discharge standards (20 mg/1). Full-scale clarification units, with
tube or plate settlers, may reduce the chemical dosage required to
achieve this effluent quality. This process is considered viable for
consideration as a polishing treatment for suspended solids
removal.
CONCLUSIONS
The laboratory bench scale study using an anaerobic fixed-film
reactor for the treatment of sanitary landfill leachatc
demonstrated the amenability of the process and its application
for the stabilization of leachate.
The landfill raw leachate was rapidly adopted by the sludge
used in the anaerobic fermentation system. Hydraulic detention
times as short as 7 days and organic loads up to 2.2 kg COD/m'
day resulted in a 94% removal of COD. The gas produced in the
system was composed of 60-70% combustible material. Based on
the rate of gas production, we estimated that more than 99% of
the COD removed was converted to CH4 gas.
The process performance with respect to COD removal and gas
production at various loading conditions is shown in Table 16.
The removal of COD and the rate of solids production were very
steady. The response of process performance to changes in
organic or hydraulic loads was reflected in the rates of methane
production; however, the efficiency of COD removal remained
unaffected for the conditions tested.
The performance of the activated sludge system for treating the
effluent from the anaerobic filter was much better than anaerobic
treatment alone. With a loading of 0.90 g COD/g MLVSS, the
system attained removals of 80 and 95% for COD and NH, - N,
respectively, within 48 hr of aeration. The estimated rate of sludge
formation in the aerobic system was 0.41 g/VS/g COD removed.
The series of tests conducted in the bench-scale treatability
study have confirmed that the landfill leachate can be treated with
the use of conventional biological and physical/chemical pro-
cesses. The effluent qualities achievable are comparable to the ef-
fluent of wastewater treatment plants with secondary treatment
facilities. Since the processes tested are all conventional
technology, the equipment required for use is commercially
available.
Based on the results of each test included in this treatability
study, a combination of anaerobic and aerobic treatment in series
alum coagulation could achieve more than 99% removal of COD
from raw leachate.
REFERENCES
I. Boyle. W.C. and Ham, R.K., "Biological Trealabilily of Landfill
Lcachatc." JWPCI. 46. 1974, 860.
2. Cook, O.M. and force, G.G., "Aerobic Biostabilization of Sanitary
Landfill Leachate," JWPCI. 46, 1974, 380.
3. Uloih, V.C. and Max ink, "Aerobic Biotreatment of a High Strength
Lcachalc," J. Environ. EnK. Dt\ . ASCE. EE4, 1977, 841.
4. Chain, E.S.K. and DcWalle, F.B.. "Evaluation of Leachate Treat-
ment." U.S EPA Report, Washington, D.C.
5. Zapf-Gilzc, R.. "Temperature Effects on Biostabilization of Leach-
ate," J. Environ. Eng. Di\'.. ASCE. EE4, 1981. 653.
6. Ehrig, H.J. and Slegcman. R.. "Treatment of Sanitary Landfill
Leachate: Biological Treatment." Waste.
7. Force, E.G. and Reid, V.H., "Anaerobic Biological Stabilization of
Sanitary Landfill Leachatc," Technical Report URY TR 65-73-CG17,
1973.
8. Wu, V.C. and Kennedy. J.C., "Anaerobic Treatment of Landfill
Leachate by an Upfloxx Two-Stage Biological Filter," Proc. Fixed-
Film Bioiechnol. Seminar, 1984.
9. Ho, S., Boyle. W.C. and Ham. R.K.. "Chemical Treatment of
Leachate from Sanitary Landfill." JWPCF. 46. 1974, 1776.
142 TREATMENT OF HAZARDOUS WASTES
-------
On-Site Versus Off-Site Treatment
of Contaminated Groundwater — An Evaluation
of Technical Feasibility and Costs
Kent L. Bainbridge
Daniel W. Rothman, P.E.
URS Company, Inc.
Buffalo, New York
ABSTRACT
The Pollution Abatement Services, Inc. (PAS) site is an aban-
doned hazardous waste incinerator site located in the City of
Oswego, New York. Groundwater contamination has resulted
due to leakage from a surface impoundment, buried tanks and
drums—both above and below ground. The principal ground-
water contaminants include volatile organic compounds, organic
acids and other low molecular weight organics, and nickel, which
also appears to be organically bound.
Groundwater recovery and leachate collection were included in
the site remediation plan. Alternatives considered for treatment
of the groundwater/leachate were: (1) hauling to an off-site com-
mercial treatment facility, (2) conveyance to the local POTW and
(3) on-site treatment. The relative merits of these alternatives were
evaluated on the basis of the degree of treatment technically
achievable and the relative cost.
Critical to the evaluation was the identification and demonstra-
tion of technology capability for treatment of the groundwater/
leachate. Initial laboratory treatability studies demonstrated that
biological treatment followed by carbon adsorption was required
for cost-effective treatment, while carbon adsorption alone was
not effective. Further studies demonstrated the feasibility of the
treatment system on an intermittent basis—a necessary require-
ment based on the expected low rate of groundwater/leachate
recovery.
In this analysis, the costs and ultimate selection of a ground-
water/leachate treatment alternative were sensitive to several fac-
tors including: (1) the wastewater character, (2) the wastewater
volume and rate of recovery and (3) the degree of treatment re-
quired. On-site treatment was offered a higher degree of treat-
ment than would be achievable at the local POTW and a lower
cost than would be expended for hauling to an off-site commer-
cial facility.
INTRODUCTION
The Pollution Abatement Services, Inc. (PAS) site, located
near the eastern limit of the City of Oswego, New York (Fig. 1), is
an abandoned hazardous waste facility which is listed among the
top 10 priority sites on the U.S. EPA's National Priorities List.
During the period from 1982-84, URS Company, Inc. was con-
tracted by the New York State Department of Environmental
Conservation (NYSDEC to perform a Remedial Investigation/
Feasibility Study (RI/FS) of the site. One result of the field in-
vestigation was the determination that the groundwatei underly-
ing the site and leachate taken from two on-site drainage ditches
was highly contaminated by organic priority pollutants.
Hydrogeological data collected during the RI/FS supported the
conclusion that all subsurface contaminant migration was being
Figure 1
Location of the Pollution Abatement Site
intercepted by two creeks which bounded the site on three sides.
Based on these findings, the RI/FS report recommended specific
remedial measures for subsurface cleanup and closure of the PAS
site. These measures included construction of a leachate collec-
tion system, installation of a groundwater recovery system and
treatment of the groundwater/leachate.1
As part of the subsequent Remedial Design Study, laboratory
treatability studies were conducted to identify process re-
quirements to determine the level of treatment attainable and to
develop design and cost data. Results of the treatability studies
are reviewed in this paper. Estimates of the cost of on-site treat-
ment also are presented. Finally, the on-site and off-site treatment
methods are compared to identify the most cost-effective treat-
ment alternative.
TREATMENT OF HAZARDOUS WASTES 143
-------
SITE HISTORY
The facility known as Pollution Abatement Services, Inc.
(PAS) was constructed and put into operation in 1969-70. A high
temperature, liquid chemical waste incinerator was the principal
unit installed by the owners. Throughout its active life, PAS ex-
perienced continuous operating problems, numerous air and
water quality violations and mounting public opposition.
During its operating period (from 1970 through 1977), large
numbers of drums containing various chemical wastes were col-
lected and stored on-site. Liquid wastes also were stored in
lagoons.
Beginning in 1973, a series of incidents including liquid waste
spills and overflowing of lagoon wastes into the adjacent White
Creek led to the involvement of the U.S. Coast Guard, the U.S.
EPA and the New York State Department of Environmental
Conservation (NYSDEC). This involvement included a number
of limited and temporary remedial actions during the period from
1973-1976. In 1977, PAS was abandoned.
During the several years immediately following abandonment,
a number of emergency remedial actions and preliminary in-
vestigations were undertaken at the PAS site including draining
and filling the waste lagoons and removing approximately 3,000
leaking barrels and several waste storage tanks from the site. Late
in 1981, contract documents for the surficial cleanup of PAS were
prepared, including the demolition and disposal of on-site
facilities, the removal of the remaining (approximately 8,000)
drums from the site and the drainage and disposal of approx-
imately 80,000 gal of liquid chemical waste from 10 bulk storage
tanks. This surficial cleanup was completed in October 1982.
In November 1982, URS Company, Inc. (URS) entered into a
contract with NYSDEC to evaluate alternative remedial measures
for final cleanup of buried wastes, contaminated soil and ground-
water and site closure.
TREATMENT ALTERNATIVES
Alternatives considered for treatment of the contaminated
groundwater/leachate were: (1) hauling to an off-site commercial
treatment facility, (2) treatment at a local POTW and (3) on-site
treatment. Two commercial hazardous waste disposal facilities
are located approximately 180 miles from the PAS site. The
Oswego East Sewage Treatment Plant, which provides primary
and secondary treatment, is located just 1 mile from the site.
A preliminary estimate indicated that on-site treatment would
be less costly than off-site disposal at a commercial facility. For
this reason, laboratory treatability studies were conducted: (1) to
determine treatment process requirements, (2) to measure the
degree of treatment achievable and (3) to develop design and cost
data.
TREATABILITY STUDIES
Available hydrogeological data indicated that the strata
underlying the PAS site would be low yielding and, on this basis,
it was estimated that groundwater pumping would yield only ap-
proximately 24,000 gal per each 1-3 week period. Since a 24,000
gal basin remains on the site and could be used for pumped
groundwater storage, it was anticipated that the treatment system
would be operated on an intermittent basis with treatment being
performed whenever the storage basin was full. A chemical
analysis of this contaminated groundwater is given in Table 1.
As part of the PAS remedial design project, laboratory
wastewater treatability studies were conducted to determine
specific treatment process requirements and capabilities. These
studies have been described previously.' The key findings are
summarized below:
• Only a minor fraction ( <1%) of the wastewater total organic
carbon (TOC) was removed by air stripping. This result indi-
Table 1
Wutewaler Analysis'
HetK,W CMoridf
Acetone
trjns-I ,?-Olchloroeth«nr
?-Butinonr
flenjene
4-Methyl -?-Penunone
Toluene
fot«l I/lenet
Phenol
Aniline
2-Hetnylphrnol
4-Melhylphfnol
2,4-OI«thylphenol ,
Tentatively Identified luo/l)
[thy] Bentene
?-3-Olethyl Oilrine
4.Methy1-?-Pent«nol
Hrtn/l Sen/me
N.N-Olethyl Forvumldr
Bulinolc Add
l-Hethoiy Cthinol
f thylbenjene
I,4-Olmethyl Benrene
Pennnolc Acid
2-tthyl Butinolc Acid
1-Methyl-?-PyrrolIdlnon*
N.N-Dlnelhyl Benien««lne
?-fthyl Heitnolc Acid
2.3-OlMlhyl Phenol
Beniene Acetic Acid
3-Mrlhyl Beniolc Acid
Convention!! PjrMgte
Sul'ite
Phosphate
H-NH3
TICK
«U«llnlty -
H«rdnesi K« )
Iron
Klckel
COO
TOC
BOO.
IDS5
Concentration
11.000
42.000
3.500
14,000
I .WO
17.000
4.300
1.600
5.700
5.600
2.300
15.000
3,400
620
28.000
6,400
8.400
5,400
14,000
2.900
940
2.400
6.300
2.300
7.MO
4.700
4.200
20.000
1,700
5. POO
12
5.5
80.7
80.7
1.500
2.150
96
2.6
1900
1020
1100
6400
Samples analyzed were 50:30 mixtures of kachate.
A complete analysis of the US EPA Contract Lab Protocol Hazardous Substances was per-
formed. Only the hazardous substances found at concentrations above analytical detection
limits are reported in this table In addition, the ten nonhazardous lubstance list compounds
present in each fraction (volatile, acid and base.'neutraf) at greatest concentration were tenta-
tively identified and quantified. These compounds are also reported in this table.
All concentrations in jig/ unless otherwise indicated.
cated that very little of the TOC consisted of volatile organic
contaminants. Therefore, air stripping was determined to be
unnecessary.
• Nickel was not removed by hydroxide precipitation but was re-
moved by activated carbon adsorption. From these data, it was
concluded that nickel was organically bound or complexed.
• Adsorption of the organic contaminants by activated carbon
was enhanced at acidic pH values.
• Wastewater COD was reduced from a concentration of 4,000
mg/1 to approximately 450 mg/1 (89°7o removal) by carbon
adsorption at pH 3.
• Carbon requirements per 24,000 gal batch of wastewater were
computed to be 6,030 Ib based on an effluent quality of 800
mg/1 COD (i.e., 20% breakthrough of the influent COD con-
centration). This usage is equivalent to a carbon loading of
0.106 Ib COD/lb activated carbon.
• With a period of 1 week acclimation, biotreatment achieved a
COD removal of 85%. Rate studies showed that this level of
treatment could be achieved in less than 8 hr retention time. A
final effluent quality of less than 50 mg/1 COD was achieved by
treating the bioreactor effluent with activated carbon.
On the basis of the batch bioreactor studies, we concluded that
a treatment process consisting of biological treatment followed by
activated carbon polishing was feasible, would be more cost-
effective and would produce a much higher quality effluent than
carbon adsorption alone. For these reasons, further laboratory
studies were subsequently performed to develop the necessary
design and cost data for a batch biological process to be operated
intermittently. The results of these studies, which have been
reported previously,' are summarized below.
144 TREATMENT OF HAZARDOUS WASTES
-------
Pretreatment Requirements
• The PAS waste has no toxic or inhibitory effect on biological
treatment. Consequently, pretreatment of the wastewater to
remove substances (such as iron) potentially inhibitory to the
biological treatment process is unnecessary.
• Approximately 8 hr of aeration were required to achieve opti-
mum COD removal at 25 °C.
• No difference in bioreactor performance was observed when
wastewater feed volumes in the range of 10-70% of the total
reactor volume were used.
Nutrient Requirements
• The addition of nutrients (i.e., nitrogen and phosphorous) to
the wastewater is not required to achieve optimum biological
removal of COD.
• The addition of nitrogen and phosphorous actually resulted in
a higher total suspended solids concentration.
Effect of Extended Lags Between
Bioreactor Operation
• Little difference was observed in the COD removal efficiency
of the batch bioreactors when operated with idle periods of
1-3 weeks.
• The rate of COD removal is inversely proportional to the length
of the idle period.
• The cycle time required to achieve optimum COD removal in-
creased from 8-11 hr when the bioreactor was operated at
1 week idle periods, to 25 hr when the idle period was 3 weeks.
Effects of Temperature on Bioreactor Performance
• During the initial weeks of bioreactor operation, the effluent
COD was inversely proportional to the bioreactor temperature.
• After several weeks of intermittent operation (1 week idle per-
iod), the effluent COD level achieved at the different tempera-
tures (i.e., 25 °C, 15 °C and 5°C) began to converge.
• Bioreactor performance as measured in terms of COD removal
rate and efficiency is adversely impacted by cooler tempera-
tures. The exact temperature at which performance begins to
degrade is not known, but lies somewhere between 5°C and
15 °C.
Effects of Powdered Activated Carbon
Addition to the Bioreactors
• Powdered activated carbon (PAC) addition substantially en-
hances COD removal rates under conditions of extended bio-
reactor idle periods and low temperature.
• COD removal efficiency increases as the PAC dosage is in-
creased.
• The time of PAC addition during the reactor cycle does not
affect the effluent quality achieved.
• When PAC is added to the bioreactor, the COD removal me-
chanism involves simultaneous adsorption and biodegradation.
Oxygen Supply Requirements
• Oxygen uptake rates are greatest in the initial phases of the
reactor cycle and decrease rapidly with time.
• As the length of the idle period between bioreactor react per-
iods is increased, the rate of oxygen uptake during the initial
phase of the react cycle decreases.
Sludge Production
• The net biological sludge production in sequenced batch bio-
reactors operated at 25-50% fill ratios was minimal or negative.
• Total bioreactor sludge production was found primarily to be
a result of inorganic precipitates. This production averaged
800 mg/1 solids of wastewater treated in bioreactors operated
at 50% fill, nonpretreated and 1 week idle periods.
• The use of PAC in the bioreactor increased the amount of
sludge produced in direct proportion to the PAS dosage.
Activated Carbon Column Studies
• Results of batch adsorption isotherm studies indicated very
little difference in adsorption characteristics at pH 3 and 7.5.
Based on this result, no pH adjustment of bioreactor effluent
is necessary to achieve optimum COD removal by carbon col-
umn polishing.
• The carbon inventory requirement for a granular activated
carbon column is a function of the allowable breakthrough
level. Effluent from a reactor operated at a 50% fill and 1 week
idle period was polished to determine carbon requirements.
For breakthrough levels of 20, 30 and 40% of the carbon col-
umn influent, carbon requirements of 2,670, 1,460 and 700 lb/
24,000 gal of wastewater treated were calculated.
• Approximately 95% COD removal was achieved using a se-
quenced batch bioreactor followed by carbon column pol-
ishing.
• Essentially complete removal of priority pollutants was
achieved by biological treatment followed by granular activated
carbon adsorption.
DEGREE OF TREATMENT ATTAINABLE
The effluent quality obtainable is a function of the raw waste-
water concentration, the operating conditions (length of idle time,
temperature, etc.) and the specific treatment process employed. A
summary of the effluent quality achieved, in terms of commonly
used measures, is presented in Table 2.
Table 2
Summary of Achievable Effluent Quality
Sample Identification
Raw Wastewater
B1o eactor Effluent ^
1 React cycle/wk
1 React cycle/2 wks
1 React cycle/3 wks
15°C
1 react cycle/wk
• 5°C
1 react cycle/wk
• 1000 mg/1 PAC
1 react cycle/wk
t 2000 mg/1 PAC
1 react cycle/wk
( 5000 mg/1 PAC
1 react cycle/wk
• 1 React cycle/wk
GAC column polishing
Temp
(°C)
..
25
25
25
15
5
25
25
25
25
COD
(mg/1)
1900-2700
.
360-410)51
385-410 ,{
370-4401''
450-550
600-750
150-200
110-140
80-90
100
TOC
(TO/1)
1020
140
128
128
__
.-
58
37
24
--
BODc
(mg/n
1050-1130
12-22
17
--
--
1-5
2-4
1-2
--
TSS
(mg/IJ
-.
62
99
262
69
65
57
--
20* breakthrough v°'
Notes:
(1) For bioreactors operated at 50% fill.
(2) Equilibrium COD.
(3) Bioreactor effluent contained 465-580 mg/1 COD.
COST OF ON-SITE TREATMENT
ALTERNATIVES
Based on the treatability study results, seven treatment alter-
natives were identified for cost analysis. The central process in
each of these schemes is biological treatment in a batch reactor.
These seven alternatives vary primarily with respect to the size
of the activated carbon column units incorporated and the quan-
tity of PAC used. These alternatives are identified below. For our
purpose, carbon breakthrough was defined as the time at which
the COD concentration in the carbon column effluent reached
100 mg/1.
Treatment Processes
• A batch biological reactor followed by a granulated activated
carbon (GAC) system sized for 30 min retention time at a flow
rate of 100 gal/min (System I-A). (Carbon breakthrough ap-
proximately every 4th batch treated.)
TREATMENT OF HAZARDOUS WASTES 145
-------
• A batch biological reactor followed by a GAC system sized for
30 min retention at a flow rate of 27 gal/min (System I-B).
(Carbon breakthrough approximately every batch treated.)
• A batch biological reactor followed by a GAC system sized for
60 min retention at a flow rate of 100 gal/min (System I-C).
(Carbon breakthrough approximately every 8th batch treated.)
• A batch biological reactor with 1,000 mg/1 powdered activated
carbon (PAC) addition and a GAC system sized for 30 min re-
tention at a flow rate of 100 gal/min (System II-A). (Carbon
breakthrough approximately every 9th batch treated.)
• A batch biological reactor with 1,000 mg/1 PAC addition and
a GAC system sized for 30 min retention at a flow rate of
11 gal/min (System II-B). (Carbon breakthrough approxi-
mately every batch treated.)
• A batch biological reactor with 1,000 mg/1 PAC addition and a
GAC system sized for 60 min retention at a flow rate of 100
gal/min (System Il-C). (Carbon breakthrough approximately
every 18th batch treated.)
• A batch biological reactor with 5,000 mg/1 PAC addition and
no GAC system (System III).
Cost Estimates
Cost estimates were based on the following assumed operating
conditions:
• Bioreactor volume (approximately 50,000 gal) is twice the feed
volume to accommodate a 50:50 ratio of mixed liquor to waste-
water feed.
• One batch (24,000 gal) of wastewater is processed each week.
• Groundwater/leachate is pumped directly to the bioreactor so
that the bioreactor is filled and ready for the reactor cycle to
begin when the operator arrives.
• Twelve hours aeration per batch.
• Two hours settling time per batch.
• Discharge time varies according to GAC system.
• A treatment system operator is available locally on an on-call
basis.
• Minimum effluent quality required is 100 mg/1 COD.
• GAC column breakthrough is 20<% of COD influent (500 mg/1
COD reduced to 100 mg/1 COD) for systems not using PAC
(Systems I-A, I-B and I-C).
• GAC column breakthrough is 40% of COD influent (250 mg/1
COD reduced to 100 mg/1 COD) for systems using 1,000 mg/1
PAC addition (Systems II-A, II-B and II-C).
• The treatment system is housed in a temperature-controlled
building.
Information regarding reactor size, sludge production, carbon
requirements and oxygen requirements that affected capital and
operating costs was determined on the basis of the treatability
study results. Capital costs were derived on the basis of vendor
quotes and annually published cost manuals.4'5
COST OF OFF-SITE ALTERNATIVES
The cost of contract hauling and disposal at a commercial
facility is a function of the particular firm that is contracted.
Costs for this alternative were based on quotes from three com-
mercial waste management firms located within 180 miles of the
PAS site.
No basis could be established for deriving the cost of treatment
of PAS groundwater/leachate at the local POTW. The city does
not have a surcharge system for treatment of high-strength
wastewater. Furthermore, this POTW has a history of operating
problems. To alleviate this problem, the city has established a
policy that no "scavenger wastes" will be accepted at the plant.
Due to this circumstance, no further consideration was given to
treatment at the local POTW as an off-site alternative.
COMPARISON OF ALTERNATIVES
A summary of the capital and annual operation/maintenance
(O&M) costs for each on-site and off-site treatment alternative
considered is presented in Table 3. These costs were developed on
the assumption that PAS groundwater/leachate would be
recovered and treated at a rate of one 24,000 gal batch per week.
The actual rate at which the groundwater/leachate will be
recovered is not known.
Table 3
Cost Summary of Treatment Alternatives
Treatment
Alternative Description
Biological Treatment &
GAC Column Pol ishing:
A I-A
B 1-8
C I-C
Capital
Cost (S 1986)(1)
564,700
500,000
707,700
Operating
Annual Operating Cost Per
Cost ($ 1986)1Z) Gallons (S 1986)13'
155,800
179,000
163,800
0.125
0.143
0.131
Biological Treatment with
1,000 mg/1 PAC Addition J
GAC Column Polishing:
D II -A
E II-B
F II-C
G Biological Treatment with
5,000 mg/1 PAC Addition
H Contract Haul ing/Of f-si te
Disposal
589,100
506,200
723,900
515,100
115,500
120,900
127,200
499,200 - 1,123,000
(549,0001
0.40 - 0.90
(0.44)
Notes:
(1) Cost for facilities to treat wastewater in 24,000 gal batches.
(2) Cost to treat 24,000 gal/week.
(3) Cost/gal based on 1,248 million gal/yr (i.e., 52 wks x 24,000 gal/wk).
Consideration of the quantity of groundwater/leachate to be
treated with time is essential to a comparison of the costs of the
various treatment alternatives, since these costs are a function of
both the quantity of wastewater to be treated and the length of
time over which treatment is to be performed. The total quantity
of groundwater/leachate requiring treatment has been estimated
to be 3 million gallons based on hydrogeological data collected
during the RI/FS and the Remedial Design Study. To address the
various rates at which this wastewater may be recovered, four
scenarios have been developed (Table 4).
Table 4
Hypothetical Rates of Groundwater/Leachate Recovery and Treatment
Remaining years
(1)
Scenario
1
2
3
4
Year
1
1
1
1
Quantity
1 MG
1 MG
1 MG
1 MG
Cycles
41.7
41.7
41.7
41.7
Period
8.8 days
8.8 days
8.8 days
8.8 days
Years
2-3
2-5
2-7
2-10
Ouan
1 MG
0.5
0.33
0.22
tity
/YR
HG/YR
MG/YR
MG/YR
Per
41.
20
13
9
Year
.7
.8
.9
.2
Per-
8.8
17.5
26.3
39.4
iod
days
da v*.
davi
flays
(1) Assumes 24,000 gal per cycle.
A present worth analysis of each treatment alternative as a
function of the four recovery rate scenarios is given in Table 5.
The following observations are based on a review of this analysis:
• All on-site treatment alternatives (A through G) are more cost-
effective than the off-site alternative (H).
• On-site alternatives which incorporate the smaller sized GAC
columns (B,E) are more cost-effective than alternatives in-
corporating larger columns (A,C,D,F).
146
TREATMENT OF HAZARDOUS WASTES
-------
Table 5
Present Worth Analysis of Treatment Alternatives
(1) (2)
Treatment
Scenario Alt.
1 A
B
C
0
E
F
G
H
2 A
B
C
D
E
F
G
H
3 A
B
C
D
E
F
G
H
4 A
B
C
D
E
F
G
H
(3)
Years
of
Treatment
3
3
3
3
3
3
3
3
5
5
5
5
5
5
5
5
7
7
7
7
7
7
7
7
10
10
10
10
10
10
10
10
(4)
Capital
Cost
(S*10>)
564.7
500.0
707.7
589.1
506.2
723.9
515.1
t
564.7
500.0
707.7
589.1
506.2
723.9
515.1
B
564.7
500.0
707.7
589.1
506.2
723.9
515.1
0
564.7
500.0
707.7
589.1
506.2
723.9
515.1
f>
(5)
1st Yr.
Operating
Cost ($xlOJ)
125.0
143.0
131.0
92.5
96.9
102.0
87.4
400.0
125.0
143.0
131.0
92.5
96.9
102.2
87. 4
440.0
125.0
143.0
131.0
92.5
96.9
102.0
87.4
440.0
125.0
143.0
131.0
92.5
96.9
102.0
87.4
440.0
(6)
Remaining Yrs
Operating
Annual Cost
Ax(ixlO')
125.0
143.0
131.0
92.5
96.9
102.0
87.4
440.0
62.5
71.5
65.5
46.2
48.4
51.0
43.7
220.0
41.2
47.2
43.2
30.5
32.0
33.7
28.8
145.2
2;. 5
31.5
28.8
20.4
21.3
22.4
19.2
96.8
(7)
Total
Present
Uorth
(5x10=)
875.8
855.9
1033.8
819.3
747.4
977.6
732.7
1095.9
858.3
835.9
1015.4
806.3
733.7
963.5
720.4
1033.9
841.5
816.9
997.9
794.0
721.0
950.0
708.6
975.5
822.1
794.8
977.4
779.9
705.7
933.8
695.0
906.8
Notes:
Column 1: Scenario (Refer to Table 7-2)
Column 2: Treatment Alternative (Refer to Table 7-1)
Column 3: Total number of years that a treatment system is operated on-site
Column 7: Calculated on basis of:
Present worth of capital and annual O&M at year n
= (P/A, i, n) x (O&M) + Capital
= (1 + i) n - 1
x O&M + Capital
i (i + i)n
where i = interest rate per year
n = number of years
• Biological treatment enhanced by a 5,000 mg/1 dose of PAC
(Alternative G) is more cost-effective than any of the alterna-
tives (A through F) which include GAC columns for effluent
polishing.
CONCLUSIONS
Based on laboratory treatability studies, biological treatment
with activated carbon polishing was determined to be the most ef-
fective treatment method. These studies also demonstrated the
feasibility of using a batch bioreactor operated intermittently.
The disadvantage of lower COD removal rates when the bioreac-
tor was operated at low temperature and increased idle periods
was overcome by the addition of powdered activated carbon.
Results of a present worth analysis showed that on-site treatment
will be much more cost-effective than off-site treatment/disposal
at a commercial facility.
REFERENCES
1. "Site Investigations and Remedial Alternative Evaluations at the
Pollution Abatement Services (PAS) Site in Oswego, New York,"
Draft Final Report, prepared for New York State Department of
Environmental Conservation by URS Company, Inc., Jan. 1984.
2. "Engineering Design Report for Site Remedial Measures Pollution
Abatement Services Site, Oswego, New York," prepared for New
York State Department of Environmental Conservation by URS
Company, Inc., May 1985.
3. "Evaluation of Alternatives for Treatment of Groundwater/Leachate
at the Pollution Abatement Services (PAS) Site in Oswego, New
York," Final Report, prepared for New York State Department of
Environmental Conservation by URS Company, Inc., Oct. 1985.
4. Building Construction Cost Data 1985, published by the Robert Snow
Means Company.
5. Dodge Manual for Building Construction, Pricing and Scheduling,
published by McGraw-Hill Information Systems Co., Vol. 2.
TREATMENT OF HAZARDOUS WASTES 147
-------
Optimization of Free Liquid Removal
Alternatives in the Closure of Hazardous Waste
Surface Impoundments
David K. Stevens
Black & Veatch
Kansas City, Missouri
ABSTRACT
A flowchart is presented to serve as a guide in determining the
optimal alternative for removal of free liquids during the closure
of hazardous waste surface impoundments. The flowchart ad-
dresses each part or element within a closure alternative as a
discrete decision process. This flowchart allows separate evalua-
tion of each phase of removal such as preclosure engineering,
removal and treatment and ultimate disposal. Criteria used to
select the best option within an element include technological
feasibility, environmental soundness, regulatory acceptance and
unit cost. Once the best elements for each phase of free liquid
removal are determined, they are recombined to form the optimal
removal alternative.
Important factors such as in situ reduction and provisions for
mixed wastes are discussed. Removal equipment and treatment
processes are presented as a function of the waste type. Ultimate
disposal options such as discharge to surface water vs. discharge
to sewer, on-site vs. off-site treatment and disposal and process
reuse and recovery are included. To illustrate the use of the
flowchart, three case studies of actual surface impoundment
closures are presented. Optimal free liquid removal alternatives
are developed for impoundments containing heavy metals, heavy
metals and degreasers, and oily wastes.
INTRODUCTION
There are tens of thousands of hazardous waste lagoons, pits or
ponds in the United States.' These types of storage facilities or
surface impoundments are used for continuous or intermittent
storage of process water, stormwater and spent products from the
chemical and manufacturing industries. RCRA or CERCLA can
require closure of an impoundment when soil and groundwater
contamination is discovered or if the impoundment poses an en-
vironmental risk. Stricter liner requirements for impoundments
under RCRA are also forcing an increased number of closures.
Given the trend toward greater environmental protection, increased
enforcement and stricter regulations, it is certain that impound-
ment closures and their associated remedial actions will continue
at a rapid pace.
Regardless of the RCRA or CERCLA regulations governing
closure, removal of free liquids is required as one of the first steps
in an impoundment closure. The removal process can present a
number of technological and regulatory problems as a large
volume of liquid hazardous waste must be removed from the site
which frequently has no suitable direct outlet. The free liquids
often have a high solids content, exhibit extremes of pH and
viscosity, have concentration gradients and are stratified mixtures
of both inorganic and organic chemicals. These technological
problems combined with the regulatory pressures for tight closure
schedules and strict air and water discharge limits require careful
engineering of the free liquid removal processes. Recognizing the
inherent problems in dealing with liquid hazardous wastes in sur-
face impoundments, the flowchart directs the user in the selection
of the optimum free liquid removal alternative.
FLOWCHART DESCRIPTION
In assembling the flowchart elements into a system of logical
steps, the universe of processes used for free liquid removal were
identified. Once the available processes were identified, they were
categorized into three elements: preclosure engineering; removal
and treatment; and ultimate disposal. Figure 1 shows the basic
outline of the flowchart. To determine the optimal removal alter-
native, each process within an element is subjected to a set of deci-
sion criteria developed from recent CERCLA guidance1 and ac-
tual closure experience. Through evaluation of each process ac-
cording to the developed decision criteria, the best option for that
process can be identified. By combining the best options within
each element, the optimal removal alternative is determined.
Flowchart Elements
The flowchart contains three major elements: preclosure
engineering; removal and treatment; and ultimate disposal (Fig. 1).
Pre-closure engineering addresses the processes for flow diver-
sion, waste characterization and the removal alternative's impact
CLOSURE
DECISION
PRE-CLOSURE ENGINEERING
• Flow Div0nion
• Waste Characterization
• Facility Cloaure Plan
REMOVAL AND TREATMENT
• In Situ Reduction
• Pumping and Dredging
• Treatment
ULTIMATE DISPOSAL OR
REUSE/RECOVERY
• On Site
• Off Site
IMPOUNDMENT
READY FOR
FINAL CLOSURE
Figure 1
Overview of Flowchart for Determining
Optimal Liquid Removal Alternatives
148 TREATMENT OF HAZARDOUS WASTES
-------
on the overall facility closure plan. The removal and treatment
element is composed of three processes: in situ reduction, liquid
removal and treatment prior to ultimate disposal. In situ reduc-
tion considers whether the hazard or volume of the free liquid
should or can be reduced while in the impoundment before actual
removal. The removal section addresses the equipment used in
removing the waste, such as pump type and construction
material. Treatment alternatives include physical and chemical
treatment of the free liquids for reduction or elimination of the
hazard from the waste before ultimate disposal. The treatment
process and the ultimate disposal option often are synergistic, re-
quiring concurrent evaluation. The ultimate disposal element
reviews the options for final deposition of the liquids such as sur-
face water, publicly owned treatment works or an RCRA facility.
Careful consideration always should be given to those processes
which allow reuse or recovery of the liquids.
Decision Criteria
The decision criteria employed for evaluation of each closure
element are expressed in both cost and non-cost factors. Cost fac-
tors include the cost for present value of operating and
maintenance costs and the capital expenses. Non-cost factors in-
clude technological feasibility, environmental soundness and
regulatory acceptance. Technological feasibility requires that the
technology under consideration be proven at the design scale or
that pilot-scale results show a good chance for success in the field.
Environmental soundness is a measure of the exposure of the
chemical constituents to the environment during the removal
operation. This decision can be a judgmental one as in some cases
environmental exposure occurs only for a short time and does not
persist once the removal is complete. Regulatory acceptance is the
third criterion. This factor indicates whether the closure processes
under consideration are acceptable to the governing regulatory
agency. Acceptance or non-acceptance depends on the pertinent
regulations and the perceived environmental risk.
PRECLOSURE ENGINEERING
During development of an RCRA closure plan or a CERCLA
feasibility study, several preliminary concerns must be addressed
before the actual free liquid removal alternatives can be fully
developed. Among these concerns are options for flow diversion,
requirements for complete waste characterization and the re-
moval alternative's impact on the overall impoundment closure
plan. In the flowchart, Figure 2, these concerns are taken into
consideration in the preclosure engineering element. Carefully ad-
dressing each of these concerns is required for optimization of the
free liquid removal alternative.
Flow Diversion
The initial step in implementation of an impoundment closure
plan is to eliminate the source of incoming waste. This is not a
major concern for inactive impoundments or impoundments used
for low-volume intermittent waste disposal. At active impound-
ments receiving continuous or semi-continuous waste streams
such as process water or contaminated stormwater, careful con-
sideration is required as dedicated treatment plants and long-term
operation become potential design factors.
CLOSURE
DECISION
PRE-CLOSURE ENGINEERING
GENERATED SLUDGE DISPOSAL
HAZARD
REDUCE
HAZARD
- Biological
• Clu-oncnl
• Physical
REMOVAL AND
VOLUME
i 1
REDUCE
J VOLUME
• PhvMC.ll
1
TREATMENT
REMOVE
LIQUIDS
j
PR E TREATMENT
CHEMICAL
PHYSICAL
TREATMENT
ULTIMATE D
ONSITE
DISPOSE ONSITE
• Reuw; —
• Recovery
• Land Apply
ISPOSAL
DISPOSE
• POTW
• Surfiicu WuU-r
• RCRA Facility
1
CONTINUE
CLOSURE
Figure 2
Free Liquid Removal Flowchart
TREATMENT OF HAZARDOUS WASTES 149
-------
Waste Characterization
Accurate information on the chemical constituents in the im-
poundment is essential, including identification of any variances
in the composition as a function of depth. Each layer can require
special consideration and different removal procedures; for exam-
ple, a floatable oil layer can cover a relatively clear aqueous layer
which in turn covers a gelatinous metal hydroxide sludge. The im-
portant physical parameters include pH, viscosity and solids con-
tent. The pH influences pump selection and treatment processes.
Viscosity and solids content influence the pump type and the
volume of liquid available for removal by pumping alone. Sludges
with solids content greater than 8-10% must be removed either by
first mixing with the other impoundment liquids to lower the
solids content, or by leaving the high solids content sludge in
place and dewatering by evaporation or by the addition of a
solidifying agent.
Impact on the Overall Closure Plan
Consideration of the free liquid removal alternative's impact on
the overall closure plan is imperative for optimizing the impound-
ment closure. Very often other remediation activities at the facil-
ity can influence selection of the free liquid removal alternative.
These considerations are site-specific and are best illustrated
through examples. Table 1 shows examples of the impact of the
selected closure plan on the free liquid removal alternative.
Table 1
Impact of Facility Closure Plan
on Free Liquid Removal Alternatives
Waste Type
Heavy Metals
Oily Wastes
PCB Oils
Facility Closure Plan
Groundwater Treatment
Program
Reclamation of Wastes
Soil Incineration
Free Liquid
Removal Alternative
Same Treatment
Process Employed
Free Liquids Re-
claimed by Same
Process
Free Liquids
Incinerated
REMOVAL AND TREATMENT
In Situ Reduction
In situ reduction is a process within the treatment and removal
element (Fig. 2). It involves the reduction of the hazard or the
volume of the waste present in the impoundment before actual
removal. Such techniques can be used to facilitate final removal
and disposal of the waste.
In dealing with very corrosive wastes such as spent pickle liquors
with pH C2 or caustic washes with a pH > 12, it is advantageous
to add a neutralizing agent to the wastes in the impoundment. Ad-
dition of bases also can be used to remove toxicity such as the
alkaline precipitation of heavy metals at pH 9-10. In small acidic
ponds, a 50% solution of sodium hydroxide is best as it is easily
pumped at moderate ambient temperatures and offers a high
degree of neutralizing power. For larger acidic ponds or when
cost is an important factor, calcium hydroxide and calcium car-
bonate should be examined. Multi-stage treatment with several
neutralizing materials has been investigated.' For alkaline ponds,
concentrated sulfuric acid is generally best as it is inexpensive, of-
fers a high degree of neutralizing ability and does not pose a
serious material handling problem. Other possibilities for in situ
hazard reduction are the use of oil booms, reduction/precipita-
tion techniques and activated carbon or polymer addition for
organics adsorption.
The other in situ alternative is to reduce the volumetric quantity
of free liquids in the impoundment. The two primary mechanisms
are evaporation and seepage. Metal-containing aqueous wastes
are usually more amenable to evaporation or seepage than organic
wastes. Although a limited number of sites actually experience a
high net evaporation rate, judicious scheduling of the liquid
removal step during periods of low rainfall and high temperature
can effectively reduce the total volume of free liquids. For exam-
ple, at a West Coast manufacturing facility the volume of water in
a stormwater retention pond varied from 3.5 million gallons dur-
ing January to 2.5 million gallons during August. Determining
whether seepage is a viable method of volume reduction is a less
straightforward calculation than determining the evaporation
potential.' The use of impoundments as seepage basins for con-
centrating wastes has been documented to be very effective with a
limited impact on ground water quality.'
Free Liquid Removal
Physical removal of the free liquids is accomplished best by
pumping. If the waste has a high solids content, mechanical, hy-
draulic or pneumatic dredging may be required. This paper
discusses only pumping of free liquids. Selection parameters in-
clude the type, construction materials and size of pumps. The
pump type and construction materials usually depend on the
waste material and its physical characteristics. Pump size is
governed by the closure schedule and treatment system capacities.
Table 2 delineates the generally available pump types and con-
struction materials as a function of the waste type.
Table 2
Pump Types and Construction Materials
as a Function of Waste Type
Physical Description of Waste
Pump T)pe
Aqueous
Vi.scous or Oily
Sludge. 1-5% solids
Sludge. 5-8% solids
Chemical Description
of Waste
Aqueous
Oils, phenols
Solvents
Heavy metab
Basic pH >I2
Acidic pH < 2
Centrifugal
Rotary lobe, progressive cavity
Centrifugal
Posiihe displacement
Construction Material
Cast iron
Cast iron, possibly bronze fittings
Bronze fitted cast iron
Stainless steel
SS ASTM 743. CN-7M
SS ASTM 743. CN-7M (3«*o Mo)
Treatment
At this stage in the evaluation process, the free liquid treatment
and disposal processes are developed. Selection of the treatment
process is based on both the waste characteristics and the
availability of ultimate disposal mechanisms. Relationships be-
tween the disposal processes and treatment requirements are sum-
marized in Table 3.
Table 3
Relationship Between Disposal Processes and Treatment Requirements
Disposal Processes
Off-site disposal
On-site disposal
On-sile reuse or recovery
Publicly owned treatment works
Process reuse
Surface water
Groundwatcr rcinjection
Treatment Requirements
Usually requires no treatment
except suspended solids removal
Suspended solid removal often
required
Treatment to drinking water
standard required
150 TREATMENT OF HAZARDOUS WASTES
-------
Since the treatment process and the ultimate disposal method
represent the largest single capital cost items involved in free liq-
uid removal, all alternatives need to be critically reviewed. Con-
sideration should be given to alternatives which provide a poten-
tial for reuse or combination with other treatment needs.
Processes available for the treatment of aqueous hazardous
waste are numerous.6 The major limitation to the use of many of
these processes is that many are not cost-effective or not proven at
the field scale. New technologies and innovative processes are best
explored when a long-term use treatment plant is required, such
as in groundwater remediation programs or surface and process
water treatment systems. In other cases, leasing of mobile treat-
ment units or low-cost temporary equipment should be con-
sidered. Table 4 shows the technologies available for treatment as
a function of the waste type.
Table 4
Treatment Technologies as a Function of Waste Type
Waste Type
Acidic pH<2
Basic pH >12
Heavy metals
Low level
organic
solvents
High level
organic
solvents
Oil
Dedicated
On-site
Equipment
Alkali addition
Acid addition
Precipitation/
decollation
Ion exchange
Adsorption
Air stripping
Oxidation
Carbon adsorption
Incineration
Separators
Incineration
Filtration
Mobile
Equipment
alkali addition
Acid addition
Precipitation/
flocculation
Adsorption
Air stripping
Oxidation
Carbon adsorption
Incineration
Separators
Incineration
Filtration
Temporary
System
alkali addition
Acid addition
Precipitation/
flocculation
Adsorption
Carbon adsorption
Separators
Filtration
ULTIMATE DISPOSAL
Direct Off-Site Disposal
In direct off-site disposal of free liquids, the liquids are taken
directly from the impoundment to an off-site facility. Potential
off-site facilities include publicly owned treatment works and
RCRA treatment facilities. If the concentrations of the hazardous
compounds are within sewer discharge limits, the liquids can be
pumped directly into the sewer system. If the sewer line is not
within economical pumping distance, alternative means of
transportation to the treatment plant must be explored. A most
practical means is to employ vacuum trucks or similar tanked
vehicles. Transportation costs can be as much as $1.50 per
unloaded mile and $3.50 per loaded mile. Sewer discharge rates
vary widely depending on waste type, quantity and treatment
facility. When the pollutant concentrations exceed the influent
pretreatment standards and direct off-site disposal is a potential
alternative, a list of available facilities should be compiled. Many
available directories list RCRA facilities, and the RCRA hotline
also can provide a list of currently permitted facilities. Transpor-
tation costs are similar to the amounts given above. Disposal costs
can be from $0.10/gal for low level contaminated aqueous
streams to over $l/gal for liquid injection of highly organic
wastes. Deep well injection, generally for acids and bases only,
costs approximately $0.20/gal.
Treatment Followed by Off-Site Discharge
Employing a pretreatment process expands the number of off-
site disposal alternatives to include surface water discharge and
groundwater injection. Groundwater injection is predominantly
used only in conjunction with other groundwater remediation
programs. Discharge to surface water requires an NPDES permit
as well as very low contaminant concentrations. For metals, some
type of filtration will be required to attain the discharge limit.
On-Site Disposal
The alternatives to on-site disposal include construction of an
on-site RCRA permitted facility. For most closures at industrial
manufacturing facilities, the closed system with reuse and
recovery is the most advantageous. At uncontrolled sites, on-site
disposal of the liquids usually is confined to treatment processes
such as landfarming, evaporation or incineration.
GENERATED SLUDGE DISPOSAL
Final consideration for free liquid alternatives is given to the
disposal of sludges generated during a treatment process. Sludge
is generated during alkaline precipitation of metals, filtration or
solidification/stabilization processes. Potential disposal alter-
natives include off-site disposal in an RCRA permitted landfill,
on-site disposal in a specially constructed RCRA permitted land-
fill and disposal in the closed impoundment.
CASE STUDIES
Three case studies of actual surface impoundment closures are
presented here to illustrate some of the aspects in determining the
optimal alternative for free liquid removal. The following cases
are discussed:
• Stormwater retention pond containing hexavalent and trivalent
chromium
• Process water storage lagoon contaminated with chlorinated
solvents and nickel
• Earthen pit used for oily waste disposal
Chromium Contamination
A manufacturing facility employed an earthen pond to store
stormwater runoff from its product storage area. The pond was
being closed because of its ineffective liner and the presence of
chromium in the underlying groundwater.
Preclosure Engineering
The pond currently was being used for storage of stormwater,
and an alternative mechanism Tor handling future runoff was
needed. Since manufacturing and storage operations were to con-
tinue at the site and no practical method of eliminating the con-
taminated runoff existed, a permanent treatment and disposal
solution was required. An ion-exchange treatment system with
surface water discharge was determined to be the best alternative
for addressing the runoff. Ion-exchange produced a very low level
concentration of chromium in the effluent and allowed reuse of
the recovered chromium. Ion exchange also added the advantage
that no hazardous waste sludge was generated during treatment as
would have been produced using traditional chemical reduction
and precipitation.
Results of the pond sampling activities showed that approx-
imately 400,000 ft 3 of free liquids containing 5-10 mg/1 hex-
avalent chromium and 1-5 mg/1 trivalent chromium were in the
impoundment. Underlying the free liquids were 1000 ft3 of a
chemical floe sludge containing 1000 to 5000 mg/1 total
chromium. The-low chromium concentrations in the free liquids
permitted use of the current ion-exchange system for free liquid
treatment. The remaining sludge could be solar dried and fixed in-
place using a solidification agent. Use of the on-site treatment
system was an attractive solution for treatment as there was no
negative impact on the other closure activities and no additional
permitting was required for surface water discharge.
TREATMENT OF HAZARDOUS WASTES 151
-------
Removal, Treatment and Ultimate Disposal
Given the overwhelming advantages in using the on-site ion-
exchange treatment system, the free liquids were removed using
an available contractor's pump, treated to NPDES requirements
and disposed via the force main to the nearby surface water
stream. Judicious closure scheduling during the summer months
reduced the volume of free liquids from 400,000 ft3 to 260,000 ft3.
This timing reduced pumping time from 14 days to 9 days and
saved approximately $20,000 in treatment expense.
During the alternative review phase, contractor proposals were
solicited for removal and treatment of the free liquids. The con-
tractor proposed to treat the liquids using sulfite reduction fol-
lowed by alkaline precipitation. For comparison purposes, the
decision criteria evaluation of the two treatment alternatives is
shown below:
lowered below discharge requirements. Different treatment
schemes were developed and compared using the four decision
criteria. The results are summarized below.
Alternative
Contractor
Bid
Use of On-site
Treatment
System
Technology
Feasibility
Feasible
Feasible
Environmental
Impact
Sludge
Generated
Process Reuse
Regulatory
Acceptance
Acceptable
More Favorable
Acceptance
Unit Cost
S/1000 gal
22.33
5.50
Degreasers and Contamination
A metal plating and fabrication facility used an earthen im-
poundment for disposal of process rinsewaters. Groundwater
contamination by the degreasers was forcing closure of the im-
poundment.
Preclosure Engineering
Due to the continuation of fabrication operations at the plant,
a permanent treatment solution was needed for the waste process
rinsewaters. Through water-use reductions and process modifica-
tions, the rinsewater flow rate was reduced from 300 to 50
gal/min. A 50 gal/min treatment system was put on line early dur-
ing the closure plan development. Treated rinsewater was dis-
charged to nearby surface water per the NPDES permit require-
ments.
The impoundment contained approximately 130,000 ft3 of free
liquids, 8,000 ft3 of pumpable sludge and 2,000 ft3 of unpump-
able sludge. The nickel and degreaser concentrations in the free
liquids ranged from 30 to 60 mg/1 and 0.2 to 0.9 mg/1, respective-
ly. Both sludge layers contained higher levels of contaminants
than the free liquids, with the unpumpable sludge layer contain-
ing the bulk of nickel and degreaser.
Due to the depth of degreaser contamination in the soil, the im-
poundment was closed with contaminated soils in place. Labora-
tory studies were conducted on both sludges to determine if they
were amenable to stabilization. If adequately stabilized, the
sludges could be disposed of by leaving them in the impound-
ment. Tests showed that the addition of portland cement effec-
tively formed a stable and solid mixture. Using the stabilization
process, only the free liquids needed to be removed and disposed;
the pumpable and unpumpable sludges were stabilized and left in
place.
Removal, Treatment and Ultimate Disposal
The volume of free liquids in the impoundment precluded the
use of the process water treatment system; the time required for
treatment using the system would exceed the required 180-day
schedule. An alternative mechanism was required. Due to the low
level of contaminants in the free liquids, consideration was given
to the use of a temporary on-site system consisting of filtration
and possibly pH adjustment. Laboratory tests confirmed that
with pH adjustment and settling, the contaminant levels were
Alternative
Coagulation
& Settling
Pressure
Filtration
Gravity
Filtration
Technology
Feasibility
Feasible
Feasible
Feasible
Environmental
Impact
Minimal
Minimal
Minimal
Regulatory
Acceptance
Acceptable
Acceptable
Acceptable
Unit Cost
5/1000 gal
11.25
8.75
7.50
It was decided that gravity sand filtration would produce the
best quality effluent at the lowest cost. To enhance operation of
the filtration equipment, care was taken to maintain a low level of
suspended solids in the free liquids. A floating intake was
employed to minimize disturbance of the underlying sludge. In
addition, when the suspended solids concentration exceeded 400
mg/1, the pump speed was decreased to further minimize resus-
pension. Backwash water was held on-site, and the settled
material was returned to the impoundment after all the free liq-
uids were removed.
Oily Wasles
An earthen pit located on a farmer's property had been used
for the disposal of spent cutting and hydraulic oils. Closure was
proceeding as a generator-directed and funded mitigation action.
The site functioned with little control, and many accounts of un-
authorized dumping had been recorded. However, one major
generator had taken prime responsibility for closure instead of
waiting for a potential enforcement action.
Preclosure Engineering
The impoundment was receiving intermittent waste streams only,
and additional waste inputs were eliminated by banning all future
waste disposal practices. Results of the pond sampling activities
showed that three distinct layers were present: an oil layer of ap-
proximately 10,000 ft 3, an aqueous layer of 60,000 ft3 and a
viscous oily sludge layer of 4,000 ft3. The aqueous layer contained
between 0.2 and 0.5% total hydrocarbons and could not be dis-
posed as a non-hazardous waste. However, analytical analysis of
the oily layer showed that it potentially could be reclaimed.
Removal, Treatment and Ultimate Disposal
In this closure, several site-specific constraints precluded many
of the normally available options; no surface water or sewer line
was readily available. Due to the uncontaminated nature of the oil
it allowed the potential for recovery. The oil layer was removed by
vacuum trucks and taken to an oil recovery plant for recox cry and
reuse. Since no outlet for the treated aqueous layer was readily
available, an off-site disposal alternative was required. The
aqueous layer was removed by vacuum trucks and taken to the
generator's manufacturing facility where it was treated in their
own industrial wastewater treatment plant. Although the trucking
costs were substantial, the oil recovery and generator treatment of
the aqueous waste offset a portion of this cost. It was determined
that disposal of the viscous sludge was accomplished best by land-
filling rather than by incineration, since incineration was an order
of magnitude in cost higher than landfilting.
CONCLUSIONS
A flowchart has been presented which guides in determining the
optimal alternative for removal of free liquids from hazardous
waste surface impoundments. The flowchart addresses each ele-
ment of the closure process as a separate entity, whereas in-
152 TREATMENT OF HAZARDOUS WASTES
-------
dividual decision criteria are applied to each element before com-
bining the elements to form the optimal alternative. The major
elements include: preclosure engineering; removal and treatment;
and ultimate disposal.
General trends identified from the case studies presented in this
paper are:
• Identify site-specific constraints early during closure plan de-
velopment. The combination of free liquid treatment and dis-
posal options with soil and groundwater remediation programs
can substantially reduce the cost for free liquid removal.
• Use recovery and reuse whenever possible; this alternative is at-
tractive environmentally and to regulatory agencies.
• Employ temporary on-site treatment measures when levels of
contaminants are low. This process is less expensive than off-
site disposal at RCRA facilities.
• When available, use on-site process water or groundwater
waste treatment equipment to treat the free liquids.
REFERENCES
1. Wyss, A.W., et al., "Closure of Hazardous Waste Surface Impound-
ments," prepared by Acurex Corporation for the U.S. EPA, Sept.
1982.
2. "Guidance on Feasibility Studies Under CERCLA," prepared by JRB
Associates for the U.S. EPA, June 1985.
3. Hale, F.D., Murphy, C.B. and Parrat, R.S., "Spent Acid and Plat-
ing Waste Surface Impoundment Closure," Proc. Third National
Conference on Management of Uncontrolled Hazardous Waste Sites,
Washington, DC, Nov. 1982, 195-201.
4. McWhorter, D.B., "Seepage in the Unsaturated Zone: A Review,"
Proc. Seepage and Leakage from Dams and Impoundments, May
1985, 200-219.
5. Looney, B.B., "Surface Impoundment Legacy: Field Studies," paper
presented at the Hazard Materials Management Conference, June
1985.
6. Sittig, M., "Pollutant Removal Handbook," Noyes Data Corpora-
tion, Park Ridge, NJ, 1973.
TREATMENT OF HAZARDOUS WASTES 153
-------
Assessment of Chemical Treatment Technologies
and Their Restrictive Waste Characteristics
Hamid Rastegar, Ph.D.
James Lu, Ph.D.
Chris Conroy
Jacobs Engineering Group, Inc.
Pasadena, California
ABSTRACT
The applicability of a particular treatment technology to a
given liquid waste stream depends upon the physical, chemical
and biological characteristics of the waste. Waste stream char-
acteristics which restrict a waste's applicability to a particular
method of treatment are presented. With this information, the
most effective treatment technology for a waste stream can be
determined. The technologies evaluated include chemical oxida-
tion and reduction, chlorinolysis, chemical fixation, ion ex-
change, neutralization and precipitation.
INTRODUCTION
The land disposal of many hazardous wastes may be banned
within the coming years under the RCRA amendments. For those
wastes which will be banned from landfill disposal, it is neces-
sary to determine alternative methods of treatment and disposal.
A wide variety of treatment technologies exists today which
can be used to treat such wastes. Treatment technologies can be
classified as either chemical, physical, biological or thermal treat-
ment. In this paper, the authors evaluate the applicability of
chemical treatment technologies for hazardous liquid waste
streams based on the wastes restrictive characteristics.
This analysis is part of an extensive study by the U.S. EPA to
determine the applicability of alternative waste treatment tech-
nologies. The most promising types of each chemical treatment
technology are discussed. The major restrictive waste characteris-
tics determining treatment applicability for each also are pre-
sented. Finally, the final product or residue from each treatment
technology and the limitations of each are evaluated.
CHEMICAL TREATMENT TECHNOLOGIES
Several chemical treatment technologies are widely used for
waste treatment. The most effective technologies discussed here
include:
Chemical Oxidation
Chemical Reduction
Chlorinolysis
Fixation
Ion Exchange
Neutralization
Precipitation
The major restrictive waste characteristics as well as the type of
applications for each of the above technologies are given below.
A summary of the major restrictive waste characteristics is given
in Table 1.
Chemical Oxidation
Chemical oxidation is a process which raises the oxidation state
of chemical species, converting the wastes into less-hazardous or
non-hazardous forms. The oxidation is accomplished by the
addition of an oxidizing agent which is reduced.
Chemical oxidation has been used most commonly to treat cya-
nide wastes from electroplating operations. The most widely used
oxidants for this application are chlorine gas and calcium and
sodium hypochlorite.
Table 1
Chemical Treatment Technologies Restrictive Waste Characteristics
Chemical Treatment
Technology
Major Restrictive Waste
Characteristics
Chemical Oxidation
Chemical Reduction
Chlorinolysis
Fixation
Ion Exchange
Neutralization
Precipitation
Physical form, oil and grease content, sus-
pended solids content, viscosity.
Physical form, oil and grease content, sus-
pended solids content, viscosity.
Total solids content, oxygen, sulfur, and
organic content, water content.
Physical form, organic content, ionic com-
position, sulfate content.
Suspended solids content, metal content,
organic content, oxidizing agents, physical
form.
Dissolved solids content, physical form.
Physical form, viscosity, metal solubility.
Other oxidants for cyanides include ozone, hydrogen peroxide,
potassium permanganate and chlorine dioxide. A wide variety of
organics and inorganics also can be treated by oxidation. The
end products of cyanide oxidation are nitrogen gas and carbon
dioxide. The end products of the oxidation of organics are either
simpler organics or carbon dioxide and water; oxidized metals
form metal hydroxides which can be precipitated out of solution
by pH adjustment.
The major restrictive characteristics of a waste which are used
to determine the applicability of chemical oxidation include the
physical form, oil and grease content, suspended solids content
154 TREATMENT OF HAZARDOUS WASTES
-------
and viscosity of the waste. Chemical oxidation is most effective
for dilute aqueous wastes, although certain gaseous wastes also
can be treated. Since oxidizing agents will indiscriminantly attack
all oxidizable material, the oil and grease content and the oxidiz-
able suspended solids content should be low, preferably below
1%. For liquid solutions, viscosities much greater than that of
pure water are restrictive because of the mixing problems en-
countered with highly viscous solutions.
Chemical Reduction
Chemical reduction lowers the oxidation state of chemical spe-
cies, converting them into less-hazardous or non-hazardous
forms. The reduction is accomplished by the addition of a reduc-
ing agent which is oxidized.
Chemical reduction is used most commonly to convert hexa-
valent chromium into trivalent chromium, which then is precip-
itated out of solution as chromium hydroxide. Mercury also can
be removed from solution by reducing mercury (II) to its elemen-
tal state. Sulfur dioxide and sulfite salts are the most widely used
reducing agents for chromium. Sodium borohydride is an effec-
tive reducing agent for mercury and lead.
The major restrictive waste characteristics for chemical reduc-
tion include the wastes' physical form, oil and grease content,
suspended solids content and viscosity. These are the same re-
strictive characteristics as for chemical oxidation, since chemical
oxidation and reduction are similar processes.
Chlorinolysis
Chlorinolysis involves the reaction between chlorine and liquid
chlorinated hydrocarbons at temperatures of 500 °C or greater
and high pressure to form carbon tetrachloride, hydrogen chlor-
ide and other by-products. In the Chlorinolysis process, the waste
first is pretreated to remove all solids and any unwanted com-
pounds which could result in the formation of unwanted reac-
tion by-products. Then, the purified waste undergoes reaction
with chlorine, and the reaction products are separated and recov-
ered by distillation.
The waste to be treated by Chlorinolysis ideally should con-
tain only liquid chlorinated hydrocarbons. The presence of other
compounds may produce toxic reaction by-products, which may
include carbon tetrachloride, hydrogen chloride and phosgene.
The restrictive waste characteristics include the solids oxygen,
sulfur organic and water content. As mentioned previously, the
presence of compounds other than chlorinated hydrocarbons is
unwanted.
Fixation
Chemical fixation is a process in which hazardous substances
are stabilized and solidified in a form which is considered non-
hazardous and safe for ultimate disposal. Various types of ma-
terials can be mixed with a waste to form a solid material which
effectively immobilizes the waste. The process is intended to be
permanent, allowing the safe disposal of the solidified mass to a
landfill or other permanent disposal facility.
Fixation is used for both inorganic and organic solids and
sludges. It is possible to treat liquids, although the effectiveness
of this is less certain. Fixation also is used for low-level radioac-
tive waste. Fixation of inorganic wastes is accomplished by solid-
ifying the waste in a cement, pozzolanic or lime based material.
Fixation of organic wastes usually requires the use of a glassifica-
tion, organic polymer or thermoplastic technique.
The major restrictive waste characteristics which determine the
applicability of fixation include the wastes' physical form,
organic content, ionic composition and sulfate content. Fixation
is most effective on solids, and sludges. The greater the moisture
content of the waste, the more difficult fixation is, although
liquids can be treated. A high organic content may limit or cause
problems when fixating an inorganic waste, because the organics
may interfere with the solidification of the solid product. Simi-
larly, high ionic compositions and high concentrations of sulfate
interfere with stabilization and solidification.
Ion Exchange
Ion exchange is the reversible interchange of ions between a
liquid and a solid phase. Ions in a liquid waste stream and ions on
the surface of an ion exchange resin are exchanged, purifying the
waste stream while concentrating the waste constituent on the
resin.
Ion exchange resins are either cationic or anionic. Cationic
resins exchange positive ions such as H+ or Na+ for negative
ions in solutions such as SQ-3. Anionic resins exchange negative
ions such as OH - and Cl - for positive ions in solution, such as
metal cations.
After an ion exchange resin has reached its capacity for ex-
changing ions, it is regenerated so that it can be used again.
Cationic ion exchangers are regenerated by passing dilute acid
solution over the resin; anionic resins are regenerated with a weak
base. During the regeneration process, a solution containing a
high concentration of the original waste constituent is produced.
This concentrated waste then must be treated or disposed of prop-
erly.
The restrictive waste characteristics for ion exchange include
suspended solids, metal content and organic contents, oxidizing
agents concentration and physical form. Ion exchange is most
effective for very dilute aqueous wastes. High concentrations of
solids, organics and oxidizing agents interfere with the exchange
of ions between the solution and the resins.
Neutralization
Neutralization involves the addition of an acid or a base to a
solution to adjust the pH, usually to between 6 and 9. Neutral-
ization often is performed in connection with other treatment
processes, such as oxidation or reduction. It can be either a final
treatment or a pretreatment step.
Alkaline neutralization is commonly accomplished using lime,
sodium hydroxide or soda ash. Acidic neutralization utilizes
hydrochloric, sulfuric or nitric acid. Neutralization is most applic-
able for acidic or alkaline aqueous wastes, although it also can be
used for certain organic liquids. Neutralization is a relatively
simple but important waste treatment process and most often is
followed by precipitation of metals from solution. This technol-
ogy can be used effectively for most aqueous wastes, although
wastes with very high dissolved solids content may result in the
formation of complexes which are difficult to remove from solu-
tion.
Precipitation
Waste removal by precipitation involves the addition of a
chemical substance to alter the equilibrium affecting dissolved
and suspended solids solubility. Precipitation normally requires
pH adjustment, followed by the addition of a precipitating agent,
resulting in flocculation and coagulation of the solids in solution.
Precipitation is used most commonly to remove heavy metals
from aqueous wastes. The precipitation process produces a sludge
composed of metal hydroxides, metal carbonates or metal sul-
fides as well as the precipitating agent used. In some instances,
precipitation can be used for organic-based liquids, although this
application is very limited due to sedimentation problems in vis-
cous media.
TREATMENT OF HAZARDOUS WASTES 155
-------
The major restrictive waste characteristics limiting the use of
precipitation include voscosity, physical form and metal solubility
constant. As mentioned before, highly viscous solutions, such as
many organic solvents, may not be applicable for treatment by
precipitation because of difficulties in settling the metals or other
particles out of solution. For highly soluble species, pH adjust-
ment is needed to minimize solubility and enhance precipitation.
CONCLUSIONS
The applicability of seven chemical treatment technologies for
hazardous liquid waste treatment has been examined. It was
found that the proper choice of treatment depends upon an
understanding of the restrictive waste characteristics for each
technology. Once the waste characteristics which restrict the use
of a particular treatment technology are known, proper pretreat-
ment or other treatment methods can be chosen.
156 TREATMENT OF HAZARDOUS WASTES
-------
Treatment Technologies for Hazardous Materials
Jeffrey M. Thomas
DETOX, Inc.
San Francisco, California
Phillip T. Jarboe
U.S. Ecology
Louisville, Kentucky
ABSTRACT
The nature and quantities of hazardous materials and the
various technologies for treating them have grown rapidly in re-
cent years. The efficiency and cost-effectiveness of technologies
such as physical/chemical treatment, metal precipitation,
biological treatment and carbon adsorption, vary enormously
depending on the application requirements. In most instances,
these treatment processes have been utilized to treat large volumes
of mixed matrix wastes.
Transferring these technologies for effective use in ground-
water treatment, with relatively low flow volume and a range 'of
contaminants from a single substance to a complex mixture, re-
quires special considerations in the design and operation of a
treatment program. The process design for these remedial pro-
grams must utilize the optimum mix of available technologies to
generate the highest degree of treatment at a reasonable cost.
This paper evaluates the process design requirements and
economic considerations for application of these processes to
groundwater treatment. It also examines the unique design and
operating conditions encountered in groundwater treatment. The
concept of "life-cycle" design is addressed to demonstrate the ef-
ficiency of optimizing the particular attributes of the different
technologies available.
INTRODUCTION
There is a broad range of potentially applicable technologies
for the elimination of hazardous materials. However, for each
material a specific review must be performed to select the most
appropriate form of treatment. During this review, consideration
must be given to the type of the contamination, organic versus in-
organic; the extent of the contamination, surface soil contamina-
tion alone, vadose zone, groundwater or any combination of
these; also to be considered are the physical and chemical char-
acteristics of the contaminants, the relative concentration of the
contaminants and the interactive effects the contaminants may
have with each other. In most instances after careful evaluation, it
will be found that a mix of appropriate technologies can be imple-
mented to produce the most efficient, cost-effective solution to
treat the contamination.
Most of the conventional treatment technologies are intended
for use on waterborne contaminants present in fairly constant
concentrations. One of the characteristics of most groundwater
contamination problems is their relatively short term nature (2-5
year treatment programs). These treatment programs initially will
be treating a relatively high concentration of contaminants. With
time, however, the concentration of these contaminants will
decrease. This concentration decrease is due to the removal of
chemicals as a result of the treatment, as well as the dilution effect
of clean groundwater passing through the zones of contamination.
As contaminated water is withdrawn from the ground (in either
the aquifer or vadose zone) clean water will either be drawn in
from the surrounding areas of the aquifer or, in the case of the
vadose zone, clean water used for flushing the contaminants will
displace the chemical compounds present in the interstitial spaces.
The net effect is a continuous, sometimes dramatic, decrease in
contaminant concentrations. This changing chemical composition
creates difficult design requirements for remedial actions. The
treatment programs implemented must be flexible enough to ac-
commodate the high concentrations initially encountered as well
as successfully treat the lower concentrations near the completion
of the program. In fact, most programs will see the highest con-
centrations of contaminants withdrawn over a relatively short
time period during the initiation of the treatment program. This is
a most important consideration as the efficiency and operating
costs of the treatment process will be seriously affected.
TREATMENT TECHNOLOGY REVIEW
There are several types of treatment processes which can be ap-
plied to the remediation of groundwater contamination. These
can be divided into three categories: physical/chemical treatment
of inorganic constituents; physical/chemical treatment of organic
materials and biological treatment of organic compounds. It is
important to understand the difference between the use of in situ
treatment as a technique for decontaminating groundwater and
soils and the use of process systems for the same purpose. When
using in situ techniques, materials (oxidants, microorganisms,
etc.) are introduced into the treatment area via wells, borings or
percolation beds. Under these conditions, there is relatively little
control over the reactions or the environmental conditions under
which they are occurring. Analysis of the performance of these
techniques is difficult to ascertain and most certainly involves a
time lag between the start of a program and any result which may
be reached.
In our approach to treating contaminated groundwaters and
soils, the contaminants and contaminated waters are withdrawn
from the ground and treated in surface reactors. This is the case
for all the technologies discussed in this paper. In most instances,
there is a requirement for pumping liquid into or out of the
ground. Since the surface reactors allow a much greater degree of
control over the process and better measurement of the results, we
find this to be much more appropriate for the treatment of con-
taminated materials. In some cases, the combination of both tech-
niques will yield the most benefit and should be employed.
PHYSICAL/CHEMICAL
TREATMENT—INORGANIC
The inorganic compounds most frequently encountered in
groundwater contamination are heavy metals. These materials
TREATMENT OF HAZARDOUS WASTES 157
-------
usually are not particularly mobile in the environment. Most
metals exist as a cationic species and have a tendency to adsorb
onto soil particles. The characteristics of some soils also allow an
"ion exchange" process to occur, thus removing the metals from
groundwater passing through a formation.
Hydroxide Precipitation
The metals which are found in extracted groundwater can be
treated most effectively through precipitation processes. Most of
the heavy metals are insoluble as a metal hydroxide or as a metal
sulfide. If they are not in a chelated form, the metals can quite
readily be removed by using pH adjustment and creating the
metal hydroxide. Typically, the minimum solubility of most
heavy metals is in the pH range of 7.5 to 9.5.
The steps necessary to determine metal removals involve ex-
perimentally determining the solubility of a metal (or metals) at
various pH levels by neutralizing a sample of the contaminated
water using lime or caustic soda. An insoluble precipitate will be
formed which can be physically removed, and the soluble concen-
trations of the metal(s) can be directly measured at each pH level.
In a full-scale treatment process, the groundwater will be
withdrawn and neutralized to the optimum pH level. The metal
hydroxide precipitate will be removed from the liquid stream via
flocculation and sedimentation. The clarified, metal-free effluent
then can be discharged. In many instances, this process must be
performed prior to processing for organic compound removal.
Sulfide Precipitation
The process of sulfide precipitation of metals works in a similar
fashion. The minimum solubility of any given species will vary
somewhat from that exhibited during hydroxide precipitation;
however, the soluble concentration of the metals after sulfide
precipitation will be considerably lower, as much as two or three
orders of magnitude lower than with hydroxides. For most treat-
ment program requirements, the levels which can be achieved via
hydroxide precipitation are more than adequate.
Chrome Removal
The process design for the removal of hexavalent chrome re-
quires an additional step prior to the precipitation process.
Chrome, in the hexavalent form, will not precipitate as a metal
hydroxide. It must first be reduced to the trivalent state using a
strong reducing agent such as sulfur dioxide or sodium bisulfite
under acidic pH conditions. The pH then can be raised using
either lime or sodium hydroxide and precipitation of the metal
will be effected.
System Design
The design criteria used in these treatment programs usually
will follow those which have long been applied in the metal
plating and processing industries. The systems applied to ground-
water treatment should be designed with ease of installation and
operation as major considerations. Once the treatment process is
developed and implemented, there is generally little in the way of
process variables which need to be controlled. Automatic pH con-
trol systems can be used to control neutralization; fiocculant feed
rates need to be monitored and adjusted periodically, but once
these systems are on line, there usually is little which can be ad-
justed. The use of lamella separators for liquid-solid separation
can substantially reduce the space required for this operation and
allow easy installation and operation of the system.
The use of metal precipitation is the first step in processing con-
taminated waters containing both heavy metals and organic com-
pounds. It is important to remove the metals at this point to
achieve and maintain high degrees of treatment efficiencies in
subsequent groundwater treatment processes.
PHYSICAL/CHEMICAL
TREATMENT—ORGANIC
Downhole Recovery
The category of physical/chemical treatment of organic com-
pounds is a broad one. The characteristics of groundwater con-
tamination are such that in many instances the compounds are
present in very high concentrations or as a pure material floating
on or settling below the groundwater. The first type of treatment
process to investigate would be recovery of pure product. Every
effort should be made to remove as much pure product as pos-
sible using either downhole recovery systems or surface
separators. The nature of most organic materials renders them
only partially soluble in water, and many compounds will
physically separate from water quite readily. There are a number
of downhole systems now available for recovering hydrocarbon
products from both the top and the bottom of a contaminated
aquifer.
The use of simple separation equipment at the surface also can
be implemented. This may consist of rope-type skimmers or even
separator tanks using tubes and plates. The cost of the treatment
program is substantially reduced when relatively pure product can
be recovered as opposed to having to process the material. In
many instances, the recovered product will have some economic
value and may partially offset some of the treatment costs.
There are also physical/chemical treatment processes for use on
low concentrations of organics in the liquid stream. Most notable
among these are air stripping, carbon adsorption, peroxide oxida-
tion and UV/ozone treatment. These are the traditional tech-
niques for removing (or destroying) organic constituents from
contaminated groundwater.
Air Stripping
Air stripping of organic compounds uses the volatility
characteristics of certain organic species, such as chlorinated
solvents, to eliminate them from an aqueous stream. The con-
taminated groundwater is pumped to the top of a column which
contains an inert packing. The water flow is downward from the
top of the column against a countercurrent air flow from the bot-
tom of the unit. The air exhausted from the column contains the
volatile organic compounds and is simply exhausted from the air
stripping column via exhaust ports at the top of the column.
Air stripping is an effective method for removing low concen-
trations of highly volatile compounds. The units are quite inex-
pensive to operate and require a minimum of operating attention.
There are, however, potential problems associated with their use
in groundwater treatment programs. Certain compounds do not
have particularly high volatility characteristics and do not readily
strip from water. In some instances, this problem can be over-
come by increasing the temperature of the contaminated water,
thus increasing the volatility of the organics, but heating can
become quite expensive.
Another consideration is the potential for fouling the internal
surfaces of the air strippers. The groundwaters which are heated
in the strippers generally will contain bacteria capable of
degrading the compounds present. Under the environmental con-
ditions found in an air stripper, it is possible to create substantial
biological growth on the internal packing surfaces as well as on
the mist eliminating packing at the top of the towers. In some in-
stances, this growth can lead to excessive maintenance re-
quirements for the operating units.
Another major consideration when evaluating the use of air
stripping is the ultimate fate of the stripped compounds. In the
past, these compounds have merely been vented to the at-
mosphere. This is not truly a treatment technique for the elimina-
tion of the compounds but rather a relocation of the problem
from the water to the air. There currently is a move by the
158 TREATMENT OF HAZARDOUS WASTES
-------
regulatory agencies in many areas to substantially reduce or elim-
inate the air emissions from these units. This control by air pollu-
tion authorities will mean implementing complex and expensive
emission control technology on the air streams exiting these units.
Activated Carbon
Carbon adsorption has been used to purify water for over 100
years. It is based upon the natural process of adsorption—that is
the natural affinity of a gas or liquid to be attracted to and held at
the surface of a solid. This attractive characteristic is due to the
surface tension of the solid holding the molecules in water at its
surface. Activated carbon particles are characterized by very high
surface area to volume ratios. The carbon particle has a structure
with a series of pores whose diameters decrease as the pores ex-
tend into the particle so that the small molecules will migrate fur-
ther into a carbon particle than a large molecule.
The adsorption of the molecule onto the carbon surface is a
physical process consisting of three steps: (1) diffusion of the
molecules through the liquid phase to the carbon particle; (2) dif-
fusion of the molecules through the macropores to the adsorption
site; and (3) adsorption of the molecule to the site. The
characteristics of the adsorbate molecule will determine the rate
of adsorption and, ultimately, the efficiency of the process.
Carbon adsorption has gained a very favorable reputation as a
means of removing a broad range of organic compounds from
water streams. It is quite effective at removing low level concen-
trations of organics, particularly chlorinated solvents. The ef-
ficiency of this process varies enormously, from as low as 3<7o to
as high as 25% utilization efficiency. Activated carbon is an ex-
tremely effective method for polishing treated effluents to remove
the last traces of contaminants.
Hydrogen Peroxide
Hydrogen peroxide is used in waste treatment applications as
an oxidizing agent. It is a moderately strong oxidizer of organic
compounds and is quite capable of oxidizing a broad range of
compounds at almost any concentration. The reaction involves
the formation of a free radical hydroxyl group. Hydrogen perox-
ide is a highly reactive chemical which reacts quite readily with
most organic materials (except saturated alkanes). Most reactions
involving the use of peroxides require the use of an iron catalyst
to effect the reaction.
Peroxides can be difficult to handle from a safety perspective
and are by nature unstable. Also, the reactions involved are quite
exothermic and can create significant hazards when used to treat
higher concentrations (>500 mg/1) of organic concentrations.
The most appropriate use for peroxide oxidation may be for the
treatment of highly biologically refractory materials, highly inter-
mittent waste flows or extremely low concentrations of organics
in a polishing operation.
UV/Ozone Treatment
The last major category of physical/chemical treatment tech-
niques for organic compounds is UV/ozone. UV/ozone treat-
ment technology seems to offer a great deal of potential for the
treatment of low concentrations of halogenated solvents. The
ozone is the source of a highly reactive free radical oxygen species.
This highly reactive free radical readily cleaves the double bonds
found in aromatic and unsaturated aliphatic compounds.
UV light increases the reactivity of the target molecules, par-
ticularly those bearing carbon-halogen bonds. The wavelength of
the ultraviolet light is in the absorption wavelength of this bond
and excites the bond to the point of breaking and cleaving the
halide from the carbon atom.
Some of the limitations of this technology are: the cost in
generating ozone; developing efficient UV light systems; reducing
the necessary contact times in the reactors and thus the number of
lamps and the electricity to operate them; and improving the
transfer of the generated ozone gas into the water stream.
The combination of UV and ozone is an excellent method for
treating extremely low levels of organics in water. This technology
is still in its development phase and has not been used in more
than a few pilot applications. One outstanding characteristic of
this technology and of the biological processes is that they both
are destructive technologies, converting hazardous organic com-
pounds into non-hazardous mineralized products of carbon di-
oxide and water.
BIOLOGICAL TREATMENT—ORGANICS
The use of the biological process in waste treatment involves
the creation of a growth environment in which bacteria are
capable of metabolizing organic compounds for energy for cell
maintenance and reproduction. Essentially, this process is the
conversion of a soluble organic compound into an insoluble
organic (cell protein) which can be physically removed from a
water stream. This technology is truly a destructive process, as the
end products of the biological process are carbon dioxide, water
and additional bacterial cells (sludge).
In order to accomplish this reaction, it is important to provide a
proper growth environment with consideration of inorganic
nutrients (nitrogen and phosphorous), pH, dissolved oxygen (for
aerobic processes) and several other process variables. The design
and operating conditions for a system of this type can be quite ex-
pensive using conventional activated sludge technology. Recent
developments in bioreactor and process technology have given
this process the opportunity to be one of the most useful tech-
niques for the treatment of hazardous materials.
The use of biological treatment for degrading organic chemicals
has been in practice for some decades now. The use of this tech-
nology on hazardous materials, and in particular on groundwater
contaminants, has not been particularly extensive until very
recently.
Activated Sludge
The process of activated sludge treatment is an operator-
intensive process which is sensitive to variations in influent
characteristics and operating conditions and requires close, careful
operator attention. One of the key elements in activated sludge
treatment is the concentration of large numbers of
microorganisms in a reaction vessel that effectively use an organic
substance for a food source. In order to concentrate these
organisms, liquid-solid separation must be effectively and care-
fully controlled. The use of fixed-film biological reactors instead
of the suspended growth activated sludge process has eliminated
the requirement for extensive operator attention and reduced the
critical nature of liquid-solid separation as part of the process.
Fixed Films
The use of a fixed medium surface allows the microbes to
become attached to a surface over which their food source
(substrate) passes. The organisms form a biological film into
which their substrate, nutrients and oxygen diffuse. In this
fashion, the organisms merely grow to a film thickness consistent
with the availability of substrate. As the concentration of
substrate diminishes, the rate of microbe growth decreases.
In a submerged fixed-film bioreactor (rotating biological con-
tactor), the release of air in a diffusion system beneath the
medium in the tank provides the means of both oxygen transfer
and mixing energy. This mixing energy creates a relatively high
velocity of air and water moving past the biological film. This
movement provides a means for shearing the accumulated bio-
logical growth from the fixed-film and for preventing bridging of
biological growth and plugging within the medium.
TREATMENT OF HAZARDOUS WASTES 159
-------
DIFFUSION OF
k;) OXYGEN
;:7 NUTRIENTS
•../ SUBSTRATE
TO MEDIA SURFACE
j] THROUGH FILM
* \
] SLIME LAYER TOO
A THICK
i . OXYGEN AND
'•H SUBSTRATE DO
';;,.! NOT REACH INNER
j I/ FILM
•A ANAEROBIC STATE
'•''\ CAUSES SLOUGHING
** AT MEDIA FILM
f
b
\
}
(
\
)
n
SLOUGHED
\\ t. INTERFACE / PARTICLE TO
\NAEROBICBIOFILM \ / EFFLUENT
PLASTIC MEDIA
ANAEROBIC LAYER
— INCREASING GROWTH »
Figure 1
Biological Film Growth and Sloughing
This mechanism of sloughing (Fig. 1) of the biomass is essen-
tially self-regulating. As the film becomes thicker due to the
growth of the microbes, its ability to remain attached to the
medium is reduced. Since the velocity of the air and water mixture
remains constant, the shear forces are constant. When the film
becomes thick enough, the shear forces are greater than the
adhesive forces and the film sloughs off the medium surface.
The effective concentration of microorganisms within the reactor
tank would be the equivalent of 6,000-8,000 mg/1 of mixed liquor
suspended solids in an activated sludge system. These biotreaters
can be placed within a fairly small reactor volume and are quite
readily preassembled and installed at a treatment location.
A minimum of operator expertise is required for these systems,
and they are becoming increasingly popular as methods of
destroying the hazardous materials. The biological process tends
to complete the degradation reaction, yielding carbon dioxide.
Perhaps 99% of all known organics can be effectively degraded
using biological treatment.
Slightly Contaminated Streams
Recent developments in biological treatment technology now
enable low concentration waste streams to be biologically treated.
In the past, all biological reactors were designed using the growth
characteristics of the biological populations as the limiting factors
to their effectiveness. It was not considered possible to use
biological processing on many wastes at a concentration contain-
ing less than 50 mg/1. This limitation is due to the nature of
biological growth dynamics, the effect of biological washout from
the reactors, loss of settling properties and other engineering con-
siderations. Low concentration reactors now have been developed
using acclimated populations for removal of low concentration
( <$ mg/1) contaminants to low fig/\ (1-3 /ig/1) effluent levels.
These reactors are portable, require almost no operator attention
and appear capable of operating several months on the low in-
fluent concentrations before needing to be reacclimated. This
technology probably will displace activated carbon in many in-
stances for the polishing of low level organic compounds.
Soil Treatment
The use of the biological process also has been successfully
adapted to the treatment of contaminated soils. It is possible, using
the proper engineering designs, to create a growth environment
for microorganisms in a soil environment. This process requires the
same environmental conditions as the treatment of a wastewater
stream, but they can be created and maintained quite readily.
The availability of commercial bacterial inocula has substan-
tially increased the ease with which the soils can be treated, and
these techniques offer substantial potential for eliminating a wide
range of soil contaminants such as pentachlorophenol, creosote,
hydrocarbons, pesticides and many other types of compounds.
The economics of soil treatment are also very favorable, and once
again this type of treatment is truly a destructive technology.
The current state-of-the-art in groundwater treatment technol-
ogies allows us to use these technologies now. However, the use of
these technologies as part of a remedial program must be carefully
scrutinized to ensure that they are consistent with the economics
and operating considerations of each individual program. For in-
stance, the use of a carbon adsorption process for treatment of
organics at a very high concentration (>100 mg/1) can create
more problems than it will solve. There are instances where highly
biodegradable compounds present in a carbon system have cre-
ated high degrees of biological activity within the columns and
have subsequently generated substantial amounts of sulfides in
the carbon column effluent. These sulfides then required an addi-
tional peroxide process to eliminate the odor and corrosion prob-
lems associated with them.
ECONOMICS
The economics of the various treatment technologies cover a
considerable range. The cost of any particular process can also vary
considerably depending on the degree of treatment required and the
time frame of the treatment program. The technologies ranked by
cost in descending order from most to least expensive are: peroxide
oxidation, activated carbon, biological and air stripping.
The costs vary as a function of contaminant concentration; as a
rule the cost will be higher for higher concentrations (on a $/gal
treated basis); the efficiency of each process (i.e., an easily stripped
compound) will cost less to treat per gallon than a less volatile
compound. The duration of the treatment program also will have
a significant effect on the economics of treatment since a longer
term program will allow for amortization of the equipment over a
longer lifespai. Also, some process equipment will have a higher
residual value ai the end of the program and can be sold or reused
in other applications. The following example compares the
economics of treating a waste stream via three different processes:
air stripping, carbon adsorption and biological/carbon adsorption.
Design Basis
Flow = 5 gal/min
Organic concentration = 300 mg/1 methyl ethyl ketone
Operating period = 2 years
The economic estimates are based on the process achieving 1
mg/1 MEK and 100 ng/\ total other organics in the effluent. The
cost for all capital equipment, operations and maintenance are in-
cluded in this analysis.
Carbon adsorption = $0.065/gal
Biological/carbon = $0.042/gal
Air stripping = $0.032/gal
The economics involved in the treatment of any material can
vary enormously depending on the specific requirements of the
application but, in general, the above figures are relatively ac-
curate. As the concentration of the contaminants to be treated in-
creases, the biological process becomes more economical. As the
volatility of the materials increases, the cost of air stripping im-
proves somewhat; however, this cost is difficult to lower much
further.
Other economic analyses have been performed on a $/lb re-
moved basis and have been reported to be as low as $0.40-0.50/lb
removed biologically to as high as $7.50-$10.00/lb removed with
carbon adsorption. A true economic evaluation of the actual costs
can be performed only on a specific case-by-case basis.
160 TREATMENT OF HAZARDOUS WASTES
-------
The most important consideration when considering the ap-
propriate technology is to optimize the cost advantage of each
process. As can be seen from the economics, there are many
possible combinations of technologies which can be put together
to produce a highly efficient, cost-effective treatment process.
TECHNOLOGY TRANSFER
The treatment requirements for hazardous materials in a
groundwater contamination problem are unique for several
reasons:
• Relatively low and/or variable flow conditions
• Varying influent concentrations
• Short time frame projects
• Ease of operation and simplicity of design
• Stringent discharge regulations
The design and operating requirements for these treatment
systems have to reconcile all these factors and at the same time of-
fer an efficient cost-effective system. The most important con-
sideration in designing these systems is to insure the systems will
provide the necessary degree of treatment throughout the lifetime
of the project while adjusting to the changing influent conditions.
Initially, the flow volume in these programs is generally quite
low. They rarely exceed 100-150 gal/min and quite often are in the
10-20 gal/min range. The actual flow over the course of the treat-
ment program may vary considerably due to seasonal conditions,
the transmissivity characteristics of the aquifer and percolation
rates of the overlying soils.
The concentration of the contaminants will also vary with a
general trend to decreasing concentrations. There periodically
may be increased contaminant levels as pockets of material are
freed, but the concentration will decrease over the life of the treat-
ment program. It is important to use a process design that will
cover the entire life-cycle of the project.
The time frame of most groundwater cleanup programs is
generally between 2-5 years. Most treatment systems for perma-
nent operation are designed on a lifespan of 10-20 years. It is
much easier to justify the costs of installing a major process on a
long-term system than on a relatively short-term system. It is most
important to minimize the amount of capital required to install a
short-term remedial project.
The systems being used in these treatment situations are
generally unfamiliar to many of the industries and individuals in-
volved in cleanup programs. There are many industries which,
until recently, merely discharged their wastewaters to public treat-
ment facilities and, as a result, have limited experience in waste-
water processing and the actual operation of wastewater treat-
ment facilities. In many instances, there is a minimal amount of
manpower available for operating a treatment system, or the loca-
tions of the treatment facilities are remote and not frequently
attended.
Taking all these factors into account, it becomes apparent that
the design of a treatment facility to be used in these situations is
indeed quite different than a typical wastewater treatment plant.
It is important to optimize the benefits of each technology. For
instance, it can be quite costly to provide a carbon adsorption
system to treat 50 gal/min of groundwater contaminated with 300
mg/1 of adsorbable solvent which will ultimately contain only a
few mg/1 of organics near the end of the treatment.
Preferable is a combined system using either air stripping or
biological treatment before the carbon system to reduce the
organic loading on the carbon and hence reduce the amount of
carbon which will be used. Both air stripping and biological treat-
ment are less expensive than the use of activated carbon for the
removal of high concentrations of organics.
In most cases, using modular, life-cycle designs will tend to
lower the costs and increase the efficiency of the tretment process.
For instance, the use of two fixed-film bioreactors to handle a
given flow and loading is preferable over using one large unit. As
the concentrations decrease, one of the units can be taken off line
and the operating costs and maintenance requirements will be
decreased substantially. Similarly, a carbon tank which is sized
for a certain flow and organic concentration may become an
operational problem if the flow and concentrations in the influent
drop substantially. Problems with carbon treatment include: (1)
channeling and carbon bed may develop biological growth which
can cause anaerobic conditions and subsequent sulfide problems.
When properly designed for life-cycle operation, the treatment
systems can provide a highly efficient, cost-effective process.
CASE HISTORIES
Phenol Contaminated Groundwater Treatment
An example of life-cycle design is the following treatment pro-
gram on the U.S. Gulf Coast. At the initiation of treatment, the
influent organic concentration was approximately 1,300 mg/1 as
TOC with phenol concentration at 400 mg/1. Over the lifespan of
the treatment program, it is known that the concentration of the
influent stream will decrease substantially until it is below 100
mg/1 TOC. It would be quite difficult to design and operate a
system that functions as effectively on 1,300 mg/1 as on 100 mg/1.
Obviously, activated carbon treatment alone would be pro-
hibitively expensive and would still require the disposal of large
quantities of spent carbon. The use of an air stripping process
also is not possible since the organic compounds present are not
readily strippable. In order to meet all the needs of this particular
program, the life-cycle design as shown in Figure 2 was im-
plemented.
I 300 ma/I TOC
Carbon
Adsorption
Aeration Aeration—
V
arifier
Fixed-Film
h
Blowers I
< 900 mq/1 TOC
< 300 mo/1 TOC
Carbon
Adsorption
« 100 ma/1 TQC
Carbon
Adsorption
Figure 2
Life-Cycle Design for Phenol Contaminated Oroundwater Treatment
TREATMENT OF HAZARDOUS WASTES 161
-------
In this installation, the aeration tanks were converted storage
tanks which, when coupled with a clarifier, were used as an ac-
tivated sludge process. As the concentrations decreased only one
tank was used for the treatment program. Eventually, that tank
was also removed from service and the fixed-film biotreater was
the remaining biological unit in operation. Throughout the dura-
tion of this program, the process design and equipment used were
changed to meet the prevailing conditions.
The cost of operating the system and the attention required for
its operation also diminished with time. Throughout the design of
the treatment process, those technologies used were applied to
provide the most economical operation. The activated sludge pro-
cess (using existing equipment) provided bulk organic removal;
the fixed-film unit provided additional treatment (below about
300 mg/1 TOC the activated sludge process would not be par-
ticularly effective) to reduce the organic loading to the carbon ad-
sorption system; and, ultimately, the carbon was used to remove
the lowest level of contaminants.
The nature of this program required that the effluent be treated
to a TOC concentration less than 18 mg/1 so that it could be used
for reinjection to force additional contaminated water toward the
withdrawal wells. Once biological activity was established in the
system, TOC was consistently biologically reduced by 75-85%.
Effluent phenol levels were consistently reduced by 99%. This
biological oxidation of organics substantially reduced the amount
of carbon required to attain the 18 mg/1 TOC effluent require-
ment. The influent to the system has decreased to less than 800
mg/1 TOC, and parts of the activated sludge system have been
taken off line. There is now a requirement for the removal of
arsenic from this particular waste. An arsenic precipitation system
is being added so that the arsenic can be co-precipitated with iron.
This system is being added at a minimal cost.
The economics involved in this particular case were based on
the use of carbon adsorption alone as compared to the cost of us-
ing biological processing. Assuming 23,000 gal/day flow at a
TOC of 1,300 mg/1, 249 pounds of organics need to be removed
from the water. Using a carbon cost of $0.75/lb and an operating
efficiency of 1 Ib organics removed/10 Ib carbon, the operating
cost for just carbon is $l,875/day. This cost does not include the
cost of capitalizing the equipment or other costs of operation such
as electricity and personnel.
The anticipated cost for the operation of a biological system
alone, including capitalization of the equipment, electricity, but
not personnel, is approximately $0.46/lb of organic material
removed. This is $115/day plus the cost of carbon for polishing
the effluent (assuming 50 Ib/day TOC is removed via carbon pol-
ishing) of $370/day. The total costs of operating this system are
$485/day. This system has been in operation for over two years,
resulting in a cost savings of $834,000. The actual costs of daily
operation should continue to diminish over the remainder of the
program.
Chemical Disposal Facility Leachale
The optimization of several technologies in combination has
been investigated to treat the leachate from a chemical disposal
facility. It has been determined that a plume of high TOC
material was migrating off-site and required treatment. A process
design using carbon adsorption alone and in conjunction with
biological treatment was evaluated in a pilot program.
Initially, 300 gal of sample were collected from the well at the
site. Analysis revealed 2,000 mg/1 TC (to compensate for in-
organic carbon, this value was reduced by 10% to compute a
TOC value). N,N- dimethyl formamide and propionic acid were
specifically identified.
A series of four carbon columns, each designed for a surface
loading of 2 gal/min ft2, processed the waste liquid. The utiliza-
tion efficiency of the carbon was found to be 25% (1 Ib organics
removed/4 Ib activated carbon used). The apparent attainable ef-
fluent level was 90 mg/1 as TC. The carbon was highly efficient in
its removal capabilities in this instance and was quite able to lower
the contaminant levels in the effluent. Even at this high efficien-
cy, the cost for treatment in replacement carbon costs alone is
about $4/lb organics removed.
A second series of evaluation was performed to determine the
viability of using biological treatment in conjunction with the ac-
tivated carbon in order to attain similar effluent levels at a re-
duced cost and, in addition, provide a destructive treatment of the
contaminant compounds. A simple biological reactor was de-
signed to provide degradation of the bulk of the organics present.
This reactor was followed by carbon adsorption for final polish-
ing and removal of refractory organic compounds. The biological
reactor was able to remove an average of 76% of the organic con-
centration of the liquid. The carbon adsorption system removed
the remainder of the TOC down to a level of 40-70 mg/1 with the
carbon system providing a utilization efficiency of about 16%.
The economics for the treatment of this waste stream break down
as follows:
Treatment System
Activated Carbon
1 gal/min
2 gal/min
Biological/Activated Carbon
1 gal/min
2 gal/min
Cost
(S/lb TOC removed)
8.50
6.90
4.50
3.50
These costs include both capital and operating expenses for the
treatment program.
On the basis of this economic analysis and the needs of this par-
ticular treatment program, a combination of biological/carbon
treatment likely will be used at this facility.
CONCLUSIONS
The adaption of conventional treatment technologies for the
treatment of hazardous materials in groundwater situtations is
clearly possible. However, the use of these technologies requires a
design basis unique to each application as well as an economic
justification unique to each application. Each application must be
evaluated independently to determine which technology or com-
bination of technologies is most appropriate for producing the
desired level of treatment on a cost-effective basis.
BIBLIOGRAPHY
Envirosphcre Co., "Evaluation of systems to stabilize waste piles or
deposits," Dec. 1983.
Nyer, E.K., Groundwater Treatment Technology', Van Nostrand Rein-
hold Company, New York, NY, 1985.
Wagner, K. and Kosin, Z., "In Situ Treatment," Proc. of the 6th Na-
tional Conference on Management of Controlled Hazardous Waste Sites,
Washington, DC, Nov. 1985, 221-230.
162 TREATMENT OF HAZARDOUS WASTES
-------
Update on the PACT Process
Harry W. Heath, Jr.
E.I. du Pont de Nemours & Company, Inc.
Chambers Works
Deepwater, New Jersey
ABSTRACT
This paper is an updated discussion of a 40 million gal/day
plant which uses the du Pont developed PACT* process to pro-
vide combined secondary/tertiary treatment to industrial waste-
water.
Reasons for selecting the PACT process are discussed. Special
features of the plant are reviewed including a 40 ton/day multiple
hearth furnace to regenerate powdered activated carbon and an
on-site, double-lined, secure hazardous waste landfill where
primary sludge from the plant is deposited. Operating data are
presented which demonstrate the effectiveness of the PACT pro-
cess including data on removal of priority pollutants.
Because of under-utilization of the plant in recent years, a suc-
cessful outside wastewater treatment business has developed. This
aspect is discussed as an example of the financial, technical and
environmental advantages of using a large, advanced treatment
facility to treat industrial wastewaters from a variety of sources.
INTRODUCTION
This paper is an updated case study of a 40 million gal/day
design industrial wastewater treatment plant (WWTP); several
papers discussing the initial years of operation of this plant were
presented in the early 1980s. This WWTP was built to treat the
waste from du Font's large Chambers Works site located on the
Delaware River estuary in southern New Jersey. The Chambers
Works site produces a large variety of organic intermediates,
mainly substituted aromatic compounds; tetraethyl lead; Freon
fluorocarbons; textile treating chemicals; fuel additives; and
miscellaneous other products. During the initial years of WWTP
operation, the site also made dyes, dye intermediates, sulfuric
acid and isocyanates. In addition, the WWTP treated acidic waste
from du Font's adjacent Carney's Point nitrocellulose plant,
since shut down.
Because the WWTP was greatly under-utilized by 1983, the site
began to actively seek outside aqueous wastes for treatment to
help reduce the fixed cost burden of the WWTP on the remaining
operations on the site. The last section of the discussion reviews
various aspects of this very successful venture.
The PACT process was invented specifically to treat the
original Chambers Works-Carney's Point waste stream. It added
powdered activated carbon to a conventional activated sludge
aerator and achieved a higher degree of treatment than could be
obtained with activated sludge alone. Table 1 summarizes the
composition of the waste stream expected to be treated during the
design stages of the WWTP.
*PACT is an acronym for Powdered Activated Carbon Treatment. The
PACT technology is now owned by Zimpro, Inc., which is making the
process available under license. "PACT System" is a registered
trademark of Zimpro, Inc.
Table 1
Design Load (Average) for Chambers Works WWTP
Flow
Soluble BOD
Color
Acidity
Dissolved Organic Carbon
Total Dissolved Solids
Total Suspended Solids
26,400 gal/min
88,700 Ib/day (280 mg/1)
1000 APHA
454,000 Ib/day* (1430 mg/1)
65,000 Ib/day (205 mg/1)
2000 to 5000 mg/1
80,000 Ib/day** (258 mg/1)
*Acidity expressed as CaCo^ equivalent throughout this paper.
**lncluded 56,000 Ib/day by product solids from lime neutralization.
This wastewater was salty, acidic (pH between 1 and 2) and
highly colored. In terms of organic concentration, the wastewater
was only medium strength; because of the high flow, the absolute
pounds loading was very large. Similarly, the total amount of acid
to be neutralized was high. The loading of dissolved organics was
highly variable, both in terms of composition and amount, in part
due to the batch nature of many Chambers Works processes. The
organic compounds to be treated included many not normally
amenable to biological treatment.
The first and third sections of the discussion review the factors
leading to selection of the PACT process and the subsequent per-
formance of the plant. The second section is a description of the
equipment, including the multiple hearth regeneration furnace for
activated carbon, to our knowledge the only use of this equipment
for powdered carbon regeneration in the world; the double-lined,
secure hazardous waste landfill for storage of primary sludge is
also described.
TECHNICAL AND ECONOMIC CONSIDERATIONS
IN SELECTING THE PACT PROCESS
The PACT process developed from technical studies of treat-
ment processes for Chambers Works-Carney's Point wastewater.
Compared to a conventional activated sludge process, the PACT
process with a dosage of 150 mg/1 virgin carbon at an 8-day
sludge age:
• Gave consistent BOP removals of over 95%
• Increased DOC removal from 62% to 85%
• Reduced color
• Minimized foaming during aeration
• Improved sludge settling and filtration properties
• Protected the microorganisms from shock loadings of organics
Data from typical continuous laboratory unit tests are shown in
Table 2. Table 3 shows similar data demonstrating the effect of
increasing carbon dose on system performance. The carbon used
in these tests had an iodine number between 800 and 1,000. Sig-
TREATMENT OF HAZARDOUS WASTES 163
-------
nificantly different performance would have been obtained with
carbon of different quality, although any PACT unit generally
will outperform a parallel activated sludge unit. Note in Table 2
that color actually was increased by activated sludge treatment.
By 1970, when design work on the Chambers Works WWTP
started, it had become obvious that proposed limits on DOC and
color would require some form of tertiary treatment. Review of
available technology showed the most promising alternatives to
PACT was granular carbon columns located either before or after
a conventional biological reactor. These options were studied in
detail.
Table 2
Typical Laboratory PACT Unit Performance*
Feed Concentration
BOD5 296 mg/l
DOC 203 mg/l
Color 1960 APHA
Effluent PACTJJnll Biological Unit
BOD5: Cone.
BOD5: Removal
DOC: Cone.
DOC: Removal
Color: Cone.
Color: Removal
Operating Conditions
Sludge Age, days 8.0 8.0
Carbon Dose, mg/l 148 0
Temp. Range, °C 15-24 15-24
Aerator Hydraulic
Res. Time (hr) 7.4 7.3
* Data for time period 1/22/77 thru 5/6/77.
Table 3
Laboratory Data Showing Effect of Carbon Dose
Effluent
11.5 mg/l
96.1%
29.1 mg/l
85.3%
395 APHA
79.9%
26.1 mg/l
91.2%
91. 2 mg/l
62.1%
2280 APHA
-16%
Soluble BOD5, mg/l
BOD Removal
DOC, mg/l
DOC Removal
Color, APHA
Color Removal
Feed
111
86
820
Activated
Sludge
7.5
93.2%
31.9
63%
700
20%
PACT
25 mg/
Carbon
Dose
6.6
94.1%
21.7
75%
320
63%
PACT
100 mg/l
Carbon
Dose
6.1
94.5%
13.1
85%
170
80%
Data represent three months' operation of continuous, laboratory 7 5 I units. Temperature range
for all units was 18 to 25°C, and hydraulic resident lime was 8 hr
In laboratory studies, neither process with carbon columns per-
formed as well as PACT. When carbon columns were placed after
the activated sludge unit, in the more conventional arrangement,
the biological system constantly experienced upsets caused by
shock loads of organics. With carbon columns before the biologi-
cal unit, the carbon was used inefficiently. Much of its com-
paratively expensive adsorption capacity was occupied by easily
biodegradable compounds. Further, as the composition of the
feed changed, it had a chromatographic effect on the carbon col-
umn. Thus, there were surges in the concentration of specific
organics in the effluent from the column, and the biological unit
did not receive the expected protection from organic shocks.
Investment costs for the PACT system were lower than for
either of the carbon column systems. Very generally, the invest-
ment difference was caused by the added expense for separate car-
bon columns as well as the expense for twice the clarification
capacity in the biological portion of the carbon column systems.
Because of the superior settling properties of PACT sludge, the
PACT system requires less clarifier capacity than conventional ac-
tivated sludge systems.
PACT process economics were compared with a conventional
activated sludge process for the Chambers Works, even though
the latter was not a viable alternate in this case. Cost estimates for
PACT secondary/tertiary treatment vs. secondary treatment by a
conventional activated sludge plant were within 10% of each
other. The extra PACT investment for carbon handling and
regeneration facilities was mostly offset by the additional clarifier
capacity and sludge disposal facility needed for the activated
sludge plant.
In effect, the PACT process at Chambers Works was expected
to provide tertiary treatment quality effluent at a cost close to that
for secondary treatment alone. In one piece of equipment, the
aerator, the PACT process destroyed most of the organic wastes
relatively cheaply by biological action, while using expensive car-
bon efficiently to remove only difficult to biodegrade or normally
non-biodegradable substances.
DESCRIPTION OF THE WWTP
Figure 1 is a flow diagram of the WWTP. It is easiest to con-
sider the plant in two sections: (1) primary section, where the
acidic wastewater is neutralized, most heavy metal removal occurs
and primary solids are settled, filtered and disposed of in the se-
cure landfill; and (2) secondary/tertiary PACT treatment section
where color and dissolved organics are removed.
Primary Treatment
The acidic water is neutralized with lime in a single stage
neutralization using three, stirred 200,000-gaI reactors
operated in parallel. The WWTP receives powdered lime by
rail car or truck. Powdered lime is stored in 4 large silos and is
slaked to 8-10% concentration as needed. Lime slurry is fed to
the neutralizes by an automatic control system that maintains
pH at any preset level. The WWTP currently consumes ap-
proximately 60 tons/day of dry lime to treat 100 tons/day
acid. The plant has enough capacity to treat 240 tons/day acid.
The highest monthly average load treated to date is 220 tons/-
day.
The neutralizes overflow to four primary clarifiers, each 1
million gallons in size. The rectangular clarifiers operate in
parallel and are 230 ft long, 55 ft wide and 12 ft deep. Solids in
the wastewater feed, as well as byproduct solids formed during
neutralization, settle and are removed as a 6-10% slurry in the
clarifier underflow. After filtration to a 45 to 50% solids cake,
the solids are hauled to the secure landfill. Two or three
(depending on the load) high-pressure (210 lb/in2 at end of cy-
cle) large recessed chamber filter presses make approximately 7
tons of wet press cake per 30 to 60 min cycle.
The WWTP currently generates from 100,000 to 200,000
Ib/day (dry basis) solids, equivalent to 43,000 to 85,000
ydVyear of landfill volume. The solids are mainly inorganic,
primarily calcium and magnesium salts, as well as silica com-
pounds from river water silt. These solids also contain small
amounts of heavy metals and a variety of organic compounds.
The solids are an inevitable byproduct of the primary treat-
ment process, and because they are primarily inorganic com-
pounds, the only feasible disposal method is landfilling.
164 TREATMENT OF HAZARDOUS WASTES
-------
Figure 1
Du Font's Chambers Works Wastewater Treatment Plant
KEY:
RR TRACK
CARBON
UNLOADING
11, iRRrrc.Ki, i.
111111111.1 h
©©©
AB =
AS =
C
CAW
F -
FP
FS
FT -
HP
L
LST =
NE -
P
RAS =
1°ST =
Afterburner
Acid Storage
Carbon Slurry Tank
Carbon Acid Wash Tank
Flocculator
Filter Press
Flow Splitter
Fuel Storage Tank
200 PSIO Filter Feed Pump
Lime Storage Silo
Lime Slurry Tank
Neutralization Tank
8,000 GPM Waste Water Feed Pump
Recycle Activated Sludge Pump
Primary Sludge Hold Tank
CLARIFIER
1,000,000 GAL
A diagram of the landfill is shown in Figure 2. It is located on
the Chambers Works site, and primary sludge is hauled there by
truck. A double liner of a chlorosulfonated polyethylene
material, Hypalon, covers the entire bottom and part of the sides
of the landfill. Collection pipes between the liners serve as leak
detectors and drain to sumps outside the landfill. The sumps are
sampled on a regular basis for any signs of contamination. Above
the top liner is a similar collection system for leachate which is
pumped back to the WWTP. The landfill ultimately will become a
70-ft high pyramid with a 15-acre base.
As outer sections of the landfill are filled, the edges and top are
covered with a 2-ft layer of essentially impervious, permeability of
less than 1 X 10-7 cm/sec, clay and 12 in. of top soil. Newer sec-
tions of the landfill also may be required to have a Hypalon liner
under the clay cap.
A controlling factor in the operation of the primary filter
presses is that the filter cake must meet rigid soil stability criteria.
These solids have to support the weight of heavy earth-moving
equipment. Further, the allowable slope of the landfill sides, and
hence the volumetric capacity of the landfill, is closely regulated
to assure a large safety factor against any possible slippage along
the slopes.
As backup protection, there are 26 monitor wells located
around the landfill; and in the very unlikely event of a double
liner failure, interceptor wells would be utilized to prevent any
harm to the environment while the problem was corrected.
PACT Secondary/Tertiary Treatment
PACT treatment starts at the flow splitter which feeds
neutralized primary effluent to the aerators. The activated carbon
is added here as a 1 Ib/gal aqueous slurry. The three, 185-ft
diameter, 4,000,000 gal aerators operate in parallel and provide 5
to 9 hr hydraulic resident time.
Air is used to suspend the mixed liquor suspended solids
(MLSS) which are an approximate 50/50 mix of biomass and ac-
tivated carbon. Initially, there were approximately 1150 static
mixers in each aerator to improve oxygen transfer. These aerators
have since been removed, leaving just air holes on the bottom of
the injection air laterals. The BOD loading to the WWTP was
never high enough to require the added gas transfer claimed for
the mixers. At low BOD loadings more air was required than
needed for the bacterial metabolism alone; the extra air flow was
needed to keep the MLSS well mixed, as the mixers appeared to
hinder gross solids circulation.
After aeration, water is fed through two parallel, 172-ft
diameter, 2,500,000 gal clarifiers. Feed enters through a
centerwell, flows through a flocculating zone and then beneath a
ring into the annular area bounded by the clarifier wall. The
treated wastewater overflows into a circumferential trough and
then to the plant effluent trench. It goes to a settling basin where
it mixes with approximately 1 Vi times its volume of non-contact
cooling water before discharge to the Delaware River estuary.
TREATMENT OF HAZARDOUS WASTES 165
-------
V TOP SOIL
2-CLAY
TOP LINER
BOTTOM LINER
1
\
• c
J<
7
LEACHATE SUMP
fs^
Figure 2
Secure Landfill
Rim drive bridges support rakes which push settled solids to the
center of the clarifiers; from there the solids flow via underground
lines to the recycle activated sludge (RAS) pumps. Two 78-in.
diameter screw pumps of 10,000 to 12,000 gal/min capacity each
are provided to pump the clarifier underflow or RAS back to the
aerators. Solids may be wasted from either the RAS stream or the
more dilute aerator discharge.
The WWTP can regenerate activated carbon from the wasted
MLSS. When the plant is regenerating carbon, the wasted slurry
is pumped through a thickener. Underflow from the thickener at
5 to 12% solids concentration is pumped to a 250 lb/in2 filter
press. The 64-in. diameter press has 112 recessed chambers 1 in.
deep. Each press cycle produces approximately 12,000 Ib of 38 to
45% solids press cake. This press is similar to the two other
presses used in primary service. In fact, one of the other two
presses is a "swing" press which also can be used to filter PACT
sludge. When not regenerating carbon, all three presses may be
used in primary service. The filtered PACT sludge is conveyed by
a series of drag flite, screw and belt conveyors to the top of a 26-ft
diameter, 40-ft high multiple hearth regeneration furnace.
There are five hearths in the furnace. Hot gases flow up the fur-
nace countercurrent to the sludge. The furnace has a rotating
centershaft with two rabble arms extending over each hearth.
Filter cake drops onto the top hearth and is raked from hearth to
hearth down through the furnace. From the fifth (and last)
hearth, the incandescent regenerated carbon drops into a water-
filled quench tank.
Some of the furnace's unique design features reduce upward
gas velocity and minimize entrainment of carbon particles. The
general principle of operation is that water is evaporated on the
first and second hearths at gas temperatures of 480 to 700°C. On
the third hearth, biological solids and adsorbed organics arc
volatilized by 750 to 870 °C gas. On the bottom two hearths, at gas
temperatures of 870 to 1020°C, the powdered carbon is thermally
regenerated in the presence of water vapor.
Off-gas from the first hearth passes through a two-stage water
scrubbing system and then to an oil-fired afterburner which
destroys odors and any organics at temperatures of 650 to 760°C.
Continuous water flow to the product quench tank is adjusted
to give a product slurry concentration of approximately 5%. The
regenerated carbon is washed in acid to remove inorganic ash.
Washings from this step are returned to primary treatment where
removed ash precipitates and is trapped in the primary sludge. The
washed, regenerated carbon slurry is transferred to a 60,000-gal
storage tank prior to recycle to the aerators. Three of these tanks
are provided to hold virgin and/or regenerative carbon.
In recent years, organic loading to the WWTP has decreased;
therefore, less activated carbon is needed. It has become more
economical to use only virgin carbon on a throw away basis. Cur-
rently, waste MLSS are returned to the influent steam to primary
treatment. The carbon and biomass are removed in underflow
from the primary clarifiers, making up about 10% of the filtered
solids deposited in the secure landfill.
WWTP PERFORMANCE
During the first years of operation, the PACT process exceeded
expectations in terms of BOD and color removals and approx-
imately equalled design in DOC removal. Data for early years are
shown in Table 4; in this period, the system was most heavily
loaded. Performance in subsequent years has been comparable,
but inlet DOC and BOD concentration have been significantly
lower. The flow rate also has decreased to approximately 20,000
gal/min.
166 TREATMENT OF HAZARDOUS WASTES
-------
Table 4
PACT System Performance, 1978 & 1979
Flow (gal/min)
BOD
Inlet Soluble BOD5, mg/1
Effluent Soluble BOD5, mg/1
Removal (%)
Color
Inlet Color, APHA
Effluent Color, APHA
Removal (%)
Aerator
Aerator MLSS (mg/1)
Sludge Age (days)
Carbon
Carbon Dose (mg/1)
Virgin (°7o)
Regenerated (%)
24-Month
Avg.
25,000
171
6.7
96
1,530
483
66
25,000
41
120
53*
47*
Design
Flowsheet
26,400
280
21
93
1,000
500
50
8,000
170
30
70
*The low ratio of regenerated to virgin carbon reflects start up problems with the filtration and
regeneration systems. In 1980 and 1981, the ratio was 70% regenerated/30% virgin at similar
total carbon dose.
The initial BOD load was less than expected because the mix of
organics to be treated shifted and became less biodegradable than
anticipated. The design ratio of BOD/DOC was 1.3 to 1.4, which
was the then current ratio in Chambers Works waste. However,
extensive waste reduction efforts prior to WWTP startup re-
moved a disproportionate amount of biodegradables. As can be
seen in Table 4, the BOD/DOC ratio in 1978 and 1979 was ap-
proximately 1.0.
The activated carbon in the aerator allowed the WWTP to treat
this difficult waste. One result of the lower BOD load was that the
minimum air flow to the aerators was set by the amount needed to
keep the biomass and carbon well suspended. This was about 13
ftVmin per static mixer. It had been expected that minimum air
flow would be set by the need to maintain a minimum dissolved
oxygen concentration of 0.5 mg/1 in the aerator. As was men-
tioned earlier, this was the reason the static mixers were removed
from the aerators in 1981. This allowed the air flow to be reduced
from 15,000 ftVmin per aerator to 8,000 ftVmin or lower, while
still maintaining 0.5 mg/1 oxygen levels.
The influent was more highly colored than was assumed during
design. However, color removal efficiency was better than antici-
pated, and the site was able to meet permit standards for color.
Average DOC removal was equal to design, so effluent quality
was slighly better than predicted, reflecting the lower influent
concentration.
In terms of meeting NPDES permit limits, variation in perfor-
mance is often more critical than long term average. Most permits
specify a daily maximum limit as well as a monthly average, and
consistently meeting the daily maximum is often the more de-
manding task for the process. The consistency of the PACT pro-
cess at Chambers Works is outstanding. Figures 3 and 4 are
histograms showing the distribution of feed and effluent DOC
and soluble BOD respectively in 1978 and 1979. Note the narrow
distribution of effluent concentration, particularly of BOD.
Analyses for the removal of U.S. EPA-designated priority
pollutants are in Table 5. Volatile organics and acid extractable
compounds generally are removed very well; base neutral com-
pounds are removed with a lesser amount of success. Some metals
removal occurs across PACT treatment, although the process is
not designed for this purpose.
N = 721
Avg = 170
Std D6V = 506
Max = 278
Mm = 80
90% CL
Log Normal
10
110 160
mg/llter DOC
4->
Q)
3
£
111
/
/
/.
/I
ill
, 90% CL
/Log Normal
y /
k /
V
i 1 1
10 60 110 160
mg/liter DOC
N
Avg
stci Dev
Max
Mm
— i
210
= 721
= 312
= 99
= 106
= 108
1 1
260
52%
60%
68% 76% 84% 92% 100%
% DOC Removal
Figure 3
DOC Histogram (1978 and 1979 Data)
A major reason for selection of PACT was the protection af-
forded the biosystem from upsets caused by shock loadings of
organic wastes. Because of the large number of batch processes at
Chambers Works, it was not impossible for the WWTP to see
20,000 to 50,000-lb slugs of various organics in a 1- to 4-hr period.
The biosystem handled documented spills to the WWTP of
50,000 Ibs of an aromatic diamine and 30,000 Ib of o-toluidine
with no ill effects. No equalization is provided before the WWTP
because pilot plant studies showed no significant improvement in
performance with equalization.
The demise of dyes manufacture on the site ended a major
source of organic shock loads. However, the increasing amounts
of trucked in outside wastes actually have made the overall prob-
lem more severe. A 5,000 gal truck is discharged into the 20,000
gal/min influent stream within 30 min. The PACT process has
tolerated the resultant shock loads when wastes with relative high
DOC content are received.
Initially, the WWTP operated at higher sludge ages, 20 to 60
days, than the 8 to 15 days specified in design. Organic removals
were improved at higher sludge age, with the result that the
WWTP operated at an average carbon dose of 120 mg/1. This
compares with a design dose of 170 mg/1. Because of lower
organic loadings, since 1983 the plant has operated at sludge ages
of 8 to 30 days with virgin carbon doses of 10 to 50 mg/1.
Because of the initial high solids concentrations in the aerators,
10,000 to 30,000 mg/1, the secondary clarifiers were heavily
overloaded. They were designed for a solids flux rate of 63 Ib/day
TREATMENT OF HAZARDOUS WASTES 167
-------
200 300 400
mg/llter BOD
500
90% CL
Log Normal
N
Avg
Std Dev
Max
Min
600
543
67
75
50
01
200 300 400
mg/llter BOD
500
500
n
0
Q>
5
N
Avg
Std Dev
Max
Min
2 60
= 543
= 961
= 36
= 999
= 66
68
1
76
90% CL
Gamma
84
% BOD Removal
f
100
£. , ** ~ LJ HI JL U
4-Nitroph
Phenol
METALS
Antimony
Arsenic
Beryllium
Cadmium
Chromium
Copper
Lead
Mercury
Nickel
Selenium
Zinc
Figure 4
BOD Histogram (1978 and 1979 Data)
Table 5
Priority Pollutant Removals Across PACT System
VOLATILE ORGANICS*
Benzene
Carbon Tetrachlorlde
Chlorobenzene
Chloroechane
Chloroform
Ethylbenzene
Methyl Chloride
Tetrachloroethylene
Toluene
1,1,1-Trlchloroe thane
Trichloroethylene
Trichlorofluoromethane
BASE-NEUTRAL EXTRACTABLES*
1,2-Dlchlorobenzene
2,4-Dinttrotoluene
2,6-Dinitrotoluene
Nitrobenzene
1,2,4-Trichlorobenzene
ACID EXTRACTABLES*
2-Chlorophenol
2,4-Dinitrophenol
Fetd
0>g/D
105
94
1720
280
201
41
1770
24
519
13
41
155
259
1900
1640
454
523
11
161
1020
489
24
19
1.4
5.6
26
72
112
.02
77
29
602
Concentration
Effluent
0-g/D
0.9
1.4
30
12
21
1.7
Nil
1.7
1.7
0.6
1.9
3.0
120
243
575
2
169
1.6
5.0
10
38
24
21
1.0
5.4
15
75
54
.01
59
23
360
Removal
(*)
99
99
98
96
90
96
99
93
99
95
95
98
54
87
65
99
68
85
97
99
92
Nil
Nil
29
4
42
Nil
52
50
23
21
40
The values in Table 5 arc averages of data whtch sho» significant day-to-day variations in
removals. Results were obtained by GC MS analyses of plant sample* and are subject to analytical
error. The figures represent only estimates of average percent remcnab for these materials.
ft2 at an underflow concentration of 3.6%. They actually
operated at solids flux rates from 200 to 350 Ib/day ft2 at
underflow concentrations of 4 to 8%. The carbon in the sludge
greatly enhanced the thickening abilities of the clarifiers.
Part of the reason for such heavy solic4, loading were initial
problems with the PACT sludge filter prf ^s and, surprisingly, the
conveying system to carry filtered slurge into the regeneration
furnace. During 1978 and 1979, the inability to remove solids
from the liquid train by normal wastage methods forced solids to
exit the system as total suspended solids (TSS) in the treated ef-
fluent overflowing the secondary clarifiers. Effluent TSS aver-
aged about 100 mg/1 vs. a design level of 30 mg/1. These solids
settled in the polishing basin and ultimately were dredged to the
primary treatment section of the WWTP.
From 1978 through March 1982 when the regeneration furnace
was shut down, there was a steady series of mechanical im-
provements in the PACT sludge filtration and conveying systems.
Not all portions of the system reached design rates, and some added
and/or redesigned equipment had to be installed.
The expected improvement in sludge filterability with the addi-
tion of carbon to the biomass was achieved. Filter cake solids
averaged 45% vs. a predicted 35% to 38% level. Press capacity
was below forecast because of: (1) lower dosage of carbon needed
with longer sludge age operation, and hence less carbon in the
sludge to aid filtration, and (2) operating problems with the filter
press, particularly with cloth life, cloth changing procedure and
high pressure feed pumps. The original steel press plates are being
replaced with lighter weight fiberglas or plastic plates that are
easier to dress and are expected to dramatically improve PACT
sludge filtration if the furnace resumes operation.
As anticipated, the regeneration furnace incurred relatively
high maintenance costs. There have been several failures of he
brick hearths. The combination of the above, plus numerous
design problems of the type often associated with a pioneering
chemical processing venture, limited furnace in time to just over
50% in the first two years of operation. In time of furnace opera-
tion improved from 55% in 1978-79 to 66% in 1980-81. Much of
the furnace downtime was caused by problems with the feed con-
veyors and not with the multiple hearth furnace itself.
When operating, the furnace produced an average of 24,000
Ib/day regenerated carbon in 1978-79. In 1980-81, production was
raised to 38,000 Ib/day. Recovery of virgin carbon properties
averaged just over 60% for the high activity carbons needed for
Chambers Works waste. Weight yield recovery was over 80%
after acid washing. As expected, the acid washing step was needed
to prevent buildup of inorganic ash on the carbon with successive
regenerations. However, during the last months of furnace opera-
tion, ash levels increased in regenerated carbon. Revisions to the
acid washing system to better remove iron and silica were sche-
duled when the decision was made to shut down the furnace.
Despite significant equipment improvements, it appears the
solids train will continue to incur relatively high maintenance
costs. As a result, alternate technologies have been considered.
168 TREATMENT OF HAZARDOUS WASTES
-------
One which shows great theoretical promise is wet air oxidation. It
would eliminate both filtration and sludge handling and replace
the furnace with a high pressure reactor. Whether this high pres-
sure/lower temperature liquid phase treatment can regenerate
Chambers Works spent carbon to the quality necessary to treat
this type of waste is not certain. Experience at other locations sug-
gests it may be able to do so, however there has not been suffi-
cient justification yet to do the large scale pilot studies necessary
to test the process for Chambers Works.
OUTSIDE WASTE BUSINESS
Operating experience at the Chambers Works WWTP over the
past 4 years has demonstrated the advantages of transporting
aqueous wastes to a large, central, advanced wastewater facility
for treatment. In the early 1980s, du Pont began aggressively
seeking outside wastes for treatment as it became apparent that
the WWTP was going to be under-utilized because of declining
production at the site. A marketing study showed this large
PACT plant had unique advantages in treating outside wastes.
Because of the great volume of wastewater being treated, the
WWTP had essentially an unlimited volumetric capacity for
trucked waste. Four tankers simultaneously discharging 5,000 gal
each in only 20 min increase influent flow less than 5%. Highly
acidic wastes are no problem. The WWTP has treated 2,500 gal of
almost pure chlorosulfonic acid in only 4 hr by direct discharge in-
to the influent trench with no release of fumes.
The tremendous dilution provided by over 20,000 gal/min flow
is a major protection against organic shock loads from a tank
truck discharge. The PACT system was selected specifically
because of its ability to handle varying loads of industrial waste.
Therefore, this plant is almost ideally suited to treat batch
discharges of aqueous wastes containing organics, especially those
which would pass through or be inhibitory to a conventional
biological treatment system.
The Chambers Works already had logistics in place to handle a
large number of truck deliveries each day and excellent interstate
highways leading to the site. This was vital because of the volume
of incoming tank trucks. Deliveries of over 50 trucks a day are
routine.
Because of other activities on the site, there were "in-house"
environmental and technical groups to support this activity. A
major technical effort is needed to evaluate wastes proposed for
treatment both in terms of their suitability to the PACT process
and the estimated costs of treatment. A strict protocol must be
established to approve a waste for treatment, and subsequent
deliveries must be monitored closely to assure that the waste ac-
tually received is what was contracted for. Because of the amount
of hazardous waste treated, there are frequent contacts with state
and federal regulatory authorities regarding operation of the
WWTP. A state inspector is at the site every 1 to 2 weeks, and it is
important that a competent environmental staff is available to
represent the plant.
From a regulator's point of view, it is advantageous to treat
wastes at a central site. The disposition of a large number of in-
dividual wastes can be monitored conveniently by a single inspec-
tor. The single, large plant usually will be better controlled than a
variety of small separate facilities, often operated by those less
than expert in waste treatment. The quality of the treated effluent
can be assured by a detailed analytical evaluation of a single ef-
fluent stream; such analyses might be impractical and prohibitive-
ly costly for a large number of individual sites.
Another advantage of this site is the receiving water. Chambers
Works main outfall discharges into an estuary, so salt content of
the stream is not critical. There are no downstream withdrawals
for drinking water purposes. The volume of flow at the Delaware
River estuary is immense, and so for a given absolute quantity of
residual waste the effluent concentration will be very low. Any
aqueous waste treatment facility will discharge some residual
material. It is an obvious advantage if the receiving water flow is
high; then the resultant concentrations of any residual pollutants
are low, and there is no adverse impact on downstream water use.
Since 1980, the outside waste business has grown rapidly. Cur-
rently, over 15% of the organic load to the WWTP is from non-
Chambers Works wastes. Over 50% of the outside waste
originates in New Jersey; 35% comes from Pennsylvania,
Delaware and Maryland. However, it has been cost-effective to
ship aqueous wastes surprisingly long distances. There are routine
contracts originating as far away as Maine, Florida and Col-
orado. The freight-logical area has been extended by offering rail
car service to transport wastes, and the potential exists for barge
deliveries.
Wastes from a variety of sources have been treated. Some are
listed in Table 6. No one type of waste makes up more than 20%
of the total.
Table 6
Typical Industrial Wastewaters Treated at Chambers Works
• Tank Truck & Tank Car Washings
• Pharmaceutical Wastes
• Water from Oil-Water Separation Processes
• Textile Treating Wastes
• Metals Treating Wastes
• Bio-sludges from Industrial Wastewater Treatment Plants
• Electronics Industry Waste
• Landfill Leachates
• Lagoon Cleanups
• Miscellaneous Chemical Process Wastes
• Latex Wastes
• Chemical Cleaning Rinse Waters
• Waste Acids or Bases
• Food Processing Wastes
• Paint & Dye Wastes
Initially, treatment of outside waste was limited to streams
which could be discharged directly into the WWTP influent as
rapidly as they flowed from the truck. This limitation precluded
certain aqueous wastes. For instance, wastes containing sulfide or
cyanide could not be accepted because of the possibility of releas-
ing toxic H2S or HCN gas when the waste mixed into the acid in-
fluent.
The WWTP has limits on most of the heavy metals. It is
capable of removing those present in the original wastewater feed
to meet NPDES permit limits. Initially, some outside wastes
which were evaluated had to be rejected because they contained
metals in amounts such that permit limits might have been ex-
ceeded had the waste been added directly to the WWTP influent.
Recently, heavy metals pretreatment facilities have been in-
stalled and are now being expanded to allow the WWTP to accept
aqueous waste streams containing higher amounts of heavy
metals. Similarly, the plant is developing treating techniques for
waste containing higher concentrations of sulfide and cyanide.
Facilities also are being installed for oil-water separation, with
the aqueous phase to be discharged to the WWTP and the oil to
be handled separately. Other expansions are being studied which
involve separate pretreatments which themselves generate
aqueous waste that could be handled by the WWTP. This might
give du Pont a competitive advantage over other stand alone
facilities.
TREATMENT OF HAZARDOUS WASTES 169
-------
This outside waste treatment has been successful enough that
du Pont is expanding the business beyond the Chambers Works
WWTP. In recognition of their expanded charter, the outside
waste business group was recently renamed du Pont Environmen-
tal Services and is actively seeking other waste treatment oppor-
tunities.
CONCLUSIONS
• PACT is an excellent, advanced treatment process for many
wastewaters containing normally non-biodegradable or slowly
biodegradable materials. The process can treat highly colored
wastes. At du Font's Chambers Works chemical complex, a
PACT plant removes over 96% of the BOD and over 80% of
the DOC from 30 to 40 million gal/day of wastewater.
• A PACT process is resistant to toxic upset, can tolerate shock
loads of organic compounds and is tolerant of rapid changes in
waste composition.
• The presence of carbon in the mixed liquor suspended solids in
a PACT system dramatically increases settling in the clarifier
and improves filtration rate. Solids concentrations of approxi-
mately 40% have been achieved with PACT sludge in a conven-
tional pressure filter.
• Conventional wastewater treatment equipment can be used for
a PACT system with a minimum amount of design changes.
• Powdered activated carbon has been successfully regenerated
from PACT sludges in a multiple hearth regeneration furnace
at a rate of 20 tons/day. Maintenance costs for the entire fil-
tration, conveying and regeneration unit are high, and to date
the best system in time has been about 65%.
• The Chambers Works PACT plant is successfully treating a
large variety of off plant wastes delivered by tank truck or
railroad car. Over 30% of the current organic load to this plant
is from outside wastes. Particular, demonstrated advantages of
a large central wastewater treatment are:
(1) An advanced treatment process is used.
(2) High quality "in house" high quality technical and en-
vironmental staffs can be afforded.
(3) It is sited on optimum receiving waters.
(4) It is easier for regulatory authorities to monitor the dis-
posal of the wastes.
(5) It can treat aqueous, organic waste streams which might
upset conventional municipal plants.
170 TREATMENT OF HAZARDOUS WASTES
-------
Treatment of PCB-Contaminated Soil
in a Circulating Bed Combustor
D.D. Jensen, Ph.D.
D.T. Young
GA Technologies Inc.
Process & Energy Systems Division
San Diego, California
ABSTRACT
A trial burn of PCB-contaminated soils was completed in GA
Technologies' 16-in. Pilot Plant Circulating Bed Combustor
(CBC). More than 4000 Ib of soil containing 1% PCB were
treated in three identical 4-hr runs at 1800°F. The results showed
excellent compliance with TSCA requirements. Destruction and
removal efficiencies (DREs) were greater than 99.9999%, and
PCB in combustor ash was less than 200 ppb. No chlorinated
dioxins or furans were detected in the stack gas, bed ash or fly
ash. In addition, no significant concentrations of other Products
of Incomplete Combustion (PICs) were detected. Combustion
efficiencies were greater than 99.9%, with CO concentrations
less than 50 ppm and NOX concentrations less than 75 ppm. Par-
ticulate emissions were generally below 0.08 grain/dscf, and HC1
emissions were maintained below 4.0 Ib/hr by introducing lime-
stone directly into the combustor.
These results demonstrate that the CBC is an environmentally
acceptable means of treating contaminated soil containing PCB
and other organic wastes. In addition, the high thermal efficiency
of the CBC, the absence of afterburners or scrubbers and the
use of simple feed systems make CBC treatment competitive with
soil removal and transport to landfills and other potential treat-
ment/disposal options.
INTRODUCTION
Of the many hazardous chemicals found in waste sites around
the country, perhaps none has received the scrutiny conferred on
polychlorinated biphenyls (PCBs). This group of 209 synthetic
chlorinated organic compounds found wide use as a dielectric
fluid in utility transformers and capacitors and as a high-tempera-
ture heat transfer medium.' However, because of their exception-
al resistance to degradation in the biosphere and apparent toxic-
ity, the manufacture and sale of PCBs were banned in 1976 for
virtually all purposes. The control, treatment and disposal of
PCBs were mandated by TSCA and currently are handled
through the U.S. EPA's Office of Toxic Substances.
Until recently, it has been common practice to remove PCB-
contaminated soils for burial in a secured landfill. However, this
option is becoming less desirable since landfill costs are escalat-
ing, the number of available landfill sites has decreased and gen-
erators or potential responsible parties (PRPs) retain the liability
associated with the contaminated soil, even in a secured landfill.
Treatment of PCB-contaminated soil by incineration in the CBC
can eliminate or significantly reduce the potential liability of gen-
erators or PRPs at a cost competitive with current landfill prices.
CBC units are designed to burn a wide variety of fuels such as
coal, peat, wood, municipal wastes and oil while treating haz-
ardous wastes. Over 25 units are operating or under construction
worldwide. Three units are currently in operation in the United
States. In 1983 GA began concentrating its efforts on the applica-
tion of GBC technology to incineration of hazardous wastes.
Table 1 presents examples of wastes that have been burned in the
CBC. It was the successful treatment of this diversity of wastes
that provided assurance that PCBs could be destroyed in a CBC
at a lower temperature than used in conventional incinerators.
Table 1
Circulating Bed Test Results1 for the Destruction of Hazardous Wastes
Waste
Carbon Tetrachlorlde
Freon
Malathlon
Dichlorobenzene
Aromatic Nltrlle
Trichloroethane
Form
Liquid
Liquid
Liquid
Sludge
Tacky
Liquid
Destruction
Efficiency, J
99.9992
99.9995
>99.9999
99.999
solid >99.9999
99.9999
HC1
Capture, %
99.3
99.7
99
99
Ca/Cl
Ratio
2.2
2.14
1 .7
1 .7
1 Results obtained in GA pilot plant CBC.
CBC DESCRIPTION
The CBC is a new generation of incinerator that uses high
velocity air to entrain circulating solids in a highly turbulent com-
bustion zone. This design allows combustion along the entire
length of the reaction zone. Because of its high thermal efficien-
cy, the CBC is ideally suited to treat low heat content feed, in-
cluding contaminated soil. Figure 1 shows the major components
of a CBC for soil treatment.
Soil is introduced into the combustor loop at the loop seal
where it immediately contacts hot recirculating soil from the hot
cyclone. Hazardous materials adhering to soil are rapidly heated
when introduced into the loop and continue to be exposed to high
temperatures throughout their residence time in the CBC. Upon
entering the combustor, high velocity air (14 to 20 ft/sec) en-
trains the circulating soil which travels upward through the com-
bustor into the hot cyclone. Retention times in the combustor
range from 2 sec for gases to approximately 30 min for larger
feed materials (^1.0 in.).
The cyclone separates the combustion gases from the hot solids
which are returned to the combustion chamber via a proprie-
tary non-mechanical seal. Hot flue gases and fly ash pass through
a convective gas cooler and on to a baghouse filter where fly
ash is removed. Filtered flue gas then exhausts to the atmosphere.
Heavier particles of purified soil remaining in the combustor
lower bed are removed slowly by a water-cooled ash conveyor
system.
TREATMENT OF HAZARDOUS WASTES 171
-------
COMIUSTOR
LIMESTONE
FEED
SOIL
FEED
COOLING
WATER
Figure I
Schematic Flow Diagram of Circulating Bed Combustor for
Soil Treatment
ASH
CONVEtO*
SYSTEM
As a consequence of the highly turbulent combustion zone,
temperatures around the entire combustion loop (combustion
chamber, hot cyclone, return leg) are uniform to within ±SO°F.
The uniform low temperature and high solids turbulence in the
CBC also help avoid ash slagging encountered in other types of
incinerators.
Acid gases formed during destruction reactions are rapidly cap-
tured by limestone added directly into the combustor. The reac-
tion of limestone and HC1, released during PCB incineration,
forms dry, benign calcium chloride. The rapid combustion and
quick neutralization of the acid gases within the combustion
chamber eliminate the need for afterburners and add-on scrub-
bers to complete destruction and acid gas capture. Emissions of
CO and NO, are controlled to low levels by excellent mixing, rela-
tively low temperatures (14SO to 1800°F) and staged combustion
achieved by injecting secondary air at higher locations in the com-
bustor. Because of its efficient combustion and highly turbulent
mixing, the CBC is capable of attaining required DREs for both
hazardous wastes (99.99%) and toxic wastes (99.9999%) at
temperatures below those used in conventional incinerators
(typically > 2000 °F).
TEST DESCRIPTION
A variety of requirements are imposed prior to and during a
PCB trial burn.' The key target of a trial burn is to ensure that
PCB DREs are ^99.9999% at the operating conditions chosen
for the incinerator. In addition, the concentration of PCB in ash
from the unit must not exceed 2 ppm. The potential formation of
PICs also is carefully evaluated, with particular attention given
to polychlorinated dibenzo-p-dioxins (PCDDs) and polychlori-
nated dibenzo-p-furans (PCDFs). The combustion efficiency of
the unit must be » 99.9fo, and particulate emission must not ex-
ceed 0.08 grain/dscf.
Figure 2
Pilot Plant CBC and Feed Preparation/Handling Equipment
172 TREATMENT OF HAZARDOUS WASTES
-------
The CBC trial burn was carried out in GA's 16-in. pilot plant
unit shown in Figure 2. This is the smallest CBC offered by GA
for commercial application. The glovebox in the foreground was
used to prepare contaminated soil prior to transport to the feed
system and CBC shown behind the glovebox. Soil treated in the
test was obtained from a former chemical processing site known
to contain pockets of PCB up to 6000 ppm as well as other
organic and inorganic wastes. To ensure that the CBC would be
permitted to treat all likely site concentrations of PCB, uncon-
taminated soil from the site was "spiked" with liquid PCB to
10,000 ppm. Spiking was carried out by blending a 50:50 com-
mercial mixture of PCB "1248" and trichlorobenzene with a rib-
bon blender in 1000 Ib lots. Approximately 4000 Ib of soil were
spiked for the three burns required by the TSCA trial burn
permit.
While the CBC was maintained at 1800T using natural gas
as the auxiliary fuel, several barrels of clean soil from the site
were treated in the combustion system prior to the addition of
spiked soil. During this time, all operating parameters and sys-
tem components were confirmed to be in the required operating
ranges. Process parameters monitored included:
• Temperature around the loop
• Pressure drop across the loop
• Soil feed rate
• Primary air flow
• Secondary air flow
• Loop seal air flow
• Total air flow
• Methane flow
• CO concentration
• CO2 concentration
• Excess oxygen level
• NOX concentration
Spiked soil was pneumatically transported to a bunker and
screw feeder. Soil feeding, limestone addition and stack gas
monitoring were started simultaneously. A U.S. EPA Modified
Method 5 sample train3 was used to sample stack gas emissions.
Table 2
PCB Trial Burn Operational Data and Test Results
Parameter
Teat Duration, hr
Operating Temperature, °F
Soil Feed Rate, Ib/hr
Total Soil Feed, Ib
PCB Concentration in Feed, ppm
DRE, J
PCB Concentration
' Bed Ash, ppm
Fly Ash, ppm
Dioxin/Furan Concentration
Stack Gas, pp
Bed Ash, ppm
Fly Ash, ppm
Combustion Efficiency, J
Acid Gas Release, Ib/hr
Particulate Emissions,
grain/dscf
Excess Oxygen, J
CO , ppm
C02. >
NOX, PPB
TSCA
Requirement 1
-1 1
1800
328
1592
11 ,000
>99.9999 99.999995
<2 0.0035
<2 0.066
ND
ND
ND
>99.9 99.91
<1.0 0.16
<0.08 0.095(b)
'
>3.0 7.9
35
6.2
26
Test Number
2
1
1800
112
1321
12,000
99.999981 99
0.033
0.0099
ND
ND
ND
99.95
0.58
0.013
6.8
28
6.0
25
3
1
1800
321
1711
9,800
.999977
0.186
0.0032
ND
ND
ND
99.97
0.70
0.0021
6.8
22
7.5
76
In addition, a separate Volatile Organic Sampling Train (VOST)4
was used to sample for volatile organic PICs. Feed, bed ash and
fly ash samples also were gathered throughout the test (see Fig. 1
for sample port locations). Three identical tests of spiked soil (4
hrs each) were carried out over two days in late May 1985. Each
test was observed and/or audited by U.S. EPA personnel or rep-
resentatives. All feed, ash and stack gas samples were subsequent-
ly analyzed for PCBs, PCDDs and PCDFs. Fly ash, bed ash and
stack gas samples also were analyzed for other PICs (both vola-
tile and semivolatile). Stack gases were analyzed for fly ash and
chloride release as well.
RESULTS
Table 2 presents a summary of the trial burn operational data
and test results gathered during the tests. Near-identical con-
ditions were maintained for each test. In each case, PCB DREs
were well in excess of the U.S. EPA-required 99.9999%. PCB
concentration in the bed ash and the fly ash was well below the
required 2 ppm and, in fact, did not exceed 200 ppb. No PCDDs
or PCDFs were detected in the stack gas, bed ash or fly ash. Com-
bustion efficiencies were greater than 99.9%, and acid gas release
was well below the required 4 Ib/hr. Paniculate emissions were
generally less than the required 0.08 grain/dscf. Only the grain
loading from the first test, obtained from a 2-hr makeup test
after the completion of Tests 1 through 3, showed a value slight-
ly higher than the limit. This higher than normal value is attrib-
uted to off-normal process conditions for the baghouse (i.e.,
excessive blowback air pressures along with a higher-than-normal
number of blowback cycles). Nitrogen oxides and CO levels re-
mained low as a result of the staged combustion utilized in the
CBC and the relatively low combustion temperature (1800T).
STACK
FLUE GAS
COOLER
8AGHOUSE
COMBUSTOR
ASH REMOVAL
FORCED
DRAFT
FAN
ND Not detected.
(^Derived from 2-hr makeup test.
Figure 3
Isometric of Site-Assembled Circulating Bed Combustor
TREATMENT OF HAZARDOUS WASTES 173
-------
These results demonstrate that the CBC is an effective means of
destroying PCBs contained in a soil matrix without the need for
high temperatures, afterburners or wet scrubbers. In particular,
the absence of undesirable combustion byproducts (i.e., dioxins
and furans) is important in ensuring that effective treatment of
soil can be accomplished in an environmentally acceptable man-
ner.
These results confirm the design of GA's transportable CBC
shown in Figure 3. The combustion and all other plant com-
ponents are designed as modular units which can be transported
by truck or rail. These units are assembled at the site into an
operating unit in 4 to 6 weeks. The major components of this
CBC plant include the combustor loop, feed system, pollution
control and air induction equipment. GA's 36-in. transportable
CBC is capable of processing up to 18,000 Ib/hr of dry soil on a
24-hr basis requiring an operating crew of only two persons per
shift. Soil treatment costs may be as low as $100/ton at a large
site. For smaller sites or sites having unique treatment require-
ments, costs may approach $400/ton.
CONCLUSIONS
The results of the PCB soil trial burn in GA's CBC demonstrate
compliance with TSCA requirements: stack emissions are well
within regulatory requirements and bed ash and fly ash contain
PCBs well below the regulatory requirements. The superior
thermal efficiency, high throughput and small staffing require-
ments of the CBC provide a soil treatment option that is cost-
competitive with landfill disposal while at the same time reducing
overall liability of the generator or PRP.
REFERENCES
I. SCS Engineers, Inc., "PCB Disposal Manual," Electric Power Re-
search Institute, Palo Alto, CA, Report No. CS-4098, June 1985.
2. "Polychlorinated Biphenyls (PCBs) Manufacture, Processing, Distri-
bution in Commerce and Use Prohibition," 40 CFR 761.70.
3. "Test Methods for Evaluating Solid Waste," U.S. EPA Report
SW-846, 2nd Edition, 1984.
4. "Proposed Sampling and Analytical Methodologies for Addition to
Test Methods for Evaluating Solid Waste: Physical/Chemical
Methods (SW-846, 2nd Edition)." U.S. EPA Report PB85-103026,
1984.
174 TREATMENT OF HAZARDOUS WASTES
-------
Treatment of Hazardous Waste Leachate
Judy L. McArdle
Michael M. Arozarena
William E. Gallagher, P.E.
PEI Associates, Inc.
Cincinnati, Ohio
Edward J. Opatken
U.S. Environmental Protection Agency
Hazardous Waste Engineering Research Laboratory
Cincinnati, Ohio
ABSTRACT
The nature of hazardous waste leachate varies greatly from site
to site and over time. Thus, leachate treatment systems must be
designed on a case-by-case basis. This paper presents profiles of
seven unit treatment operations which can be applied in varying
combinations to effect the desired degree of contaminant re-
moval: (1) equalization (to dampen influent flow and concentra-
tion fluctuations); (2) air stripping (to remove ammonia and vola-
tile organic compounds); (3) precipitation/flocculation/sedimen-
tation (to remove soluble heavy metals and suspended solids);
(4) neutralization (to adjust pH); (5) activated sludge treatment
(to remove aerobically biodegradable organic matter); (6) carbon
adsorption (to remove dissolved organics and certain inorgan-
ics); and (7) reverse osmosis (to remove dissolved inorganics and
high molecular weight organics). These technology profiles em-
phasize process applicability and limitations, treatment effective-
ness and capital and operating costs.
INTRODUCTION
Leachate is generated by the percolation of water (from precip-
itation, groundwater flow or liquid wastes) through a waste dis-
posal site. The leachate can be expected to contain soluble toxic
components of the waste as well as soluble chemical and biochem-
ical reaction products. Leachate collection systems are designed
to collect this contaminated liquid and to channel it away from
the disposal site before it can contaminate the surrounding soil,
groundwater or surface water. The collected leachate then can be
treated to reduce the level of contaminants prior to its discharge
into a stream, to groundwater recharge or to a municipal or in-
dustrial wastewater treatment system.
The composition and volume of leachate vary highly from site
to site and over time. Factors contributing to this variability in-
clude: the nature of the waste; the age of the land disposal unit;
the amount of precipitation; and the porosity, permeability and
adsorption characteristics of the soil. Table 1 presents leachate
characterization data from three land disposal sites.
The high strength and widely varying nature of hazardous
waste leachate complicate its treatment. Under contract to the
U.S. EPA's Hazardous Waste Engineering Research Laboratory,
PEI Associates, Inc., Cincinnati, Ohio, is preparing a compre-
hensive handbook to guide in the planning and design of leach-
ate treatment systems. Technology profiles emphasizing process
applicability and limitations, treatment effectiveness and capital
and operating costs are being developed for both demonstrated
and potential technologies. Six highly developed unit treatment
operations (equalization, air stripping, precipitation/flocculation/
sedimentation, neutralization, activated sludge tretment and car-
bon adsorption) and one potentially applicable technology (re-
verse osmosis) were selected for presentation here from the more
than 20 technologies to be profiled in the handbook.
Table 1
Leachate Characterization Data: Concentration of Contaminants
Parameter
Biochemical oxygen demand
(5-day)(BOD5)
Chemical oxygen demand (COD)
Total organic carbon
Total suspended solids
Total dissolved solids
pH, s.u.
Alkalinity (as CaCOj)
Hardness (as CaCOj)
Total Kjeldahl nitrogen
Ammonia-nitrogen
Nitrate-nitrogen
Nitrite-nitrogen
Total phosphorus (as P)
Phosphate (as P)
Sulfate
Sulfide
Chloride
Calcium
Magnesium
Sodium
Potassium
Cadmium
Chromium
Copper
Iron
Nickel
Lead
Zinc
Mercury
Love
Canal*
NRd
5,900-11.500
1,800-4,300
200-400
15,700
5.6-6.9
NR
NR
5
1
<0.1
<0.1
<0.1-3.2
<0.1
240
<0.1
9,500
2.500
NR
1,000
NR
0.01
0.27
0.54
30-330
0.24
0.3-0.4
0.48
<0.001
Stringfellpw
Acid Pits
<60
3,600
NR
6,300
24,900
3.37
NR
NR
30
10
14.0
<0.5
NR
2.5
NR
NR
300
400
540
NR
NR
NR
NR
NR
NR
NR
NR
NR
NR
G.R.O.W.S.
Landfillc
4,500-13,000
11,200-21,800
NR
500-2,000
11,200-14,200
6.6-7.3
4,800-5,700
3,100-5,100
600-1,700
700-2,000
NR
NR
NR
2.6-3.0
110-680
NR
3,100-4,800
650-900
250-650
1,200-1,500
950-970
0.04-0.10
0.16-0.43
0.32-0.44
180-380
0.55-2.0
0.4-0.8
8.7-31
0.005-0.012
Range
'60-13,000
3,600-21,800
1,800-4,300
200-6,300
11,200-24,900
3.3-7.3
4,800-5,700
3,100-5,100
5-1,700
1-2,000
<0. 1-14.0
<0.5
<0.1-3.2
<0. 1-3.0
110-680
<0.1
300-9,500
400-2.500
250-650
1,000-1,500
950-970
0.01-0.10
0.16-0.43
0.32-0.54
30-380
0.24-2.0
0.3-0.8
0.48-31
<0. 001-0. 012
aReference 1.
bReference 2.
cReference 3.
dNR = not reported.
eAll data except pH are in mg/L
TREATMENT OF HAZARDOUS WASTES 175
-------
FLOATING AERATOR
"Vf
77/777
Figure 1
Equalization Basin Geometry
UNIT TREATMENT OPERATIONS
Equalization
The objective of equalization is to dampen influent flow and
concentration fluctuations to improve performance of down-
stream processes. Because the composition and volume of leach-
ate from a given landfill fluctuate with time, equalization should
be considered in the planning and design of all leachate treat-
ment facilities.
Equalization basins may be of concrete, steel or (lined) earthen
construction. With in-line equalization, the entire daily flow
passes through the basin; with off-line equalization, only that
portion of the flow outside of an acceptable concentration range
is diverted into a basin. When the goal is equalization of the
pollutant mass loading (as in the treatment of hazardous waste
leachate), the in-line arrangement must be used. Typical basin
geometry is illustrated in Figure 1.
Mixing and aeration of equalization basins should be provided
to prevent both the deposition of solids and the onset of septic
conditions. As with any process using aeration, volatile com-
ponents can be stripped from the wastewater, which results in air
emissions of potentially toxic substances.
Table 2 presents capital and operating and maintenance (O&M)
equalization costs. These costs are based on the use of a concrete
basin with a 24-hour retention time. In general, capital costs of
equalization systems are moderately high because of the size of
the basins. For large basins, power costs for operating the aera-
tion system contribute significantly to the overall O&M costs.
Table 2
Leachate Treatment Costs*
(1985 dollars)
Unit operation
Equalization
Air stripping
Prec/floc/sed
Neutral 1zat1on
Activated sludge
Carbon adsorption
Reverse osmosis
25 gal/mln
Capital
101,000
34,500
115,000
10,000
137,000
79,400
582,000
04MD
0.422
0.472
2.66
0.09
0.792
2.35
6.92
50 gal/m1n
Capital
118,000
76,000
140,000
15.000
183.000
104,800
885,000
04H
0.243
0.236
1.62
0.08
0.517
1.72
3.60
100 gal/min
Capital
144.000
127,000
191,000
23,000
242.000
202,800
1,342,000
OiM
0.172
0.118
1.03
0.07
0.328
1.18
2.10
aReference4.
bS/lOOOgal based on 300 day/yr operation.
Air Stripping
Air stripping is a mass-transfer operation that uses forced air to
remove pollutants from a liquid phase. Applications of this pro-
cess include the removal of ammonia from wastewater and, more
recently, the removal of volatile organic contaminants from
drinking water. In the treatment of leachate, air stripping can
be used ahead of biological processes (to reduce toxic concen-
trations of ammonia) or ahead of adsorption processes (to reduce
the organic loading and thereby extend the life of the sorbent).
Among the various aeration devices that can be used in air
stripping are diffused aerators, mechanical aerators, spray tow-
ers and countercurrent flow packed towers. Packed towers pro-
vide the greatest gas-liquid interfacial area for mass transfer but
the shortest contact time. Removal efficiencies are dependent on
the height of the tower and the air-to-liquid ratio.
Figure 2 is a schematic diagram of a packed-tower air strip-
ping unit. The unit consists of a vertical column with randomly
dumped packing on a packing support. Common packings in-
clude the Raschig ring, Berl saddle, Intalox saddle and Pall ring.
Liquid is distributed uniformly on the top of the packing with
sprays or distribution trays and flows downward by gravity. Air
is blown upward through the packing and flows countercurrently
to the descending liquid. Depending on the concentration and
volatility of the contaminants, air-to-water volumetric ratios may
range from 10 to 1 up to 300 to I.' Air stripping of ammonia is
carried out at elevated pH (above 11).
AIR PLUS VOLATILES
LEACHATE
OEMISTER
LIQUID DISTRIBUTOR
PACKING
LIQUID REDISTRIBUTE
PACKING
PACKING SUPPORT
EFFLUENT
BLOWER
Figure 2
Schematic Diagram of Air Stripping in a Countercurrent Packed Tower
The volatile organics or ammonia stripped from the leachate
during treatment in a packed tower subsequently may be removed
from the off-gas to prevent air emissions of toxic compounds
or, possibly, for economic recovery. It also may be necessary to
pretreat the leachate for removal of suspended solids and dis-
solved metals that are oxidized to an insoluble form when aerated
to prevent fouling of the packed bed.
Capital and O&M costs for air stripping, presented in Table 2,
are based on the use of a countercurrent packed tower and in-
clude influent pumping and an ammonia recovery section. Costs
for labor, materials and power comprise the O&M costs of this
system. Because air requirements, and hence power require-
ments, increase at lower temperatures, operation of packed
towers at temperatures below freezing may be cost-prohibitive.
176 TREATMENT OF HAZARDOUS WASTES
-------
Precipitation/Flocculation/Sedimentation
The most common method of removing suspended solids and
soluble heavy metals (including arsenic, lead, cadmium and
chromium) from leachate is by combined precipitation/floccula-
tion/sedimentation. Precipitation involves the addition of chem-
icals to leachate to form insoluble precipitates from soluble spe-
cies. Flocculation promotes agglomeration of suspended solids,
which makes them easier to remove. Sedimentation is the removal
of suspended particles by gravity settling.
The processes of precipitation, flocculation and sedimentation
can be carried out in separate basins, as shown in Figure 3, or in a
single basin (e.g., an upflow solids-contact reactor-clarifier) with
separate zones for each process. Precipitation requires rapid mix-
ing (10 to 60 sec) to disperse the chemical, whereas flocculation
requires slow and gentle mixing (15 to 30 min) to promote particle
contact. Ancillary equipment requirements for these processes in-
clude mixers/paddles, chemical storage tanks and chemical feed
pumps. Packaged plants are available for low flow rates (10,000
gal/day to 2 million gal/day), but these may require extensive
modifications to enable them to handle the high level of precipi-
tated solids characteristic of hazardous waste leachates.
CHEMICAL
PRECIPITANTS
CHEMICAL
FLOCCULANTS
RAPID-MIX TANK
M M
PAOOLES
L>
43D-
-D
i
FLOCCULATION CHAMBER
SEDIMENTATION TANK
Figure 3
Schematic Diagram of Precipitation/Flocculation/Sedimentation
Metals can be precipitated from leachate as hydroxides or sul-
fides through the addition of lime or ferric sulfide. Hydroxide
precipitation with lime at high pH is the most commonly used
method. The presence of complexing agents such as ammonia or
cyanide in the leachate, however, will inhibit precipitation of
some heavy metals as hydroxides. Flocculants such as alum, ferric
chloride or polyelectrolytes often are added to the rapid-mix tank
along with the chemical precipitant to reduce repulsive forces
between particles and bring about particle aggregation and
settling.
Precipitation/flocculation produces large amounts of wet
sludge that may be hazardous because of its heavy metals con-
tents. This sludge, which is settled in a sedimentation unit, must
be treated further (e.g., by dewatering or fixation/stabilization)
before it is iandfilled.
Fluctuating leachate quality requires frequent jar testing to de-
termine appropriate chemicaraosages and removal efficiencies.
Available data indicate that precipitation/flocculation/sedimen-
tation can provide good removal of suspended solids (80 to 90%)
and moderate removal of BOD5 (40 to 70%) and COD (30 to
60%),' The additional treatment required will depend on effluent
discharge limitations.
Table 2 presents costs of a packaged precipitation/floccula-
tion/sedimentation system that includes a granular filtration unit.
Operation of these systems is labor-intensive, as reflected by the
moderately high O&M costs.
Neutralization
Neutralization involves the addition of an acid or a base to an
aqueous waste stream to adjust its pH to the desired level (usual-
ly between 6.0 and 9.0). Neutralization may be required as a pre-
treatment step to optimize the performance of downstream treat-
ment, to protect pH-sensitive processes (particularly biological
teatment operations) or as a final step to meet effluent criteria.
The technology is inexpensive, highly developed and widely used
in the treatment of hazardous waste leachate.
Typically, neutralization is carried out in completely mixed
corrosion-resistant tanks as shown in Figure 4. The tanks may be
operated in batch or continuous mode; however, the latter is only
suitable for flow rates greater than about 100,000 gal/day. The
neutralization process usually is controlled automatically by feed-
back, feedforward or multimode controllers. Common bases used
for neutralization include lime, calcium hydroxide, caustic, soda
ash and ammonium hydroxide; common acids include sulfuric
acid, hydrochloric acid and nitric acid. Reagent selection is based
on cost, speed of reaction and the reaction byproducts formed.
Equalization usually precedes neutralization in the treatment
process train to control fluctuations in influent flow and con-
centration. The neutralization process can produce heat and toxic
gases (ammonia, hydrogen sulfide and hydrogen cyanide) if the
chemicals are not properly added to and mixed with the waste.
Treatment to remove precipitated solids (flocculation/sedimen-
tation or filtration) often is required after neutralization. If the
removed solids are hazardous, secure disposal will be required.
Table 2 presents costs of a neutralization system consisting of
an agitated tank with a 3-min retention time, a metering pump
for acid or caustic and a pH control loop and valve. Both capital
and O&M costs are low in comparison with other unit treatment
operations.
pH CONTROLLER
ACID OR CAUSTIC-
2rl
FEED
PUMP
LEACHATE
• MIXER
" EFFLUENT
MIXING TANK
Figure 4
Schematic Diagram of Neutralization
Activated Sludge
Activated sludge is a biological treatment process in which an
active mass of microorganisms (bacteria, protozoa, rotifers and
fungi) is used to stabilize biodegradable organic matter under
aerobic conditions. Activated sludge can reduce concentrations of
a wide variety of organic compounds, including many toxic and
hazardous compounds. It is widely used in municipal and indus-
trial wastewater treatment applications (typical BODj concentra-
tion of 200 mg/1) and can effectively treat leachate with organic
concentrations one to two orders of magnitude higher (up to
10,000 mg/1 BOD5."
Conventional activated sludge systems, illustrated in Figure 5,
include two stages. The first stage involves aeration of the waste
in an open tank and maintenance of an active biomass. Aeration
is accomplished by mechanical-surface, diffused-air or sparged-
turbine aerators. The second stage entails separation of the solids
in a secondary clarifier. A portion of the solids is recycled back to
the aeration basin to maintain the desired concentration of organ-
isms, and the remaining portion is wasted.
TREATMENT OF HAZARDOUS WASTES 177
-------
AERATOR
LEACHATE
i
C
>
, I
D
AERATION TANK
RECYCLED SLUDGE ._
EFFLUENT
Figure 5
Schematic Diagram of Conventional Activated Sludge
Variations of the conventional activated sludge system have
been developed to provide greater tolerance for shock loadings,
to improve sludge settling characteristics and to achieve higher
BOD5 removals. Process modifications include complete-mix,
step aeration, extended aeration, contact stabilization and the use
of pure oxygen.
Because of the sensitivity of biological systems, pretreatment
requirements for high-strength leachate are extensive and may in-
clude equalization to buffer hydraulic and organic load varia-
tions; sedimentation/flotation to remove suspended solids, oil
and grease; neutralization to adjust the pH to near neutral; and
nutrient addition to provide adequate levels of nitrogen, phos-
phorus and trace elements. Acclimation of the biological system
to the influent waste stream is necessary prior to full-scale opera-
tion of the process. The presence of refractory or biologically
toxic compounds (e.g., ammonia or heavy metals) in the leachate
may necessitate the use of other physical/chemical processes in
conjunction with biological treatment.
Process residuals from activated sludge treatment of leachate
include waste activated sludge, which may be high in metals and
refractory organics, and air emissions of volatile organic com-
pounds that are stripped from the waste during aeration.
In general, biological processes such as activated sludge treat-
ment are the most cost-effective means for removing organics
from high-strength leachates. Table 2 presents costs of a conven-
tional activated sludge system consisting of an aeration tank with
a 6-hr retention time and a mixed liquor suspended solids con-
centration of 2000 mg/1, a secondary clarifier and a sludge re-
cycle pump. The O&M costs of activated sludge treatment are
comprised principally of labor and materials.
Carbon Adsorption
Carbon adsorption is a separation technique for removal of dis-
solved organics and certain inorganics (e.g., cyanide and chrom-
ium) from aqueous waste streams. This well-developed process
has numerous full-scale applications including treatment of
domestic and industrial wastewaters, cleanup of spilled haz-
ardous wastes. In leachate treatment applications, carbon adsorp-
tion can be used in the pretreatment, intermediate or polishing
steps.
Treatment of leachate by carbon adsorption involves passing
the waste stream through beds of granular activated carbon. Acti-
vated carbon is an amorphous form of carbon characterized by
a large internal surface area. Contaminants are adsorbed from the
waste onto the carbon surface by physical and chemical forces.
When the adsorptive capacity of the carbon has been reached,
regeneration or disposal of the spent carbon is required.
Most carbon adsorption systems use cylindrical pressure vessels
arranged in series or parallel and operated in a downflow, up-
flow or pulsed-bed mode. The downflow mode can handle a high-
er concentration of influent suspended solids than can the other
modes; however, frequent backwashing of the carbon beds may
be required to prevent excessive pressure drop. A schematic dia-
gram of an alternating, two-column, downflow system is pre-
sented in Figure 6.
PRETREATED
LEACHATE
COLUMN IN
ADSORPTION «•
STAGE
BACKWASH PLUS
CONTAMINANTS
EFFLUENT
Figure 6
Schematic Diagram of Alternating, Two-Column, Downflow,
Carbon Adsorption System
Carbon adsorption is applicable to the treatment of many
toxic and refractory organics; consequently, it often is used in
combination with biological treatment. Because the carbon beds
are susceptible to clogging, leachate must be pretreated to re-
move high concentrations of suspended solids, oil and grease. In
general, influent concentrations are limited to 10,000 mg/1 total
organic carbon, 1,000 mg/1 total inorganics, 2,000 mg/1 sus-
pended solids (downflow), 50 mg/1 suspended solids (upflow)
and 10 mg/1 oil and grease.' With proper pretreatment, removal
efficiencies greater than 99"% can be obtained.'
Table 2 presents capital and O&M costs of a carbon adsorp-
tion system. The O&M costs of this system are governed by the
organic loading, which is assumed to be 10 Ib organics per 100 Ib
of carbon. These costs do not include the costs of disposal or re-
generation of the spent carbon. For plants using less than 200
Ib/day of carbon (or, equivalently, for plants treating less than
800,000 gal/day of leachate), on-site regeneration of carbon
probably is not economical.' Most leachate treatment facilities
will fall into this range.
Reverse Osmosis
Reverse osmosis is a membrane separation technique which
primarily is utilized for the demineralization of water. Poten-
tially, this technique could be used as the final polishing step in
the treatment of leachate to separate dissolved solids [inorganic
salts and high (>120) molecular weight organics] from secon-
dary treatment effluent. With current technology, reverse osmosis
is not practical for the treatment of raw leachate because of rapid
fouling of the membrane and poor selectivity for low molecular
weight organics.'
In reverse osmosis, water is separated from dissolved solids in
solution by filtering it through a semipermeable membrane at
pressures of 200 to 1200 Ib/in1.' Basic components of a reverse
osmosis unit are the membrane, a membrane support structure,
a containing vessel and a high-pressure pump; these are illus-
trated schematically in Figure 7.
All commercially available membranes are structured asym-
metrically, with a thin (0.1 to 1.0 /un) dense surface ("skin")
supported by a porous substructure. This design promotes high
178 TREATMENT OF HAZARDOUS WASTES
-------
water transport across the membrane.8 Cellulose acetate and aro-
matic polyamides are common membrane materials.
Membrane support structures are of four types: tubular, spiral-
wound, hollow-fiber and plate-and-frame. The tubular configur-
ation, which consists of a rigid-walled porous tube lined on the
inside with a membrane, is recommended for use with waste-
water effluents.6 Reverse osmosis units can be arranged in parallel
(to increase hydraulic capacity) or in series (to effect the desired
degree of separation).
For optimal performance, feed water to the reverse osmosis
unit must be pretreated to remove gross amounts of solids and to
prevent fouling by scaling or biological growth inside the device.
In addition, compounds that are incompatible with the mem-
brane must be removed during pretreatment. Because the com-
position and strength of hazardous waste leachates vary widely,
pilot testing will be necessary to determine pretreatment require-
ments and separation performance. A major disadvantage of this
technology is that it produces a highly concentrated reject stream
that requires further treatment or disposal.
Table 2 presents costs of a reverse osmosis unit operating at
800 lb/in2. The O&M costs of this unit do not include the costs of
disposing of the concentrated reject stream. Regardless, reverse
osmosis is an extremely expensive technology and should be con-
sidered only for leachate treatment applications when an effluent
of extremely high quality is required.
PRESSURE VESSEL
HIGH-PRESSURE
PUMP
-»• EFFLUENT
'SEMI PERMEABLE
MEMBRANE
CONCENTRATED
REJECT STREAM
Figure 7
Schematic Diagram of Reverse Osmosis
CONCLUSIONS
The highly variable, complex nature of hazardous waste leach-
ate often requires that several unit treatment operations be
applied in combination to effect the desired degree of contami-
nant removal. Equalization is used to buffer influent flow and
concentration fluctuations and to improve the performance of
more costly downstream operations. Air stripping, neutraliza-
tion and chemical precipitation often are used ahead of biolog-
ical treatment (activated sludge) to remove biologically toxic com-
pounds such as ammonia and heavy metals and to adjust the pH
to near neutral. Carbon adsorption and reverse osmosis are used
as polishing steps to remove refractory organics and to meet final
effluent criteria. Unit treatment costs range from $0.07/1000 gal
for neutralization to $2.10/1000 gal for reverse osmosis.
REFERENCES
1. Shuckrow, A.J., Pajak, A.P. and Touhill, C.J., "Management of
Hazardous Waste Leachate," U.S. EPA Publication No. SW-871,
1982.
2. Copa, W.M., et al., "Powdered Activated Carbon Treatment
(PACT™) of Leachate From the Stringfellow Quarry," Proc. of the
Eleventh Annual Research Symposium on Land Disposal, Remedial
Action, Incineration and Treatment of Hazardous Waste, EPA-
600/9-85-028, Sept. 1985, 52-65.
3. Steiner, R.L., Keenan, J.D. and Fungaroli, A.A., "Demonstrating
Leachate Treatment: Report on a Full-Scale Operating Plant," U.S.
EPA Publication No. SW-758, 1979.
4. U.S. EPA, Handbook, Remedial Action at Waste Disposal Sites,
EPA-625/6-82-006, 1982.
5. Metcalf & Eddy, Inc., Briefing: "Technologies Applicable to Haz-
ardous Waste," prepared for the U.S. EPA Hazardous Waste Engi-
neering Research Laboratory, Cincinnati, OH, May 1985.
6. Metcalf & Eddy, Inc., Waste-water Engineering: Treatment, Disposal,
Reuse, 2nded., McGraw-Hill, Inc., New York, NY, 1979.
7. Chian, E.S.K. and DeWalle, F.B., "Evaluation of Leachate Treat-
ment, Vol. II: Biological and Physical-Chemical Processes," EPA-
600/2-77-186b, 1977.
8. Applegate, L.E., "Membrane Separation Processes," Chemical Engi-
neering, 71, 1984,64-89.
TREATMENT OF HAZARDOUS WASTES 179
-------
Manuscript Withdrawn at the Request
of The PQ Corporation
180 BARRIERS & WASTE SOLIDIFICATION
-------
Manuscript Withdrawn at the Request
of The PQ Corporation
BARRIERS & WASTE SOLIDIFICATION 181
-------
Manuscript Withdrawn at the Request
of The PQ Corporation
182 BARRIERS & WASTE SOLIDIFICATION
-------
Manuscript Withdrawn at the Request
of The PQ Corporation
BARRIERS & WASTE SOLIDIFICATION 183
-------
Manuscript Withdrawn at the Request
of The PQ Corporation
184 BARRIERS & WASTE SOLIDIFICATION
-------
Manuscript Withdrawn at the Request
of The PQ Corporation
BARRIERS & WASTE SOLIDIFICATION 18~5
-------
Innovative Techniques for the Evaluation
of Solidified Hazardous Waste Systems
Harvill C. Eaton
Marty E. Tittlebaum
Frank K. Cartledge
Hazardous Waste Research Center
Louisiana State University
Baton Rouge, Louisiana
ABSTRACT
Hazardous wastes solidified or stabilized by mixing with
cementing materials are described as complex composiie materials
with microslruclurcs and microchemisiries which determine their
properties. The behaxior of these materials and their possible
long-term effect on the environment can be determined by ex-
amining them with the tools common to studies in materials
science and engineering. Several of the microscopic and micro-
analytical techniques useful in these investigations are described
and selected results are shown. These techniques, together with
measurements of bulk physical properties, currently are being
utilized by the present authors to determine the mechanisms of
waste stabilization. The results are being used to design effec-
tive methods to handle real-world organic and inorganic haz-
ardous industrial wastes.
INTRODUCTION
Certain types of hazardous wastes can be converted into a
manageable form by solidification. For example, aqueous sludges
containing inorganic metal ions can be combined with various
cementing mixtures to form solids which are often disposable in
a secure landfill. Several advantages are potentially afforded by
solidification, including little or no free-standing liquids and re-
duced hydraulic pressures which might cause failure of the liner
system. If the permeability of the solidified waste is low, then
the leachate from monolithic structures may be free of toxic ions
after their initial depletion from exposed surfaces.
In the most desirable situation, the metal ions react chemically
with the matrix thereby forming a complex which is insoluble in
aqueous media and which is stable when leached by solutions with
a pH near those common in the environment or in a landfill.
Therefore, under the ideal situation, the waste is rendered effec-
tively non-hazardous. Consequently, solidification/stabilization
is potentially an important management alternative which may be
attractive for many hazardous wastes.
Unfortunately, little is known about the mechanisms of waste
solidification and even less about stabilization. Until recently,
system design was essentially empirical and based on industrial re-
search which usually remained proprietary. The relevant informa-
tion in the open literature dealt with nuclear waste stabilization'
or cement admixtures2 and, although useful, it more often than
not concerned the interaction of cementing matrices with com-
pounds which are not important to hazardous chemical waste
management.
Recently, however, it has been recognized that advances can be
made toward understanding the mechanisms of solidification and
stabilization by regarding the products as "complex composite
materials" with a "microstructure".1 Mechanistic information
which might result from this viewpoint is essential to establishing
u fundamental basis for the a priori design of a solidified or stabil-
i/cd waslc.
A material science point of view suggests that the measurable
properties (e.g., unconfined compressive strength and toxic ion
leachability) of the waste/matrix systems depend on the micro-
structure and the microchemistry. The experimental techniques
commonly used by the materials scientist offer ways to examine
the waste/matrix systems which may provide information about
the mechanisms of solidification and/or stabilization.
In this paper, the authors discuss the techniques that they have
found useful in their recent studies of solidified inorganic sludges
and of solidified organics.' The methods can be conveniently
grouped into those which provide morphological information and
those which provide chemical information. They are briefly de-
scribed and representative results are presented.
SAMPLE PREPARATION
The authors' investigations primarily have involved mixtures of
hazardous chemical wastes and Type I Portland cement. Portland
cement was chosen because it is a common component of many
solidification formulae which have been proposed and of many
which currently are being used by industry. Mixtures of pure
Portland cement and hazardous wastes are unlikely to be environ-
mentally, technically or economically practical, but they do offer
some of the laboratory control that is important to a basic study.
Other, more complicated, matrices presently are being examined
in the authors' laboratories as part of a study of interference
mechanisms. The work is funded by the U.S. EPA and the U.S.
Army Corps of Engineers.
A second application of the methods is part of an investigation
of test methods for evaluating real solidified wastes. The latter is a
cooperative study involving the U.S. EPA, Environment Canada
and several vendors in the United States, Canada and abroad.
Weighed amounts of waste and water were mixed with 10 g of
Portland cement, contained in capped vials and stored at room
temperature in a darkened room to avoid photochemical reaction
of the organics. Samples were allowed to cure for various times
ranging from 1 week to over 1 year.
Three organics have been investigated: p-bromophenol, p-
chlorophenol and ethylene glycol. The reasons for selection of
these compounds and detailed descriptions of their morphology,
microchemistry and leaching behavior have been described else-
where.5'6
MICROSTRUCTURE
The microstructure of a solidified waste depends on the follow-
ing factors:
186 BARRIERS & WASTE SOLIDIFICATION
-------
• The chemical composition of the waste
• The chemical composition of the solidifying matrix compon-
ents (cement, fly ash, clay, etc.)
• Relative amounts of waste and matrix components
• Amount of water
• Extent of mixing
• Temperature of mixing
• Time and temperature of set
In the studies performed by the authors, these factors have
been considered and maintained constant when appropriate, even
though in actual practice most of the factors can vary consider-
ably. It will be important, therefore, for researchers to estab-
lish how strongly these factors affect the gross properties of the
final product.
Optical Microscopy
Microstructure can be observed at several microscopic levels.
For example, entrained air can form bubbles ranging in size from
0.1 to 1 mm in diameter. These bubbles may be important in the
development of mechanical strength in the solidified product. Air
bubbles also can be the source of carbon dioxide to the matrix.
Carbonation can affect the hydration of cement paste by promot-
ing the formation of calcium carbonate which then decomposes to
a mixture of calcium carbonate and hydrated silica.7 Microstruc-
tural features of this size are resolved easily with the optical
microscope.
There are several types of optical microscopes. They are con-
veniently categorized according to the branch of science which
they serve (e.g., the biological microscope, the petrographic
microscope and the metallographic microscope). The petro-
graphic microscope and, to a lesser extent, the metallographic
microscopes are the most useful for examining microstructures in
solidified and stabilized wastes.
'.'.
/"^ -..;
VfeP
100^, "Vf
; «>w
' O
Figure 1
An Optical Micrograph of a Solidified P-Bromophenol Sample. A
Bimodal Grain Structure is Revealed.
Thin sections can be prepared, using methods standard to geo-
logical practice, so that light can be transmitted directly through
a small piece of solidified material. Using this method, the grain
structure can be observed. Figure 1 is an optical micrograph
which reveals the polycrystalline structure of solidified p-bromo-
phenol. The Portland cement matrix material is apparent in the
microstructure with small grain-like hydration phases. There is
an equiaxed grain structure which is clearly visible at 100X mag-
nification, and the presence of several mineral phases is apparent.
If polarized light is used, the crystalline and noncrystalline com-
ponents of the solidified system can be distinguished. It also is
possible to do phase analysis of the crystalline components us-
ing index of refraction liquids. It is, therefore, in principle, pos-
sible to determine the compounds formed due to the presence of
the hazardous waste. In practice, this determination can be very
difficult and the researcher likely will more successfully identify
than locate the waste containing phases. This identification is
illustrated in Figure 2 where the arrow indicates a mineral, i.e.,
crystalline, phase which contains hazardous p-bromophenol.
u.
Figure 2
A Higher Magnification Optical Micrograph of the Material Shown in
Figure 1. The Arrow Points to a Phase Which is Concentrated in
P-Bromophenol.
When only morphological information is required, such as in
studies of air entrainment, the metallurgical microscope is use-
ful. With the metallurgical microscope, light is directed down the
microscope tube and is reflected off of the specimen surface.
Reflected light travels back up the tube and through the ocular to
be detected by the operator's eye or to be recorded on a photo-
graphic plate or film. The chief advantage in the use of the
metallurgical microscope over the petrographic microscope is in
specimen preparation. Since light is not transmitted through the
specimen, only a polished surface is required.
Scanning Electron Microscopy
Microstructure on a smaller scale is apparent when the several
minerals formed upon hydration of Portland cement are con-
sidered. For example, calcium hydroxide crystallites are formed
and often are observed as hexagonal plates several microns in
extent. Even smaller crystallites of ettringite are observed in
young cement pastes. Ettringite usually appears as rod or needle
shaped, single-crystals. They contain sulphur and are only a few
microns long. The ettringite group of minerals form the AFt
(A = Al, F = Fe and t = three moles of CaO.SO3) phases.'
Both the calcium hydroxide and the AFt phases are observed best
by scanning electron microscopy (SEM).
In the scanning electron microscope, a primary electron beam
is focused onto a specimen surface and rastered over a small area.
Primary electrons interact with the waste and matrix atoms con-
tained in the specimen producing additional electrons. These elec-
trons are released from the specimen atoms themselves and can be
BARRIERS & WASTE SOLIDIFICATION 187
-------
released through several mechanisms. The two electron genera-
tion mechanisms particularly useful for waste studies are secon-
dary electrons and backscattered electrons. A secondary electron
detector or a backscattered electron detector placed near the spec-
imen can be used to produce an image of the variations in their
generation as the primary beam is rastered across the specimen
surface. The utility of the SEM is related to the following char-
acteristics of its use and image contrast mechanisms:
• Magnification variable from 10X to ca. 100X with a resolution
of up to 5 nm or less
• Magnification variable
• Large depth of field
• East of specimen preparation
Specimens can be prepared by grinding and polishing in a man-
ner identical to that used for the preparation of specimens for
the petrographic microscope or for the metallurgical microscope.
The authors, in fact, frequently use the same specimen for both
types of examination. If three mutually orthogonal sections are
examined, the researcher usually can eliminate the possibility of
mistaking textured for uniform structure. Polished sections must
be carefully studied, however, to eliminate artifacts caused by the
preferred removal of certain phases during grinding and polish-
ing. Fractured surfaces can be useful in assuring that this is not
occurring. Figure 3 is an example of a scanning electron micro-
graph of solidified ethylene glycol.
Figure 3
A Scanning Electron Micrograph of a Solidified Ethylene Glycol
Sample
Transmission Electron Microscopy
When Portland cement is a component of the solidification or
stabilization matrix, a major hydration product is calcium sili-
cate hydrate (CSH). CSH is a poorly crystalline material which,
like the AFt phases, is comprised of several different compounds
which are mineralogically similar. When viewed by the optical
microscope or the SEM, CSH-appears as the dominant or matrix
phase of a solidified hazardous waste. The SEM sometimes is
able to resolve Its morphology but frequently reveals it as a fea-
tureless, but sometimes nodular or grainy, material in which well
developed crystals like calcium hydroxide or AFt grow.
The transmission electron microscope (TEM) is the instrument
of choice for studying the morphology of CSH and other very
small structures. The TEM is capable of resolving features as
small as a few angstrom units in cementitous materials. Unfor-
tunately, specimen preparation usually is more difficult than for
either optical or scanning electron microscopy.
A stationary electron beam is focused onto a specimen which
intersects the optical axis of the instrument. The transmitted
beam is scattered and produces an image with contrast which is
primarily a function of the mass-thickness of the noncrystalline
components. Crystalline components scatter the specimen in
more complex ways due to diffraction. Figure 4 is a transmission
electron micrograph of a small crystalline phase in a solidified
waste.
Figure 4
A Transmission Electron Micrograph of a Ca(OH)j Crystal (the large
dark structure at (he center of (he micrograph) in a Solidified Ethylene
Glycol Sample
MICROCHEM1STRY
Bulk analytical instruments are an important and necessary
part of any investigation of the solidification and stabilization of
a hazardous chemical waste. If mechanistic information is desir-
able, it is useful to have chemical analyses of the small phases
present in the solidified material. Consequently, special methods
must be used.
Electron Diffraction
The high resolution images obtainable from the transmission
electron microscope and the ability to photograph diffraction pit-
terns make the TEM very useful for solidification/stabilization
research. Chemical identification of mineral components of the
solidified waste is possible by analysis of the electron diffraction
patterns.
Figure 5 is an example of the results which can be obtained
from examination of a solidified hazardous waste by transmission
electron microscopy. The dark structure indicated by the arrow is
a calcium hydroxide crystal as revealed by the diffraction pattern.
Energy Dispersive and Wavelength Dispersive
X-Ray Analysis
Specimens examined in the SEM emit x-rays in addition to elec-
trons. The x-ray energies are characteristic of the elements from
which they come. This means that they can be used to determine
the elemental composition of the specimen. The method is known
as energy dispersive x-ray analysis (EDS). Most EDS spectromet-
ers will detect elements with atomic number 11 (Na) or greater.
188 BARRIERS & WASTE SOLIDIFICATION
-------
Figure 5
A Transmission Electron Micrograph of a Ca (OH)2 Crystal in a
Solidified P-Bromophenol Sample. The Diffraction Pattern was Used to
Confirm the Composition of the Crystal.
Those spectrometers equipped with a windowless detector can de-
tect even lighter elements and are particularly useful for carbon
and oxygen determination. A similar technique is commonly asso-
ciated with electron probe microanalysis (EPMA) and is known as
wavelength dispersive x-ray analysis (WDS).
It is obvious that EDS and WDS would be useful in studies of
inorganic metal ion stabilization, but in certain cases they also can
be used in studies of organic waste solidification and stabiliza-
tion. The authors have used the techniques to study the partition-
ing of selected organics to the hydration products of Portland
cement. Two organics, p-bromophenol and p-chlorophenol, were
selected because they were molecules which contained a heavy ele-
ment which could be detected by EDS and WDS. The crystalline
phase shown by the arrow in Figure 2 was determined by WDS to
contain large amounts of Br and, therefore, to be rich in organic.
X-Ray Power Diffraction
X-ray powder diffraction (XRD) is very useful for studies of
waste solidification. A bulk analytical technique, XRD often can
be used to identify the exact chemical compounds present in the
solidified system. This technique is generally familiar, so it will
not be discussed in this paper.
CONCLUSIONS
Studies of hazardous waste solidification/stabilization are com-
plicated by the fact that cementing mixtures are used as cementing
matrices. Upon hydration, a complex composite material is
formed. The microstructure of this material has structure at levels
which vary from that visible to the eye to that which can only be
resolved by electron microscopy. The properties of the solidified
system are determined by the microstructure and the microchem-
istry.
Incorporation of hazardous wastes into the matrix by mixing
prior to hydration introduces the possibility that the wastes them-
selves are chemically involved with the hydration reactions and
therefore are stabilized by the process. In addition, there is a pos-
sibility that the wastes are physically contained by the matrix in
pores or interstitial spaces. These sites of physical entrapment
may be so small that they are difficult to locate even by micro-
scopic methods.
These characteristics suggest the use of microscopic and micro-
analytical methods for studying the mechanisms of waste solidifi-
cation and stabilization. The various methods used by the present
authors have been discussed. Some of the results of these studies
have been shown in the form of photomicrographs which reveal
selected characteristics of cement/organic waste solidification
mixtures.
ACKNOWLEDGEMENTS
The authors would like to thank the U.S. EPA, Environment
Canada, the U.S. Army Corps of Engineers and the LSU Haz-
ardous Waste Research Center for financial support of this work.
Special thanks go to the following individuals for their sugges-
tions, conversations and guidance: Carlton Wiles, Charles
Meshni, Phil Malone, Jerry Jones, Trevor Bridle, Pierre Cote and
Roger Seals. Special thanks go to our students: Marie Walsh,
Donna Skipper, Devi Chalasani and Amitava Roy.
REFERENCES
1. Cooper, J.A., Cousens, D.R., Lewis, R.A., Myhra, S., Segall, R.L.,
Smart, R. St. C., Turner, P.S. and White, T.J., "Microstructural
Characterization of Synroc C and E by Electron Microscopy," /.
Am. CeramicSoc., 68, 1985,64-70.
2. Ramachandran, V.S., Concrete Admixtures Handbook: Properties,
Science, and Technology, Noyes Publications, Park Ridge, NJ, 1984.
3. Cartledge, F.K., Chalasani, D., Eaton, H.C., Tittlebaum, M.E.
and Walsh, M., "Combined Techniques to Probe Chemical Inter-
actions of Organics with Solid Matrices," Proc. of the Meeting of the
American Chemical Society, Miami, FL, May 1985.
4. Eaton, H.C., Walsh, M.B., Tittlebaum, M.E., Cartledge, F.K. and
Chalasani, D., "Microscopic Characterization of the Solidification/
Stabilization of Organic Hazardous Wastes," Proc. of the Energy-
Sources and Technology Conference and Exhibition, Dallas, TX,
Feb. 1985, American Society of Mechanical Engineers Paper, No.
85-Pet-4.
5. Walsh, M.B., Eaton, H.C., Tittlebaum, M.E., Cartledge, F.K. and
Chalasani, D., "The Effect of Two Organic Compounds on a Port-
land Cement-Based Stabilization Matrix," See these Proceedings.
6. Chalasani, D., Cartledge, F.K., Eaton, H.C., Tittlebaum, M.E. and
Walsh, M.B., "The Effects of Ethylene Glycol on a Cement-Based
Solidification Process," See these Proceedings.
7. Suzuki, K., Nishikawa, T. and Ito, S., "Formation and Carbona-
tion of C-S-H in Water," Cement and Concrete Research, 15, 1985,
213-224.
8. Ramachandran, V.S., "Cement Science," in Concrete Admixtures
Handbook: Properties, Science, and Technology, V.S. Ramachan-
dran, Ed., Noyes Publications, Park Ridge, NJ, 1984, 8.
BARRIERS & WASTE SOLIDIFICATION 189
-------
Site Characteristics and the Structural Integrity
of Dikes for Surface Impoundments
Jey K. Jeyapalan, Ph.D., P.E.
Ernest R. Hanna
Wisconsin Hazardous Waste Management Center
University of Wisconsin
Madison, Wisconsin
ABSTRACT
Before a land disposal facility is developed for hazardous
wastes, the site characteristics need to be determined. The infor-
mation obtained for the geological and hydrogeological features
is carefully used in the design of the surface impoundment.
Special design provisions need to be chosen if the site conditions
are poor.
The damage due to the failure of dikes containing the wastes in
surface impoundments poses severe risk of contamination of sur-
rounding areas. The requirements of RCRA seek to ensure the
structural integrity of dikes by requiring certification of these
structures by registered engineers. The evaluation of site char-
acteristics and the structural integrity of dikes is a complex sub-
ject. For the benefit of those involved in dike certification and
review of permit applications, the essential steps involved in these
evaluations are summarized in this paper.
INTRODUCTION
A surface impoundment design requires adequate data on the
site characteristics, geometry of the dike and properties of the
soils forming the dike. In addition, information on groundwater
conditions is required to complete the design. Thus, for proper
water disposal techniques, the identification of various features at
the site is an important step. Subsequent to the determination of
site characteristics, a detailed laboratory testing program needs to
be undertaken to determine the remaining engineering properties
of soils.
The geometry of the problem is defined based on specifications
or field survey results, and the structural integrity of dikes is
determined using the appropriate analysis procedure. The struc-
tural adequacy needs to be ensured for different stages of the life
of the surface impoundment and shear strength properties ap-
propriate for these stages need to be used in the analyses. Thus,
adequate understanding and recognition of all fundamental prin-
ciples of geotechnical engineering and hydrogeology are necessary
to successfully complete the above steps of engineering analyses.
SITE CHARACTERISTICS
The first step is the site reconnaissance in which all pertinent in-
formation is obtained. An efficient and economical way to begin
the site investigation is to review all available literature. This pro-
cess involves a review of past and present land uses, review of site
aerial photography and assessment of the site topography,
cultural landmarks, surface hydrology and surficial geology. The
location, accessibility, general area topography, past and present
ownership and/or use of a site are determined from maps ac-
quired from the local assessor's office and/or from the U.S. Geo-
logical Survey (USGS). The USGS and the Soil Conservation Ser-
vice (SCS) may be utilized as sources of preliminary information
regarding the surficial geology of the site under investigation.
Site-study reports, as well as previous plans and specifications,
may be available from area contractors, consultants and govern-
ment agencies. Any construction activity in the project area
should be visited. A site survey should be performed to familiarize
the engineer with the area. High water marks and vegetation can
indicate the nature of the soil. Recent changes in topography also
can be helpful in design procedures.
Geological conditions and project objectives will dictate the
type of drilling needed to meet the job requirements. It is not
always possible or cost-effective to use a single drilling method to
complete a project. This is one reason why, prior to specifying a
drilling method, an analysis of the project requirements is done.
Selection of a drilling method most suited to the particular job
can be based on the following factors:
• Type of formation
• Depth of drilling required
• Availability of drilling equipment
• Special and/or other specific requirements
The use of a backhoe or shovel is a very simple and quick
means of determining shallow soil conditions and shallow
groundwater conditions. Advantages are the low cost, mobility
and ease of operation. The main disadvantages are the depth
limitation and sample disturbance.
Surface Geophysical Techniques
The application of geophysics to hazardous waste management
combines proven technology with new technology to form an in-
tegrated systems approach. This process is efficient and cost-
effective. Several types of geophysical exploration techniques are
now available for a rapid evaluation of subsoil characteristics.
Definite interpretation of the results requires correlation with
standard subsurface exploration methods. Geophysical methods
can be divided into two categories: (1) surface techniques and (2)
downhole techniques. A few of these surface geophysical tech-
niques are described below.
Ground Penetrating Radar
Ground penetrating radar (GPR) uses high frequency radio
waves to acquire subsurface information. Energy is radiated
downward from a small antenna which is moved slowly across the
surface of the ground. The energy is reflected back to the receiv-
ing antenna, where variations in the return signal are continuously
recorded. Radar responds to changes such as bedding, cementa-
tion, moisture, clay content, voids, fractures, intrusions and man-
made objects. Depth of penetration is site-specific.
190 BARRIERS & WASTE SOLIDIHCATION
-------
GPR is limited in depth by attenuation. Better overall penetra-
tion is achieved in dry, sandy or rocky areas. Poor results are ob-
tained in moist, clayey or conductive soils. Radar penetration to
30 ft is common.
Electromagnetic Measurements
The electromagnetic method provides a means of measuring the
electrical conductivity of subsurface soil, rock and groundwater.
Electrical conductivity is a function of the type of minerals in the
formation, its porosity and permeability and the fluids which fill
the pore space. Often the conductivity of the pore fluids
dominates the measurements.
Natural variations in subsurface conductivity may be caused by
changes in soil moisture content, pore fluid conductance, depth
of soil and thickness of formation layers. The absolute values are
not necessarily diagnostic; it is the variations that are significant.
Seismic Refraction and Reflection
Seismic refraction and reflection techniques are used to deter-
mine the thickness and depth of geologic layers by using the
velocity of seismic waves which is an indication of the soil and
rock's engineering properties. Reflection can be used effectively
to depths ranging from about 50 ft to several thousand feet.
Structural and formation boundaries can be detected. Refraction
commonly is applied to shallow investigations up to a few hun-
dred feet. Depth to bedrock and the layer thickness of the overly-
ing unconsolidated material can be determined.
Seismic waves transmitted into the subsurface travel at dif-
ferent velocities in various types of soil and rock. The waves travel
in all directions from the source and are reflected and refracted
when they impinge on an interface. The combination of pattern
and velocity will affect travel time. An array of geophones on the
surface measures the travel times of the seismic waves from the
source to the geophones. The refraction and reflection techniques
use the travel times of the waves to provide data on subsurface
layers, including the number of layers, the thickness of the layers
and their depths.
Resistivity Method
The resistivity method is used to measure the electrical resistiv-
ity of the section which includes the soil, rock and groundwater.
The method may be used to determine lateral changes and vertical
cross-sections.
The procedure requires that an electrical current be injected into
the ground by a pair of surface electrodes. The resulting potential
field is measured at the surface between a second pair of elec-
trodes. The subsurface resistivity can be calculated by knowing
the separation and geometry of the electrode positions, applied
current and measured voltage.
In general, most soil and rock minerals are highly resistive,
hence the flow of current is conducted primarily through the
moisture-filled pore spaces within the soil and rock. Therefore,
the resistivity of soils and rocks is predominantly controlled by
the porosity and the permeability of the system. Electrical sound-
ing is used to reveal the variations of apparent resistivity with
depth. Horizontal profiling is used to determine lateral variations
in resistivity.
In Situ Measurement Techniques
The determination of soil properties by in situ measurement
techniques also has advanced greatly. A few of these methods are
described below.
Standard Penetration Test
The standard penetration test is the oldest and most widely ac-
cepted method for obtaining shear strength properties. This in-
volves counting the number of blows required of a 140-lb hammer
dropped from a height of 30 in. to move a split spoon sampler 12
in. through the soil. The resultant number is termed the "N"
count and is at present the most widely used index of subsurface
soil conditions. Advantages of the standard penetration test in-
clude economy of use, simplicity of procedure and widespread ac-
ceptance.
Cone Penetration
Cone penetration devices are useful because of simplicity of
testing, reproducibility of results and ease of correlating results to
other test results. The drawback of the test is that no soil sample is
obtained.
The test procedure involves advancing vertically, at a constant
rate, a penetrating cone of about 10 cm2 base area and about 60°
apex angle. The force needed to penetrate a given distance in a
soil is used in measuring shear strength.
Vane Shear
In contrast to methods which determine strength parameters
from empirical correlation, the vane shear test attempts to
measure undrained shear strength directly. The test procedure is
to advance a vane configuration to a desired soil depth and
measure the applied torque as the vane is rotated at a constant
rate.
Shearing resistance is considered to be mobilized on a cylin-
drical failure surface of rotation, corresponding to the top, bot-
tom and sides of the vane assembly. The vane shear test possesses
the advantages of economy of use and reduced soil disturbance.
Pressuremeter
The pressuremeter is used by lowering the apparatus into a
borehole or, if the machine is self-boring, by drilling to the re-
quired depth. Increments of hydraulic pressure are applied to the
test cell which acts against the borehole wall. The resulting defor-
mation of the borehole wall is determined by monitoring the
volume change of the test cell.
The relationship between pressure increment, volumetric ex-
pansion and time is examined to determine strength
characteristics. No soil sample is obtained if the self-boring
pressuremeter is used.
Iowa Borehole Shear Test
The Iowa borehole shear test is performed by lowering a shear
head consisting of two opposing horizontally grooved shear plates
to the required test position in an uncased borehole. The two
shear plates are expanded until seated in the borehole walls at a
selected pressure. After a time period allotted to allow for con-
solidation to occur, the shear head is either pulled upward or
pushed downward at a steady rate of 2 mm/min. The required
forces for shearing are measured, and the shearing stress is plotted
against the normal pressure. By performing a number of tests at
different seating pressures, the soil strength can be calculated.
Soil strength may be expressed in a number of ways, and the
particular strength parameters to use in design will depend on the
nature of a specific problem.
SHEAR STRENGTH OF COHESIVE SOILS
The shear strength behavior of cohesive soils differs consider-
ably from that of cohesionless soils. This section presents the
details of the shearing resistance of cohesive soils under various
loading and drainage conditions. In order to obtain meaningful
results from a laboratory test program for the shear strength
parameters of cohesive soils, it is necessary to test good quality
samples representative of the in situ soils or compacted soils pro-
posed for the project. For in situ soils, the samples to be used
BARRIERS & WASTE SOLIDIFICATION
191
-------
should be undisturbed as far as possible and must be tested at the
in situ water content and dry density. For compacted soils, it is
necessary to match the following three factors between the
laboratory samples and the soils used in the field:
• Method of compaction
• Dry unit weight, 74
• Water content, u
It is necessary to relate shear strength to compaction at an
elementary level in this section. A typical compaction curve show-
ing the dry unit weight-water content relation for a cohesive soil is
given in Figure l(a). The variation of shear strength with water
content is shown in Figure l(b). Although a cohesive soil can be
compacted at the same dry unit weight at points A and B in Figure
l(a), the water content at these points would differ considerably.
Thus, the shear strength of the compacted soil at A would be
higher than that at B. Therefore, it is necessary to prepare
samples in the laboratory at values of water content and dry unit
weight comparable to those of the field compacted soils. Further-
more, the structure of a compacted clay is highly dependent upon
the method of compaction, and the structure has a significant in-
fluence on the shear strength and drainage properties of the clay.
Therefore, it is also important to use a laboratory compaction
procedure that would yield compacted test samples with a struc-
ture similar to that of the field compacted soils.
Compaction
Curve
Water Content , oo
Figure 1
Compaction Curve
Shearing Resistance, Cohesion
The shearing resistance of cohesive soils differs from that of
cohesionless soils due to the following differences in soil factors:
• Particle size
• Particle shape
• Permeability
• Angle of internal friction
• Plasticity
The sizes of particles forming cohesive soils are significantly
smaller than those of cohesionless soils. Cohesionless soils are
made of angular to rounded shapes of particles, whereas cohesive
soils are made of platey shaped particles. Cohesive soils drain
much more slowly than cohensionless soils due to the fact that the
coefficient of permeability of cohesive soils is lower than that of
cohesionless soils. The angle of internal friction of cohesive soils
generally is slightly lower than that of cohesionless soils. Further-
more, the small particle sizes of cohesive soils have attractive
forces giving the plastic behavior for cohesive soils; this factor
does not apply to cohesionless soils. Due to the above factors, the
shear strength behavior of cohesive soils is expressed by the equa-
tion:
t = c + crtan
(1)
where c is cohesion and is the angle of internal friction.
The Mohr-Coulomb failure envelopes for clays and sands are
shown in Figure 2 for comparison. The value c indicated on the
shear stress, also referred to as the "Cohesion Intercept," is due
to the attractive forces between the platey shaped particles form-
ing cohesive soils such as clays.
Mohr-Coulomb
Envelope for Sands-
Mo hr -Coulomb
Envelope for Sands
Normal Stress, o
Figure 2
Strength Envelopes
Pore Pressure Parameters
The low permeability of clay causes pore water pressure to
build up during laboratory or field tests and control the shear
strength properties measured for use in design computations.
Therefore, it becomes necessary to understand the mechanisms of
pore water pressure changes in cohesive soils before making an at-
tempt to understand the shear strength behavior of cohesive soils.
When a sample of clay is loaded in a triaxial test initially by an
equal all around confining pressure, 0$, followed by a deviator
stress, a, -<73, a total pore pressure change, Au, occurs as shown
in Figure 3, under undrained conditions. The above two types of
192 BARRIERS & WASTE SOLIDIFICATION
-------
cr 3 + (a j -a 3 )
Pore
Au
Total
Pressure
Change
Au
A u3
O
•=0=
Pore Pressure
AU3=B03
Change
(a,
a3 )
Au
Pore Pressure CMange
= A( o"!- 0-3 )
Figure 3
Pore Pressures
loading independently cause two increments of pore pressures,
Au3 and Aui3, given by:
Au = Au3 +
where:
and
J3 =
(2)
(3)
(4)
Au13 = A(ff] - CT3)
as shown in Figure 3. In Equations 2 and 4, B and A are referred
to as the "Pore Pressure Parameters." By knowing these pore
pressure parameters, the excess pore water pressures developed in
a triaxial test can be estimated. Conversely, by measuring the
changes in pore pressures, Au3 and Aui3, in a laboratory^ test for
known values of a3 and (al - a3), the values of B and A can be
computed.
The pore pressure parameter, B, is a function of the coefficient
of volume compressibility, mv, of the soil sample shown in Figure
4(a) and the degree of saturation, S, indicated in Figure 4(b). The
coefficient of volume compressibility increases as the stiffness of
the soil skeleton decreases and, as a result, the value of B in-
creases for the same value of saturation as shown in Figure 4(b).
When the sample is dry due to air being far more compressible
than the soil skeleton, the pore water pressures developed are zero
and, therefore, the value of B is regarded as zero. Since the pore
water will carry the entire confining pressure under undrained
loading conditions for a saturated sample of clay, the pore
pressure parameter, B, becomes 1.0. The pore pressure
parameter, A, is a function of stress ratio, oj/cr;, as shown in
Figure 4(c). For normally consolidated clay, A is approximately
1.0; for overconsolidated clays, it is between -0.3 and 0.8.
Au
1.0
0.5
a)
10
^
to
a,
50 100
Saturation, S(%)
1.5
\<
-0.3
024
Stress Ratio, ai /o 3
Figure 4
Pore Pressure Parameters
Unconsolidated-Undrained Test
The Unconsolidated-Undrained, U-U, Quick or Q Test usually
is performed to obtain strength parameters, c and >, correspond-
ing to a fast loading of the soils at sites where drainage is poor. In
this test, the soil sample is trimmed and placed in a thin mem-
brane as shown in Figure 5(a). The confining pressure, a3, is ap-
plied to simulate the confinement available to the sample in the
field while the drainage value is closed. The sample is sheared
while the drainage valve is closed, not allowing any excess pore
pressures built up in the sample to dissipate, as shown in Figure
5(b).
Since it is impossible to saturate the sample without con-
solidating, pore pressures are never measured in the U-U test. The
Mohr-Coulomb envelope for a saturated sample of clay in a U-U
test is horizontal, as shown in Figure 6(a), since the deviator stress
at failure is independent of the confining pressure applied to the
BARRIERS & WASTE SOLIDIFICATION 193
-------
O 1
Sample of Soil
o 3
Valve Closed
o i
0 3
Valve Closed
(Pore Pressures
are never
measured)
Figure 5
Unconsolidated-Undrained Test
sample. For a saturated sample of clay yields the U-U test gives
the following strength parameters:
c = Su
= O
(5)
(6)
where Su is the undrained shear strength of the clay.
Since the deviator stress is independent of the confining
pressure, it is possible to perform a test with a confining pressure
of zero and obtain the undrained shear strength, Su> of the soil
from the diameter of the failure Mohr's circle as shown in Figure
6(b). This test is referred to as the "Unconfined Compression
Test," where the sample is sheared by a compressive load, a^
without the use of a confining cell. The diameter of the Mohr's
circle qu, in Figure 6(b) is called the "Unconfined Compressive
Strength" of the sample.
When a partially saturated sample of clay is sheared in a U-U
test, the size of the Mohr's circle increases with increasing confin-
ing pressure as shown in Figure 7. Once the confining pressure
reaches a high enough value to cause the pore air to dissolve in
pore water, the sample becomes saturated. Beyond this value, the
sample yields a horizontal Mohr-Coulomb envelope as shown in
this figure. Thus, in order to obtain the strength parameters, c
and >, corresponding to a confining pressure, a3 (of interest in the
partially saturated range), it becomes necessary to draw a tangent
at <73 = a'j on the envelope and read values for c* and ' as shown
in Figure 7.
The strength parameters obtained from a U-U test are useful
for analyzing geotechnical engineering problems where the soils
did not consolidate appreciably and the drainage conditions were
poor. An example of such an application is the slope stability
evaluation for the dike for the end-of-construction loading condi-
tion.
in
to
0)
1-1
4-1
0
Confining
Cell
Unconsolidated-Undrained Envelope
c = S . « = 0
r ZOLA
Normal Stress, a
Mohr Circle for the
Unconfined Compression
Test with O3 = 0
Normal Stress,o
Figure 6
Strength Envelopes
I/)
ID
in
Soil is
Pa_rt ially
Saturated
Soil is Saturated
* for 03 03 *
O 3 03
A A
Normal Stress, o
03-03
Figure 7
Strength Envelopes
194 BARRIERS & WASTE SOLIDIFICATION
-------
Consolidated-Undrained Test
The Consolidated-Undrained, C-U or R Test usually is per-
formed to obtain strength parameters, c and , corresponding to a
fast loading of the soils at a site where drainage is poor and the
site has undergone consolidation. In this test, the drainage valve is
left open during the application of the initial confining pressure,
'under drained loading conditions. The rate
of loading in this test is slow enough to dissipate the excess pore
w
10
0)
1-1
4-1
CO
Total Stress
Envelope
Effective Stress
Envelope
Effective Normal Stress,a
Figure 9
Strength Envelopes
water pressures as they build up during the application of the
shear load. The strength parameters obtained from this test pro-
vide the data necessary for long-term stability calculations of
slopes and embankment dams.
EVALUATION OF STRUCTURAL
INTEGRITY OF DIKES
It is necessary that the dike be stable for the following loading
conditions:
• End-of-construction or short-term
• Steady seepage or long-term
• Rapid drawdown
• Seismic
The factor of safety against any form of slope failure usually is
determined using the following:
• Slope geometry
• Shear strength of soil
• Method of analysis
There are more than 30 methods of slope stability analysis.
Among these, Bishop's modified method, Ordinary Method of
Slices and Spencer's method are most commonly used. Bishop's
Method and the Ordinary Method of Slices are used with circular
failure surfaces, and the Spencer's Method is used for noncircular
slope surfaces. These methods can be applied for dike stability
evaluation using:
BARRIERS & WASTE SOLIDIFICATION 195
-------
• Hand computations
• Computer programs
• Stability charts
Hand Computations
The hand computations are almost never used in practice at the
present time for slope stability calculations. These are tedious and
time-consuming.
Computer Programs
Computer programs for IBM personal computers and main
frame computers are available from the authors for performing
dike stability evaluation. These computer programs have various
desirable options to handle even the most difficult site conditions.
Stability Charts
The stability of slopes can be analysed efficiently using the
stability charts shown in the manual by Duncan and Buchignani.'
Although the charts assume simple slopes and uniform soil condi-
tions, they can be used to obtain reasonably accurate answers for
most complex problems if irregular slopes are approximated by
simple slopes and average values of unit weight, cohesion and
friction angle are used.
MINIMUM FACTOR OF SAFETY
When detailed analyses of slope stability are performed using
the computer programs based on either Ordinary Method of
Slices, Bishop's Modified Method or Spencer's Method, a
number of slip surfaces must be examined to locate the most
critical failure surface with the lowest factor of safety. The slope
stability charts yield minimum factors of safety when properly ap-
plied. Under most conditions, the uncertainties due to approx-
imations and assumptions in the method of analysis are smaller
than the uncertainties due to inaccuracies in measuring the shear
strength. Approximations in the analysis usually amount to 15%
or less, but the margin for error in evaluating shear strength may
be considerably greater. The minimum allowable value of factor
of safety for a slope in a dike depends on several factors; some
guidance is given in Table 1 for static and seismic conditions.
Table 1
Recommended Minimum Values of Factor of Safety
Uncertainly of Strength Measurements
Costs and Consequences
of Slope Failure Small1
Cost of repair comparable to cost
of construction. No danger to human 1.25
life or other property if slope fails. (1.2)*
Cost of repair much greater than
cost of construction, or danger to 1.5
human life or other valuable (1.3)
property if slope fails.
Large'
1.5
(1.3)
2.0
or greater
(1.7 or
greater)
1 The uncertainly of ihc Mrcnglh measurement* is smallest when ihc soil conditions arc unilorm
and high quality .strength lest data provide a conslslcnt, complete and logical picture of the
strength characteristics.
' The uncertainly of (he strength measurements is greatest when the soil conditions are complex
and when the available strength data do not provide a consistent, complete, or logical piclure
of the strength characteristics
* Numbers without parentheses apply for static conditions and Ihosc with parentheses apply lor
seismic conditions
(After Duncan and Buchignani')
EVALUATION OF BEARING
CAPACITY OF DIKE FOUNDATIONS
If the foundation of a proposed dike is composed of low
strength soils, the factor of safety against any type of bearing
capacity failure should be evaluated. The bearing capacity consid-
eration will control the base width of the dike, especially when the
foundation contains a scam of weak material, as shown in Figure
10. In this figure:
q'w=7H. 0)
Using stress influence charts, the influence factor, I, is obtained
corresponding to the dimensions, a and b.
Using I, the stress, qw, applied to the seam of weak soils can be
calculated as:
Iq «
(8)
The ultimate bearing capacity of the weak material is evaluated
using the equation:
qu _ cNc
BN- t- Z"»-NC:
(9)
where Ne, N,f and Ng are bearing capacity factors obtained from
bearing capacity charts corresponding to the angle of friction, ,
of the soil. In this equation, B and y refer to the width of the
stress distribution as shown in Figure 10 and the unit weight of the
soil, respectively. The factor of safety against bearing capacity
failure is calculated using:
F
' * ~ q»
and Fs should be higher than 2.5 to 3.
(10)
1 1 1 i 1 1 i 1 1
WEAK LAYER
, B ,
Figure 10
Bearing Capacity Failure
CONCLUSIONS
The RCRA regulations require that the surface impoundments
must have dikes that are designed, constructed and maintained
with sufficient structural integrity to prevent massive failures of
the dikes. Thus, the engineers involved in certifying dikes need to
use appropriate engineering procedures to determine site
characteristics and dike integrity. Some guidance is given in this
paper to accomplish these tasks undertaken by engineers involved
in the preparation and review of permit applications for surface
impoundments, waste piles and other land disposal units.
196 BARRIERS & WA§TE SOLIDIFICATION
-------
REFERENCES
1. Bishop, A.W., "The Use of Slip Circle in the Stability Analysis of
Slopes," Geotechnique, 15, 1955, 7-17.
2. Bishop, A.W. and Bjerrum, L., "The Relevance of the Triaxial Test
to the Solution of Stability Problems," Proc. of the ASCE Research
Conference on the Shear Strength of Cohesive Soils, Boulder, CO,
1960.
3. Duncan, J.M. and Buchignani, A.L., "An Engineering Manual for
Slope Stability Studies," U.S. Berkeley Geotechnical Engineering
Report, 1975.
4. Holtz, R.D. and Kovacs, W.D., An Introduction to Geotechnical
Engineering, Prentice-Hall, NJ, 1981.
5. Hvorslev, M.J., "Subsurface Exploration and Sampling of Soils for
Civil Engineering Purposes," Corps of Engineers, Waterways Ex-
periment Station, Vicksburg, MS, 1949.
6. Janbu, N., "Slope Stability Computations," Soil Mechanics and
Foundation Engineering Report," the Technical University of Nor-
way, Trondheim, Norway, 1968.
7. Jeyapalan, J.K., Theory and Problems of Geotechnical Engineering,
McGraw-Hill, New York, NY, 1985.
8. Lowe, III, J. and Karafiath, L., "Stability of Earth Dams Upon
Drawdown," Proc. of the First PanAmerican Conference on Soil
Mechanics and Foundation Engineering, Mexico City, Mexico, 2,
1960, 537-552.
9. Mitchell, J.K., Guzikowski, F. and Villet, W.C.B., "The Measure-
ment of Soil Properties In situ, Present Methods—Their Applica-
bility and Potential," National Technical Information Service,
1978.
10. Morgenstern, N.R., "Stability Charts for Earth Slopes During Rapid
Draw-down," Geotechnique, 13, 1983, 121-131.
11. Navfac DM-7, "Soil Mechanics Design Manual," U.S. Navy Of-
fice, 1981.
12. Seed, H.B., "Considerations in Seismic Design of Earth and Rock-
fill Drums," 19th Rankine Lecture, London, 1979.
APPENDIX I — NOTATION
The following symbols used in this paper:
U-U
C-U
D
c
0
S,,
H
7w
N
Qc
Unconsolidated-Undrained
Consolidated-Undrained
Drained
Cohesion
Angle of Internal Friction
Undrained Shear Strength
Effective Consolidation Pressure
Unit Weight
Dike Slope Angle
Dike Height
Unit Weight of Shear
Blow Count
Cone Resistance
Factor of Safety
BARRIERS & WASTE SOLIDIFICATION 197
-------
Use of X-Ray Radiographic Methods
in the Study of Clay Liners
Philip G. Malone, Ph.D.
James H. May
Geotechnical Laboratory
U.S. Army Engineers Waterways Experiment Station
Vicksburg, Mississippi
Kirk W. Brown, Ph.D.
James C. Thomas
Soil and Crop Sciences Department
Texas A&M University
College Station, Texas
ABSTRACT
X-ray radiography has been widely used in soil investigations
to study the distribution of layers in soil cores and the effects of
changing conditions (loading or impact) on soil structure. X-ray
radiographic techniques also can be useful in studying clays or
clay soils used in liners.
Laboratory investigations were undertaken to demonstrate that
X-ray radiographic techniques could be used to detect density and
soil structure changes that usually accompany variations in
hydraulic conductivity of clay liners. An example of a real-time
test of a simulated bentonite and sand liner attacked with acid
lead nitrate and examples of radiographic examination of clay soil
(non-calcareous smectite) samples that have been permeated by
lead acetate or lead nitrate are presented. The changes in density
and structure can be related to changes observed in hydraulic
conductivity during permeation.
X-ray radiography easily can be applied to field samples of
soil or clay liner materials to detect density and structural changes
that occur as the liner and permeating fluid interact. X-ray tech-
niques have applications in both understanding failure mechan-
isms and forecasting liner performance.
INTRODUCTION
X-ray radiographs of soil cores have been employed in geo-
technical investigations for many years because radiographs pro-
vide a unique technique for detecting changes in the physical
arrangement of soil particles without disturbing a soil core. In
many soil cores, X-ray radiographs will detect subtle differences
in soil density that might have escaped notice during the visual
examination of core sections. The X-ray image displays varia-
tions in the absorption of X-ray photons in different parts of the
object examined. Differences in absorption can be related to
changes in sample thickness, density or chemical composition.1
When the sample geometry is controlled by preparing cylinders
or slices of cylinders for examinations, patterns observed in the
radiographs can be reliably related to the differences in sample
density and chemistry.
This study examined the use of X-ray radiographic techniques
in identifying changes in the density and composition of samples
of simulated clay or other soil liners that have been exposed to
permeants that could be encountered in groundwater beneath
an industrial or hazardous waste disposal site. Two approaches
were used: (1) examining the real-time invasion of a bentonite-
sand liner using a lead nitrate-nitric acid solution; and (2) ex-
amining the condition of soil samples after the samples have
been permeated with simulated industrial waste solutions.
PVCW€
BENTOMTC
AND —
SAND
PVC COUPUNG -
s 83mm —
|> 75 mm ^
/
. _ . '
. "
/ '
/ "
"~ , _ , __ . _ . _ . /
/
•k ~"-"~ /
% -_ \f
W " ": " \
s - ^!_ ~ - __ - . /
IOC
1
I
T
1<
k.
*71 *.
40C
mm
X) mm
t
> rr
Figure I
Cross-Section of the Test Cell Used in the Real-Time Liner Test
REAL-TIME STUDY OF LINER PERMEATION
A simulated waste pond liner was prepared by mixing one
part of a commercially available bentonite (Aquagel from Baroid
Petroleum Services of N.L. Industries, Houston, Texas) with
four parts by volume of coarse filter sand. The mixture follows
typical procedures for producing a bentonite liner.J The sand and
198 BARRIERS & WASTE SOLIDIFICATION
-------
bentonite were mixed in a dry condition and water was gradually
added with continued mixing to assure that the bentonite was
completely hydrated. The wet mixture was permitted to settle for
24 hours, and excess water was decanted. A test cell was prepared
from a 400-mm length of 75-mm ID schedule 40 PVC pipe (Fig.
1).
The bottom of the test cell was fitted with a plastic plug and a
hole was drilled through the plug so that a drainage line could be
attached. The bottom of the test cell was packed with Dacron
filter floss, and a 100-mm thick layer of coarse filter sand was
poured into the test cell. The bentonite and sand mixture was
poured in on top of the filter sand, and the test cell was gently
shaken to assure that the bentonite and sand layer settled and no
air bubbles formed in the simulated liner. Clean water was placed
on top of the bentonite and sand layer to verify that the simu-
lated liner would not leak. To initiate the testing, the clean water
was decanted, and a simulated waste formed by saturating a 10%
nitric acid solution with reagent-grade crystalline lead nitrate was
added above the bentonite and sand layer. The saturated acidic
lead nitrate solution was approximately 140 mm deep over the
simulated liner. This depth of waste is less than the 300-mm
depth permitted under disposal guidelines.3
INVADED AREA-
SUPPORT RING
1 cm
Figure 3
Print of Radiograph of Simulated Liner One Day after Application of
Acidic Lead Nitrate Solution
Figure 2
Print of Radiograph of Simulated Liner Before Application of Acidic
Lead Nitrate Solution
X-ray radiographs were made of the test cell at one to five
day intervals using a Norelco MG 3000 X-ray system equipped
with a tungsten target tube. Kodak Industrial M film was used
with a 1 to 2 min exposure. The X-ray unit was operated at 290
kV with 11 mA current.
Figure 2 shows the liner prior to testing. Breakthrough of the
simulated waste was noted after 16 days. Figures 3 and 4 show the
progressive failure of the liner after a one-day and after a 25-day
exposure to the simulated waste. When the test cell was opened,
the presence of voids in the simulated liner easily could be ob-
served (Fig. 5). The X-ray radiographs had detected shrinkage
cracks forming in the simulated liner and had accurately indi-
cated the mode of failure in the simulated liner. Lead nitrate, a
radio-opaque solution, was used initially in the tests to assure that
any invasion of the liner could be observed. The formation of the
voids in the clay reduced the usefulness of the lead solution be-
cause the lead salts lowered the contrast between the fluid-filled
voids and the surrounding soil. Where flocculation and void
formation are the expected modes of failure, there is no benefit in
using radio-opaque permeants.
BARRIERS & WASTE SOLIDIFICATION 199
-------
INVADED AREA
1 cm
Figure 4
Print of Radiograph of Simulated Liner 25 Days after Application
of Acidic Lead Nitrate
X-RAY RADIOGRAPHIC STUDY OF
PERMEAMETER SAMPLES
The evaluation of the interaction of soil liners and simulated
or actual industrial wastes generally has been performed by forc-
ing samples of selected permeants (wastes) through packed soil
columns in a fixed wall or flexible wall permeameter.' Inter-
actions are evaluated by determining if the permeability of the
packed soil is changed by the permeant. The type of interaction
that has occurred between the soil and the permeant cannot be
determined without examining the soil. For example, the
hydraulic conductivity of the liner may decrease, but the reason
for the decrease (for example, swelling of the soil or the forma-
tion of precipitates) cannot be determined without measuring the
change in sample volume (for swelling) or looking for the forma-
tion of new compounds in the clay (for precipitation). X-ray
radiographs can detect both volume changes and increased liner
density due to the formation of precipitates. More than 60 soil
samples from permeameter tests were examined to evaluate the
ability of radiographic data to assist in the interpretation of
permeability changes.' The major effects observed have included
the breakdown of the clay structure (causing an increase in sample
permeability) and the formation of precipitates (with a decrease
in permeability).
Figure 5
Photo of Bentonile and Sand Mixture after Permeation with Acidic Lad
Nitrate Solution
Samples of permeated soils were prepared using techniques
proposed for testing potential landfill liner materials.' In the
examples discussed here, six samples of the Lufkin soil (a non-
calcareous smectite) from Brazos County, Texas, were packed
into standard-design fixed wall permeameters (Fig. 6). Each
sample was compacted in three equal lifts using 25 blows per lift
from a 2.4l-kg hammer falling 300 mm. One set of three samples
was permeated with a 60% (by weight) lead acetate solution; the
other set was permeated with a 50% (by weight) lead nitrate solu-
tion in 0.1% HNOj.
After a sufficient volume of permeant had passed through the
sample to establish a trend in changes in hydraulic conductivity,
the permeameters were opened and checked for swelling. The
samples permeated with lead acetate had increased in height by
PRESSURE INPUT
PERMEAMETER
BASE
OUTLET
POROUS STONE
TEFLON TUBING
Figure 6
Cross-Section of Permeameter Used in the Study of Soil-Waste
Interaction
200 BARRIERS & WASTE SOLIDIFICATION
-------
1 cm
Figure 7
Print of Radiograph of Non-Calcareous Smectite Soil Permeated
with Lead Acetate Solution
maximum of 2.7%. Maximum swelling in the samples permeated
with lead nitrate was 1.6%. The samples were extruded and wrap-
ped in foil. X-ray radiographs were prepared using a Norelco
MG 300 Industrial X-ray System. The unit was operated at 290 kV
at 10-12 mA. Exposure times varied from 1 to 3 min using Kodak
Industrial M film.
Permeameter testing established that the lead acetate caused a
decrease in hydraulic conductivity by a minimum of 50%. X-ray
radiographs showed lead compounds were precipitating in the soil
(Fig. 7).
The lead nitrate solution caused an increase in hydraulic con-
ductivity by a minimum of over 100%. The X-ray radiographs
showed irregular connecting voids suggesting the clay had devel-
oped a flocculated or blocky structure similar to that observed
in the real-time failure study (Fig. 8).
CONCLUSIONS
X-ray radiographic techniques, previously applied to the study
of soils, can be employed usefully in the investigation of the inter-
action of wastes and clay (or clay soil) liners. Using suitable test
cells, real-time effects of waste (or simulated wastes) on sections
of liner can be observed. Soil samples that have been permeated
by simulated wastes can be X-rayed to examine the soils for den-
sity or structure changes that indicate occurring liner-waste inter-
action.
Field techniques for sampling soils to obtain undisturbed sam-
ples are available, and it should be possible to use X-ray radio-
graphic methods along with standard soil description techniques
to examine the condition of liners at landfills or waste storage
areas. X-ray examination of soil cores is completely non-destruc-
tive, and altered sections observed in the X-ray radiographs can
be documented by sampling and examining anomalous soil inter-
vals.
1 cm
Figure 8
Print of Radiograph of Non-Calcareous Smectite Soil Permeated
with Lead Nitrate Solution
ACKNOWLEDGEMENTS
The tests described and the resulting data presented herein, un-
less otherwise noted, were obtained from research conducted
under the Independent Laboratory Initiated Research Program,
Project No. 4A161102A91D Task Area 02 Work Unit 154, of the
United States Army Corps of Engineers by the USAE Water-
ways Experiment Station. Permission was granted by the Chief
of Engineers to publish this information.
REFERENCES
1. Krinitzsky, E.L., Radiography in the Earth Sciences and Soil Mechan-
ics, Plenum Press, New York, NY, 1970.
2. Matrecon, Inc., "Lining of Waste Impoundment and Disposal Facil-
ities," U.S. EPA, SW-870, Washington, DC, 1983.
3. U.S. EPA, "Hazardous Waste Management System; Permitting
Requirements for Land Disposal Facilities," Federal Register, July
26, 1982.
4. Brown, K.W. and Anderson, D.C., "Effects of Organic Solvents on
the Permeability of Clay Soils," EPA-600/2-83-016, U.S. EPA, Cin-
cinnati, OH, 1983.
5. Malone, P.O., May, J.H., Brown, K.W. and Thomas, J.C., "Use
of X-ray Radiographic Techniques in the Evaluation of Soil Liners,"
Miscellaneous Paper GL-85-14, USAE Waterways Experiment Sta-
tion, Vicksburg, MS, 1985.
BARRIERS & WASTE SOLIDIFICATION 201
-------
Closure Design and Construction of Hazardous
Wastes Landfills Using Clay Sealants
John F. O'Brien, P.E.
Lonnie E. Reese
O.H. Materials Co.
Rosewell, Georgia
Ian Kinnear, P.E.
Dames & Moore
Boca Raton, Florida
ABSTRACT
Clay sealants are employed at an increasing number of sites
securing land disposal of hazardous industrial wastes. Such seal-
ants are employed as topliners to limit infiltration and/or as bot-
tom liners to limit exfiltration and provide for leachate collection.
These sealants are employed both as primary liners and as added
security to synthetic liners. Liner design for hazardous waste land-
fills is well-regulated and quantitative analytical tools are well-
documented. However, the practical implementation of cost-
effective, buildable and reliable clay liners requires design in-
sight beyond the obvious. Key elements in the selection, design
and construction of landfill liners using clay sealants are discussed
herein. The case histories of two landfill closures are presented to
amplify the impact of the designers' assumptions on implemen-
tation.
INTRODUCTION
The national concern for environmentally secure disposition of
hazardous wastes has led to increased sophistication in disposal
technology. In land disposal, landfill design has been expedited
not only by expansive regulatory requirements, but also by ana-
lytical modeling which allows cost- and time-effective evalua-
tion of design alternatives.
Ideally, landfills securing hazardous materials are designed
prior to disposal of the materials. Such planning includes design
to limit the development of leachate and preclude exfiltration of
leachate to the groundwater. However, as a result of years of un-
regulated and unenforced disposal, thousands of small to large
"out back" sites at industrial facilities exist across the nation. It
is the closure of these existing facilities which is the subject of this
paper. These existing facilities include several distinguishing
characteristics which impact closure design and construction:
• The waste material often is poorly defined
• Stabilized or unstabilized, wastes are left in-place and closure
includes emplacement of a top liner only
• Construction of the design groundform often includes sub-
stantial regrading of unknown materials, placing a premium on
planning equipment, selection and safety.
Two case histories will amplify this discussion.
• "Site A, Southern Florida": closure of a hazardous waste land-
fill using an imported bentonitic clay as a topliner
• "Site B, South Carolina": closure of a solvent contaminated
landfill using locally available clays
Site A, South Florida
The landfill at Site A is located within an industrial facility in
South Florida. It consists of one 4-acre cell that was constructed
to a height of about 35 ft above existing grade. The landfill was in
operation from 1958 until the middle of 1984. The cell was util-
ized primarily for the disposal of paper cartons, boxes, construc-
tion debris, cafeteria waste, yard trimmings and scrap metal.
However, in the early years of filling when operating practices
and regulations were less stringent than today, unbagged asbestos
and drums containing solvents and fuels were deposited in the
landfill.
Below the cell, the groundwater is contaminated with VOCs;
the resulting contamination plume is slowly migrating eastward,
away from the landfill. Associated with the landfill closure, a
groundwater recovery well system and treatment plant will be
constructed around the perimeter of the landfill.
The landfill is bounded closely on three sides by water. Uncon-
trolled filling over the years left a very irregular groundform
with slopes as steep as 1:1.
Site B, South Carolina
Site B is located within an industrial facility in South Carolina.
Over at least 10 years of operation, the site was used as a dump
for solvents and drums. At the time of closure, the drums had
been removed and closure involved capping solvent contaminated
soils on the 5-acre site. The contaminated soils were encountered
largely in depressions in the land surface on this site.
MODELING LANDFILL PERFORMANCE
Closure design of industrial and hazardous waste landfills
usually involves the following generic evaluations:
• Determine the waste composition and characteristics at the
closure sites
• Determine regulatory and design dictated materials and
groundform requirements
• Determine liner materials alternatives based on compatibility
with the waste, availability, constructability and performance
• Complete analytical evaluations to determine optimum landfill
design based on cost and performance
Regulatory
The regulatory requirements for design of hazardous waste
landfills are well-documented by the U.S. EPA under RCRA,
published July 26, 1982. The Interim Final Rule on "Hazardous
Waste Management System: Permitting Requirements for Land
Disposal Facilities" became effective Jan. 26, 1983.' These fed-
eral criteria do not dictate either liner parameter, but do include
the criteria for bottom liners for new (not existing) landfills.
...The liner must be constructed of materials that prevent
wastes from passing into the liner during the active life of
the facility.1
202 BARRIERS & W/kSTE SOLIDIFICATION
-------
This requirement virtually requires the use of geomembranes in
lieu of any clay liner as the primary sealant of a new landfill.
In the closure design of existing facilities, closure requires a
cap to limit infiltration as well as groundwater monitoring. In the
usual case, closure of an existing hazardous site allows the de-
signer some latitude in the selection of a landfill cap, either em-
ploying geomembranes or clay liners.
The regulatory environment of closure design requires con-
sideration. To this end, it is imperative the owner and engineer
develop communication with the appropriate regulatory agency
early in the design process to assure compatibility of the design
with the regulatory expectations.
Analytical
The hydrologic processes within landfills were originally mod-
eled using water balance techniques developed by Thornwaite.3
The U.S. EPA developed the first widely used landfill model,
Hydrologic Simulation on Solid Waste Disposal Sites (HSSWDS),
in 1980. Since that time, the U.S. EPA has developed Hydro-
logic Evaluation of Landfill Potential (HELP), the most sophis-
ticated analytical model now available for landfill systems. HELP
was developed specifically for hazardous waste landfill evalua-
tions as required by RCRA."
HELP provides a quasi-two-dimensional hydrologic model of a
landfill section, applying section geometry, climatologic data and
material parameters to account for the movement of water across,
into, through and out of landfills. Figure 1 presents a generic
landfill section indicating the solutions provided by the model.
The model allows the use of geomembranes in the section, both
enhancing lateral drainage and limiting percolation. Despite the
claims of some suppliers, geomembranes (synthetic liners) should
not be considered impervious. Hydrologic modeling of a landfill
performance may be completed using any of the above tech-
niques: Thornwaite, HSSWDS or HELP. Of these models,
HELP certainly represents the most sophisticated modeling
alternative. Regardless of the approach used, some quantitative
evaluation of landfill performance should be undertaken.
PRECIPITATION
EVAPOTRANSPIRATION
RUNOFF
ATIVE COVER » » » t t T T
^,J,x„,„,*, * n L»*-,
\'/'/'/'/ /'/'/'/'//\/\ V,'/'/'*'*'}•',''/'/'/'
W'/'/'/'/'/'///Y'\'\. ''/' TOPSO L ''/'/'///.
GEOMEMBRANE
LINER
////
TO
////////
r //////PEI
///^/LINER ////// I J |
^k.^jfe^s^u«kL.' if.-.'.-. DISCHARGE
WASTE
SECTION
'LATERAL DRAINAGE LAYER •-. '• . .*;*.'*• ScJSSe'e
EXFILTRATION THROUGH
BASE OF LANDFILL SECTION
(after Schroeder, et al)
6" DRAINAGE PIPE
TO DRAINAGE DITCH
LINER ANCHOR TRENCH
Figure 1
Modeling a Landfill Section Using HELP
Figure 2
Typical Landfill Cap Employing Clay, Geomembrane and Geotextile
Sealant Alternatives
Clay sealants may be used as, or as a part of, a landfill liner.
Figure 2 shows the implementation of a geomembrane and clay
liner as a cap following stabilization of a sludge pond. Whatever
the case, the designer will have two alternatives in clay selection:
• Naturally occurring and locally available clays
• Locallv available soils improved by admixtures
Clays are chosen as sealants for their low degree of pervious-
ness. In the design of topliners, a clay barrier with a coefficient of
permeability of 10~7 cm/sec or less normally will be required to
limit infiltration.
Many naturally occurring clays can readily achieve this low
permeability. However, the designer should not presume that any
available clay may be placed to this permeability. Design assump-
tions should be correlated with careful laboratory testing. Repre-
sentative samples of the available clay should be compacted in
laboratory molds and tested to determine permeability.
In most topliners used in the closure of existing hazardous
waste sites, the permeant will be rainwater. However, if this is not
the case, a permeant representative of that which will infiltrate the
liner should be used to test the compatibility of the permeant
with the clay minerology.
Permeability of clays will vary with relative compaction. The
most commonly used compaction standard is the standard Proc-
tor, ASTM D-698. Laboratory molds compacted to varying de-
grees of relative compaction and thereafter tested to determine
permeability generally will show substantial variation. Clays com-
pacted to 95% of the maximum dry density achieved in the stan-
dard Proctor test may show an order of magnitude decrease in
permeability over a sample of the same soil compacted to 90% of
standard Proctor. Similarly, a clay compacted from wet material
of the optimum moisture content generally will have a lower
permeability than the same soil at the same relative compaction,
but compacted from dry material. Thorough testing to deter-
mine the mechanical properties of clays considered for use as lin-
ers is important not only to determine the suitability of a material,
but also to determine the range of performance to be expected
from field variations in installation.
In many areas, the naturally occurring soils will not provide the
low permeability required for liner design. In this event, soil
admixes may be used to lower permeability. Cement, as well as
industrial wastes such as fly ash and kiln dust, may be added to
soils to lower permeability. However, the brittle behavior makes
these materials undesirable to seal structures as prone to irregular
settlement as landfills.
The most common soil admix is sodium bentonite. Not com-
monly encountered in nature, sodium bentonite is mined largely
in the North Central United States. As a soil admixture, sodium
BARRIERS & WASTE SOLIDIFICATION 203
-------
VEGETATIVE COVER
TOP SLOPE 4% (TTP)
SIDE SLOPE JM. IV
IMPERMEABLE
'SEAL (4 INCHES) •
IZ" SOIL BUfFER,
REGRAOEO WASTE
OAS CONVCTANCE TRENCH
•f 12' WIDE • ?' OEEPI
Figure 3
Typical Section, Florida Site. Including Gas Conveyance
bentonite is attractive for its expansive characteristics. The addi-
tion of 5-15% (by weight) sodium bentonite to soils will lead to a
coefficient of permeability of 10 7 cm/sec or less. As with natural
clays, the selection of sodium bentonite as a soil admix should
include careful testing of representative materials.
Figure 3 presents a section of the top liner used at Site A, South
Florida. Clayey soils are almost totally absent in South Florida.
Economic evaluations of geomembranes and admix improved
local soils led to the selection of the sodium bentonite and local
sand section. Laboratory testing indicated that the sandy local
soils mixed with 13% by weight sodium bentonite and compacted
to 95% of standard Proctor (ASTM D-698) would achieve a
permeability of 10~7 cm/sec or less.
Figure 4 presents the topliner section used at Site B, South Car-
olina. At this site, the Owner's property included an area of thick
clayey residual soil about 0.5 miles from the landfill. Laboratory
testing demonstrated a permeability of 10~7 cm/sec for these
clays compacted to 95% of standard Proctor.
SOLVENT CONTAMINATED
RESIDUAL SOILS REGRADEO TO 6%
SLOPE
Figure 4
Typical Section, South Carolina Site B
Buffer and Cover Protection
Protection of clay liners from environmental damages is an im-
portant design consideration. Clay liners should, at a minimum,
include a top soil buffer to protect the material from erosion. The
buffer also should be of sufficient thickness to preclude breaking
by roots and burrowing animals. A vegetated soil cover of about
1 ft is sufficient in most cases to adequately protect the clay.
Specifications and Quality Control
Certainly the best closure planning and design is impotent with-
out carefully developed specifications and good field quality con-
trol.
The designer should be responsible for the technical specifi-
cations. In the writers' experience, many landfill closure specifi-
cations include exacting criteria for material type, degree of com-
paction and permeability. Such tightly constructed specifications
open the owner to substantial costs in change orders.
If the material to be used has been pre-determined by
laboratory testing, specifications for liner construction should be
limited to degree of compaction and material type. Additionally,
to assure good field control, a minimum compactive effort and
soil moisture may be specified (for example, "three passes of a
kneading-type compactor weighing a minimum of 8 tons, com-
pacting soils within 3% of their optimum moisture content").
If the specific clay material has not been predetermined,
specifications should be limited to general material type and
permeability requirements.
Installation of clay liners should be under the full-time
surveillance of a representative of the designer's office. At a
minimum, this surveillance should be supplemented by testing for
soil parameters detailed in the drawings and specifications to in-
clude material type, thickness, degree of compaction, moisture
and permeability.
State regulated closure criteria for Site A in South Florida (Fig.
3) dictated permeability requirements (1Q-7 cm/sec), but left
materials selection to the contractor. Quality control testing for
the liner installation included thickness, moisture, relative com-
paction and laboratory permeability of field samples. Additional-
ly, the contractor was required to submit his sodium bentonite-
local sand mix design, which was monitored in the field. The
average permeability determined from field samples was 10~8
cm/sec.
At Site B, South Carolina, liner construction specifications de-
tailed the borrow source and compaction requirements only. Pre-
testing of these soils had well-established the density and
permeability relationships of these soils. A full-time represen-
tative from the designer's office monitored compaction and liner
thickness. Additionally, it was required that this 2-ft thick liner be
placed and compacted in increments no greater than 6 in.
CONSTRUCTION CONCERNS IN CLOSURE
Closure construction of hazardous waste landfills requires con-
siderable skill, equipment and experience. Such work should be
undertaken only by a contractor experienced in the field.
Construction planning for closure involves the following
generic process:
• Equipment scoping, based on materials, earthwork and pro-
ductivity requirements
• Project safety planning
• Decontamination planning
• Contingency planning
The contractor views the closure of a hazardous waste landfill
differently than does the engineer. The good design engineer best
serves his client, the owner, if designs and specifications consider
the constructability questions the contractor faces.
Equipment Scoping
Both sites A and B present good examples of the constructor's
response in equipment and approach to particular site re-
quirements. Site A, in South Florida, was characterized by con-
strained work space and relatively steep side slopes. Site B, in
South Carolina, required effective hauling of a large quantity of
clay cover to the site.
The constrained (bounded closely on three sides by water) site
conditions at Site A, South Florida, required the landfill ground-
form to be constructed with a side slope as steep as practical. The
204 BARRIERS & WASTE SOLIDIFICATION
-------
designer selected 3 horizontal to 1 vertical (3:1). Construction of
the topliner section required the importation of about 16,000 yd3
of soil. Permit requirements required construction to be complete
in an 8-week period. The physical constraints not only dictated
the landfill groundform, but also limited the area available for
materials staging.
The steep side slopes and moderate materials requirements
complicated earthwork at Site A. All soils were placed using
12-yd3 scrapers pushed by a dozer up the sides of the landfill and
depositing soils as they glided down-slope. Materials deliveries
were carefully coordinated with placement productivity because
of the limited staging area.
The sodium bentonite-local sand liner was mixed on-site using a
pugmill to effect thorough mixing. About half-way through the
clay liner construction, mechanical problems forced abandon-
ment of this mixing technique. Thereafter, the clay and sand were
mixed using bulldozers to place the materials together in a con-
trolled area. The scrappers then picked up and placed the mix.
The 4-in. thick clay liner was difficult to place on the steep side
slopes. Measured thickness of the liner varied from 3.5 to 7 in.,
averaging about 5 in.
A self-propelled, rubber tired vibratory compactor was used to
compact the sand-clay topliner at Site A. Again, because of the
steep side slopes, the compactor was pushed up-slope, then glided
and compacted moving down-slope. The thin liner section limited
equipment alternatives. The contractor, concerned that a dozer-
towed vibratory compactor would damage the 4-in. liner with its
tread, elected to use the rubber tired compactor. Adequate com-
paction was achieved in a single pass, which alleviated the
awkwardness of compaction,
Experience at Site A indicated that 3:1 slopes are about the
steepest which may be practically constructed, limiting equipment
and earthwork control. Indeed, flatter slopes are more desirable
from a construction standpoint.
At Site B, South Carolina, emplacement of the clay liner re-
quired moving about 30,000 yd3 of clay from the borrow source 1
mile (by road) from the site. To optimize placement, a high pro-
ductivity earthwork program was established, moving 3,000 yd3
per day. Self-loading 12-yd3 scrapers were used to pick up,
transfer and place the clay. Kneading compactors ("sheepsfoot"
compactors) towed behind bulldozers were used to compact the
clay to specifications.
Project Safety Planning
Safety should be given the highest priority during any work
with hazardous wastes. Personnel involved in cleanups at both
Sites A and B were participants in comprehensive medical
monitoring programs—both that on-going by the contractor and
project specific medical monitoring.
No closure of a hazardous waste landfill should be completed
without a Site Safety Plan (SSP) developed by a Certified In-
dustrial Hygienist (CIH). This document not only advises person-
nel of hazards and potential hazards caused by exposure to the
landfill materials, but also establishes site safety protocol. At a
minimum, this protocol includes definitions of levels of protec-
tion required for different areas of the site as well as site monitor-
ing procedures and emergency equipment and procedures.
Both Sites A and B required personnel to wear full face air-
purifying respirators and skin protection afforded by disposal
suits and gloves. Working in this wear (particularly in summer) is
quite fatiguing and should be associated with increased breaks
and liquid intake. Project planning should consider the associated
loss in productivity.
Both Sites A and B were attended by a full-time health and
safety officer. A portable decontamination trailer, including
showers, and contingency decontamination equipment was util-
ized on each site. Daily "tailgate" safety sessions were conducted
throughout the projects to address safety concerns for each day's
planned work.
Equipment and Personnel Decontamination
Equipment and personnel decontamination is required follow-
ing any exposure to hazardous materials. The best decontamina-
tion program is, of course, one which minimizes the contact of
equipment and personnel with the hazardous wastes. However, as
this exposure is not always avoidable, equipment and personnel
decontamination should be provided for.
Equipment decontamination procedures should be such that no
decontamination is trafficked off-site. Equipment decontamina-
tion normally is undertaken on a specially constructed pad de-
signed to collect and store any wash water.
A convenient personnel decontamination area should be pro-
vided. At this station, workers should have easy access to all safe-
ty equipment required for work on the site. It is here that workers
remove street clothes and put on protective clothing before enter-
ing the site.
CONCLUSIONS
Landfill design has progressed to the point where a number of
liner alternatives may be evaluated quickly and cost effectively.
While quantitative methods to model landfill performance are
available, practical implementation of cost-effective and
buildable liners requires design insight beyond the obvious.
Closure of existing hazardous waste facilities should be completed
by contractors experienced in the field. Experience not only in the
work but also in the associated personnel protection and decon-
tamination concerns, should be closure imperatives.
REFERENCES
1. U.S. EPA, Federal Register, July 26, 1982, 32274-32388.
2. U.S. EPA, 40 CFR 264.301(a)(l).
3. Thornwaite, C.W. and Mather, J. R., "The Water Balance," Clima-
tology, 8, No. 1, 1955.
4. Schroeder, et al., "The Hydrologic Evaluation of Landfill Perfor-
mance (HELP) Model," Municipal Environmental Research Labora-
tory, Office of Research and Development, U.S. EPA, Cincinnati,
OH, 1984.
BARRIERS & WASTE SOLIDIFICATION 205
-------
Soil Liners for Hazardous Waste Disposal Facilities
D.C. Anderson
K.W. Brown and Associates, Inc.
College Station, Texas
ABSTRACT
The Hazardous and Solid Waste Amendments of 1984 con-
tained a requirement for double liners containing both a flexible
membrane liner and compacted soil liner. The soil component can
be used either as the sole secondary liner or as the lower com-
ponent in a composite secondary liner. In either case, soil liners
must be at least 3 ft thick with a demonstrated field permeability
of less than or equal to 1 x 10- ' cm/sec. Factors affecting field
permeability of soil liners include soil characteristics and liner
construction methodology. Some soil characteristics which can
yield low permeability liners (e.g., high plasticity) also can give
rise to the potential for shrinkage and permeability increases.
One of the most important considerations in the selection of
methods for soil liner construction is the destruction of clod struc-
ture during compaction. After the selection of a specific construc-
tion methodology, the best way to achieve design specifications
is by implementating a sound construction quality assurance
program.
INTRODUCTION
Soil liners are a required component of double liner systems
for hazardous waste disposal facilities (Fig. 1). This requirement
was established in the Hazardous and Solid Waste Amendments
of 1984 and further defined in the Minimum Technology Guid-
ance (MTG) on Double Liner Systems.' In this paper, the author
discusses ways to incorporate soil liners into these liner systems
and factors most greatly affecting the field permeability of these
liners.
All double liner systems are required to have primary flexible
membrane liners (FML). The MTG provides two options for the
design of secondary liner systems, both of which include soil lin-
ers. Following is a discussion of the soil liner components to these
liner systems.
Soil Liner Component to
Secondary Composite Liners
One of the designs suggested in the MTG incorporates a secon-
dary composite liner composed of an upper FML and a lower
compacted soil liner (Fig. 1A). In the configuration suggested in
the MTG, the FML and soil liner are sandwiched together to min-
imize the potential for lateral flow between the liners. It has been
suggested that this configuration would limit leakage through de-
fects in the FML to the rate at which leachate could flow into the
area of soil liner directly under the FML defect. If this theory is
correct, composite liners should be able to reduce leakage from
disposal facilities to a much greater extent than either an FML or
soil liner alone.
PRIMARY FML
-SECONDARY FML
UMSATUSATED ZONE
SECONDARY PRIMARY
LEACHATE LEACKATE
COLLECTION—, (COLLECTION
-PRIMARY FVL
UNSATURATEC 20NE
Figure 1
Double Liners for Hazardous Waste Landfills Must Have a Primary FML
and Either an (A) Composite, or (B) Thick Compacted Soil
Secondary Liner
Requirements for the soil component to secondary composite
liners are given in Table 1. The traditional requirement for a
permeability of less than or equal to 1 x 10-' cm/sec must be
verified in the field.' Laboratory permeability tests are no longer
considered adequate to verify actual field performance.
Secondary Soil Liners
The other secondary liner design given in the MTG is that of a
thick compacted |